Reforming The Public Expenditure Management System (MTEF)

Public Disclosure Authorized
35107
Reforming the Public Expenditure

Management System:
Medium-Term Expenditure Framework,
Performance Management, and Fiscal Transparency
The World Bank and Korea Development Institute

Conference Proceedings (March 2004)
Contents
Welcoming Remarks by Choogsoo Kim (President, KDI)
Keynote Speech by Byung-il Kim (Minister of Planning and Budget)
Conference Program
PART I. MEDIUM-TERM EXPENDITURE FRAMEWORK
CHAPTER 1. Developing a Medium-Term Expenditure Framework ......... 1

1. Overview .................................................................................................. 1
2. Stage 1: Macroeconomic and public sector envelopes ....................... 2
3. Stage 2: High-level policy: aligning policies and objectives under
resource constraints ................................................................................ 7
4. Stage 3: Linking policy, resources, and means by sector ................... 9
5. Stage 4: Reconciling resources with means ........................................ 12
6. Stage 5: Reconciling strategic policy and means ............................... 14
References ................................................................................................... 16
CHAPTER 2. MTEF Implementation in Korea .............................................. 17

1. The Starting Point of Reform ............................................................... 17
2. The Vision of MPB ................................................................................. 18
3. What Lies Ahead ................................................................................... 21
CHAPTER 3. Key Issues for Introducing MTEF and Top-down Budgeting

in Korea ...................................................................................................... 23
1. Introduction ............................................................................................ 23
2. Background and features of MTEF initiatives in Korea ................... 24
3. Key issues to be addressed ................................................................... 28
4. Some suggestions for implementation strategy ................................ 35
References ................................................................................................... 39
CHAPTER 4. Successful Installation of MTEF to the Korean Fiscal System .... 41

1. Introduction ............................................................................................ 41
2. General features of MTEF ..................................................................... 43
3. Specific issues for the implementation of MTEF in Korea ............... 49
ii Contents
4. Concluding Remarks ............................................................................. 55

References ................................................................................................... 56
CHAPTER 5. MTEF in Sweden .......................................................................... 57

1. Background ............................................................................................ 57
2. Foundations of budget preparation: level 1 ....................................... 58
3. Allocation: level 2 .................................................................................. 60
4. From formulation to enactment ........................................................... 64
5. Execution ................................................................................................ 66
6. In-year changes in the budget .............................................................. 67
7. Problems and outstanding issues ........................................................ 67
CHAPTER 6. The Swedish Budget Process: Expenditure Targets and Top-

down Decisions ......................................................................................... 71
1. Background ............................................................................................ 71
2. Two budgetary targets, the expenditure ceiling and the surplus
target ....................................................................................................... 73
3. The framework model ........................................................................... 76
4. Effects of the reformed budget process .............................................. 80
5. Criticism and problems in the budget process .................................. 81
6. Concluding remarks .............................................................................. 82
CHAPTER 7. Medium-term Financial Planning in the Federal Republic of

Germany ..................................................................................................... 85
1. Introduction ............................................................................................ 85
2. Principles of medium-term financial planning ................................. 85
3. Participants and procedures in the medium-term financial planning
of the German federal government .................................................... 87
4. Medium-term financial planning in the budget preparation
procedure ................................................................................................ 89
5. Conclusion .............................................................................................. 92
CHAPTER 8. Medium-term Perspectives in Budgeting: The U.S. Case .... 93

1. Introduction ............................................................................................ 93
2. Background ............................................................................................ 93
3. Foundations of budget preparation: macrofiscal .............................. 95
4. Strategic Allocation ............................................................................... 99
5. Policy and budgets .............................................................................. 103
6. Legislature ............................................................................................ 104
7. In-year Effect ........................................................................................ 105
8. Discussion and overall assessment ................................................... 105
Contents iii
PART II. PERFORMANCE MANAGEMENT
CHAPTER 9. An Overall Framework for Considering Performance

Management ............................................................................................ 109
1. Overview .............................................................................................. 109
2. Performance in the public sector ....................................................... 109
3. Improving performance through the budget process .................... 112
4. Results of performance reforms and implications for the design of
performance-enhancing reforms ....................................................... 116
References ................................................................................................. 119
CHAPTER 10. Introducing Performance Management .............................. 121

1. Performance management framework ............................................. 121
2. Current performance Management Practices in Korea .................. 130
3. Developing a performance management framework .................... 138
References ................................................................................................. 145
CHAPTER 11. Performance Management in the Health Sector ................ 147

1. Need for performance evaluation in the health field ..................... 147
2. Types of performance evaluation in the health field ...................... 148
3. Objectives of public health intervention .......................................... 149
4. Special characteristics of performance evaluation in the health field .... 151
7. Examining indicators for healthcare evaluation: experience of U.S. ... 158
8. Summary ............................................................................................... 167
References ................................................................................................. 168
CHAPTER 12. Performance Management in Education: A Call for

Paradigm Reform in Korea ................................................................... 169
1. Introduction .......................................................................................... 169
2. Performance in education ................................................................... 170
3. Performance-management in education .......................................... 172
4. A call for paradigm reform ................................................................ 181
References ................................................................................................. 186
CHAPTER 13. Establishing a Performance Management Strategy:

Governing for Results ........................................................................... 187
1. Introduction .......................................................................................... 187
2. Time-phasing of basic steps in implementing performance
management ......................................................................................... 190
3. Roles and responsibilities .................................................................. 197
iv Contents
4. Designing a performance management system .............................. 209

5. Implementing a performance management system ....................... 216
References ................................................................................................. 225
CHAPTER 14. Results Based Budgeting in Australia ......................................... 227

1. Outline ................................................................................................... 227
2. The fiscal framework ........................................................................... 227
3. The outcomes and outputs framework ............................................ 229
4. What have we learned? ....................................................................... 238
CHAPTER 15. Performance Management and Budgeting: a Case Study of

Sweden ..................................................................................................... 241
1. Country Context .................................................................................. 241
2. Origins of the reform and the historical context ............................. 242
3. Elements of performance management and budgeting in Sweden .... 244
4. Implementation .................................................................................... 251
5. Incentives .............................................................................................. 253
CHAPTER 16. Performance Budgeting in the United States .............................. 257

1. Outline ................................................................................................... 257
2. Context .................................................................................................. 257
3. GPRA ..................................................................................................... 257
4. GPRA Strategic Plan Requirements .................................................. 258
5. GPRA Performance Plans Requirements ......................................... 258
6. Key Implementing Guidance ............................................................. 258
7. Lessons and Observations .................................................................. 259
8. Future Directions ................................................................................. 259
9. Program Assessment Rating Tool ..................................................... 259
10. Coverage of the PART ....................................................................... 259
PART III. FISCAL TRANSPARENCY
CHAPTER 17. Coverage of Public Finance in Korea .......................................... 263

1. Introduction .......................................................................................... 263
2. Several concepts on the coverage of public finance ........................ 264
3. The coverage of public finance in Korea .......................................... 268
4. The coverage of public finance in the UK ........................................ 270
5. The coverage of public finance in Japan ........................................... 274
6. Conclusions: lessons from the two countries .................................. 275
References ................................................................................................. 277
Contents v
CHAPTER 18. Information Flow for Successful MTEF Operation and its
Implication for the Korea Government .............................................. 279
1. Introduction .......................................................................................... 279
2. MTEF, FMIS, and other public expenditure reforms ..................... 280
3. New information needs for MTEF operation .................................. 282
4. Inadequacy of current FMIS to support MTEF ............................... 286
5. Implementation strategy for improving FMIS and MTEF ............. 290
References ................................................................................................. 298
CHAPTER 19. Medium-Term Expenditure Frameworks, Performance, and

Transparency ........................................................................................... 299
1. Introduction .......................................................................................... 299
2. Medium-Term Expenditure Frameworks (MTEFs) ........................ 299
3. Performance budget/management ................................................... 299
4. Transparency ........................................................................................ 301
5. MTEF and performance ...................................................................... 302
6. MTEF and transparency ..................................................................... 303
7. Weaving it all together ........................................................................ 304
References ................................................................................................. 305
List of Tables
Table 4-1. The objectives and their matching policy tools of MTEF................ 45
Table 4-2. Comparison of single year budgeting and MTEF............................ 45
Table 4-3. Annual procedure of MTEF ................................................................ 48
Table 4-4. Time coverage of MTEF....................................................................... 50
Table 4-5. Fixed vs. flexible MTEF ....................................................................... 54
Table 7-1. Preparation of the budget 2005........................................................... 90
Table 9-1. Difference between performance and traditional budgeting....... 113
Table 10-1. Key Differences between evaluation and monitoring................. 122

Table 10-2. Estimation results ............................................................................. 124
Table 10-3. Systems of performance management .......................................... 131
Table 10-4. Timetable for the Performance Management System ................. 132
Table 10-5. Budget of the Ministry of Education ............................................. 134
Table 10-6. Components of Government Operations Assessment System .. 135
Table 10-7. Organizational assessment.............................................................. 136
Table 10-8. Criteria for assessment..................................................................... 137
Table 10-9. Documents required in GPRA........................................................ 140
Table 12-1. Performance test ............................................................................... 170

Table 12-2. Student engagement ........................................................................ 171
Table 12-3. Student participation in educational courses outside school..... 171
Table 12-4. Teacher evaluation ........................................................................... 172
Table 12-5. Areas of assessment, number of subjects and items, and
distribution of the office evaluation .............................................. 175
Table 12-6. Budget distributed according to the results of the office
evaluation.......................................................................................... 176
Table 12-7. Areas of assessment and contents of the school evaluation
conducted by the ministry of education and human resource
development ..................................................................................... 177
Table 12-8. Contents of the field evaluation conducted by the ministry of
education and human resource development to evaluate schools.... 178
viii List of Tables
Table 12-9. Areas of assessment and evaluation tasks for general secondary
schools of Seoul metropolitan office of education ...................... 179
Table 12-10. Accountability-Strengthening Program ...................................... 185
Table 13-1. Performance information definitions and sample indicators .... 189
Table 13-2. Possible composition of a national oversight steering committee ... 191
Table 13-3. The responsibilities of the national oversight offices .................. 199
Table 13-4. Decisions at the ministerial and agency levels ............................. 200
Table 13-5. Criteria for selecting issues for program evaluations ................. 215
Table 13-6. Suggested elements in a states strategic planning process........ 221
Table 17-1. The coverage of public finance in Korea ....................................... 270

Table 17-2. Key fiscal indicators in the UK ....................................................... 271
Table 17-3. TME by spending sectors ................................................................ 272
Table 17-4. Voted and non-voted expenditure in TME................................... 273
Table 17-5. The Coverage of Public Finance in the UK ................................... 274
Table 18-1. Analysis of budget composition..................................................... 289

List of Figures
Figure 1-1. Stage 1: Macroeconomic and public sector envelopes..................... 2
Figure 1-2. Stage 2: High-level policy .................................................................... 7
Figure 1-3. Stage 3: Linking sector policy ........................................................... 10
Figure 1-4. Stage 4: Reconciling ............................................................................ 12
Figure 1-5. Pulling it all together.......................................................................... 14
Figure 4-1. Business cycle and fiscal balance...................................................... 44

Figure 4-2. Cross-sectoral procedure for MTEF ................................................. 53
Figure 4-3. Cross-time procedures of MTEF....................................................... 54
Figure 6-1. Annual GDP growth 1980-2002 ........................................................ 71

Figure 6-2. Public sector financial balance, percent of GDP, 1980-2002.......... 72
Figure 6-3. Public sector expenditure, percent of GDP, 1980-2002.................. 82
Figure 6-4. The expenditure ceiling, percent of GDP, 1997-2006 ..................... 83
Figure 8-1. Functional Classification.................................................................. 100
Figure 10-1. BOD levels of four major rivers .................................................... 123

Figure 10-2. Accountability relationship ........................................................... 129
Figure 10-3. Performance structure.................................................................... 133
Figure 10-4. Restructuring of accounts in INS.................................................. 143
Figure 11-1. Chain of objectives in the health field.......................................... 150

Figure 11-2. Lead poisoning program ............................................................... 152
Figure 11-3. Smoking and low birth weight ..................................................... 153
Figure 13-1. Sample Performance Management System schedule for each

program/service 36 month schedule....................................... 192
Figure 14-1. Improvement in net debt ............................................................... 228

Figure 14-3. A model of the framework ............................................................ 230
Figure 14-4. Mean scores of Australian students in reading, mathematical and
scientific literacy compared with OECD average for 2000 (set to 500).... 237
Figure 14-5. Graduate starting salaries as a proportion of Average Weekly Earnings..... 237
Figure 14-6. Number of patents issued to Australian residents in Australia... 238
Figure 17-1. Stages of expenditure management ............................................. 266

Figure 17-2. Public expenditures in Korea ........................................................ 269
Figure 18-1. Stylized chart for Korean MTEF process..................................... 282

Welcoming Remarks
Choongsoo Kim
(President, Korea Development Institute)
Honorable Minister,
Distinguished participants from the World Bank and other organizations,
Ladies and gentlemen,
On behalf of the Korea Development Institute, it is my pleasure to welcome you all
to the Joint Conference of the World Bank and KDI on Reforming the Public
Expenditure Management System. I would like to give a special welcome to Minister
Byungil Kim of the Ministry of Planning and Budget who sponsored this Conference.
I want to start by thanking the organizers from the World Bank, Ministry of
Planning and Budget, and KDI for their dedicated efforts. I am pleased to see so many
leading experts from the international community who came here to share knowledge
and experience.
KDI was established in 1971 as the first government-funded research institute in
social science. At the time of its inception, KDI focused on policy issues related to the
Five-Year Economic Development Plans, which served as a basis for economic
management in those days. As Korea made a democratic transition in the 1980s, social
issues such as welfare and balanced regional growth came to the fore, and were
accorded a greater weight in KDIs research agenda. This change was consistent with
its mission as a policy-oriented research institution devoted to the development in
the broad sense of the term. In fact, KDI is distinguished from other government-
funded research institutions for its extensive coverage of research areas and its
comprehensive approach.
An important area of research throughout this period has been fiscal policy. Fiscal
policy addresses both macroeconomic and microeconomic issues. It is an important
determinant of the macroeconomic stability of a nation in the long and the short term; a
prudent management of government liabilities is critical for the long-term price
stability and economic growth, and a flexible management of budget balances is
essential to dampen the short-term cyclical fluctuations.
From a microeconomic perspective, fiscal policy is concerned about reallocating
national resources across different sectors. Government policies are implemented in
most part through expenditures on education, defense, infrastructure investment,
income security, health, environmental protection, and others. These sectoral policies
are collated against each other and prioritized in the annual budget process. In this
way, fiscal policy works as a disciplining mechanism to impose an overall resource
constraint on sectoral demands and maximize the efficiency of public expenditure.
The recent economic crisis that hit the Korean economy at the end of 1997 left an
unmistakable mark in every policy area, including the fiscal policy. The budget balance
deteriorated substantially, and government liabilities doubled in a matter of two years.
However, the rapid recovery in 1999 and 2000 contributed to the regained tranquility in
xii Welcoming Remarks
public finance, with the balance showing surpluses and the debt-to-GDP ratio
stabilizing at around 20 percent in the following years.
But challenges still remain. Revenue growth is expected to slow down as the
potential growth rate of the Korean economy declines. At the same time, pressures for
increased spending are building up to prepare the Korean economy for the intensifying
global competition, support the financial-sector restructuring, and finance the
expanding welfare programs. In addition, public demand for a transparent
management of taxmoney is increasing.
Faced with these challenges, we found that a thorough reexamination of the public
expenditure management system was called for. We need a system that enables us to
implement fiscal policy in a strategic way to fulfill all of its promises; contributing to
the long- and short-term macroeconomic stability and maximizing the allocative and
operational efficiency of public expenditure.
Three issues were selected for our study; (1) medium-term expenditure framework,
(2) performance management, and (3) fiscal transparency. Each of them represents a
significant departure from the past mode of expenditure management. They are
complementary to and interdependent on each other.
The Ministry of Planning and Budget has already made some progress in
introducing elements of the needed reforms in these areas. Our study builds on these
efforts and aims to provide directions for further improvement. It began early last year
and has been carried out in collaboration with the World Bank, with the financial
support from the Ministry of Planning and Budget.
This Conference was organized to present our major findings and gather a broad
range of inputs from many sources. It will provide a forum for debating policy options
for reforming the public expenditure management system. I hope at the end of the
Conference, we can have a better understanding on this matter. Any contribution
toward providing new thoughts and insights will prove to be both important and
meaningful, and will be incorporated into our final report to the Ministry of Planning
and Budget. The report will be published next month in Korean.
Now with my welcoming remarks complete, let me turn the floor over to Minister
Byungil Kim of the Ministry of Planning and Budget.
I wish you all an informative and productive conference.
Thank you.
Keynote Speech
Byung-il Kim
(Minister of Planning and Budget)
1. Introduction
Dr. Choong-Soo Kim, President of KDI,
Mr. Bill Dorotinsky, Lead Public Sector Specialist of the World Bank, and
distinguished guests from around the world!
It is my great honor to have this opportunity to be the keynote speaker at today's
joint conference of the World Bank and KDI, and talk about fiscal reform in Korea.
I would like to begin by extending my compliments to the staff of the World Bank
and KDI for their devoted research on Reforming the Public Expenditure Management
System, initiated by the Korean Government last year.
In addition, it is my pleasure to welcome the participants from various countries
and international organizations such as the World Bank and IMF, who have travelled a
long way to be here today.
Thank you for your participation.
2. Korea's efforts to enhance the public expenditure management system

2.1. Changes in Korea's fiscal management environment
One of the recent global trends is that the role of public finance as a policy tool is
becoming increasingly important, as the scope of government intervention in the
economy becomes smaller.
In addition, budget deficits and national debt can have direct effects on the
national economy under the open economic system, and are important factors in
deciding national credit ratings. Therefore, in reality, the influence of public finance on
national economy is significant.
Under such circumstances, Korea's fiscal management environment is changing
rapidly.
Internally, economic and social structures are changing and the demand of the
citizens is becoming more diverse.
Externally, the security situation in the Korean Peninsula is changing, and the
world economy is shifting into bloc economies such as FTAs, constantly giving rise to
new demands for public expenditure.
In the short-term, there are demands for building a Northeast Asian Economic
Hub, and R&D for the development of new growth-driving forces. Moreover, the need
for additional spending in the agricultural sector restructuring is rising, to meet
international standards including the opening of the agricultural market.
In the long-term, public expenditure in the areas of public pension, health
insurance, and welfare for the aged population is expected to increase substantially due
to rapid population ageing.
xiv Keynote Speech
On the other hand, the revenue situation is expected to become more difficult, as
the Korean economy gradually turns into a stable low-growth economy.
In response to such changes in the fiscal management environment, it is time to
make substantial changes in the public expenditure management system to enhance the
productivity of public finance.
2.2. The participatory government's efforts to reform the public expenditure

system and management practices
The Korean government is implementing 3 fiscal reform tasks since the inauguration of the
Participatory Government, to enhance the efficiency of public spending while maintaining fiscal
soundness.
Firstly, in contrast to previous medium-term fiscal plans that were formulated
irregularly, a National Fiscal Management Plan was formulated last year, to be revised
and updated every year, allowing the efficient allocation of national resources from a
mid to long-term perspective.
The Plan sets aggregate objectives of national fiscal management based on a 5-year
macroeconomic outlook and projected public expenditure requirements, and presents
sectoral resource allocation priorities and spending size by budgetary programs.
Furthermore, a Top-down budgeting system will be introduced this year.
Accordingly, expenditure ceilings will be set using the total budget size and sectoral
and ministerial resource allocation sizes presented in the National Fiscal Management
Plan.
After an agreement on sectoral and ministerial ceilings has been reached, each line
ministry will autonomously formulate their respective budgets within the expenditure
ceilings. The budget will be finalized after a final consultation with the Ministry of
Planning and Budget.
In response to such changes in the fiscal management system, a Performance
Management System is being introduced at the same time, to ensure accountability
against expanded autonomy of the line ministries.
Last year, performance goals and indicators have been developed for some of the
budgetary programs in 22 line ministries. Beginning this year, the system will be
gradually expanded to encompass more ministries and programs, with its full
introduction scheduled for the year 2008.
2.3. Revising the coverage of public finance to meet international standards
Turning to the issue of the coverage of public finance and fiscal transparency,
Korea's coverage of public finance has been rather limited compared to
international standards. This gave rise to the problem of making it difficult to perform
international comparisons in fiscal size and balance, expenditure by areas, and national
debt.
The Korean government has formulated a plan to revise the coverage of public
finance, and will begin implementing the plan this year. As a result, it will be possible
to assess and analyze budgetary activities in accordance to international standards.
In this process, detailed reviews of international standards and country cases will
be of great help.
I would like to take this opportunity to request the valuable advice and
information from the international experts here at the conference.
Keynote Speech xv
3. Conclusion
This international conference will offer an opportunity to discuss a variety of
policy recommendations and proposals. I trust that these will provide momentum for
reforming the public expenditure management system in Korea, and serve as references
for fiscal reform in other countries as well.
Distinguished guests from around the world!
Thank you again for your participation.
Although your stay in Seoul will only be a week long, I hope you will take time to
enjoy the beautiful atmosphere of spring in Korea.
May your stay in the Land of the Morning Calm leave you with pleasant and
unforgettable memories of the Korean culture.
Thank you.
Conference Program
March 15, 2004 (Monday)
Issue 1: Medium-Term Expenditure Framework
09:00 10:00 Registration and Networking
10:00 10:30 Opening Ceremony

Welcoming remarks: Choongsoo Kim (President, Korea Development Institute)
Keynote speech: Byungil Kim (Minister, Ministry of Planning and Budget,
Republic of Korea)
10:30 11:00 Break
11:00 12:00 Session 1-1: Conceptual Framework

Moderator: Hong Tack Chun (Vice President, Korea Development Institute)
Topic 1: Developing a Medium-Term Expenditure Framework
Presenter: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)
Discussants: Seok-kyun Hur (Associate Fellow, Korea Development Institute)
Saravuth Sok (Deputy Director, Ministry of Economy and Finance, Cambodia)
12:00 14:00 Lunch
14:00 15:30 Session 1-2: Introducing MTEF in Korea

Moderator: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)
Topic 2: Key Issues for Introducing MTEF and Top-Down Budgeting in Korea
Presenter: Soomin Park (Director, Ministry of Planning and Budget, Republic of Korea)
Dong Yeon Kim (Senior Public Sector Specialist, World Bank)
Discussants: Allan Gustafsson (President, Mapsec, Sweden)
Youngsun Koh (Fellow, Korea Development Institute)
15:30 16:00 Break
16:00 17:00 Topic 3: Successful Installation of MTEF to the Korean Fiscal System
Presenter: Seok-kyun Hur (Associate Fellow, Korea Development Institute)
Discussants: YoungJoo Lee (Deputy Director, Ministry of Finance and Economy,
Republic of Korea),
Dang Huu Phap (Deputy Director, Ministry of Finance, Vietnam)
19:00 20:30 Welcoming Dinner (invitees only)

Hosted by Yangkyoon Byeon (Vice Minister, Ministry of Planning and Budget, Republic of Korea)
xviii Conference Program
March 16, 2004 (Tuesday)
09:10 10:10 Session 1-3: International Experiences in MTEF

Moderator: Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Topic 4: MTEF in Sweden (1)
Presenter: Allan Gustafsson (President, Mapsec, Sweden)
Discussants: Hongsang Jung (Director, Ministry of Planning and Budget,
Republic of Korea)
Mulia Nasution (Head of State Financial Accounting Agency,
Ministry of Finance, Indonesia)
10:10 10:20 Break
10:20 11:20 Topic 5: MTEF in Sweden (2)

Presenter: Niklas Frank (Deputy Secretary, Swedish Parliament)
Discussants: Junghun Cho (Public Sector Specialist, World Bank)
Munkyu Bang (Director, Ministry of Planning and Budget, Republic of Korea)
11:20 11:30 Break
11:30 12:30 Topic 6: MTEF in Germany

Presenter: Astrid Lbke (Deputy Head of Division, Federal Ministry of Finance, Germany)
Discussants: Ki-Baeg Park (Research Fellow, Korea Institute of Public Finance)
Ratanak Hav (Deputy Director, Ministry of Economy and Finance, Cambodia)
12:30 14:00 Lunch
14:00 15:00 Topic 7: MTEF in US, UK and Other Countries

Discussants: Changho Lee (Director General, The Embassy of Korea in US)
Sanya Praseuth (Deputy Director, Ministry of Finance, Lao PDR)
15:00 15:30 Break
15:30 17:00 Panel Discussion: MTEF and Top-down Budgeting

Moderator: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)
Discussants: Eonseog Song (Director, Ministry of Planning and Budget, Republic of Korea)
Seok-kyun Hur (Associate Fellow, Korea Development Institute)
Astrid Lbke (Deputy Head of Division, Federal Ministry of Finance, Germany)
Niklas Frank (Deputy Secretary, Swedish Parliament)
Jambaa Jargalsaihan (Director-General, Ministry of Finance and Economy, Mongolia)
Conference Program xix
March 17, 2004 (Wednesday)
Issue 2: Performance Management
09:30 10:30 Session 2-1: Conceptual Framework

Moderator: Ilho Yoo (Professor, KDI School of Public Policy and Management)
Topic 1: Overall Design and Framework of Performance Management
Presenter: Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Discussants: Ho-seung Moon (Director, Board of Audit and Inspection,
Republic of Korea)
Younghee Lee (Staff, Presidential Committee on Government
Innovation and Decentralization, Republic of Korea)
10:30 11:00 Break
11:00 12:00 Session 2-2: Performance Management in Korea

Topic 2: Performance Management in Korea
Presenter: Youngsun Koh (Fellow, Korea Development Institute)
Discussants: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)
Jangro Lee (Senior Deputy Director, Ministry of Planning and
Budget, Republic of Korea)
12:00 14:00 Lunch
14:00 15:00 Moderator: Allan Gustafsson (President, Mapsec, Sweden)

Topic 3: Performance Management in the Health Sector
Presenter: Heesuk Yun (Associate Fellow, Korea Development Institute)
Discussants: Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Eui Kyung Lee (Senior Researcher, Korea Institute for Health and
Social Affairs)
15:00 15:30 Break
15:30 16:30 Topic 4: Performance Management in the Education Sector

Presenter: Ju Ho Lee (Professor, KDI School of Public Policy and Management)
Discussants: Cheonsik Woo (Fellow, Korea Development Institute)
Kioh Jeong (Professor, Korea National University of Education)
xx Conference Program
March 18, 2004 (Thursday)
09:10 10:10 Session 2-3: International Experiences in Performance Management

Moderator: Niklas Frank (Deputy Secretary, Swedish Parliament)
Topic 5: Strategy for the Introduction of Performance Management System
Presenter: Harry Hatry (Research Associate, The Urban Institute, US)/
Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Discussants: Byoung Sik Park (Professor, Dongguk University)
Ui-hwan Tcha (Director General, Office for Government Policy
Coordination, Republic of Korea)
10:10 10:20 Break
10:20 11:20 Topic 6: Performance Management in Australia

Presenter: Richard Murray (Executive Director, The Treasury, Australia)
Discussants: Choi, Joon Ook (Fellow, Korea Institute of Public Finance)
Jin Park (Professor, KDI School of Public Policy and Management)
11:20 11:30 Break
11:30 12:30 Topic 7: Performance Management in Sweden

Presenter: Allan Gustafsson (President, MAPSEC, Sweden)
Discussants: Park, No Wook (Fellow, Korea Development Institute),
Dang Huu Phap (Deputy Director, Ministry of Finance, Vietnam)
12:30 13:30 Lunch
13:30 15:00 Topic 8: Performance Management in US and UK

Presenter: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)/
Joel Turkewitz (Senior Public Sector Specialist, World Bank)
MuliaNasution(HeadofStateFinancialAccountingAgency,MinistryofFinance,Indonesia)
15:00 15:15 Break
15:15 17:00 Panel Discussion: Lessons from International Experiences on Performance Management
Moderator: SungIl Lim (Senior Fellow, Korea Research Institute for Local Administration)
Discussants: Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Ohbong Kwon (Director, Ministry of Planning and Budget, Republic of Korea)
Saravuth Sok (Deputy Director, Ministry of Finance and Economy, ambodia)
Sanya Praseuth (Deputy Director General, Ministry of Finance, Lao PDR)
19:00 21:00 Dinner (invitees only)

Hosted by Choongsoo Kim (President, Korea Development Institute)
Conference Program xxi
March 19, 2004 (Friday)
Issue 3: Fiscal Transparency
09:10 10:10 Session 3-1: Transparency in Public Expenditure Management

Moderator: Bill Dorotinsky (Lead Public Sector Specialist, World Bank)/
Topic 1: Coverage of the Government Sector in Korea
Presenter: Dong Suk Oak (Professor, Incheon University)/
Discussants: Bum-Ki Woo (Presidential Committee on Government Innovation
and Decentralization)
Dongho Hong (Director, Ministry of Planning and Budget, Republic
of Korea)
10:10 10:20 Break
10:20 11:20 Topic 2: Information Flow under the Medium-Term Expenditure Framework
Presenter: Junghun Cho (Public Sector Specialist, World Bank)
Discussants: Seok-kyun Hur (Associate Fellow, Korea Development Institute)
Geum-Nam Kim (Senior Deputy Director, Ministry of Finance and
Economy, Republic of Korea)
11:20 11:30 Break
11:30 12:30 Session 3-2: Integrated Approach to Public Expenditure Reform

Moderator: Joel Turkewitz (Senior Public Sector Specialist, World Bank)
Topic 3: Relationship between MTEF, Performance Management, and Fiscal
Transparency
Batsukh Batmunkh (Vice Director, Mongolia Development Institute)
12:30 13:00 Wrap Up and Closing Remarks

Bill Dorotinsky (Lead Public Sector Specialist, World Bank)
13:00 14:30 Lunch

LIST OF PARTICIPANTS
Batsukh Batmunkh
Vice Director
Mongolian Development Institute
Government Building-3
Baga Toiruu-44
Ulaanbaatar-20a, Mongolia
Phone: (976) 11-322311
Fax: (976) 11-324518
Email: [email protected]
Munkyu Bang
Director
Industrial Public Finance Division
Ministry of Planning and Budget
Republic of Korea
Phone: (82) 2-3496-5045
Fax: (82) 2-3480-7980
Yangkyoon Byeon
Vice Minister
Republic of Korea
Phone: (82) 2-3496-5100
Fax: (82) 2-3496-5104
Joon Ook Choi

Research Fellow
Research Division
Korea Institute of Public Finance
Phone: (82)-2-2186-2221
Junghun Cho
Consultant
Public Sector Group
The World Bank
Phone: (1) 202-458-0661
Fax: (1) 202-522-7132
xxiv List of Participants
Hong Tack Chun

Vice President
Korea Development Institute
Phone: (82) 2-958-4011
Fax: (82) 2-960-4570
Bill Dorotinsky
Lead Public Sector Specialist
Public Sector Group
The World Bank
Phone: (1) 202-473-7189
Fax: (1) 202-614-1065
Niklas Frank
Deputy Secretary
Committee on Finance
Swedish Parliament
100 12 Stockholm, Sweden
Phone: (46) 8-786-5935
Fax: (46) 8-786-6171
Allan Gustafsson
President
Mapsec
Kungsbron 1, D4, Stockholm
Sweden
Phone: (46) 708-24 24 34
Fax: (46) 8-506 362 01
Harry P Hatry
Research Associate
Office of Public Affairs
The Urban Institute
2100 M Street, NW
Washington, DC 20037
USA
Phone: (1) 202-261-5709
Ratanak Hav
Deputy Director
Department of Economic Integration and ASEAN
List of Participants xxv
Ministry of Economy and Finance

St. 92, Wat Phnom, Phnom Penh
Cambodia
Phone: (855) 12-555-929
Fax: (855) 23-430-745
Dongho Hong
Director
Medium-Term Fiscal Planning Division
Republic of Korea
Phone: (82) 2-3496-5028
Fax: (82) 2-3480-7625
Seok-kyun Hur
Associate Fellow
Division of Public Finance and Social Development
Phone: (82) 2-958-4239
Fax: (82) 2-958-4090
Jambaa Jargalsaihan
Director-General
Economic Policy and Planning Department
Ministry of Finance and Economy
210646 Ulaanbaatar
Mongolia
Phone: (976) 11-991-18345
Fax: (976) 11-32-02-07
Kioh Jeong
Professor
Korea National University of Education
Phone: (82) 43-230-3420
Fax: (82) 43-233-6366
Hongsang Jung
Director
Public Funds Management & Research Division
Republic of Korea
xxvi List of Participants
Phone: (82) 2-3480-7963

Fax: (82) 2-3480-7976
Byungil Kim
Minister
Republic of Korea
Phone: (82) 2-3496-5000
Fax: (82) 2-3496-5010
Choongsoo Kim
President
Phone: (82) 2-958-4001
Fax: (82) 2-960-2952
Dong Yeon Kim

Senior Public Sector Specialist
WBIPR
The World Bank
Phone: (1) 202-458-2633
Fax: (1) 202-676-9810
Geum-Nam Kim
Senior Deputy Director
Government Finance Information & Management Division
Treasury Bureau
Republic of Korea
Phone: (82) 2-503-9131
Fax: (82) 2-3679-6107
Youngsun Koh
Fellow
Division of Public Finance and Social Development
Phone: (82) 2-958-4096
Fax: (82) 2-958-4090
List of Participants xxvii
Ohbong Kwon
Director
Fiscal Analysis Division
Republic of Korea
Phone: (82) 2-3496-5047
Fax: (82) 2-3480-7653
Changho Lee
Director General, Ministry of Planning and Budget
The Embassy of Korea in the U.S.A.
Phone: (1) 202-939-5638
Email: [email protected] / [email protected]
Eui Kyung Lee

Senior Researcher
Department of Health Policy
Korea Institute for Health and Social Affairs (KIHASA)
Phone: (82) 2-384-3085
Fax: (82) 2-353-0344
Jang-ro Lee
Senior Deputy Director
Fiscal Analysis Division
Republic of Korea
Phone: (82) 2-3480-7657
Fax: (82) 2-3480-7653
Ju Ho Lee
Professor
KDI School of Public Policy and Management
Republic of Korea
Phone: (82) 2-3299-1016
Fax: (82) 2-968-5072
Younghee Lee
Fellow
Local Tax Division
Korea Research Institute for Local Administratioin
Staff
Presidential Committee on Government Innovation and Decentralization
xxviii List of Participants
Phone: (82) 2-3488-7331

Fax: (82) 2-3488-7364
YoungJoo Lee
Deputy Director
Treasury Division
Republic of Korea
Phone: (82) 2-2110-2560
Fax: (82) 2-503-9282
SungIl Lim
Senior Fellow
Division of Local Finance
Korea Research Institute for Local Administration
Phone: (82) 2-3488-7321
Fax: (82) 2-3488-7306
Astrid Lbke
Deputy Head of Division
Division A 4
Directorate-General of Budget
Federal Ministry of Finance
Wilhelmstrasse 97, D-10117 Berlin
Germany
Phone: (49) 1888-682-1341
Fax: (49) 1888-682-3690
Ho-seung Moon
Director
Corporate Policy and Research Division
The Board of Audit and Inspection
Republic of Korea
Phone: (82) 2-2011-2420
Fax: (82) 2-2011-2848
Richard Murray
Executive Director, Fiscal and Corporate
The Treasury, Australia
Langton Crescent, Parkes ACT 2600,
Australia
List of Participants xxix
Phone: (61) 2-6263-3745

Fax: (61) 2-6263-3360
Mulia P Nasution
Head of State Financial Accounting Agency
Ministry of Finance
Jl. Lapangan Banteng Timur 2-4
Jakarta Pusat 10710
Indonesia
Phone: (62) 21-345-0959
Fax: (62) 21-345-0959
Dongseok Oak
Professor
Incheon University
phone: (82) 32-770-8538
Byoung Sik Park

Professor
Department of Public Administration
Dongguk University
President
Korean Association for Policy Analysis and Evaluation
Phone: (82) 2-2277-4809
Fax: (82) 2-2277-6478
Jin Park
Professor
Republic of Korea
Phone: (82) 2-3299-1027
Fax: (82) 2-3299-1240
Ki-Baeg Park
Research Fellow
Research Division 1
Phone: (82) 2-2186-2227
Fax: (82) 2-2186-1065
xxx List of Participants
No Wook Park
Research Fellow
Center for Fiscal Analysis
Phone: (82) 2-2186-2267
Fax: (82) 2-2186-2320
Soomin Park
Director
National Resource Allocation Improvement Team
Republic of Korea
Phone: (82) 2-3496-5035
Fax: (82) 2-3480-7651
Dang Huu Phap

Deputy Director
State Budget Department
Ministry of Finance
8 Phan Huy Chu, Hanoi
Vietnam
Sanya Praseuth
Deputy Director General
Accounting Department
Ministry of Finance
Lanexang Avenue, P.O.Box 46
Vientiane, Lao PDR
Phone: (856) 21-252462
Fax: (856) 21-217451
Saravuth Sok
Deputy Director
Economic and Financial Policy Planning and Monitoring Unit
Ministry of Economy and Finance
St. 92, Wat Phnom, Phnom Penh
Cambodia]
Phone/Fax: (855) 23-430-745
Eonseog Song
Director
Construction & Transportation Budget Division
List of Participants xxxi
Republic of Korea
Phone: (82) 2-3496-5146
Fax: (82) 2-3480-7675
Michael L. O. Stevens
Consultant
Africa Region
The World Bank
Phone: (1) 202-473-7493
Ui-hwan Tcha
Director General for Policy Analysis and Evaluation
The Office for Policy Analysis and Evaluation
The Office for Government Policy Coordination
Republic of Korea
Phone: (82) 2-3703-3915
Fax: (82) 2-734-8795
Ha Duy Tung
Senior Officer
International Cooperation Department
Ministry of Finance
8 Phan Huy Chu, Hanoi
Vietnam
Phone: (84) 4-933-0106
Fax: (84) 4-933-2351
Joel A. Turkewitz
Senior Public Sector Specialist
WBIPR
The World Bank
Phone: (1) 202-458-2365
Cheonsik Woo
Fellow
Division of Knowledge Economy
Phone: (82) 2-958-4139
Fax: (82) 2-962-7810
xxxii List of Participants
Bum-ki Woo
Director
Fiscal and Tax Team
Presidential Committee on Government Innovation and Decentralization
Phone: (82) 2-3703-3217
Ilho Yoo
Professor
Republic of Korea
Phone: (82) 2-3299-1058
Heesuk Yun
Associate Fellow
Public Investment Management Center
Phone: (82) 2-958-4689
Fax: (82) 2-962-9143
Part I
Medium-Term Expenditure Framework

CHAPTER 1
Developing a Medium-Term Expenditure Framework
by
Bill Dorotinsky, The World Bank
1. Overview
Moving towards a Medium-Term Expenditure Framework (MTEF) involves

building upon the current budget and policy decision-making process, with an aim to
overcome some of the current incentives in the process and improve public
expenditure decisions and management. This paper will explore the stages of a MTEF
process, the role of some of the key institutions in the MTEF process, and references
international practices to illustrate points. The paper is not exhaustive, but tries to
provide more insight into a good budget formulation process, and some considerations
for countries to keep in mind as they seeks to strengthen budget formulation. Some
aspects of improved budget formulation have already been addressed in the main
paper, and this Annex seeks to pull these together into a broader context.
Generally, a MTEF tries to integrate policy with resource allocation decisions in a
multi-year context to better reflect the effects of decisions on public finances, and
generally lead to improved policy decisions. It also involves structuring the decision-
making process to encourage economy and efficiency, and focus spending and policy
on attaining the governments objectives. So a MTEF could be said to be about strategic
allocation of resources in line with government priorities and the opportunity costs of
decisions.
A very simplified outline of a MTEF process can be represented in three steps1:
(1) Estimating total available resources for the public sector over a multi-year
period, usually within a multi-year macroeconomic and fiscal policy
framework.
(2) Estimating the actual cost of current government policy and programs by
sector, again in a multi-year context.
(3) Reconciling the information in steps 1 and 2 to align policies with available
resources, and using this information as a basis for improved policy and
funding choices, and ultimately improved outcomes.
This simplified framework provides the general thrust of the MTEF concept, but is
necessary to break the process into greater detail to explore the dynamics of the
process. Actual decision-making processes and institutions will vary by country, but
for ease of exposition, an MTEF process can be divided into five general operational
stages:
1 Drawn from Muggeridge (1997).

2 Part I. Medium-Term Expenditure Framework
Stage 1: Macroeconomic and public sector envelopes.

Stage 2: High-level policy: aligning policies and objectives under resource constraints.
Stage 3: Linking policy, resources, and means by sector.
Stage 4: Reconciling resources with means.
Stage 5: Reconciling strategic policy and means.
In general terms, Stages 1 and 2 are about developing macroeconomic estimates for
the economy, and shaping public fiscal policy within the multi-year framework. Stage 3
is about sectoral (e.g. health, education, welfare, etc.) policy development. Stages 4 and
5 are about reconciling competing claims on public resources with available resources.
2. Stage 1: Macroeconomic and public sector envelopes

Figure 1-1. Stage 1: Macroeconomic and public sector envelopes
Macroeconomic Estimates
Revenue Estimates
Fiscal Policy
Expenditure Estimates
(current services)
Expenditure Estimates z Affordable/sustainable

(current law, Normatives) Fiscal
z Monetary and Fiscal
Policy
z Debt and Deficits
Aid flows
All in multi-year context
2.1. Macroeconomic forecasts
Stage 1 begins with the development of sound, conservative macroeconomic

forecasts for budget development. It is important to distinguish between political
macroeconomic estimates estimates of desired or hoped for growth and sound
estimates for budget development. The former are optimistic estimates prepared for
public consumption, to convey the governments goals, and to build optimism in the
general public over the economic direction of the government. This is serving primarily
a political purpose. For a variety of reasons, these are not good estimates to use in
developing the budget. Being optimistic, they will lead to excessive spending relative
to likely revenues, and higher deficits and debt. More importantly, they are not
estimates that will influence the public or private sector investors (foreign or domestic).
Investors have many sources of information, and generally governments own
estimates will be the least credible and least relied upon for investor decision-making.
Moreover, the simple fact that more economic researchers are using consumer
Chapter 1. Developing a Medium-Term Expenditure Framework 3
confidence surveys as leading indicators of economic direction suggests that it is

consumer sentiments that serve as early indications of trends rather than government
economic estimates that heavily influence citizen perceptions. 2
2.2. Institutional measures to support sound macro-estimates
Countries use a variety of techniques to assure sound economic estimates, some

during the budget development stage (ex ante to budget enactment) and some after
budget passage (ex post to budget passage) to assure alignment of the budget with best
estimates or emerging trends. Among ex ante good practices are: comparison with
private or other forecasts during forecasts development3; vetting of forecasts through
academic or non-government fora or blue-ribbon panels4; early release of forecasts to
Parliament and the public; using the median forecasts of several reputable non-
government organizations as the government forecasts 5 ; technical rules (publicly
announced) for making the forecasts more conservative6; economic updates prior to
budget enactment that allow modifications to the budget proposals to fit emerging
trends;7 inclusion in the budget of a table comparing the key economic variables of the
government forecast with several reputable private or public forecasts, with
explanations for significant variation for median forecasts; and having multiple
government forecasts that could be the basis of budget development (e.g. Central Bank
or legislative budget office, the latter being more applicable to presidential systems
with separation of powers where the legislature can enact a budget entirely different
from the executives).8,9
2 For a longer discourse on the adverse effects of overly optimistic forecasts, see Dorotinsky (1999).
3 Current practice in Poland, Canada, and U.S..
4 Also current practice in Poland.
5 For example, the United States Government does not have a macroeconomic forecasting office for budget
development. Rather it purchases six leading blue-chip and academic forecasts which form the basis of
forecasts. Economists from the Office of Management and Budget, Council of Economic Advisors, and
Department of the Treasury gather to evaluate the forecasts with respect to the most recent Government data
and expected trends, and collectively agree on the most likely evolution of key variables. The variables are then
fed into a mini-model to produce all those variables needed for producing the budget.
6 Canada, for example, had systematically over-estimated economic growth and public revenues during the
1980s, contributing to their annual deficits. To compensate, Canada averages several non-governmental
forecasts, then adds 50-100 basis points to interest rates to assure a conservative forecasts, and runs these
through a small model to produce the forecast variables for budget formulation.
7 Finland prepares economic forecasts four times each year: in February for the Government decisions on
aggregate spending limits; in May as a mid-year update; in August for the budget discussions; and a smaller
exercise in November to confirm the estimates used in budget development.
8 Many countries have independent Central Banks that produce their own forecasts for monetary policy,
but these are rarely used in budget formulation. In the U.S., the Federal Reserve and Executive branch forecasts
are informally shared at a technical level to try to assure some consistency in the basis of policy formulation by
both institutions. Also, the Congressional Budget Office also produces its own independent set of economic
estimates for Congress, which can serve as the basis of Congress budget development.
9 Coordination, at least informally, of Central Bank estimates and expected monetary policy with
Government economic estimates and fiscal policy, is important to assure a sound basis for both monetary and
fiscal policy. Doing so informally, at a technical level, can preserve the independence of the Central Bank and
still assure Government fiscal policy is fully informed and sound. Developing a budget expecting moderate
monetary policy can be easily undermined where the central bank adopts a tight monetary stance, driving up
Good practices include: reserves set-aside during budget formulation to account

for some variation in forecasts after budget passage10; mid-fiscal year economic updates
after budget passage, combined with required proposals to bring the budget into line
with emerging trends during the fiscal year; formal procedures for passing
supplemental budgets during the fiscal year, and use of these procedures 11 ; and
subsequent annual budgets which include a comparison of prior year budget forecasts
with actual economic variables to assess the quality of the estimates.
Countries process the economic estimates themselves in different ways: Ministry
of Finance staff share the estimates only with the Minister of Finance for information
purposes or for general approval to move to the next stage of budget formulation;
economic estimates are shared with Cabinet for general information and blessing; or
economic estimates are sent by the Government to Parliament (or general public) early
in the budget cycle for information and transparency. While distribution of the
economic forecasts varies, it is accepted practice that policy officials not modify the
economic estimates or alter the technical details, and certainly not for political
expediency. The estimates need to be the best, objective estimates possible by technical
staff in order to form a sound basis for setting economic policy and for budget
development.
2.3. Revenue envelope
Once the macroeconomic estimates are prepared, sound estimates of government

revenue are developed. Equally critical as sound macroeconomic estimates are
conservative revenue estimates based on collection experience. Overly optimistic
revenue estimates can also lead to excessive spending. Even if the budget will contain
proposals for tax increases or improvements in collection rates, it is valuable to
separately identify the base estimate for revenues without policy or administrative
changes and those revenues expected because of policy changes or planned
improvements in compliance and collection rates. The latter could be further divided
into revenues expected with a high degree of certainty (e.g. new revenue from a
1 percent increase in the sales or income tax, or foreign aid under already signed
agreements) and revenues with low certainty of materializing (e.g. promised foreign
aid, revenues from entirely new taxes or fees, and improved collection rates from
improved enforcement). By differentiating the different sources of revenue and their
likelihood of materializing, better judgment can be used in setting fiscal policy as well
as in ex post assessment of the quality of estimates used in budget formulation. Of
course, all revenues must be considered, including user fees, administrative fees,
ministry own revenues, etc., to provide a complete picture of government financing
levels and needs.
interest rates and lowering economic activity.

10 Canada also builds a reserve into the budget that can only be used for variations from forecast and
unexpected events. If unspent at year end, these funds automatically pay off outstanding debt. Sweden
similarly uses unspent reserves for debt reduction, but rather than conservatively adjusting the macroeconomic
estimates Sweden reduces its public expenditure target below its estimated affordable expenditure estimates to
build a margin of safety. These reforms provide a systematic bias towards conservative expenditures and a
prudent use of unspent funds for debt reduction.
11 Sweden has two formal supplemental budgets, one in the spring accompanying the spring fiscal policy
update and one in the fall accompanying the budget proposals for the following year. In the U.S., supplemental
requests can be transmitted at any time, but are discouraged unless large and of high importance.
External aid flows must also be included to provide the comprehensive revenue
picture. These deserve some special treatment in the estimates for several reasons:
counterpart funds must be identified and available from domestic resources; the aid
flows themselves may be highly unpredictable, and it is important to assure that
spending does not occur in advance of receipt of aid funds that may not materialize; in
selected cases, domestic expenditure must occur before the aid flows are released as
reimbursements, and these must be identified as well; the future cost for operating and
maintenance of infrastructure financed by foreign aid must be factored into the budget,
and this is more difficult if the budget does not even contain all aid revenues; ideally,
aid inflows would match government sectoral priorities rather than driving them; for
foreign loans, debt service costs must be factored into the multi-year plan.
Also important is identification of tax expenditures special exemptions from general
taxation for special interests, industries, regions, etc. These are revenues foregone, and
future estimates should assume current laws or policies remain in place, and model the
multi-year trends in tax expenditures for appropriate factors such as economic growth. Tax
expenditures should be explicitly identified for each type of tax revenue, allowing policy
officials to assess the benefit of this tool.
2.4. Expenditure estimates
With conservative economic estimates in hand, the Ministry of Finance can prepare
broad expenditure estimates based on current policy, and explore expenditure
sensitivity to changes in economic estimates. The objective of this exercise is to explore,
at a relatively aggregate level of spending, the trends in spending and gauge
affordability over several years. The multi-year aspect allows the Ministry of Finance
an early warning indicator if some programs are likely to grow beyond affordable
limits, as well as to explore general policy options (e.g. pay policy) and their impact on
spending trends (as well as explore policy changes that might liberate resources in
current and future years for new initiatives). It is important to explicitly note that this
covers ALL public spending trends consolidated central government in GFS
terminology. If only budgetary central government is included, the picture is only
partial, and does not allow exploration of all sources of spending, identification of
over-all trends that may require Government attention, and will lead to suboptimal
policy decisions.12
These initial estimates are often termed baseline estimates, and while different
countries use different bases for calculating them, the general concept is to forecast
future spending assuming continuation of the current spending levels or laws that is,
assuming no changes in policy or spending trends, what do the out-years look like?13,14
12 While more attention in projections and sensitivity analysis will be paid to consolidated central government,
especially extrabudgetary and social funds, trends in subnational government revenues, spending, and deficits
also need to be reviewed. Excess subnational spending can directly threaten the central budget, and monitoring
trends early for possible Government intervention is important for long-term fiscal health.
13 The U.S. develops baseline estimates on current law: for mandatory spending (spending that occurs
through statute, rather than annual budget bills), the forecast assumes no change to laws, and uses economic
estimates and demographic trends to extrapolate spending to the budget year plus nine additional years. For
discretionary spending (spending that occurs only through the annual budget), the most recent enacted budget
is extrapolated out nine years using basic economic assumptions supplied by the budget office, typically the
consumer price index for wage and non-wage expenditures.
14 Australia uses the concept of current policy, focusing more on the policy commitments of the current
A useful exercise would be to prepare two broad baseline estimates: one assuming
current funding levels of discretionary spending and current laws for mandatory spending as
a current policy baseline; and one assuming current normatives and other laws with implicit
commitments were fully funded (which may require spending ministry assistance, and
which may in fact also represent the current annual requests from spending agencies). The
exercise would be instructed to identify the policy gap or difference between current levels
of public spending and those required to fulfill policy commitments as contained in
normative laws (and gauge the realism of normative sector policies).
With conservative revenue estimates in hand, the Ministry of Finance can prepare a
fiscal policy paper for presentation to Cabinet. This document draws together the results of
the modeling and forecasting exercises into an understandable report of trends and policy
implications for senior government officials and the public. The paper explores
implications of monetary and fiscal policy for economic growth, as well as trends in fiscal
policy, deficits and debt projections under current trends, and broad options for changing
policies. It serves as a useful vehicle if clear in presentation and readily assessable to
non-economists for educating senior policy officials of basic economic issues, and
linkages between government policies and outcomes (economic, fiscal, etc.). The paper
should include recommendations of MoF on a path for sustainable or improved economic
growth, including aggregate revenue and expenditure paths (including deficits and debt
policy).
The fiscal policy paper would be presented to Cabinet for discussion, with the
objective of obtaining some Cabinet decision on over-all government revenue,
expenditure, and deficit policy or framework. The focus of the decisions should be on
the policies for revenues and expenditures, the path over the budget and subsequent
years, to attain government objectives. The discussion should generally not turn on the
economic forecasts themselves, especially if these represent best estimates.
Countries follow different practices in releasing the fiscal policy and expenditure
estimates to the Parliament and public. Finland releases this information to the public
and parliament for informational purposes, while Sweden actually submits for passage
by parliament the fiscal policy, expenditure envelope, and allocations to sectors. In
Sweden, this was done to bind Parliament to the expenditure totals and sector ceilings,
and prevent unproductive amendments that added to expenditure rather than
encouraged efficient choices within sectors. In other countries, such as the U.S., the
fiscal policy, expenditure targets, and sector funding levels are not released until the
formal budget is proposed.
government over several years. Termed forward estimates, Australias multi-year forecast separately identifies
revenue policies, including tax expenditures, as well as general spending. For a good example of a comprehensive
multi-year document, see Australias 2002-2003 Budget Paper No. 1 at http://www.budget.gov.au/2002-03/bp1/html/.
3. Stage 2: High-level policy: aligning policies and objectives under resource

constraints
After the fiscal policy envelope is developed setting the aggregate public revenues
and the expenditure ceiling, decisions need to be made on the allocation of this ceiling
across sectors (e.g. health, education) and ministry (which may be the same as sectoral
or not, depending on the budget classification system). These indicative ceilings may
have been provided in the fiscal policy paper by the MoF, or may be addressed as a
separate exercise. Sweden combines discussions on the fiscal policy, expenditure
envelope, and sector ceilings into one two-day Cabinet retreat, after which the
approved levels are sent to Parliament for ratification. Each ministry is then bound to
meet the ceiling set by Cabinet and Parliament.
Some sectoral allocations may have developed in the course of the fiscal policy
discussions,15 where key sectoral spending trends may have been identified as threats
to policy or fiscal stability. The remainder of funds need to be allocated, and the
allocations would reflect at least general policy directions of Government (e.g.
additional education spending), and specification or modification of Government
objectives, in broad terms, for each sector.
A good starting point for setting the indicative ceilings for sectors and ministries is
the second year of the prior budgets multi-year forecast, adjusted for known
modifications in Government priorities and policies, improved spending estimates,
major revisions in revenues and expenditure targets, etc. Some countries also reduce
sector ceiling allocations and create a central policy reserve that is used to fund new
initiatives after spending requests are received in Stage 4.
Figure 1-2. Stage 2: High-level policy
Fiscal
Fiscal Envelope
Envelope
SectoralResource
Sectoral ResourceCeilings
Ceilings
Setting Ministry
Ministry Resource
Resource Ceilings
Ceilings
Strategic Policy
Priorities Under
Resource Limits
(downsizing, or
expansions)
15 In Finland and The Netherlands, the allocation stage is easier during budget preparation as many of the
policy priorities have been set in government coalition agreements. Absent coalition governments, it is not clear
these pre-specification of policy priorities would continue.
Within sectors, it is important to distinguish between mandatory and discretionary

spending in making allocations. Unless major policy changes are expected for
extrabudgtary funds and mandatory programs, the best estimate of spending for these
programs would receive its full allocation. Discretionary programs would receive
residual amounts for the sector perhaps lower than desired. This stage can generate
much debate on growth of entitlement spending and its crowding-out of other
government spending, and helps clarify the tension in these programs. Box 1-1
describes the Brazilian approach to making these decisions.
Also important for sectoral allocation is a review of revenue earmarking and the
constraints it imposes on reallocation of spending across sectors or programs.
Once approved by Cabinet, these indicative ceilings for each sector/ministry are
transmitted to Ministries, usually as part of the annual budget call circular that initiates
the budget process and provides instructions for ministry budget submissions.
Whether the indicative ceilings are actually cabinet-approved can differ by
country. In the U.S., indicative ceilings are not approved by the President, and thus do
not carry much weight with the spending ministries. Moreover, in the U.S. system,
even if the ceilings were presidentially-approved, it is not very likely there would be
adverse consequences to a cabinet secretary that submitted spending requests in excess
of the ceiling. It certainly is not a serious enough transgressions to warrant dismissal
from cabinet. The US Budget Office could try to impose some sanctions or negative
incentives for such behavior, but it currently does not, and such efforts would be
difficult to sustain in the US context. In Parliamentary systems, Cabinet sanction can be
important for compliance, as it represents a collective Government decision, which
each Cabinet Member is more bound to follow.16 Regardless, adherence to indicative
ceilings by ministries in developing budget proposals makes budget preparation far
easier in following stages.
16 For a more thorough treatment of cabinet dynamics, see Manning, Barma, Blondel, Pilichowski, and
Wright (1999).
Box 1-1. Brazilian Budget Process
The broad functional steps in the Brazilian budget formulation process are:
Define the macro-economic projections, estimate public revenues, and plan/set

expenditure and debt levels.
Develop indicative sectoral limits based on analysis of current operating costs, programs,
and activities. The ceilings include several subsets of estimates:
a) Personnel costs A detailed personnel registry provides information on wages, new

hires, and early retirements. This is a known number.
b) Debt Part of debt service for loans related to each Ministries programs are allocated
to each Ministry and part to the Ministry of Finance. The budget office knows all
contracts, and knows this number.
c) Activities, continuing The cost of continuing the current programs is estimated by
the budget office, and refined by the line Ministry. This includes administrative costs in
support of personnel (utilities, etc.), plus mandatory activities (health, pensions, etc.)
d) Activities, new or expansions Line Ministries develop these proposals and submit to
the Ministry of Planning for review.
e) Projects, continuing Line Ministries estimate the cost of maintaining the economic
viability or continuing on-going projects.
f) Projects, new or expansions Line Ministries estimate the cost of expansion of projects
or of new projects.
Within the available resource estimates, subsets a, b, and c are treated as priorities in the
annual budget process. Subsets d, e, and f are treated as residual: if resources are available
to allocate, they first go to new or expanded activities, then to continuation of projects, and
finally to new or expanded projects.
Note that items d through f are only the marginal cost of the activity or project. The
personnel costs and associated overhead costs (running costs) are separately funded in a
and c.
This process assures the basic obligations of the state debt service, mandatory social
benefits, personnel (which are fixed in Brazil) and fixed administrative costs, as well as
transfer payments, receive funding first.
4. Stage 3: Linking policy, resources, and means by sector

4.1. Preparing budget proposal
Having received sector or ministry indicative ceilings, each line Ministry must
develop its budget proposal. The expectation is that each Ministry would take the
resource ceiling as given, and allocate spending among its activities and programs to
attain its objectives.
This is particularly important to assure harmonization of normative laws and
sectoral and ministry strategic plans and attendant objectives, organization and
staffing, with the resources available. These decisions include both capital and current
spending, encouraging ministries to think strategically about their programs, to

evaluate what are the necessary labor and capital inputs to attain an objective or
implement a program.
Figure 1-3. Stage 3: Linking sector policy
Sector Policies and Objectives Reconcile policy, laws,

resource limits
Program priorities
relative to objectives
Sector programs Assess modalities for
objective
Evaluate Production
Function
Under ceiling
Sector Budget Request
Above ceiling
Again, multi-year
This is important for maximizing the effectiveness and efficiency of spending. An

example may illustrate the point. Assume a health ministry wants to develop a health
policy for the country, and it convenes expert groups as well as various groups
involved in the sector, such as the Chamber of Doctors, hospital administrators, nurses
organizations, etc. Collectively these groups will use their best professional judgment
in devising an optimal health policy. This policy may include specific targets of
number of beds per capita, physicians per hospital, etc. It may also specify vehicles for
advancing the policy, such as the nature and type of health benefits included in the
national health insurance plan, or assume certain modes of service delivery (e.g.
expansion in government-owned clinics). Now, when it comes to financing the policy,
it is commonly the case that the entire package is unaffordable. If the policy is retained
as formulated, the best that can be hoped for is, for arguments sake, 60 percent of the
total cost of the policy. The result is that the policy cannot be implemented, objectives
cannot be achieved, and even the reduced funding level may be too low for individual
elements of the policy to be effective (e.g. funding for vaccines may be too low to
assure adequate coverage of the population).
In contrast, if the over-all funding level were known in advance, framing the
policy discussion, entirely different objectives, programs, or means of production may
be explored. Instead of government-owned clinics, perhaps the government would
contract with private clinics to provide services at lower cost. Or, vaccination programs
would be considered essential and fully funded. The package of health benefits
covered under the health insurance plan might be adjusted. The point is that both
objectives and means of achieving the objectives will be different under different
resource constraints.
A second example, more narrow, also illustrates the point. During the 1980s, the
U.S. Consumer Product Safety Commission, an agency charged with regulating
consumer products to assure their safety, trained staff in offices across the U.S. by
sending specialists from Washington, D.C. to provide the training. As the budget
became tighter, the agency changed its means of delivering training. Instead of sending
trainers from Washington and incurring travel and lodging expenses, the agency made
training videos in Washington, and mailed these to its offices. Evidence suggested this
means of training was as effective as the live trainers. However, the agency used this as
an example of the terrible effects of the budget restrictions. The Budget Office used this
same case to illustrate the point that such evaluation of activities and improvements in
efficiency are exactly what an agency should be doing. Different levels of funding
influence choice of means of production.
Ministries may have begun to develop their budgets prior to receiving the call
circular, and such early budget development can assure that ministry budgets are
carefully developed. Ministries may have undertaken their own strategic planning
exercise, or undertaken a special review of the effectiveness or continued usefulness of
policies, programs and projects. These reviews and plans should be undertaken, either
as part of the annual budget cycle, or prior to it and then fed into the resource
allocation decisions in the budget process.
4.2. Resource allocation flexibility
Different countries provide different degrees of freedom to ministries to allocate

spending among alternatives. In the Australia and New Zealand New Public
Management model, ministries are assumed to allocate resources optimally among
programs, and are given great discretion in their choices. Similarly, Sweden provides
great latitude for allocating resources to ministries. In these systems, the Ministry must
justify and defend its choices, and bears responsibility for the consequences of their
choices. This transparency and accountability provides some incentive for sound
allocation choices.
In other systems, such as Finland, fewer degrees of freedom are offered. The
Ministry of Finance not only has indicative ministry or sector ceilings, but also
indicative spending targets for programs within the sector ceiling. Spending ministries
that want to reallocate across programs face difficulty doing so because of the Ministry
of Finance targets.17
In the United States, the incentives work differently. Spending Ministries budget
requests form the Presidents Budget, and the President is responsible for the policies
contained therein. Allocations that would cause interest groups to strongly object
would fall upon the President to advocate or defend, rather than the minister
responsible. The dynamics promote gamesmanship, where spending ministries
propose budget reductions in areas they know would never be approved by Congress,
and reallocate those resources to ministry priorities, which may not be the Presidents
priorities. As a result, the US Budget Office pays very close attention to the proposals
of ministries. While in principle the ministry can propose any change, in practice, the
proposals are carefully scrutinized for practical effect, incidence, and likelihood of
being enacted, as well as technical justification.
17 OECD (2002), p. 6, paragraph 18.

An important issue is the extent to which the spending ministries have the policy
and program analysis skills to undertake the reviews of activities and laws, assess their
effectiveness and cost, and develop options for the responsible Minister. These are
skills that have heretofore not been in great demand, and may require training to
further develop.
At the end of the stage, the ministries will submit their budget requests to the
Ministry of Finance. The format of the requests will vary by country, but something
more than tables of numbers with a cover memorandum is appropriate. Text
explaining the policies contained in the proposals, changes in policy from prior years,
significant reallocation decisions, new objectives, all should be included. As countries
move towards performance budgeting, the budget requests should include the output
or outcome targets associated with the resource levels and allocations contained in the
request. Moreover, simple identification within the request of on-going activities and
programs from new programs assists in evaluating the proposals.
5. Stage 4: Reconciling resources with means

After receiving the requests, the requests must be reconciled with the resource
ceilings, competing demands and policies reconciled, and final policy and funding
decisions reached. This stage can be contentious or productive depending on how the
process is structured (extent to which ceilings were respected), whether budget
retrenchment or expansion is the trend, and Ministry of Finance relations with
spending ministries.
Figure 1-4. Stage 4: Reconciling
Vetting
Sector Ceilings assumptions
policies,
Technical estimate
Assessment Challengin
Sector Requests and modalities
Reconciliation productio
functio
Evaluate
performanc
Revised requests and/or decision papers
Assuming the sector ceilings and Sector requests match, the Ministry of Finances
role becomes one of providing critical analysis of the agency requests, and conducting
a process that leads the Ministries feel they had a fair hearing for their requests.
Ministries should be questioned in detail about their requests, especially the
assumptions and data underlying their requests. It is also useful to conduct budget
hearings, where the Ministries have an opportunity to meet with government officials
to make the best possible case for their budget request. These hearings can be time
consuming, but are important to give the Ministries a further stake in the budget
process. The less Ministries feel they have had a fair chance to make their case, the less
likely they are to accept the results of the process or to support the approved budget
during budget execution.18
Generally, the Ministry of Finance needs to vet the spending requests, probing
whether recurrent spending is needed and ministries have adequately reviewed
existing programs, and whether new programs are well-developed. How the MoF
interacts with spending ministries can influence outcomes. The MoF should seek to
engage in a dialogue with ministries to assess whether they have in fact considered the
questions of mix of inputs, whether there are benchmarks for assessing staffing
patterns, whether there are alternate means or modes of achieving the objectives, and
should use these issues to develop a budget. It is the dialogue that is important the
questions and answers. These issues must regularly be raised, or they will go
unaddressed: thus, current staffing and production patterns will remain fixed, and only
the increment for new programs or activities will be considered. For example, the MoF
may provide additional funds to spending ministries for official pay raises but may
provide no funds for promotions. Assuming there is a salary cap per ministry, this
puts a slight but palpable pressure on spending ministries to economize on salaries and
the wage bill, examining where savings can be achieved within the wage bill, to
provide some level of promotions. It also forces economical decisions on the number of
promotions, which may differ from the normal practice of automatic promotions
based on seniority and tenure (perhaps substituting performance evaluations and merit
as the basis of promotions).
Ministry of Finances performance assessment should include reviews of capital
spending and progress in capital projects, as well as assessments of whether external
and internal audit findings have been addressed by each ministry. The assessment can
cover many aspects of performance, and the results should be fed into the MoF
recommendations on funding. In the U.S., policy papers with funding options are
prepared for each program by the budget office, and these papers are used by senior
policy officials in the administration to make budget decisions. These papers integrate
findings on agency performance along many aspects in arriving at a funding decision.
If the sector requests and ceilings are not the same, differentiation between
recurrent and new spending as suggested above allows the Ministry of Finance to
quickly assess if recurrent spending is over target, and respond appropriately to the
source of the problem.
In some countries, review of budget requests, especially recurrent spending, is
given higher priority and more formal attention. The U.K.s former Public Expenditure
Surveys served as an annual functional review of ministry programs. In the early
1990s to facilitate its fiscal retrenchment, Canada instituted program reviews under the
auspices of a cabinet committee to review spending proposals in detail.
In addition to Ministry of Finances vetting of requests, this stage serves as a forum
for the various new funding proposals to compete for resources. If the Government sets
18 The Swedish Ministry of Finance does second-guess spending ministry budget submissions, especially
on social transfer programs, mainly through seeking additional technical details on the estimates. The MoF does
not directly challenge the requests. The Ministry budget requests are used to develop the Ministry of Finances
recommendation to cabinet, but does not directly respond to each Ministry on their request. The MoF is
prepared to detail its views at the two-day cabinet retreat where MoF allocations are below Ministry requests.
aside some resources during the sector ceiling stage for new funding proposals, these
funds would be used to fund the best of the new spending initiatives. (If no reserve
was established, and instead government allocated all available funding to the sectors,
giving some priority to some sectors over others, then competition between proposals
may have occurred within the sector.)
As noted earlier, this stage may proceed with varying degrees of cooperation or
conflict between MoF and spending ministries. In Australia, the process is a
professional dialogue where both parties to the discussion bring their unique skills and
perspective to the table to improve the projects and proposals. Ministries themselves
revise the requests based on discussions with MoF, and may reconsider the means
being proposed to attain an objective.
At the end of the day, the MoF must make some recommendations, and forward
these on to Cabinet for final approval.
6. Stage 5: Reconciling strategic policy and means

Figure 1-5. Pulling it all together
Policy official
Policy dialogue
assessment Modify baselines
and
reconciliation Finalize decisions
Revised
Requests, Budget Proposal
Unresolved
Issues
This final stage will take markedly different forms depending on the structure of
the budget process. In the U.S., after decisions have been taken, results are passed back
to each Ministry, which has an opportunity to appeal the results to the budget office. If
issues are not resolved at that level, they may appeal upward, eventually to the
President. Not many issues reach this stage. After appeals are resolved, then ministries
must revise their budget request to be in alignment with the Presidents policy. This
process of appeals is intended to assure some due process to Ministries in funding
decisions, but makes for a contentious process.
In Australia, MoF discussions with line ministries are less contentious, and more in
the nature of a management consultant (MoF) advising a client (line ministry). Line
Ministries take on board the MoF suggestions they agree with, and revise their budget
requests to come into alignment with MoF.
Line ministries of course are busy revising their requests, preparing their own
justification material, and even drafting legislation that may be required to support the
policies contained in the budget.
A primary role for the MoF in this stage is resolving any loose-ends in the decision
process, finalizing or refining numbers and policies to assure consistency. MoF efforts
move into production of documents and tables and graphs, getting everything ready
for final review and approval by Government and transmission to Parliament.
A further MoF role in many countries is preparation of communications strategy
and material to clearly explain decisions, elaborate the rationale for choices taken, and
explaining the directions for budget policy. An effective communication strategy, well
executed, is a critical factor in the success of the budget proposals and sustainability of
reforms. Careful attention to preparing policy officials to answer questions and defend
the proposals, as well as to the positive messages that need to be communicated to the
public, would increase the likely success and acceptance of any budget proposal.
References
Dorotinsky, Bill, It Flows Downhill, US Treasury, 1999.

Manning, Barma, Blondel, Pilichowski, and Wright, Strategic Decision-making in Cabinet
Government: Institutional Underpinnings and Obstacles, Sector Study Series, The World
Bank: Washington, D.C., October 1999.
Muggeridge, Elizabeth, Mozambique: Assistance with the Development of a Medium-Term
Expenditure Framework, Draft Report, June 1997.
OECD, Budgeting in Finland, 2002.
CHAPTER 2
MTEF Implementation in Korea
by
Soomin Park, Ministry of Planning and Budget
1. The Starting Point of Reform
Why did MPB look for a new resource allocation system?
However Earlier the better

Fiscal Crisis No Potential issues to be addressed
( populating ageing, unification costs )
Efficiency
Stiff portfolio in macro & micro levels
of Public Yes - economic sector: over 20% for 30 year

Expenditure - large spending but unsatisfactory
improvement in some sectors
Fiscal 22 special accounts and 59 public funds

Yes
Transparency highly fragmented and compartmentalized
(OECD)
Weaknesses in the present fiscal system
Bottom-up Too much workload on MPB from reviewing

individual programs ( 6,000 or more )
Oriented
Budgeting Little time for macro analysis
Insufficient role of line ministries, field expertise is

Central Fiscal not absorbed into the budgeting process
Authority-led
Asymmetry of information on cost between MPB
and line ministries
Lacks long-term perspectives,

Incremental budgeting
Yearly Budget
Little consideration on business cycle
2. The Vision of MPB
Where do we want to go?
Top-down Budgeting
Bottom-up Oriented
( Strategic Resource Allocation based on Macro
Budgeting
Yearly Budget Medium-Term Expenditure Framework
Central Fiscal Cooperation and Role Sharing between MPB and
Authority-led Line Ministries

Chapter 2. MTEF Implementation in Korea 19
The Korean Model (still evolving ...)

( Mar~May ) (Jun~Aug)
Line Ministries MPB
As is
Budget Requests
Y+1
( by programs )
To be (Jan - Apr) (Apr) (Apr - Jun) (Jun - Aug)
MPB Line Ministries MPB

Cabinet
Meeting
Total Ceiling Budget

Medium- Consultation Y+1
Sectoral Formulation
Term Plan and Review
Ceilings Within Ceilings
Y+5
Stage 1: Establishing Medium-Term Plan (National Fiscal Management Plan)
Supporting Strategic National Management
Analyzes 5-year Fiscal Balance & Total Expenditure Levels

- Gives way to Medium-term Budget Balance
Sets National Policy Goals & Priorities for resource allocation
Proposes Sectoral and Ministerial Ceilings & Major National Projects

Stage 2: Cabinet Meeting
Melting pot of Policies Making and Resource Allocation
In-depth Discussions on Sectoral & Ministerial Ceilings
- 5-year fiscal targeting

- Resource allocation by national priorities
- Review Major Policies upon Future Resource
Allocation Projection
Based on MPB Proposal
Stage 3: Line Ministries Budget Formulation
Change of Role and Mindset
Autonomous Resource Allocation within Ceilings
- Own priorities by sectors

- Using field expertise in budgeting
Yearly 5 years
MPBs Stand Alone Decision-making
All the Line Ministries Decision-making
- Provides Budget Formulation Manual

Chapter 2. MTEF Implementation in Korea 21
Stage 4: MPBs Consultation & Review
Player Coordinator
Reviews Spending Plans of Line Ministries
- Decisions from the Cabinet Meeting

- Conformity to the Formulation Manual
Reflects Changes after Cabinet Meeting
- Macro Variables, unexpected issues
3. What Lies Ahead
What are we expecting?

(The Starting Point)
F is c a l z Long-term Perspective
S o und ne s s z Symphony of Sustainability & Economic Stabilizing
z Macro Analysis Strategic Resource Allocation

Efficiency of z Role Sharing Self Expenditure Restructuring
P u b li c
E x p e n d it u r e z Cost Information Sharing Extinction of Excessive Budget
Request ( 20190% Increase )
F is c a l z A new Consolidated Structure over special accounts and public

T r a ns p a r e n c y funds
The road ahead
1. Not enough long-term fiscal forecasting background
2. Early stage of Performance Management System
3. Potential difficulties in gaining consensus of the National Assembly
Our roadmap
2003 2004 2008
Planning Introduction Adjustment Full

& Design establishment
Gradual introduction
Maintaining strong fiscal discipline
Surpassing global standards
CHAPTER 3
Key Issues for Introducing MTEF and Top-down Budgeting in Korea
by
Dong Yeon Kim, The World Bank
1. Introduction
Budget reforms usually reflect specific political, social, and economic problems
and issues. The rationale to undertake fiscal reform is invariably one of tackling
problem areas and addressing identified weaknesses. Even traditional line-item
budgeting, for example, was itself an evolutionary reform in response to serious
concern about the lack of adequate spending control and substantial corruption. So too
were performance budgeting, program budgeting, and zero-based budgeting.
Medium-term expenditure framework (MTEF), along with top-down budgeting,
has drawn particular attention from many OECD and developing countries over the
last decade or so.19 MTEF, by integrating long-term perspectives into budgeting, is a
practical tool to control fiscal risk and facilitates effective allocation of resources
through strengthened strategic decision-making in resource allocation. For example,
Sweden adopted MTEF to contain and reduce fiscal deficit. Since the introduction of
MTEF, fiscal balance has improved from a deficit of 10.8% of GDP in 1994 to a surplus
of 4.8% in 2001. 20 On the other hand, U.K. intended to increase public investment,
which continued to decrease. Through MTEF, public investment has increased from 20
UK billion pounds in 1997 to 31 billion pounds in 2003.21 Recognizing its value, the
World Bank has advocated MTEF as part of fiscal reform to improve fiscal discipline,
strengthen effective resource allocation, and improve policy efficiency and
effectiveness. To this end, the World Bank has promoted operations to introduce MTEF
into client countries.22
Korea has initiated a series of fiscal reforms since the East Asian financial crisis
that occurred in late 1997. The new Administration, since its inauguration in 2003, has
been striving to further plans for reform, central to which is MTEF along with top-
down budgeting. The primary objective of MTEF is to integrate national policy priority
into budgeting and set up a new strategic decision-making mechanism. The reforms to
introduce these new initiatives are due to be implemented in 2004 with the aim of full
application for the 2005 budget. This change will be a turning point in Korean public
expenditure management (PEM), because it is not limited to bringing about changes to
the budgeting process, but seeks to set up a strategic decision-making mechanism
19 More than 20 OECD countries surveyed prepare multi-year expenditure estimates and 7 OECD countries surveyed
prepare a medium-term fiscal framework(http://www.oecd.org/document/61/0,2340,en_2649_ 34119_2494461_1_1_1_1,00.html)..

20 Swedish Ministry of Finance (2003), p. 7.
21 The number is expressed in nominal term. See Liz Passman (2003), p.4.
22 The World Bank has provided assistance, financial and/or technical, to at least 45 client countries as of
2003 to implement MTEF to help strengthen countries public expenditure management system.
which, through a top-down and more discretionary approach, will alter managerial
behavior and organizational culture in the government.
This paper deals with Koreas MTEF and top-down budgeting reform plan by
analyzing key issues which may affect successful implementation of reform. More
specifically, the paper identifies why MTEF is considered necessary and what new
budget process it will involve. Then, this paper addresses key challenges which Korea
may confront and provides suggestions for implementation strategy to mitigate risks
for successful reform. While this paper deals mostly with Koreas PEM system and
addresses issues pertaining to Korea, it will be useful for other countries to draw some
policy implications and lessons.
2. Background and features of MTEF initiatives in Korea

More than most countries, public finance in Korea played a critical role in
propelling economic growth during its development era. However, this arrangement
also created numerous problems such as input and growth-oriented budgeting,
restrictions of line ministries discretion, and lack of transparency in the fiscal system.23
These issues came to the fore at the time of the Asian financial crisis and prompted the
Korean government to introduce many fiscal reform measures.24 The results of these
measures, however, have been mixed. Aware of many unresolved problems, the new
Administration has initiated PEM reforms focusing on MTEF with top-down approach.
This chapter elaborates on the current problems which MTEF may address,
characteristics of the new initiatives, and the budgeting process under the new system.
2.1. Why MTEF?
2.1.1. Short-term perspective with bottom-up approach

Koreas current budget process operates on a short-term perspective, i.e. yearly
based, and is highly tilted toward bottom-up approach. Under this system, revenues
are calculated for only the upcoming year and expenditure prepared based on
anticipated revenues for that year only. Although a medium-term fiscal plan was
developed in 1999 to handle these short-comings, this plan has not been integrated into
the annual budgetary process and has failed to significantly impact budget decision-
making. The most pronounced problem in the current system thus is the lack of
medium/long-term perspective and weak linkage between national policy priorities
and budget.
Bottom-up approach tends to lead to gradual yet consistent increases across almost
all budget items. Due to this incrementalism, the structure of public expenditure has
barely changed over the past decades. For example, expenditure on economic affairs,
the biggest item by function, has remained at relatively the same level between 1970
(27.4 percent of GDP) and 2000 (25.2 percent) even though the purpose of this
expenditure, spending for economic development, has become less relevant in todays
market-oriented economy.
23MPB (2003), pp.124-125.

24Some of these reforms were the preliminary feasibility study requirement, tighter management of project
ceiling budgets, performance-oriented budgeting, reforms to the public fund system, consolidation of quasi-
taxes, and double-entry bookkeeping (MPB, 2003, pp.126-149).
Chapter 3. Key Issues for Introducing MTEF and Top-down Budgeting in Korea 25
Another concern regards the strategic decision-making mechanism in resource

allocation. At present, little high-level discussion is conducted in the executive body
between ministries and within the Ministry of Planning and Budget (MPB) regarding
how public spending should be related to policy priorities. This lack of communication
means consensus is seldom achieved during the early stage of the budgeting process.
As a result, national policy objectives receive less than proper consideration in the
resource allocation process. This also contributes to the lack of transparency in
decision-making for resource allocation and setting policy priorities.
2.1.2. Future fiscal risk

Since the early 1980s, the Korean government has maintained strong fiscal
discipline through tight fiscal and monetary policy. Notably, the deficit of the
consolidated central government budget averaged 1.5 percent of GDP in 1981-1986 and
the size of government debt was successfully restrained at a manageable level. Further,
the Korean government, having experienced fiscal deficit at 4 percent of GDP in 1998,
has made consistent effort to contain expenditure growth. Largely due to the rapid
growth in revenue arising from a dramatic rebound in the economy and well-
controlled expenditure increases, the fiscal balance returned to surplus of 1.3 percent in
2000, 2.0 percent in 2001 and 2002.25
However, there are many dormant problems, threatening fiscal soundness in the
future. Unless properly handled, these threats could burgeon beyond manageable
levels, necessitating that they be addressed sooner rather than later.
One of the most serious risks lies in social welfare demands. These represent a
major upcoming burden on Koreas public finances. Especially, major public pension
schemes have many problems such as generous benefits relative to contributions,
implying that the reserve will be depleted in a foreseeable future.26 Lee (2003) outlines
Koreas expected social welfare burden from the four public pension schemes showing
that benefits due will peak at 11.1 percent of GDP by 2035.27
This problem becomes more serious in the context of Koreas aging society. The
aged those 65 years old and above population in Korea was 7 percent of total
population in 2000 but that figure is expected to reach 14 percent by 2019. It should be
emphasized that a key dimension to the aging population in Korea is the speed with
which the demographics of society are shifting. What this fast aging phenomenon
implies is that expenditure demand will rise rapidly and on a large scale and revenue
source will decrease near in the future.
The second issue is restructuring of the financial sector. The fiscal support to
financial restructuring mostly took the form of government guaranteed loans.
Consequently, high net lending during financial sector restructuring in the aftermath of
the financial crisis has left irretrievable debts of 69 trillion won, equivalent to 12
25 The consolidated budget surplus is a r g u a b l y over-

stated due to the National Pension Fund, net lending, and obligations for repayment of restructuring bonds. A
perhaps more accurate picture of the fiscal balance appears after adjusting for these, showing fiscal balances of -
0.5 percent, 0.1 percent, and 0.1 percent in 2000, 2001, and 2002 respectively (Koh, 2003, p.10).
26 The four main pension schemes are the National Pension Scheme (NPS), Government Employees
Pension Scheme (GEPS), Private School Teachers Pension Scheme (TPS), and Military Personnel Pension
Scheme (MPPS).
27 Koh (2003) , p.13.
percent of GDP in 2002.28 This will inevitably make it harder to maintain fiscal balance
for quite a long time.
Thirdly, of no small concern is the current division of the Korean peninsular and
financial burden this situation continues to place on expenditure demands. This unique
issue, whose outcome remains very ambiguous, requires sufficient planning in terms of
public finance. The future of Korea, including possible unification with North Korea, is
unclear but suffice to say all contingencies warrant due consideration. Reducing the
negative effects of unification under any circumstances, will depend heavily on
preparation and organization prior to such an eventuality.29
Moreover, the revenue side poses as high a risk to the fiscal balance as the
expenditure side. Further aggravating the potential danger of threats is the outlook of
Koreas economy growth rate, which is expected to slow markedly in the coming years.
Reduced economic growth rates necessarily imply declining rates of revenue, which
will test Koreas fiscal discipline.
2.1.3. Inefficiency from managerial inflexibility

Under the current system, there is a lack of managerial flexibility and limited
autonomy in line ministries both in planning policies and implementing budget. It is
believed that the existing centralized, input-control based, bottom-up approach in
budgetary decision-making is one of the main causes of inefficiency and rigidity.
Currently, the MPB reviews all projects in the budget review stage, thus it
intensely influences line ministries. This induces line ministries to exaggerate budget
proposals and often leads to laborious bargaining between line ministries and the MPB.
Such circumstances also often incur enormous transaction costs and foster informality
in the public sector,30 representing a misuse of time toward unproductive efforts.
It also causes accountability problems. The link between responsibility and the
discretion to perform flexibly and autonomously may be tenuous, creating a mismatch
between policy and accountability. The lack of managerial flexibility in the public
sector has undermined governments efforts, by performance management system, to
improve accountability.
2.2. Koreas reform initiatives31
Korea is in the process of preparing public expenditure reform plans with the
following characteristics representing the main directions.
Long-term Perspective: Under the new system, the MPB will prepare a Long-term
Financial Management Plan, which makes a long term more than 10 years
projection of the overall fiscal balance and sectoral allocation. For this projection, the
MPB will factor in various macroeconomic and social variables including sustainable
debt level and tax burden, national policy agenda, and sectoral expenditure demand.
This Plan will serve as a baseline to determine fiscal balance and sectoral allocation for
28 Koh (2003) , p.1.

29 IMF estimates that it will cost 975 trillion won, equivalent to 820 US dollars, if the two Koreas were unite
d as of 2001 and North Koreas GDP will reach the same level as South Korea in fifty years(Funke and Strulik, 2
002, p.18).
30 For informality in the culture of government, see Schick (1998), pp.127-129.
31 This plan is in the process of being prepared so this section is not necessarily representative of the official
position of the Korean government.

the upcoming three years and be reviewed and updated on a rolling basis to reflect any
unexpected changes.
Top-down Budgeting: Fiscal targeting, of which sustainable fiscal balance over the
economic cycle is a core aspect, will be driven by the Long-term Financial Management
Plan and macro-economic forecasting. Based on fiscal targeting and revenue estimates,
the decision is to be made for the budget aggregate for three years. Following this, and
with consideration to policy objectives and priorities, sectoral and ministerial
expenditure limits, for three years, will be finalized.
Strategic Decision Making Mechanism: As part of the new strategic decision-making
mechanism, cabinet-level meetings will be held to discuss long-term plans, macro-
economics, budget aggregate, and sectoral limits. The main purpose of the cabinet
meeting is to establish a forum wherein consensus-building on national policy
priorities and expenditures will be made at the early stage of budget formulation. The
additional objective is to obtain strong commitment from the cabinet, and enhance
transparency through open and direct discussions. To support this meeting, the MPB
should prepare a draft fiscal target, fiscal aggregate, and sectoral ceilings. During the
preparation process, the MPB should interact closely with the President, who will
preside over the cabinet meeting, on proposals in order to foster greater coordination
and obtain critical support.
Decentralization: Currently, the MPB is heavily involved in preparing
appropriations for individual projects in the budget process, which leaves very little
room for line ministries to maneuver. In contrast, the new system will make line
ministries play a much more substantial role. Within the boundary of the sectoral
ceiling and policy priorities set by the cabinet meeting, line ministries will have more
autonomy to prepare their own budget proposals. The MPB will focus its efforts on
reviewing budget proposals to ensure that sectoral ceilings have been respected and
that proposed requests sufficiently reflect policy priorities. Increased line ministries
discretion over their budget preparation can significantly contribute to enhancing the
accountability of line ministries, because line ministries will be held accountable for
what they can control rather than for what they can not. A proper system of checks and
balances including performance management should be in place to align managerial
discretion with accountability.
Under the above direction, Koreas budgeting process will undergo fundamental
change. The changes, incorporating the introduction of MTEF, will lead to new
processes and stages outlined below:
Stage 1: Long-term financial management planning

This stage covers medium-term macro-economic forecasting, and review of social
and economic factors that may affect public finance. Various national debt levels such
as the debt ratio to GDP or total expenditure would also be reviewed at this time.
Contingent upon the findings and reviews conducted at this stage, the Long-term
Financial Management Plan should be constructed by the MPB. In the Plan, the long-
term national agenda and sectoral expenditure demands should also be checked and a
strategic, long-term resource allocation plan suggested.
Stage 2: Fiscal targeting and budget aggregate

In this stage, based on the macro-economic forecasting and the Long-term
Financial Management Plan, the fiscal target and important fiscal policy for three years
will be set. The core fiscal target under consideration should be fiscal balance policy.
Additionally, along with analyses of other important factors including revenue
estimates, the budget aggregate for three years should be suggested.
Stage 3: Sectoral and ministerial ceilings

Following stage two, the MPB should suggest strategic resource allocation in line
with sectoral and ministerial expenditure ceilings. In this stage, the national strategic
policy objective and sectoral objectives are considered and major budget projects are
classified according to strategic importance.32
Stage 4: Cabinet meetings

For this stage, a special Cabinet meeting will be convened to discuss and agree on
macro-economic forecasting, long-term plans, fiscal targeting, and expenditure
aggregates and sectoral ceilings. Consensus among cabinet members should be built
through constructive and informed discussion and unanimous support attained at the
cabinet level. It is perceived that cabinet meetings should be convened twice during a
fiscal year. The first, held in March, would discuss and agree on the fiscal target, total
size of budget, and sectoral ceilings. The second in August would update
macroeconomic conditions and review the necessity of adjustments, and finalize the
coming years budget proposal. After consensus is reached at the cabinet meeting, the
MPB should present and explain the proposal to the National Assembly.33
Stage 5: Line ministries requests

Line ministries should prepare their budget proposal so that it clearly reflects
policy priorities and respects sectoral ceilings. Through this, line ministries would have
more autonomy and discretion and the MPB, to facilitate line ministries preparation,
would provide guidelines, including criteria for common expenses. In case of
complicated projects which involve more than one ministry such as R&D and regional
development projects, the MPB could consider establishing committees, consisting of
relevant stakeholders inside and outside government, to reach an objective and
concerted decision.
Stage 6: Review and budget documentation

The MPB reviews line ministries budget requests for sectoral ceiling compliance
and policy priorities. In this stage, negotiation/bargaining between the MPB and line
ministries will still occur. Even where budget proposals remain within sectoral ceiling,
the MPB may raise questions and suggest changes to line ministries if the MPB believes
policy objectives can be better achieved with different projects/programs. The type and
frequency of discussion between the MPB and line ministries, however, will change
such that discussion will be based upon transparent and fact-based rules.
3. Key issues to be addressed

The principle of linking policy with budgeting, which MTEF embodies, is useful
and popular for many countries around the world. However, not every country has
been successful in introducing MTEF or fully realizing its benefits. One of the
commonly cited reasons for less successful MTEF in many countries, developing and
32 One possible classification scheme considered by the MPB is to sort all budget projects as A, B, or C,
where A is projects related to national strategic objectives and sector objectives, and mandatory expenditures
such as debt payment and debt interest; B is core projects in each sector; and C is projects that are classified into
neither A nor B.
33 The role of the National Assembly and legislative procedures are integral to the budgeting system and
important for reform. However, legislative issues are beyond the scope of this paper.
developed alike, is that too much attention is given to the technical aspect of MTEF and
little consideration on how to change the institutional framework, decision-making
process, and organizational culture of public sector. International experience clearly
suggests that basic principles of MTEF must be carefully adjusted to fit into country
context.
In this regard, it will be necessary to analyze key issues Korea may face down the
road of reform and suggest how Korea can prepare itself to address these challenges.
Special attention should be paid to country context in the areas of macro-economic
forecasting, strategic decision-making, setting the level of line ministries discretion,
designing incentive systems, and reforming budgetary basics such as budget structure,
scope and improving information system.
3.1. Macro-economic forecasting
Questions concerning macroeconomic forecasting are two-fold, institutional and

functional. Institutional questions are, how can a macro-economic forecasting function be set
up in the MPB under the current institutional setting, and how can capacity be developed in
the MPB? Functional questions are, how can unbiased macro-economic forecasting be
undertaken, and how can macro-economic uncertainty be overcome?
Korea has organizational complexity regarding macro-economic forecasting and
policy coordination. Currently, the Ministry of Finance and Economy (MoFE) has
macro-economic forecasting and policy coordination functions, while the MPBs duties
lie in fiscal policy and budgeting.34 In addressing organizational complexity, dramatic
change of organizational structure is likely to not only be unfeasible but also create
significant resistance to the change that MTEF will bring in. Rather, it is
recommendable to clarify roles of each ministry under the current system and set up a
coordination mechanism.
Although the MPB will conduct macro-economic forecasting under the new
scheme, its role should be limited, considering the current institutional environment.
The MPBs macro-economic forecasting function should be implemented only within
the range of fiscal policy making. This may help avoid undesirable institutional conflict
in government. Beyond that, it would be valuable for the MPB to embrace the findings
of other government agencies and organizations, and coordinate the works from all
sources for better macro-economic forecasting. To this end, forming an advisory
committee, comprising of MoFE, Bank of Korea, and public and private research
institutes might be worth considering. However, it is the MPB who should be in the
drivers seat when it comes to deciding on macro-economic forecasting, at least for
setting the baseline for the Long-term Financial Management Plan and fiscal policies.
Unbiased, honest, and realistic forecasting is essential. MPBs well-trained
technocrats and the independent advisory committee may be a way to move toward
this end. It is advisable to devise a macro-economic forecasting formula for objective
34 By way of background of Koreas institutional changes, economic forecasting and budgeting has seen
several important changes. These changes have occurred at the institutional level directly impacting the
decision-making mechanism and coordination functions. In 1994 the EPB was merged with the Ministry of
Finance to form the Ministry of Finance and Economy (MoFE). This was soon fractured leading to the
establishment of the Planning and Budget Commission (PBC) and the Office of National Budget (ONB) as
separate bodies from the MoFE. Shortly thereafter, the PBC and ONB themselves were merged to create the
MPB, leaving the MoFE and MPB as two independent ministries.
analysis and avoid pressure from outside political circles. In developing this formula, a
conservative method would be preferable to contribute to fiscal soundness.35
It is recommended to establish a unit/department within MPB responsible for
macro-economic forecasting and assessment of social/political factors that affect fiscal
policy. This new organization could comprise of one or two divisions. Having such an
organization within the ministry could allow the MPB staff to appropriately deal with
macro-issues, undergo training to enhance their capacity, and to coordinate greater
links with outside experts. The MPB may make both judgemental and predictive errors,
but will perform better over time as they build up experience.
Lastly, to cover macro-economic uncertainty, the concept of budget margin
should be created in the annual budget. The budget margin acts as a kind of buffer and
mainly aims to absorb fluctuations in the expenditure level due to changes in the cycle
and other macro-economic uncertainties.36, 37 It can be utilized in cases of forecasting
deviations and new presidential policy initiatives. The size of the budget margin may
depend on accuracy of information. The budget margin should be no larger or smaller
than what is considered enough to fulfil its role of accommodating uncertainty. It is
perceived from international experience that, over time, macro-economic forecasting
tends to become more accurate allowing the size of the budget margin to decrease.38
3.2. Strategic decision-making on budget aggregate and sectoral ceiling
How should decisions about total budget aggregate and sectoral ceilings be made and by
whom?
Regardless of the type of decision-making mechanism, consensus and commitment

at the highest level of government is a critical component for reliable fiscal targeting,
budget aggregate and sectoral ceiling levels. In some OECD countries, such as Sweden,
Norway, and the Netherlands, decisions are made through cabinet meetings. On the
other hand, in the UK, H.M. Treasury fulfils the leadership role which is enhanced
through a close relationship with the Prime Minister. Different again, in the USA there
is a close informal interaction between OMB and the President, with informal cabinet
level discussions, resulting in a strong commitment at the highest level and across the
board. In all these countries systems, the common and important characteristic is
consensus and commitment at the highest levels of government.
Korea is now strongly pursuing top-down decision-making mechanisms to link
policy and budgeting. 39 In this regard, cabinet meetings should provide the forum
where strategic decisions on fiscal targeting, total aggregate, and sectoral ceilings take
place. Through this meeting, agreement will be reached in the Cabinet ensuring
35 For example, the Canadian Department of Finance systematically revises the private sector forecasts dow
nwards as a further measure of prudence. This takes the form of the government adding 50-100 basis points (0.
5-1.0 percentage points) to the average private sector economic forecasts for interest rates and then feeding this
through its entire econometric model, thus producing lower forecast economic activity. This provides a buffer i
n order to maintain the governments fiscal objectives (OECD, 1999, p.9).
36 Swedish Ministry of Finance (2002), p.13.
37 Budget margin is different from budget reservation i n K o r e a n b u d g e t i n g

which represents a permanent allocation in the budget for unexpected demands such as natural disasters.
38 The Sweden case shows the budget margins for 1997-2001 were 3.33, 0.28, 0.20, 0.65, and 0.59 percent and
are projected for 2002-2004 as 0.04, 0.05 and 0.07 percent, respectively (Swedish Ministry of Finance, 2002, p.13).
39 However , it should be noted that MTEF consists of top-down resource allocation and bottom-up
approach for cost estimation. In this sense, the bottom-up element should not be overlooked.
commitment at the highest level. The MPB should play a key role by preparing its
proposal on fiscal policy, total budget aggregate, and sectoral ceiling for the cabinet
meeting. Furthermore, it is important that the MPB works in close consultation with
the President in preparing the proposal. From this, mutual benefit can be achieved as
the President can input his policy priority into the budget and the MPB can gain
political support.
A concern may arise over the possibility of disagreement at the Cabinet meeting,
particularly over sectoral ceilings. This concern arises because Korea has been weak in
forming consensus through consultation and negotiation processes. Added to this, lack
of negotiation experience is the growing voice from numerous interest groups as Korea
becomes a more democratized society. While valid demands should be heard and
brought to the table for discussing national agenda, many not-in-my-back-yard type
protests are becoming ubiquitous as well. This situation calls for special attention to be
paid to the design and implementation of MTEF, since the cabinet meeting is likely to
experience difficulty reaching agreement.
Given this situation, it may be advisable that the cabinet meetings adopt a two-
stage approach. The first stage would be to decide the budget aggregate, and the
second stage would determine the sectoral allocation. If these two were placed on the
table together, it would be very difficult to reach an agreement. Thus, division of these
two in decision-making becomes important as the decision on the total budget
aggregate turns the discussions on sectoral ceilings into a zero-sum game. Against all
these efforts, if cabinet meeting fails to reach an agreement, it should be understood
that the President will make the final decision, with the MPBs objective and
professional support.
3.3. Discretion of line ministries
How much discretion should line ministries be given?
A centralized budgeting process has been with Korea for a long time, with the
MPBs deep involvement in almost all facets of the budgeting process. Entrenched in
this all-under-my-control mechanism, the MPB may be reluctant to relinquish its
influence over appropriations. Meanwhile, concern over line ministries lack of
capacity and experience in areas such as prioritizing within a sector may also lead to
hesitation for greater discretion to line ministries.
Such barriers, however, can only be overcome by making steps to directly address
these issues. A new incentive system must be introduced so that MTEF can be
perceived by the MPB as being beneficial to them and that their influence over
appropriation will be upgraded to more professional and legitimate influence.
Ultimately, discretion to prepare a budget request within sectoral and ministerial
ceilings should be granted to line ministries. This may afford line ministries greater
ownership over their respective budget proposals and make them more likely to the
identify policy priorities within their sectors as well as grant managerial flexibility
resulting in greater efficiency. Through these experiences, line ministries can also build
up necessary in-house capacity to manage increased discretion over their policy and
budget more effectively and responsibly.
On the other hand, Budget proposals prepared by line ministries should be subject
to the MPBs review. This review should mainly act as a checkpoint to ensure that line
ministries proposals are in line with the national policy priority agreed on at the
Cabinet meeting and stay within the sectoral ceiling. Also the MPB should provide
guidelines, including criteria for common expenses, and perform monitoring and
coordinating roles to ensure line ministries proposal is prepared according to the

provided guidelines.
In addition, an incentive system should be developed to promote more efficient
spending for line ministries. Presently any money left over at the end of fiscal year is
returned to the General Account. Obviously, this system provides little or no incentive
for line ministries to save money. One way to alter the incentive of line ministries is to
classify any remaining funds at the end of the fiscal year into two categories; savings
and over-budgeting. The first, savings, would refer to amounts arising due to efficient
spending and genuine efforts to reduce spending and improve efficiency. For funds
falling into this category, a certain percentage of savings empirically at least more
than half of the savings should be allowed40 could be retained by the line ministry
for use in the upcoming fiscal year. The second category, over-budgeting, should be
returned to the General Account as per the current system.
3.4. New incentive system
How can the new incentive system overcome resistance to implementation, and what
should be the new role of the MPB and line ministries under this new system?
It is possible that both the MPB and line ministries will resist reforms for many
reasons. The MPB may be concerned with potential loss of influence and power
through the proposal to diminish its role in allocating resources. On the other hand,
line ministries may be skeptical about receiving more discretion in their own budget
preparation. Both the MPB and line ministries may also perceive the changes as merely
additional work without any real benefit.
Two main considerations should be given to deal with resistance to the new
system. First, some resistance may arise from a lack of knowledge regarding the real
nature and characteristics of the proposed new system; how annual budgeting process
changes under MTEF, what will be the new roles of MPB and line ministries, what
kinds of benefits can be expected from MTEF. Secondly, uncertainty over individual
incentives under the new system, if not handled appropriately, may lead to resistance.
MPB and line ministries staff may not be sure how the changes will affect their
personal interest.
For the MPB, the existing system places a large work burden on the staff and much
time is devoted to micro-level decision on individual activities.41 However, under the new
system, focus will shift to higher level decision-making, analyses of macro-economics and
national policy priorities, and preparation of draft proposals for the cabinet meetings.
Furthermore, the MPB could have the privilege to interact closely with the President. From
all these, the MPB will not lose its current standing as a prestigious ministry. On the
contrary, its influence and power may be strengthened. The MPBs influence will remain
high through monitoring and coordination over line ministries budget proposals during
the budget review stage and through performance evaluations of major projects.
On the other hand, line ministries have an opportunity to enhance their level of
influence, providing an incentive to gain greater discretion in decision making. Line
ministries, firstly, will be granted discretion in planning and prioritization within their
Allen and Tommasi (2001), p.359.

40
The number of projects per budget examiner in the MPB has jumped from 81 in 1994 to 150 in 2003; and t
41
he portion of budget allocated per budget examiner increased from 1.0 trillion won in 1994 to 4.6 trillion won in
2003 (MPB, 2003, p.8).
sectors and secondly, they will have flexibility in implementing policy and executing
budget. There will be a learning curve when introducing these changes which will
further allow line ministries to build up capacity to deal with their new role. The
timeline for granting discretion may need to be done gradually to plan a path for
granting discretion over to line ministries.
Besides, gaining greater discretion over budgetary decisions can also increase the
legitimacy of holding line ministries accountable for the performance of their output.
Under the new system, unlike the current situation where line ministries are sometimes
held accountable even though they dont have any control over resources and decision-
makings, line ministries should be accountable for the decisions they make and their
performance only.
3.5. Performance and accountability
In an effort to enhance performance, the Korean government launched a pilot

project in 1999 to implant performance-based budgeting (PBB). Under this project 39
participating agencies prepared and submitted performance plans to the MPB
specifying objectives, strategies, and indicators. They also compiled an ex post
performance report for comparison against the original plan. The impact of this pilot
project between 1999 and 2002, however, was analysed as falling short of expectation.
The MPB, recognizing that it may have been too ambitious to link performance
and budgeting, is in the process of designing an alternative, applicable management
system. In the new system, line ministries are asked to develop objectives, indicators,
and measurement methods for certain projects under the guidelines provided by the
MPB.42 In this way, line ministries will have strong sense of ownership and the MPB
can ensure performance indicators are appropriate and consistent over all ministries.
The new performance management system should ideally follow a long-term
phased approach. Since this can be a daunting task, it may be beneficial to formulate a
viable, short-term plan, which sets reasonable goals and fits within the broader
program, to facilitate early and successful implementation. Pilot projects to develop
and introduce output-based indicators would be a preliminary step followed by
expanding into all programs and developing outcome-based indicators.
Expectations are high regarding the latest tools and techniques for performance
management. Evidence from international experience, however, indicates that
performance management or PBB adds value only when the existing institutional
arrangement in the public sector is conducive to support performance orientation.
Therefore, the Korean government would need to foster an environment which
supports performance orientation while fully introducing PBB or outcome budgeting.
For example, personnel management and organizational structures that provide
incentives for improving performance should be in place along with a performance
management device.
The quid pro quo for the increase in discretion is that the line ministries be held
accountable for results. The new system should bring a fundamental change in terms of
accountability. It should make the MPB and line ministries more responsible for their
results, and not their processes. Thus, instead of input control, the focus should be on
what they achieved.
42 According to the MPBs plan, a certain number of projects, amounting to 30 percent of the total
ministerial budget, will be evaluated.

The first step to enhance accountability is to emphasize that ministers should hold
ultimate responsibility for programs operated under their discretion.43 Presently, the
Korean government has little ex post evaluation, although it has introduced some ex
ante evaluation systems, including the preliminary feasibility study. To move further
along this path, the MPB should embrace the role of watchdog by performing more
comprehensive monitoring and evaluation of projects. Setting up policy objectives for
the cabinet meeting and discussing projects objectives with line ministries during the
budget review stage will be a sound basis for this process. To further contribute to this
mechanism, budgeting and accounting systems should be linked and information
exchange between the MPB, the MoFE, line ministries, and the Board of Audit and
Inspection integrated.
3.6. Budget structure, classification, and scope
Another lesson from international experiences with MTEF is that success of fiscal
reform largely depends upon budgetary basics such as budget structure, scope,
classification, accounting, information, evaluating, and auditing. Unless these basics
are firmly set up, introducing MTEF may not generate anticipated benefits due to a
weak public expenditure management environment.
Unfortunately, many weaknesses exist in the Korean PEM system. First, the budget
structure is too complicated; a fragmented assortment comprising one General
Account, 22 Special Accounts, and 45 Funds. Second, its expenditure classification is far
too subdivided; consisting of more that 2,200 appropriation accounts (Se-Hang) and
more than 6,000 sub-accounts (Se-se-Hang). Each of them is subject to negotiation
between the MPB and line ministries. Third, the coverage of the budget is limited;
excluding some important fiscal activities of the government, like the National Health
Insurance and quasi-government activities. 44 Current coverage of the budget is not
consistent with Government Finance Statistics (GFS), an international standard for
public expenditure statistics issued by the IMF.45
The fragmented budget structure, complex classification, and limited budget
coverage cause several problems. First, it is difficult to control public financial
resources, which impedes allocative efficiency. Second, the inherent opaqueness of the
system leads to transparency and accountability issues. Third, these budgetary basics
problems also contribute to managerial inflexibility and make results orientation more
difficult to attain. These problems, all together, make it technically prohibitive to
allocate resources within sectoral limits down at the level of sub-accounts, and this
would hold even if the Korean government introduces a new system incorporating
sectoral ceilings.
Much effort to realign the budget structure, classification, and coverage are needed.
The Korean government should make all the special accounts and funds subject to the
same level of scrutiny as budgetary expenditure. And, extra-budgetary funds should be
discouraged unless very strong justification exists. The budget subdivisions, currently
classified by line items with its heavy emphasis on input control, should be more
43 A serious problem is that the tenure of ministers has been notoriously short with the average term of a
minister being less than one year, during 1993-2002 (Maeil Economic Newspaper, March 10, 2003).
44 For more details, see Koh (2003), pp.29-31.
45 It is however important to note that GFS is for statistical purposes rather than budgetary management
purposes. IMF acknowledges that countries are not obliged to prepare an annual budget completely in line with
GFS. Rather, it requires a country to ex-post prepare and rearrange expenditure according to GFS.
simplified and, in the long run, managed with expenditure classification on a

functional and program basis. Also, the consolidated budget report should embrace all
activities that fall into public spending.
There also exists an urgent need to improve the information base and establish
efficient information flows to facilitate strategic decision-making. Core functions of a
financial management system like budgeting, accounting, cash-debt management, and
auditing-reporting need to work as one entity even though it consists of multiple,
geographically dispersed organizations. The information outflow from one component
should be an inflow to all other related components so all components share a common
financial evaluation and measurement base and decision makers can make timely and
informed decisions.46 For example, as line ministries would have more discretion in
preparing their budget, the MPB needs to verify that each ministry uses the same set of
macroeconomic indicators such as the unemployment rate and inflation rate.47 Another
important benefit from the integration of financial information is enhanced
transparency of financial management procedures. Increased information flows and
accuracy strengthens the checks and balance system within administrations, between
the MPB and line ministries, and between the public sector and citizens.
4. Some suggestions for implementation strategy

For the successful implementation of MTEF, it is important to acknowledge the
unique public sector environment of a country. Consideration of country context is all
the more important for Korea, because Koreas public sector has very distinctive
features from western countries where MTEF was originally developed and practiced.
This section first briefly discusses a few aspects of Koreas public sector which may
require special attention when developing an MTEF implementation strategy.
Following this, some selected observations and suggestions regarding Koreas path to
implement MTEF and top-down budgeting are provided.
First, Korea has shown a strong track record in linking planning and budgeting in
the past. Koreas Economic Planning Board (EPB), a super-ministerial economic agency
in Korea in 1961-1994, was empowered to handle both planning and budgeting
functions under one roof for better directing economic growth. It is regarded that the
EPB played an important role in directing economic development and was a crucial
player in Koreas economic success story. The minister of the EPB was Deputy Prime
Minister and enjoyed strong political support through visible Presidential involvement.
In retrospect, Korea has, in substance, exercised MTEF without calling it MTEF.
Second, it is important to note that the current reform initiative is initiated within
the executive branch and led by a number of technocrat champions. This creates a
strong sense of ownership within the government and helps to design workable
features of reform and to implement it. Traditionally, technocrats have played very
46 An integrated information system becomes a more vital issues under Koreas divided organizational stru
cture where treasury management and budget execution reporting is under the MoFE and budget formulation
and execution is under the MPB. The current two stand-alone information systems in the two ministries need to
be seamlessly linked to each other.
47 This is not to say that the MPB should handle macroeconomic conditions for all purposes. Rather, the
MoFE should be in charge of monitoring and setting macroeconomic conditions for general purpose and the
MPBs concern should be limited to those related to budget formulation.
strong roles in Koreas public sector.48 Among the reasons to explain this predominance
of career technocrats are historical and social respects for government officials. This has
attracted many well-educated and highly trained people to join the public sector and
become career technocrats. This also means that technocrats relatively distance
themselves from political pressures. At the same time, the executive body and
technocrats need to draw together to present a broad and united front of political
support for its reforms as demands of various interests increasingly enter into the
decision-making process.
Past experience of linking planning closely with budget during the EPB period and
strong sense of ownership for the reform within the government and technocrats
creates a favorable environment for MTEF to be successful. But that is not to say that
the path of this reform will be easy. Challenges and obstacles lie ahead making the
successful implementation of the new system no guarantee. In order to overcome these
obstacles, the key issues outlined in the previous chapter need to be carefully
addressed and detailed plans prepared. Besides these key questions, below are some
observations and suggestions regarding Koreas path for implementing MTEF and top-
down approach.
Pre-conditions
There are several important pre-conditions for successful implementation of MTEF
in Korea. Among these are strong political support, the MPBs willingness and
commitment to the new system, compliance of line ministries, capacity building for the
MPB and line ministries, and development of a legal framework for the new system.
Some have been met but not all. Although not all the pre-conditions are in place, the
Korean government, rather than delay reform, should focus on putting them in place in
parallel with proceeding with reforms. Most OECD countries also had, and still have,
some shortcomings in their expenditure management. Instead of waiting till everything
becomes impeccable, they used MTEF as an instrument to alter the status quo and
improve along with strengthening policy linkage with budget. Considering its highly
skilled civil service, Korea may opt to introduce MTEF while simultaneously striving to
improve these pre-conditions.
Bureaucracy and organizational culture

Central to success of the proposed reform will be the evolution of Koreas
bureaucracy and organizational culture. Koreas civil service system is characterized by
life-long job security, seniority-based systems for promotion and non-merit based
salary, short-term tenure of high level policy makers, and lack of performance-oriented
culture, all of which leave little incentive for bureaucrats to accommodate change.
These characteristics have contributed in part to the failure of past public sector
reforms which government ambitiously tried to implement and ended up only causing
superficial changes and limited tangible results.
Fiscal reform, due to its potential impact on resource allocation, can be a strong
driving force in broader public sector reforms. However, fiscal reform is a type of
reform that requires a holistic approach; as public expenditure touches almost all
realms of public sector, meaningful change in PEM must be complemented with other
public sector reforms. Therefore, it is important to reignite various public sector reform
efforts that were initiated at the time of the Asian financial crisis but remain
48 It is indicative to point out that most political appointees in economic policy-related ministries, particular
ly the MPB and MoFE, have been actually been awarded to career technocrats over the last four decades.
incomplete. It will not only create a favorable environment for MTEF to be introduced,
but also will create a synergy effect for building a more effective and transparent public
sector.
Leadership and capacity

Strong leadership is essential in any reform. For MTEF and top-down budgeting,
direct involvement of the President and active engagement of cabinet are necessary.
Also, the MPBs strong leadership is crucial to initiate the reform. It will be particularly
essential to have key individuals in the ministry to champion change both initially
and to sustain the momentum. Expanding from this initial step, a core group should be
identified to lead and promote the initiative throughout the ministry.49
Capacity building is another important element to implant this new system. For
the MPB staff, skills in macroeconomic and policy analysis, policy prioritization,
coordination and conflict resolution, and problem-solving abilities should be built up.
Line ministries should develop capacity in planning and policy analysis, development
of implementation strategy and tools, and problem-solving abilities.
The MPB and line ministries can build up capacity through job training and
seminars for all levels of MPB and line ministry staff. Not just the technical features of
MTEF but the benefits of MTEF should be conveyed to all staff levels. More important,
however, is to create an incentive system that motivates self-development and capacity
building. Once sufficiently motivated, the MPB and line ministry staff will benefit from
the learning curve as they internalize the MTEF process.
Big bang and gradual approach

Devising a specific implementation strategy will increase the likelihood of success
of the reform. Reform can be implemented via a big bang or gradual approach, i.e.
introducing the whole plan at once or on a piecemeal basis. For Korea, a combination of
a big bang approach and gradual approach together would be more desirable. Top-
down decision-making may be adopted via a big bang approach and should see
strategic decision-making for budget aggregate and sectoral ceiling introduced together
with resource allocation policy making. However, in areas not covered in strategic
decision-making, a more gradual approach should be adopted. This includes various
budget stages, such as budget proposal preparation, review, monitoring, and
evaluation. A comprehensive long-run plan should map out smaller steps designed to
realize the gradual approach. In the process of implementation, attention should
continually be paid to meeting the pre-conditions referred to above.
Performance-orientation
Improvement of PEM requires serious focus on results achievement or
performance from government expenditures. Accordingly, MTEF and performance
management represent two sides of the same coin. Even if MTEF is successfully
introduced, it may be vulnerable unless performance-oriented management is
implemented correspondingly.
International experience shows that nurturing a performance-oriented
environment is of central importance before new tools and techniques for performance
management or PBB can be fully introduced. In the case of Korea, performance
orientation in the public sector is considerably under-developed. In the current system,
life-long tenure, seniority-based promotion and salaries are the norm. There is little
49 Muggeridge (1997), p.32.

managerial flexibility, to the extent that even ministers cannot make changes in the
organizational structure, nor do they have the discretion to hire career officials. Thus, it
is at least equally important for Korea to make efforts to realign the public sector
environment, especially the civil service system, to fit into performance-oriented
management while efforts to introduce performance management system needs to
continue.
A suggestion to exemplify the advantages of such a culture would be to select one
or two government ministries/agencies and designate them as regulation-free
organizations. All regulations would be removed, allowing ministers/CEOs full
discretion over personnel, organizational structure, management, etc. 50 At the same
time, the chosen agencies would have clear assignments and a vision to achieve the
objectives. In such an organization, performance would be a key driver for the minister
and senior officials because they would be judged on what was achieved rather than how
things were done. This should also lead to positive spill-over effects. Through this pilot
project a performance-oriented culture could spread out in the public sector.
Koreas PEM is currently in transition, having already begun some reforms to
move forward, but still many more issues remain to be addressed. Many challenges lie
ahead that must be confronted, but there is also cause for optimism. The Korean
government is dedicated to reform, initiatives have been derived and pushed from the
inside, and there are positive dynamics for change. A budget system is a
communication system, conveying symbols and signals about behavior, priorities,
intentions, commitment, and price. The new PEM system, thus, is much more than
simply fiscal reform; its about creating new value and changing behavior, making
public sector more performance-oriented, accountable, and transparent.
50 This idea is borrowed from Special Economic Zones, which are designated areas dispensed of regulations
in order to boost investment and create a free and competitive business environment for economic activities.
References
Allen, Richard and Tommasi, Daniel, Managing Public Expenditure: A Reference Book for Transition
Economies, OECD, 2001.
Anipa, Seth, Felix Kaluma, and Elizabeth Muggeridge, DFID Seminar on Best Practice in Public
Expenditure Management: Case Study of MTEF in Malawi and Ghana, Consulting Africa
Limited, 1999.
Blondal, Jon, Budgeting in Sweden, OECD, 2001.
Byeon, Yang-Kyun, Introducing A Top-down Resource Allocation System and Its Expected
Effects, in his Three Essays on the Sustainability of Government Finance and Improving
Efficiency in Resource Allocation in Korea, Ph.D. Dissertation, Sogang University, 2002, (in
Korean).
Dorotinsky, Bill, Developing a Medium-Term Expenditure Framework (MTEF) for Columbia.
Dorotinsky, Bill, Key Roles in Executive Branch Budget Formulation: Integrating the Elected Executive
into the Budget Process, 1999.
Funke, Michael and Holger Strulik, Growth and Convergence in a Two-Region Model: The
Hypothetical Case of Korean Unification, IMF Working Paper, 2002.
Jun, Taek-Seung and Ki-Paik Park, Review of Pilot Studies on Performance-based Budgeting, Korea
Institute of Public Finance, 2002 (in Korean).
Koh, Youngsun, Public Expenditure Management in Korea, Korea Development Institute, 2003.
Maeil Economic Newspaper, March 10, 2003 (in Korean).
Ministry of Finance and Economy, Consolidated Fiscal Balance of Korea, each year (in Korean).
MPB (Ministry of Planning and Budget), National Resource Allocation Improvement Plan (Draft),
2003 (in Korean).
_____, How Korea Reformed the Public Sector (1998-2002), 2003 (in Korean).
Muggeridge, Elizabeth, Mozambique; Assistance with Development of a Medium-term
Expenditure Framework, Draft Report, World Bank, 1997.
National Statistics Office, Future Population Estimates, 2001 (in Korean).
OECD, Budgeting in Canada, PUMA/SBO(99)5/FINAL, 1999.
Passman, Liz, Presentation on Public Expenditure Planning and Control, UK Treasury, November
2003, p.4.
Potter, Barry and Jack Diamond, Guidelines for Public Expenditure Management, International
Monetary Fund, 1999.
Schick, Allan, Why Most Developing Countries Should Not Try New Zealands Reforms, The
World Bank Research Observer, Vol. 13, No. 1, February 1998, pp. 123-31.
Swedish Ministry of Finance, The Swedish Medium-term Budget Framework, 2002.
_____, Budget Statement 2003.
United Nations, The Sex and Age Distribution of World Population, each year.
Wilks, Stuart, Reform of the National Budget Process in Sweden, International Journal of Public
Sector Management, Vol. 8, No. 2, 1995, pp.33-43.
World Bank, Public Expenditure Management Handbook, 1999.
CHAPTER 4
Successful Installation of MTEF to the Korean Fiscal System
by
Seok-Kyun Hur, Korea Development Institute
1. Introduction
The budget process in Korea begins in March when the Ministry of Planning and
Budget (MPB) transmits the Guide to Budget Compilation to the ministries. After
negotiations between concerned parties, the budget draft is approved by the President
and then presented to the National Assembly by October 2nd. The latter is supposed to
approve the budget no later than December 2nd , well ahead of the new fiscal year that
starts on January 1st of next year.
The current budgeting process in Korea is characterized by a lack of medium-term
strategic perspectives on financial management. Macroeconomic forecasts contained in
the Guide to Budget Compilation cover the next budget year only and no mention is
made of the out-years. A medium-term fiscal plan was published in 1999, which
included projections of fiscal aggregates and outlined priorities among the 12 broad
spending categories between 1999 and 2002. It was subsequently revised in 2001.
However, the plan remained largely indicative and, in practice, there was little formal
connection between the plan and the annual budgeting process.
Another characteristic of the Korean budgeting process is the excessive reliance on
the bottom-up approach. The Guide to Budget Compilation contains a very rough
ceiling on the aggregate spending level (for example, one-digit increase in total). It does
not provide a sectoral or ministerial allocation of the budget. As a result, budget
requests for the next year across all line ministries tend to be unrealistically large, often
implying a 20-30 percent growth from the current years budget. The MPB then enters
into a laborious period of negotiation with line ministries. The process becomes more
painful because of the extremely segmented appropriation accounts there are over
2,200 accounts (Se-Hang) and the MPB often negotiates with line ministries on sub-
accounts (Se-Se-Hang) which total more than 6,000.
The current approach has certain merits. The budget authorities have large
discretion over the total size of the annual budget and have used the power to contain
and adjust expenditure growth to changing revenue conditions. To some degree, such
myopia is inevitable in Korea where the socio-economic environment changes quite
rapidly and unexpectedly. In addition, by emphasizing input control and regularity of
budget execution, we find few cases of abuse or misuse of tax money.
On the other hand, the weaknesses of the current approach should not be
overlooked. Without a long-term view on the appropriate level of tax burden, it can
induce an ever-increasing expenditure to accommodate rising demands from various
sectors. It also fosters gradualism in budgeting and can hinder a strategic
reprioritization of spending, exactly when the strengthened control on aggregate
expenditure generates greater necessity for such reprioritization. Line ministries have
little information on the availability of resources in the future and their planning
function is severely limited. Limited planning function in turn reduces the effectiveness
and efficiency of overall public spending.
Recognizing these shortcomings, the MPB is moving toward a medium-term
expenditure framework (MTEF) and a top-down approach in budgeting. In 2003, it
specified the aggregate level of the general account for fiscal year 2004, which takes up
the largest part of central government spending. It also asked line ministries to provide
out-year projections of major program expenditures for fiscal years 2004 and 2005
together with the budget requests for fiscal year 2003. In addition, the MPB designated
a few organizations for a pilot study on top-down budgeting.
These are desirable developments, but we still have a long way to go before seeing
the successful installation of an MTEF in Korea. The MPB did not set medium-term
expenditure targets and did not provide line ministries with a sectoral allocation of the
budget and medium-term forecasts on economic variables necessary to project out-year
spending. Nor were there concrete guidelines on expenditure projections. For example,
dividing the projection into three parts for existing policies, for savings options, and
for new policy initiatives would be useful in identifying line ministries spending
priorities, but it was not attempted.
One big challenge is the extreme compartmentalization and fragmentation of the
budget, which is comprised of one general account, 22 special accounts, and 45 public
funds. These accounts and funds are intricately inter-related with each other through
complicated flow of grants and loans. Even when a sectoral allocation of the budget is
set up, it will be technically very difficult to allocate the sectoral budget to accounts
and funds.
Key questions in setting up an MTEF and introducing a top-down budgeting
process include the following: How can we produce unbiased and credible medium-
term macroeconomic forecasts? How can we overcome shortcomings of a
compartmentalized budget structure, and establish an aggregate fiscal discipline? What
are the guiding principles in setting up a medium-term aggregate expenditure level
and allocating it across sectors and/or ministries? How can we minimize resistance
and objections from line ministries to pre-determined sectoral allocation of spending?
What are the technical details needed in out-year projections of spending? How much
autonomy should be reserved for line ministries in top-down budgeting? What will be
the redefined roles of the budget authorities?
The present study aims to answer these questions and produce operational plans
to introduce an MTEF and top-down budgeting. It will first discuss general issues in
budgeting, overview MTEF as recommended by international organizations and as
practiced in foreign countries, and compare it with the current budgeting process in
Korea. It will then propose an appropriate system for Korea.
The contents of the paper are construed as follows: Section 2 discusses
motivational aspects of introducing MTEF and explores the requisites for successful
installation of MTEF in Korea while comparing the current single year budgeting
system with MTEF. Section 3 explores common pillars of MTEF system and provides
suggestions for successful installation of MTEF in Korea. Finally, section 4 concludes.
Chapter 4. Successful Installation of MTEF to the Korean Fiscal System 43
2. General features of MTEF

2.1. Objectives and means of MTEF
The pursuit of MTEF is multi-purposed: First, it aims to maintain fiscal

sustainability by reigning in excessive expansion of public expenditures51. Under the
current bottom-up budgeting system, conflicting interests among line ministries are
often resolved by aggregate budget expansion beyond what is optimal. In contrast,
under MTEF, there are two different types of ceilings a global ceiling and sectoral
ceilings. These ceilings are determined interdependently by considering medium-term
economic forecasts and sectoral demand estimates for public expenditures over the
matching period. By requiring that the global ceiling be stricter than the sectoral
ceilings, such a two-tier ceiling system can accommodate fiscal sustainability as well as
flexibility of moving public resources across sectors.
Second, MTEF enhances the efficiency of public spending by setting national
priorities. The public sector produces various kinds of public goods in an economy.
Accordingly, a reprioritization procedure is very critical in determining a production
portfolio of diverse public goods within the limited resources. Accompanied with a
top-down allocation system, MTEF enables each line ministry to make a self-motivated
sectoral budget plan based on a longer time horizon. This in turn ensures the time
consistency and effectiveness of government expenditure programs and enhances the
possibility of these programs contributing to long-run economic growth. In this
context, the role of top-down resource allocation cannot be emphasized enough.
However, the importance and complementary role of top-down allocation with respect
to MTEF does not necessarily imply that top-down allocation is the only means of
MTEF and the bottom-up approach is an absolute vice to avoid. Rather, a successful
MTEF relies on both approaches of allocation. Of course, physical resources and
information on macro economic forecasting should be delivered from top to bottom. In
contrast, critical information for allocation of fiscal resources, such as production
efficiency and demand quantity of specific public goods (or projects), should be
transferred from bottom to top. Thus, a key to successful introduction of MTEF lies in
finding harmony between top-down and bottom-up allocations.
Third, MTEF strengthens the function of the fiscal management system as an
automatic stabilizer. According to traditional public finance theory, one of the roles of
the public sector is the stabilization of an economy. The role of automatic stabilizers is
critical in that it smoothes out the production stream of public goods and reduces
swings in business cycles by absorbing shocks, which otherwise would be directly born
by private entities. Figure 4-1 shows how a well-designed MTEF reflecting the medium
term economic prospect can provide insurance to the economy by changing fiscal
positions (deficit->surplus in the figure) in the course of a business cycle52. However,
the power of stabilizers is limited because most of the aggregate shocks arising in the
economy cannot be hedged. In other words, it means that complete stabilization is not
achievable without affecting the fiscal balance.
51 The idea of MTEF was devised originally in countries (e.g. Sweden and UK) suffering from heavy
government debt. However, due to the heterogeneity of the situation each country is faced with, the
implementation of MTEF in practice varies from a country to another.
52 Even though a result differs by the time spans of data, the current year-by-year system is shown to be
pro-cyclical. We suspect that the combination of the year-by-year system and the strong emphasis on fiscal
consolidation causes the pro-cyclical feature.
Figure 4-1. Business cycle and fiscal balance
Boom Expenditure
Surplus
Revenue
Deficit
Depression
Table 4-1 describes the objectives of MTEF and matches them with relevant policy
components in MTEF. In addition we provide analogies of MTEF to a consumers
intertemporal maximization program under uncertainty. Despite fundamental
differences between the two, there are enough existing similarities to deepen our
understanding about the MTEF.
Suppose a consumer with certain duration of life lives in a world full of
uncertainties. Then, evidently his/her economic decisions should concern the
following three dimensions: time, states, and sectors. To rephrase, he/she has to choose
(1) how much to consume today instead of consuming tomorrow, (2) how many apples
to consume instead of pears today, and (3) how much more to consume tomorrow in
case of an event in comparison with the case of another event. Interestingly, each
combination of an objective and a policy tool is analogous to each dimension of the
consumers decisions.
Likewise, MTEF defines the functions of fiscal activities in a multi-dimensional
space: First, MTEF is a mechanism through which subsidiaries of a government are
encouraged to compete with others by demonstrating cost efficiency and social need
for public goods or services they produce. Second, MTEF provides insurance not only
to subsidiary organizations but also to the aggregate economy. Budget allocation in
consideration of long-term fiscal sustainability will induce precautionary fiscal savings
for a rainy day. Third, budget planning based on the longer term prospect contributes
to improvement in growth potentials through rational allocation of public resources.
Table 4-1. The objectives and their matching policy tools of MTEF
Objectives Policy tools in MTEF and their analogies

z Setting a global ceiling and sectoral ceilings on the
Fiscal Consolidation budget plan based on medium-term economic forecasts
(Cross-time substitution) z Analogous to consumption-savings decision by a private
entity
z Assigning sectoral ceilings based on national priorities
Enhancing Efficiency in Public and encouraging line ministries to allocate resources
Spending within their sectoral ceilings
(Cross-sectional substitution) z Analogous to consumption bundle decision by a private
entity
z Smoothing public expenditures across time (in a business
Automatic Stabilization
cycle) by assigning public expenditures within a
Mechanism
medium- (or long-) term framework
(Cross-state substitution)
z Analogous to precautionary savings by a private entity
Table 4-2. Comparison of single year budgeting and MTEF
Traditional Budget MTEF

(at Ministry level) (3-year-rolling program at Sector level)
AGGREGATE FISCAL Focused on short-term macro-economic Situates short-term macro-economic concerns

DISCIPLINE concerns within a medium-term macro-economic and
(to keep expenditure within the (with international agencies providing sector perspective. Involves building
means) the discipline in many countries). domestic macro-economic modeling capacity.
LINK BETWEEN POLICY, Very weak because policy choices are Policy-making tightly disciplined by resource
PLANNING, AND BUDGETING made independent of resource realities. realities. Thus a much stronger link exists
(Reflecting the governments Thus policy is not sustainable and between policy-making, planning, and
capacity and willingness to spending patterns may not reflect the budgeting. Spending reflects the stated
prioritize expenditure programs) priorities articulated by government. priorities of Government.
PERFORMNCE AND SERVICE Incentives for results in terms of Emphasis is on the delivery of agreed outputs
DELIVERY outputs and outcomes are generally low and outcomes with available resources.
(Relating to operational because emphasis is on input control. Incentives are structured to increase the
performance of all resources Little attention to the predictability of demand for evidence of good performance
human as well as financial) budget funding. (accountability for sector managers for
results).
Consequently, service delivery should
improve.
AUTONOMY OF CREDIT Generally low, because lack of Generally high because of greater discipline
MANAGERS discipline within the traditional budget in setting and enforcing hard budget
framework is translated into detailed constraints plus accountability mechanisms
input controls that makes it possible for managers to be
given more authority to determine how
agreed outputs and outcomes should be
achieved.
In contrast with the current single year budgeting system, the potential advantages
of adopting MTEF are highlighted in Table 4-2. In an ideal situation, in which there is
no information asymmetry, single year budgeting may produce the same level of
outcome as MTEF. However, reality is that information asymmetry combined with
various socio-economic uncertainties persists within a government as well as between
the government (as a whole) and the public. Furthermore, generational inequity cannot
be restrained due to dynamic inefficiency under the current single year budgeting.
Unlike other countries, which have already adopted MTEF, Korea does not have
an apparent fiscal problem at the present. The annual budget is more or less
maintained in balance and the ratio of government debt to GDP is smaller than in other
OECD countries. Inefficiency arising in the process of allocating fiscal resources exists
but is not big enough to attract the publics attention. In such a situation, it would be
very difficult to address the need for introducing MTEF not only to the public but also
to public officials. However, anticipating increasing social welfare spending following
the rapidly aging demographic structure, we have to be reminded that MTEF serves as
a precautionary measure. By taking precautionary measures before symptoms arise,
unnecessary turmoil in the economy could be avoided.
2.2. Requisites and key processes of a well-designed MTEF
In discussing the optimal design of MTEF, the following policy environments

should be taken into account: First, the major participants in this setup encompass the
budgeting authority, legislative body, line ministries (including local governments),
and the public. Among them, most of our discussion focuses on the relationships
between the budgeting authority and line ministries as well as the relationships among
the line ministries. Second, each participant has a heterogeneous motive. For example,
each line ministry has an objective of maximizing the production of (a) public good(s),
which it is assigned to produce.53 On the other hand, the budgeting authority has an
incentive to promote the nations welfare across time and across states. Accordingly,
the budgeting authority tries to enhance the production efficiency of public goods and
find an optimal composition of various public goods to satiate the publics needs. In
such an environment, a Pareto-efficient allocation would be reached in the absence of
information asymmetry and other frictions. However, the reality confirms their
existence. MTEF is a trial to overcome information asymmetry arising in the budgeting
process. In this sense, we have to consider what elements should be included in MTEF
in order to reduce social welfare loss due to information asymmetry.
Government budgets are prepared according to an annual cycle, but to be well
formulated and to contribute to high-quality and sustainable services, they must take
into account events outside the annual cycle; particularly macro-economic realities,
expected revenues, and longer-term needs of programs and of governments spending
policies. This is why annual budgeting cannot be performed properly in isolation but
has to be linked to planning, in the context of a multiyear framework.
A medium-term expenditure framework consists of a top-down estimate of
aggregate resources available for public expenditure consistent with macro-economic
stability; bottom-up cost estimates of carrying out policies, both existing and new; and
a framework that reconciles these costs with aggregate resources. It is called medium-
53 In a principal-agent context, the objective of the line ministry would be assumed to maximize the size of
budget allocated to the ministry rather than the size of production.

term because it provides data on a prospective basis, for the budget year t and for
following years t+1 and t+2.
MTEF is a rolling process repeated every year and aims at reducing the imbalance
between what is affordable and what are demanded by line ministries. MTEF does this
by bringing together policy-making, planning, and budgeting early in the budgeting
cycle, with adjustments taking place through policy changes. It involves building
domestic macro-economic and sector modeling capacity. Also, even if the whole of the
Governments budgeting system is not working well, each sector is better off managing
itself with a medium-term perspective.
A well implemented MTEF should: (i) link the Governments priorities with a
budget within a sustainable spending envelope; (ii) highlight the tradeoffs between the
competing objectives of the Government; (iii) link budgets with the policy choices
made; and (iv) improve outcomes by increasing transparency, accountability, and the
predictability of funding.
In detail, seven major requirements must be considered for MTEF implementation:
z Good macro-economic policies and reliable forecasting: Good macro-

economic analysis and forecasts are needed as a basis for a MTEF. Inaccurate
macro-economic forecasting dilutes the credibility and feasibility of MTEF.
z Adaptable fiscal policy and instruments: The MTEF approach is based on a

strong link between macro-economic and fiscal policy. Plans for future
expenditure must be based on reasonable estimates of prospective resources.
z Reprioritization and reallocation: Behind the move to MTEF is a conviction

that the annual budget by itself is a poor mechanism for shifting resources
from lower to higher priority use. A major function of an MTEF is to provide a
better mechanism for aligning budgets with policies.
z Budgetary discipline: Budget allocations must be based on a hard aggregate

budget constraint derived from what is affordable, and line ministries must
live with their budget allocations.
z Institutional conformity and absence of bias: An MTEF requires a

supportive institutional base; that is to say, one in which the various actors
use the MTEF as a framework within which expenditure decisions are made.
In particular, political decision makers must accept the MTEF as the means by
which resources are allocated.
z Appropriate parameters: Designing an MTEF requires that its parameters be

set. These parameters include the following: the definition of aggregate
expenditure to be used, the relationship between the sectoral breakdown and
the organizational structure of the government, the content of expenditure
envelopes, the appropriate price basis for estimating future expenditures, the
mechanism for its coordination with the annual budget process, and the
degree to which it is to be flexed for different scenarios.
z Transparency: Fiscal transparency and policy transparency improve the

accountability of actors engaged in the MTEF process. Fiscal transparency
means being open to the public about the structure and functions of
government, fiscal policy intentions, public sector accounts, and fiscal
projections. Policy transparency means being open to the public about what
Government intentions are in a particular policy area, which outcomes are to
be achieved, and the costs of achieving these outcomes. Also, transparency
means reporting actual performance with quality of outputs and results
achieved.
An MTEF cannot be achieved without the operation of key government processes

on which it will depend upon. For the MTEF to work, key supporting processes have to
function as in Table 4-3.
Table 4-3. Annual procedure of MTEF
CORE PROCESSES SUPPORTING PROCESSES PURPOSE

Define the aggregate resources Macro-economic analysis, To provide a realistic estimate
revenue forecasting, and of the total resources available
definition of sustainable fiscal in the medium-term to allocate
policy. to spending programs.
Formulate and cost sectoral Spending ministries formulate To show sectoral objectives,
spending plans sectoral expenditure programs programs and activities, and
(SEPs) their costs.
Politicians and other decision-
Reconcile available resources To reach agreement on
makers reconcile top-down
with sectoral spending plans medium term expenditure
constraints with bottom-up
programs.
spending demands.
Set medium-term sectoral On the basis of relevant data, To communicate to ministries
allocations decision-makers allocate the a sectoral expenditure policy
aggregate resources to sectors. constrained by aggregate
resources.
Announce sectoral expenditure Formulation of annual budget. To ensure that the budgets
limits for year one of the prepared by ministries reflect
MTEF agreed sectoral expenditure
programs.
Ensure that budget execution Accounting, reporting, and To prevent excessive deviation
is in line with budget intentions expenditure controls are used from the annual budget and
during the execution of the MTEF
annual budget.
Ensure that desired results Incentives for civil workers to To align civil workers and
are achieved perform. Ex-post audit and politicians incentives with
evaluation; public goals.
An MTEF involves a radical change in the business of budgeting. Consequently,

without political commitment, it has little chance of succeeding. Evidence also suggests
that a comprehensive approach (entire government) is preferable to introducing an
MTEF incrementally, starting with select sectors and then expanding coverage for the
entire government. This is partly a question of data gathering, analysis, and reporting;
and partly of people who can both supply and use the information, and the
institutional mechanisms for coordinating the efforts of participants.
Most of the necessary changes for successful MTEF do not require high-tech
solutions, but rather common sense and perseverance. In this sense, political support
would be critical in order to succeed in introducing MTEF once and for all. The lack of
political support is of concern because budgeting is still seen as a technical exercise
mainly driven by the Ministry of Planning and Budget (MPB). Another reason is that
politicians seem unaware of the benefits of the new process, and still rely to a large
extent on the old, incremental methods to obtain funding. Finally, the incentive for
good budget preparation remains limited, because much is lost in the implementation,
which is still driven by cash rationing.
2.3. Anticipated outcomes from MTEF
The MTEF approach stresses that expenditure management is about appropriate

policies in the medium-term, rather than about cash management in the short-term.
Where the available cash is the strongest influence on expenditure, management in the
normal meaning of the term becomes impossible. MTEF provides more predictable
program funding and therefore better management. From the adoption of MTEF, we
expect the following outcomes:
(1) Improve macro-economic balance by developing a consistent and realistic

resource framework.
(2) Improve the allocation of resources to strategic priorities among and within
sectors.
(3) Increase the commitment to predictability of both policy and funding so that
ministries can plan ahead and programs can be sustained.
(4) Provide line agencies with a hard budget constraint and increased autonomy,
and increasing incentives for efficient and effective use of funds.
Implicit in the above are two further gains: (i) to improve the linkage between
annual budgeting and medium-term considerations, such as investment plans,
borrowing capacity, changing spending policies, and priorities, and (ii) to provide
relevant information to political decision-makers on cost implications of expenditure
policies.
3. Specific issues for the implementation of MTEF in Korea

In the previous sections we discussed the rationales for the adoption of MTEF and
its general working mechanism. Thus, most of the claims suggested there are valid for
any countries. This section covers specific issues regarding the design of MTEF for
Korea. We emphasize the word specific for the following reasons. First, a more
detailed description of MTEF system is necessary. Second, country specific factors
(Korean factors) should be counted.
3.1. Coverage of MTEF
MTEF should cover the entire consolidated budget in GFS. It means that all the
resources under the discretion of the government need to be reviewed and prioritized
for spending. This comprehensive budget coverage is particularly important
considering that MTEF tries to improve resource allocation in line with the
governments priority.
3.2. Time horizon of MTEF
Usually, the time horizon of MTEF ranges from 3 to 4 years. The longer the
horizon, the less responsive the annual budgeting is to changes in the socio-economic
environments. In contrast, a shorter time horizon is not consistent with the basic
motive of MTEF, the provision of insurance to line ministries. Accordingly, an optimal
time horizon should be determined which lies in-between. In addition, the Korean
economy is exposed to various risk factors, internal and external, as evidenced in the
past. Therefore, it is recommended that MTEF of Korea cover a three-year time horizon.
Table 4-4. Time coverage of MTEF
Time Horizon 2-year 3-year 4-year 5-year

Canada, Czech, Netherlands,
Country Italy France, Germany, New Zealand USA
Sweden, UK Mexico
3.3. Macroeconomic forecasting
3.3.1. Horizon of macro-economic forecasting

The time horizon of macro economic forecasts should not be shorter than the time
horizon of MTEF. In cases of other countries, forecast horizons are usually longer than
the time horizon of MTEF, even though officially, the budgeting authority announces
macro forecasts only for the time horizon of MTEF. Under the premise of a 3-year
MTEF, macro forecasting 4 to 5 years ahead would be appropriate.
3.3.2. Who forecasts macro variables?

In case of US, the Office of Management and Budget (OMB) is mainly responsible
for forecasting, which OMB has a separate division for. The forecasting figures are
cross-checked by the Department of Treasury and Council of Economic Advisors. In
contrast, the Department of Finance in Canada collects economic prospects from
credible private entities and takes arithmetic averages of the numbers reported.
Considering that MPB of Korea does not have an in-house macro-economic
research division whereas there are many publicly-funded research institutes, the
adoption of a Canadian style approach depending on government funded research
institutes would be preferable.
3.3.3. How should the forecasts be reflected on MTEF?

Conservatism prevails. At the stage of making a medium-term expenditure plan,
conservatively discounted economic forecasts should be used. Cautioning on the side
of prudence would provide a buffer zone to the fiscal system.
3.4. Harmony of Top-down and Bottom-up in MTEF
As mentioned earlier, top-down and bottom-up approaches do not conflict each

other within MTEF. An optimal mix of top-down and bottom-up approaches should
include the following transfer processes of information and physical resources.
(1) Top-Down: information flows on macroeconomic forecasts and national

priorities, broadly itemized budget allocation and risks of performance
(2) Bottom-up: information flows on performance measurement and demand or
social need for specific public good and risk of budgeting and debt
management
(3) These duplex exchanges of information and resources are made at legally set
(or publicly guaranteed) opportunities.
3.5. Categories of Spending in MTEF
It is recommended that classification and presentation of budget estimates under

MTEF be presented in a way to link expenditures to objectives, outputs and activities
clearly. This result-oriented budget structure is, however, not a necessary condition to
introduce MTEF. A country may opt to apply existing budget classification and charts
of accounts for better sectoral allocation. Once this stage of MTEF is internalized, the
government can move on to improve budget classification and charts of accounts to
produce sufficient information in determining policy and program costs, which will
further the benefits of efficient resource allocation and financial operation.
MTEF, in its full-fledged form, should provide information on the cost of
individual programs and, ultimately, its overall impact. Under this situation, it will
serve as a basis for decision-making when programs compete for funding and
aggregating all selected program expenditure to make a sectoral budget proposal.
In many countries, including Korea, existing budget classification structures, chart
of accounts and legal requirement may not allow line ministries to be privy to this
information. If so, sectoral ceilings are set first by the central budget office using a
macroeconomic framework, and then line ministries decide on how to allocate
resources to their programs.
It is recommended that if the initial MTEF exercise becomes internalized, the
government should consider improving/modifying the budget structure and the chart
of accounts in order to build a stronger link between resources and priorities.
3.6. Rules for Selecting a Global Ceiling and Sectoral Ceilings.
3.6.1. Who sets the ceilings?

Ultimately, overall size of expenditure and sectoral/ministerial limit should be
decided/agreed at a ministerial level meeting, where the central budget office provides
a medium-term expenditure plan, usually considering multiple scenarios for different
economic and social factors.
It is important to recognize that the medium-term expenditure plan, most likely in
a multiple number of (contingency) scenarios, are technical inputs to decisions that
need to be made at the cabinet meeting on allocation between sectors. Thus, the
expenditure framework and scenarios are not the end themselves, but are tools for
improving decision making.
3.6.2. What factors determine ceilings?

As for the global ceiling, the potential GDP growth rate would be the most
reasonable guideline. However, precise estimation of the potential GDP is not an easy
task. Thus, the government revenue forecast is recommended as its proxy.
For the measurement of fiscal affordability (sustainability), the current government
debt structure should be considered as well as the revenue forecast. Most commonly
used measures for fiscal sustainability are ratios of government debts/GDP and fiscal
deficit/GDP. Since the integration of fiscal deficit over time is equal to government
deficit, these two ratios encompass the same amount of information mathematically.
According to the IMF, government debt is defined as follows:
Government debt = Central government liabilities + Local government liabilities

- Local liabilities to the central government
The above definition of government debt does not include de facto government
liabilities such as contingent ones. A true indicator of government debt should be the
sum of the above IMF definition and the following:
Non-government debt = Guarantee of Payment + Central bank debt

+ State owned companies + Public pension
The national budget accounting system varies from country to country.

Accordingly, it is very difficult to maintain comparability among data provided by
governments. For this reason, the IMF measure of government debt is widely used for
cross-country comparison, though it may underestimate the actual amount.
As an extra precaution, MTEF of Korea should adopt the measure of fiscal
sustainability that includes non-government debt going beyond the IMF criterion. In
addition, the duration of revenue forecasts and government debt structure should be
matched in order to hedge against interest rate and liquidity risks. After undergoing a
thorough examination of the revenue forecast and current government debt structure,
the MPB could determine global ceilings for the next three years.
Also, determining sectoral ceilings is not independent of the factors influencing the
global ceiling. In addition to such macro economic factors, sectoral ceilings should
reflect national priorities and information on sectoral production efficiency.
In regard to the procedures on determining sectoral ceilings, the timing and
number of communications between the budgeting authority and line ministries is
important. These procedural issues do not have a right or wrong answer. Depending
on the country specific characteristics (e.g., schedules for congressional sessions, the
organization of a government body, and its relation to the legislative body), the types
of communication or negotiating channels within the government should be defined.
Instead, more important is what information the involved parties should be equipped
with at the time of discussions over sectoral ceilings. Frictionless flow of information
should be encouraged and, at the same, conflicting demands among the line ministries
should be reconciled through the voluntary demonstration of social worthiness and
efficiency.
In regard to the enforceability of the ceilings, we recommend that global ceilings
be binding at least for the coming fiscal year while the sectoral ceilings may be
indicative. In other words, this implies that voluntary exchanges of sectoral budget
allocations among line ministries with/without mediation of the budgeting authority.
There might be concerns that allowing voluntary exchanges among the line ministries
will make the process of budget allocation deviate from national priorities. However, if
it is complemented by Performance Based Budgeting (PBB), which assigns more weight
to a program with higher national priority, the exchange scheme will ensure more
dynamic flexibility of line ministries.
Figure 4-2. Cross-sectoral procedure for MTEF
MPB Demand Ministry Ministry Ministry Ministry

Estimates A B C D
Sector 1 Q(A,1) Q(B,1) Q(C,1) Q(D,1)

Total Revenue
Global Ceiling
Sector 2 Q(A,2) Q(B,2) Q(C,2) Q(D,2)
Sector 3 Q(A,3) Q(B,3) Q(C,3) Q(D,3)
Sectoral Ministry Ministry Ministry Ministry

Ceilings A B C D
Sector1 R(A,1) R(B,1) R(C,1) R(D,1)
3.7. Time consistency of ceilings changeable or unchangeable
Under the assumption that the annual budget is prepared through MTEF process,
the previous years MTEF plan (t) does not necessarily bind the annual budget or the
MTEF of this year (t+1).54 Expenditure perspectives of two or three years included in
MTEF should be seen as guidance. Since situations almost always occur unexpectedly,
it would be optimal to adjust the original plan following the track of events. For
54 In our prior knowledge, Brazil is the only country, which strictly keeps the ceilings set by MTEF for the
whole matching time horizon.

example, in the event of a natural disaster such as a typhoon, there would be a huge
amount of unexpected public expenditures.
On the other hand, if a change in the ceiling is permitted frequently and flexibly,
the time consistency or credibility of MTEF would be affected bringing about a similar
outcome to that of the year-by-year approach.
Figure 4-3 describes the multi-year process of MTEF. C represents the MTEF
ceilings determined for the periods (t, t+1,t+2) at t-1 whereas A and B are the MTEF
ceilngs determined for the periods (t+1, t+2,t+3) at t. A and B differ by the state of an
economy from period t to t+1. Realization of different paths makes A and B deviate
from the original plan C. In order to maintain balance between flexibility and
consistency, we have to determine an acceptable degree of variation, which sets a
boundary separated by space between A and C or B and C.
Figure 4-3. Cross-time procedures of MTEF

A

B
t t+1 t+2 t+3
In practice, there are several ways to maintaining balance between flexibility and
consistency: (1) to allow adjustments among the sectoral ceilings while maintaining the
global ceiling, (2) to assign different margins of adjustment between the global ceiling
and the sectoral ceilings or among the sectoral ceilings, and (3) to decompose the
budget items into mandatory spending and discretionary spending and allow different
margins of adjustment to them. Each of the above alterations has pros and cons in
comparison with others. Hence, it is in the hands of political decision-makers to choose
one of them.
Table 4-5. Fixed vs. flexible MTEF

Fixed MTEF Flexible MTEF
UK, US (till 2002), Canada, France, Germany,
Country
Sweden, Netherlands New Zealand Czech, Mexico
Source: OECD (2003).
3.8. Performance Based Budgeting and MTEF
One of most important modules in MTEF is the provision of diverse channels for
information flow, the usefulness of which depends on its precision. Accordingly, a
logical tool for justifying a certain program, such as performance evaluation, would be
essential in MTEF. However, heterogeneity (or diversity) of public programs and
unaccountability of their hidden or intangible effects makes inter-program
performance comparison impossible.
3.9. Performance Measures for MTEF
MTEF itself is a fiscal program and is an object of performance. The two biggest
objectives of MTEF are (1) to maintain fiscal sustainability and (2) to enhance the
allocative efficiency among various sectors. However, due to the difficulty of finding a
right performance measure for inter-program comparison, in practice, performance
evaluation for MTEF is limited in checking fiscal sustainability. MTEF is not a legal
code to keep, once put into practice. Rather, MTEF is an apparatus to be evolved
reflecting past track records. The provision of performance measures for MTEF will
complete the evolutionary cycle of MTEF.
4. Concluding Remarks
The budgeting process is a collective action game. Various agents participate in the
process in the name of social welfare with hidden heterogeneous incentives. From the
standpoint of a principal, who devolves his/her decision right to the agents, optimal
mechanisms designed to coordinate interactions and conflicts among the agents would
be very critical. However, at the implementation stage informational asymmetry may
work against achieving social optimum.
The purpose of our research is to introduce a Medium-Term Expenditure
Framework and evaluate the pros and cons following its adoption to Koreas current
budgeting process. MTEF is not a perfect solution guaranteeing the efficiency of the
public sector. However, compared with the current system, MTEF is advantageous in
the following sense:
(1) MTEF is less reliant on discretionary factors and is rather rule-based. Rule-
based allocation of public resources ensures system predictability to
participants. Accordingly, the participants can pursue fiscal programs in a
more consistent way.
(2) MTEF system encourages efficiency-based competition among the conflicting
agents leading to a pseudo-market allocation. In contrast, allocation under the
current system does not provide a formal channel, through which the agents
can compete against others.
For the successful installation of MTEF in Korea, there are several points to
ponder. A substantial portion of this paper is devoted to the examination of various
issues, especially procedural ones. References to foreign country cases makes us realize
that at the stage of implementation MTEF should adapt the socio-economic
environments of an economy it is applied to. Accordingly, different schemes can be
adopted from country to country depending on the country-specific factors. Though
seemingly heterogeneous, diverse types of MTEF are designed to attain a common
goal, the provision of information channels and mechanism for early resolution of
uncertainty within a government, which is a necessary condition for Pareto efficiency
in Welfare Economics. Therefore, in designing an optimal MTEF for Korea, we have to
always ask ourselves whether a certain feature or a procedure under consideration will
contribute to reaching a pseudo market allocation of the annual budget.
References
Anderson, B., Budgeting in a Surplus Environment, OECD, 1999.

Bayoumi, Tamim, The Japanese Fiscal System and Fiscal Transparency, in (eds.) Bijan B.
Aghevli, Tamim Bayoumi, and Guy Meredith, Structural Change in Japan: Macroeconomic
Impact and Policy Challenges, 1998, Washington D.C.: International Monetary Fund.
________ and Barry Eichengreen, Restraining Yourself: The Implications of Fiscal Rules for
Economic Stabilization, IMF Staff Papers, Vol. 42, No. 1, March 1995, pp. 32-48.
Brusewitz, U.H., The Swedish Medium-Term Budget Framework, International Forum on
National Budget System, KDI School of Public Policy and Management, 2002.
Committee on the Budget of the United States Senate, The Congressional Budget Process: An
Explanation, 1988.
Congressional Budget Office (CBO), The Budget and Economic Outlook: Fiscal Years 2004-
2013, 2003.
Dorotinsky, Bill, Developing a Medium-term Expenditure Framework, Chapter 1 in this
volume.
Koh, Young-Sun, Budget Structure, Budget Process, and Fiscal Consolidation in Korea, KDI
Working Paper, 2000.
________, Introduction of MTEF, KDI Policy Forum, 2002 (in Korean).
Ministry of Finance and Economy, Consolidated Fiscal Balance of Korea, each year
Ministry of Planning and Budget, National Resource Allocation Improvement Plan (Draft),
2003 (in Korean).
OECD, Budgeting in Canada. Puma/SBO(99)5/Final, 1999.
OECD, Budgeting in the United States, Gov/Puma/SBO(2003)16, 2003.
Potter, B. and J. Diamond, Guidelines for public expenditure management, IMF, 1999.
Regeringskansliet, The Central government budget process, 2003.
Swedish Riksdag, The Riksdag and the Central Government Budget, Factsheets N0.7, 1998.
Von Hagen, Jurgen and Ian Harden, Budget Processes and Commitment to Fiscal Discipline,
IMF Working Paper WP/96/78, July 1996.
World Bank, Public Expenditure Management Handbook, 1999.
World Bank/OECD, Survey on Budget Practices and Procedures, 2003.
CHAPTER 5
MTEF in Sweden
by
Allan Gustafsson, Sweden
1. Background
1.1. Constitutional arrangements
Sweden is a constitutional monarchy and the King is the Head of State. His duties
are, however, mainly of a representative and ceremonial nature. The Constitution,
based on the principles of popular sovereignty, representative democracy, and
parliamentarism, thus begins by stating that all public power in Sweden proceeds
from the people.
Sweden has a unicameral parliament, the Riksdag, with 349 seats. Elections are by
proportional representation. The electoral system is designed to ensure a distribution
of seats between the parties in proportion to the votes cast for them nationally. A party
must gain at least 4 percent of the national vote to qualify for representation. Elections
are held every four years. The Riksdag, upon a proposal by the speaker of the Riksdag,
appoints a Prime Minister who is given the task of forming a government. The Prime
Minister chooses his/her ministers and together they make up the Swedish
Government. The Government is accountable to the Riksdag.
The Swedish Constitution consists of four separate documents. Of those, the
Instrument of Government (Regeringsformen), together with the Riksdag Act
(Riksdagsordningen), which occupies a position midway between constitutional and
regular statute law, set the overall framework for the budget and the budget process.
The structure of the budget and the budget work is regulated by a Budget Act which is
a regular statute law.
1.2. Reforms
The present structure and the procedures for preparing and deciding on the
budget are the result of a series of reforms initiated and implemented in a very short
time in the middle of the 1990s. The impetus to the reform came from the very deep
crises affecting the national economy and the government finances at the beginning of
the 1990s.
When analyzing the crises, it was found that one of the contributing factors to the
crises was a very lax budget process. One of its consequences was a seemingly
unstoppable growth in the relative size of the public sector and thus the tax burden.
Furthermore the Swedish public finances had become extremely sensitive to the swings
of the business cycle because of the dominant role of transfers to households in
government expenditure. As long as these swings were moderate, the opposing effects
on government revenue and expenditure served as automatic stabilizers. External
shocks of the magnitude experienced in the early 1990s, however, took the economy
out of bounds. The central government deficit exploded and reached 12 percent of
GNP in 1993. The gross financial debt peaked at 77 percent of GNP in 1996.
In order to reduce the swings in the budget balance but also and perhaps more
importantly to create a semi-automatic mechanism for containing the growth and
eventually reducing the size of the central government, the traditional incremental
bottom-up approach to budgeting was replaced by a very strict top-down process.
1.3. General characteristics of the Swedish public financial management
Subsequent to the reforms of the mid-1990s, public financial management now has
a three-year perspective within a top-down framework. This three-year perspective
and the top-down approach apply to budget preparation as well as to the decision-
making process in the Riksdag.
Execution of the budget is fully decentralized to the agencies responsible for
executing the some 500 appropriations. Appropriations for the administration of the
agencies are given as block appropriations, i.e. the agencies are completely free to
decide on the input-mix required to achieve the objectives established for its
operations. Objectives, targets and general reporting requirements are stated in the
yearly letter of appropriation (see further below) .
For many agencies, central financing from the Treasury is supplemented by or, in
some cases, completely made up of revenue accruing directly to them in the form of
fees and sales revenue. To give additional financial flexibility, the agencies have the
right, up to a limit, to carry over unused appropriations to the following year. They can
also, within limits, borrow against next years appropriation. Investment is financed by
internal loans issued by the Swedish National Debt Office. Payments are made directly
by the agencies and the accounts are kept according to generally accepted accounting
practice i.e. the type of accounting used private enterprises. The main instrument
against overspending is agency self-discipline. The accounts of every agency are
audited by the Swedish National Audit Office (Riksrevisionen). The audited accounts
are presented in a yearly report together with a review of activities of the past year put
in relation to the objectives and targets set in the letter of appropriation.
2. Foundations of budget preparation: level 1

2.1. Actors in budget formulation
Budget formulation involves, in one way or another: the Riksdag, the Office of the
Prime-Minister, the Ministry of Finance, the line ministries55 and the agencies.
2.2. Macro-forecasting
Macro-forecasting is mainly carried out by the Department for Economic Affairs

(DEA) at the Ministry of Finance. The work carried out in this department builds on
work by and material from the National Institute for Economic Research (NIER), the
Department for National Accounts at Statistics Sweden, the National Tax Board, the
OECD and other international institutions. For the component that refers to public
sector consumption, DEA receives input from agencies and line ministries channeled
55 The Office of the Prime Minister and the Ministries together make up the Government Offices.
Chapter 5. MTEF in Sweden 59
through the Budget Department. The Financial Markets Department within the
Ministry of Finance provides input on financial variables, notably interest rates.
Besides providing input to the forecasting work of the DEA, NIER produces its
own independent macro-economic forecasts. Macro-economic forecasts are also
produced by the major banks and other private sector institutions. The existence of
multiple forecasters serves as a check on the forecasts made by the DEA.
Short-term macro-forecasts are based on a set of partial simulation models but
outcomes of those are adjusted on the basis of judgment backed up by a set of quick
indicators. Medium-term models rely more heavily on models. DEA uses a stripped-
down version of a partial equilibrium model developed by the NIER. Work is on-going
at the NIER to develop a more sophisticated scenario-type model.
The process at DEA generates forecasts for a number of variables that are
subsequently used as inputs in the forecasting of government expenditure, for
example: nominal GNP, nominal GNI, household consumption, CPI, interest rates and
employment.
DEA produces four to six complete forecasts during a year. Two of those are
published those on which the Spring Finance Bill and the Budget Bill are based.
The ambition is to produce as accurate forecasts as possible. The political level at
the Ministry of Finance has largely refrained from trying to influence the forecasts
knowing well that any manipulation would jeopardize its credibility. The existence of
alternative forecasts those of the NIER and private sector institutions also serve as a
check on the official ones. Contrary to what is done in some other countries, for
example Canada, there is no systematic conservative bias introduced into the forecasts.
The need for a buffer to absorb external shocks is instead met by an explicit budget
margin on the expenditure side of the budget.
2.3. Revenue forecasting
Revenue forecasting is primarily the responsibility of the Fiscal Affairs

Department at the Ministry of Finance. In-year revenue forecasting is also done by the
National Financial Management Authority (ESV). Detailed information on debited
taxes and tax payments is provided by the National Tax Board.
Forecasts are made in accruals as well as in cash terms. The forecasts are made
with the help of a set of partial models one for every type of tax.
2.4. Debt policy
The management of the central government debt is the responsibility of an

independent agency, the Swedish National Debt Office (SNDO). Its objective is to
borrow at the lowest possible cost, taking into account risk. General guidelines for the
management are established by the Government on the basis of proposals by the
SNDO. Proposals for the coming year are presented to the Government in October
every year. The proposal covers a three-year period, year two and three being
preliminary. Once decided, the Board of the Debt Office sets detailed guidelines for
next years work.
For 2004, for example, the Government reiterates its long-term aim to reduce
foreign currency debt as a percentage of total central government debt. The medium
target is for foreign denominated debt to amount to no more than 20 percent at the end
of 2006. For 2004 the target for amortization of foreign debt is SEK 25 billion. However,
the Debt Office has the mandate to deviate from this target by +/- 15 billion should
conditions in the market change. The Debt Office is furthermore instructed to increase
the proportion of inflation-linked krona-denominated loans. The benchmark for the

average duration of the nominal krona and foreign-currency debt is to remain
unchanged at 2.7 years.
2.5. Expenditure forecasting and fiscal envelope
As a starting point for every budget round, a baseline is estimated. The base-line,
termed consequence assessment, is an estimate of what would be the outlay for every
single appropriation and, thus, for the aggregate budget. The baseline presumes no
change in policy and no change in activity volumes except for those that follow from
decisions made in earlier budget rounds. The baseline thus only takes into
consideration changes in factors that are beyond the immediate control of government,
such as changes in prices, wages, exchange rates, and interest rates. To increase
accuracy and fairness, a number of more focused price indices not just CPI are used.
For the appropriations for normal administrative costs of agencies, three indices are
used: for wages, for rents and for other operational costs. From the thus reflated initial
baseline estimate is then deducted a variable percentage to reflect presumed efficiency
gains. The efficiency factor, which is based on national accounts statistics, concretely a
moving average for the preceding ten years, has in recent years hovered around 1.5
percent.
For appropriations like defense and infrastructural investment other types of
specialized indices are used.
The models used to produce the baseline estimates vary a great deal in complexity.
Some are quite simple doing nothing more than reflating last years appropriation
adjusted for presumed efficiency gains. Other models, in particular those referring to
some of the entitlement programs, are more complex. Some simulations use a sample
of individual household data as its basis; others more aggregate data. All models are
partial but estimated feed-back effects are incorporated into successive forecasting
rounds.
At present, the responsibility for the different models is largely decentralized to
line ministries and sometimes to the responsible agencies. At the beginning of the
exercise, Ministry of Finance, however, makes available a set of parameters
assumptions regarding the evolution of the different price indices, GDP growth rates,
unemployment etc. to the forecasting institutions. Sometime in the future, the models
will be incorporated into the government-wide budgeting and reporting system that is
in the process of being developed.
3. Allocation: level 2
3.1. Global expenditure ceiling
Global expenditure ceilings are set for three years on a rolling basis. That is, every
year a global ceiling for year t+3 is established. Beginning 2002, the Governments firm
proposal is presented in the Budget Bill and voted on by the Riksdag in the fall.56
56 During the first years after the introduction of the new budget process, it was presented in the Spring
Finance Bill and voted on separately. The rationale for presenting and voting on the global expenditure ceiling
separately in the spring was to put focus on the longer-term fiscal issues. However, it, in practice, led to two
budget processes leaving very little time for follow-up of government programs and the proposal and the
The Riksdags decision on the expenditure ceilings are, strictly speaking, not
legally binding but in practice they have held firm.57 The over-all ceiling has become
almost sacrosanct, while its break-down into expenditure ceilings is more indicative
and finds its final form only for the up-coming year.
The process in which the global expenditure ceiling for year t+3 is set is not
entirely straight-forward, being partly technical, partly political. The starting point is a
forecast of central government revenue based on the existing tax structure and existing
tax rates plus any already decided on changes in the latter. From this revenue estimate
is deducted the business-cycle adjusted surplus target and the target budget margin in
order to arrive at the global expenditure ceiling. The present surplus target is an
average of 2 percent of GDP over the business cycle and the target budget margin for
year t+3 is 1.5 percent of total expenditure. If year t+3 is forecasted to be a boom year,
the surplus target should in principle be adjusted upwards and vice versa. In
mathematical notation:
C t+3 = R t+3 + Tt+3 t+3
where C = ceiling, R = revenue, T = taxes, and = surplus target.
Underlying this approach to the expenditure ceiling coming in from the revenue
side is, or at least was, a concern about the tax burden in general and the marginal
effects in particular. The concrete effect was, as the Swedish economy recovered from
the crises of the early-mid nineties, a dramatic lowering of the proportion taken up by
general government in the economy: from 70 percent to 55 percent measured on the
expenditure side. In the last three years, this downward trend has, however, leveled off
at about 55 percent.
The revenue projections are, however, not end of story. Expenditure is also
forecasted making in effect Tt+3 in the formula above to some extent endogenous.
Should the projections of growth in the economy, and thus of revenue, point to a
surplus beyond the desired surplus target and the target budget margin, the
Government can decide to lower taxes and/or increase expenditure. If the opposite
would obtain, expenditure will have to be curtailed and/or taxes increased.
The budget department prepares a number of scenarios which are presented to the
Minister of Finance and the Prime Minister who in the end make a political choice as to
the exact level of the expenditure ceiling.
One technical aspect of the expenditure ceilings is worth pointing out; they are
expressed in nominal terms. A nominal ceiling on government expenditure adds
credibility to the inflation ceiling established by the central bank. It is particularly
important in a country, like Sweden, where the public sector occupies such a large
share of the national economy.
The total expenditure ceiling for year t+2 may, in principle, be altered but the
Government has so far chosen to confirm the one established the year before, save for
any adjustments for technical reasons.
3.2. Expenditure area ceilings
discussion on the global expenditure ceilings have now been shifted to the fall.
57 Contrary to the Latin tradition, budgets in Sweden do not have the status of laws.
The ceiling for t+2 becomes the starting point for the process of preparing the
detailed budget the subsequent year. During the first years after the budget reform, the
Government prepared and Parliament decided on indicative expenditure ceilings in the
spring. This is no longer the case. Nevertheless, the Government, in its March retreat,
decides internally on the expenditure area allocations and those allocations serve as
guidelines for the preparation of the detailed budget proposal, i.e. the proposed
allocation to the approximately 500 appropriations.
The starting point for the allocations between the expenditure areas is the baseline
calculations or consequence assessments described above. The line ministries also
present budget requests on the basis of requests prepared by the agencies under their
jurisdiction. In these requests the ministries may argue for higher allocations to
particular areas and they are not, at least not in practice, constrained by the
expenditure area ceilings decided on by Parliament in the previous fall. What ensues is
a rather intensive negotiations process involving the line ministry and the Budget
Department within the Ministry of Finance and to the extent necessary the Minister of
Finance and even the Prime Minister. However, once the expenditure ceilings are set
by the Cabinet in March, they, in principle, have to be respected when making the final
allocation to individual appropriations presented in the Budget Bill. Some adjustment
may have to be made for technical reasons or because some fundamental assumption,
for example, on employment/unemployment, has to be revised.
The budgeting process is thus fundamentally incremental, albeit within the
confines of the global and the expenditure area ceilings. There is definitely no zero-
based budgeting taking place and there is no systematic attempt at estimating the
return on government spending in different areas. Nor is there really any systematic
use of information on results and of all other types of information that agencies are
required to feed back to the Government Offices and to Parliament.
Attempts at basing the budget process on and around a more streamlined
programmatic structure for the budget, allowing a systematic feed-back and analysis of
outputs and outcomes, have, at least for the time being, got bogged down; a new
programmatic structure has been introduced and should, in principle, be used for
reporting, but budgeting and the allocation of resources is still made in the old
structure severely limiting the usefulness of the reform.58
3.3. Conflict resolution and consensus building
One important factor which facilitates decision-making in Government is that all

decisions are taken by the Cabinet as a collective body. There is thus less scope for
bilateral negotiations between line ministers and the Minister of Finance. Just as
obtains for the decision-making process in Parliament, any minister arguing for an
increase of spending in his or her area must be ready to confront his or her fellow
ministers with proposals for where the funding for it in the form of reducing
spending in competing areas should be found.
Most decisions of a more detailed nature are taken at a lower level in the
Government Offices, either within line ministries or in negotiations between the
Budget Department and the line ministries. Despite the new top-down process, the
58 The duplication of structures will undoubtedly prove untenable and in time a replacement of one for the
other or some type of merger of the two will necessarily have to be engineered. As a trial balloon an accrual-
based shadow budget for 2004 will be presented for information purposes. The shadow budget will be broken
down by policy areas rather than by expenditure areas.
Budget Department has a tendency to get involved in rather low-level issues, issues
that presumably, given the existence of expenditure areas ceilings, could be handled
within the respective line ministries. However, well established procedures for
handling differences of position do exist; outstanding issues are raised to successively
hierarchically higher levels if need be to the Prime Minister until they are resolved.
3.4. Effect of ceilings
The top-down approach to budgeting has not completely closed the window for
intra-governmental negotiations. Although line ministries, in principle, ought to
respect the ceilings established by the Riksdag, they often do use their budget proposal
as a vehicle for demands for additional funding. These proposals can, however, not be
wildly out of line or they would not be taken seriously. The ceilings thus act as a
restraint but not as an absolute constraint on the agency and line ministry proposals.
One could also argue that, for want of a more systematic and structured way of
assessing the effectiveness of government spending, in particular of larger
programmatic aggregates, agencies and line ministries should have at least one formal
channel for arguing their case.
3.5. Final budget allocation
The proposed allocation between expenditure areas and within expenditure areas
is finally established in the Budget Bill. The final inter-expenditure area allocations
correspond very closely to those decided on in the March retreat by the Cabinet. The
difference between the two has, in recent years, been less than one percent and almost
solely in entitlement programs where changes in the macro environment and/or in
other external factors have necessitated a review of the forecasts. In principle these
changes normally increases should be absorbed by a reallocation within or between
expenditure areas. In practice they have, in most cases, led to a reduction of the budget
margin.
The differences between the provisional inter-expenditure allocations presented
for t+2 in the Budget Bill one year earlier and the eventual final allocations have tended
to be considerably bigger. In the two most recent Budget Bills, the importance of the
expenditure area breakdowns for t+2 and for t+3 have been downplayed by only
requesting the Riksdag to approve the forecasted consequence assessments rather
than to set firm ceilings. In next years budget round, the allocations for t+2 may again
be given a more definite status while those for t+3 are likely to continue to be only
indicative and based on the presumption of no substantive changes.
3.6. Roles within the administration
3.6.1. The division of responsibilities within the Government Offices

The introduction of the top-down budgeting model has, on the one hand
strengthened the role of the Ministry of Finance in regard to the major allocative
decisions. At the same time it has given the line ministries greater responsibilities for
the allocation of budgetary resources within the different expenditure areas.
In principle, one could conceive of the responsibilities for deciding on individual
appropriations being fully delegated to the line ministries. This, however, has not
happened. The Budget Department, for good or bad, still tends to get involved in many
detailed decision at intra-expenditure area level. The justification for this is a blanket
rule that the Budget Department should be involved in any decision that has financial
consequences. Generously interpreted that could mean that the Budget Department
ought to get involved in just about every single decision on the budget. On the other
hand, with the top-down decision-making mechanism in place, any effects on
aggregate spending are already effectively blocked. The departments involvement in
the details is really more a question of old habits taking a while do die. The present
culture at the Department was also very much set during the days of the financial
crises when the Departments singly most important task was to scour the budget for
potential expenditure reductions.
The Department is still largely staffed with relatively young professionals for
whom a couple years at the Budget Department traditionally have served as a catapult
to good careers in other parts of the public sector.59 The increased responsibility for the
allocation of resources between expenditure areas has not yet led to any noticeable
changes in staff composition or any systematic up-grading of staff skills, although it
ought to have.
In the line ministries, the units responsible for the coordination of budget work
have come to play a more important role. It is difficult to say whether there has been
any systematic up-grading of staff skills.
3.6.2. The role of the politically appointed officials

The Budget Bill, which consists of some 2,500 pages, is prepared by the technical
staff of the Government Offices. The political appointees in the form of Ministers, State
Secretaries, and their political advisers are, however, more or less heavily involved in
any policy issues reflected in the budget for example the allocation of resources
between expenditure areas. The Finance Minister is obviously interested in taxation
issues and he follows very closely the financial execution of the budget, given the very
substantial political capital invested in the global expenditure ceiling. In recent years
he has, because of insufficient budget margins, been forced to decide on more or less
transparent ways of slipping under the ceiling. The most important trick in volume
terms has been to transform straight expenditure into tax expenditure, i.e. tax
exemptions tied to specific activities. Common year-end measures have been to
postpone payments to the following year. In total volume terms these maneuverings
have, however, been quite limited, never surpassing one percent of the total budget.
Their cost has been more political than economic.
4. From formulation to enactment

The top-down multi-year budget introduced in the middle of the nineties stands
on two equally important legs: the top-down budget preparation process and the top-
down decision-making process in the Riksdag. The ever-increasing size of the public
sector had to do as much with the decision-making process in parliament as it did with
conscious political choices by the party that has been in power for most of the time
since the 1930s, the Swedish Social Democratic Party. It was very easy for a coalition of
parties or even of individual MPs to propose and gain acceptance for proposals for
increased spending in particular pet areas. They could do this without having to take
any responsibility for how the proposed increased spending was to be financed. The
importance of this type of dynamics became evident in the early nineties when a
coalition of presumably more fiscally conservative parties in practice was held together
59 There are signs that that may now be changing with a general tightening of the public sector job market.
largely by accommodating each single partners spending preferences at the expense

of the advocated budget discipline.
With the new process, no new proposal for spending can be put forth by members
of parliament without a matching proposal for how it is to be financed in concrete
terms what other spending is to be reduced since the global ceiling is fixed by an earlier
decision.
As explained above, the global expenditure is set on a rolling basis for three years.
Each year a new year is added and the ceiling for year t+2 is reconfirmed, possibly with
some changes mainly for technical reasons.
The decision process on next years budget begins with a new decision on the
global ceiling. This decision is taken as a package together with the allocation to the 27
expenditure areas and any changes in taxation that will affect the revenues of the
governments. The decision also includes an acknowledgment of the revenue estimates
prepared by the Government.
The package proposal is prepared by the Finance Committee and voted on by the
full house. The composition of the Finance Committee reflects that of the house, thus,
in practice, the proposal prepared by the Finance Committee corresponds to that of the
Governments except possibly for some minor details.
Once the global ceiling and the expenditure area ceilings are established, proposals
for the break-down of the expenditure ceilings into a total of some 500 appropriations
are prepared by the committees responsible for the different areas. These proposals
again, largely, correspond to those of the Governments. The detailed allocations are
voted on as a package, expenditure area by expenditure area.
In principle, the Riksdag is supreme in relation to the budget. However, the rules
of the voting in the Riksdag on the budget are such that in order for the proposals
prepared by the Finance Committee for the area ceilings and by the other standing
committees for appropriations to be overthrown, the opposition will need to agree on
one single package to be put up against that of the governments. In proportional
electoral systems, such as the Swedish one, this type of voting arrangements
strengthens the hand of the governments considerably making it possible for
minority governments to govern. In recent years, the Social Democrats have
nonetheless chosen to routinely engage in a dialogue with two opposition parties, the
Greens and the Leftist Party, in order to ensure their prior support for the budget
proposal although, arithmetically speaking, they do not need to.
Once the budget proposal is finished and transmitted to the legislature, the
executive has a very limited role as regards its enactment. The decision on the final
proposal, i.e. what is reflected in the Budget Bill, is taken as a collective decision by the
Cabinet. All ministers are thus bound by it and would not conceive of arguing against
it, at least not publicly.
As soon as all decisions on the budget are taken by the Riksdag, the line ministries,
in close collaboration with the Budget Department, prepare Letters of Appropriation
through which the agencies are formally informed of the appropriations for which they
are responsible and of the conditions, objectives, and reporting requirements associated
with those appropriations. Letters of appropriation may be prepared for single
appropriation or for a set of appropriations under the jurisdiction of one agency. The
letters of appropriation may vary a bit in character depending on the nature of the
appropriation, but when it concerns the operation of an agency, they will contain a
section on objectives, results indicators and special reporting requirements, a section on
assignments (tasks) for the coming year, and a section on the financing of the activities
of the agency. The latter includes: a) the appropriation for the year, b) an indication of
any planned use of carried-over appropriations, c) a ceiling on the appropriation credit,
i.e. the right to borrow against next years appropriation, d) an estimate of fee revenue
that will be under the control of the agency, e) a ceiling on borrowing for investment in
fixed assets, f) a ceiling on credit to meet short-term cash-flow needs, and g) any
explicitly authorized, special, longer-term or multi-year commitments.
In parallel with the preparation of the letters of appropriation by the Government,
the Riksdag compiles the final central government budget and hands it over to the
Government a few days before Christmas.
5. Execution
Once the budget is enacted and the letters of appropriation issued, the execution of
the budget is fully delegated to the agencies. For the framework appropriations
financing the administrative expenses of agencies, one twelfth of the yearly
appropriation is transferred monthly to interest-bearing accounts within the
government group account held in one of the commercial banks. 60 Payments are
executed by the agencies themselves without the intervention of any central
government treasury. Should the need for liquidity exceed what is available, it is met
by a credit facility attached to the bank accounts.61
The strict annuality of traditional budgetary principles is, thus, done away with
and all control of execution is ex post. Within the more flexible framework created by
block appropriations, carry-overs, appropriation credits, loan and credit facilities, the
agencies are expected to, and in fact generally do, behave responsibly.
The accounting is decentralized to the agencies and the accounting is done
according to generally accepted accounting principles, i.e. the principles that are used
in the private sector world wide. The general accounts are complete, i.e. they, besides
the central funding, include all other revenue such as sales income and fees, all
expenses, and assets and liabilities. The follow-up against the appropriations
sometimes called the budgetary accounting is integrated into the general accounts.62
Information from the budgetary accounts is transferred monthly to the central
government financial management information system. Complete, accrual-based
accounts for all individual agencies are prepared twice a year and consolidated into
aggregate accounts for the entire central government. This is done by the National
Financial Management Authority.
All agencies are subject to financial audits every year and performance audits are
carried out as necessary. With the creation of the new Swedish National Audit Office
under the Riksdag, the consolidated accounts will also be audited. Another mechanism
to stimulate good and responsible financial management is a yearly rating of the
agencies done by the National Financial Management Authority.
60 In the near future, the Government will tender for payments and group account services from several
banks. The Swedish National Debt Office is responsible for the oversight of government payments and is
responsible for the over-night clearing of the group-account(s).
61 As regards transfer payments, they are also channeled through sub-accounts of the government group
account but are replenished on the day of payment from the general government account controlled by the
National Debt Office.
62 This integration is, however, not entirely straightforward and transparent because the budgetary
accounts follow different principles than those of the general accounts. Efforts to fully integrate budget follow-
up into the general accounts by defining the appropriations in terms that are fully compatible with those of
accrual accounting have got bogged down.
6. In-year changes in the budget

Supplementary estimates are prepared and presented to the Riksdag twice a year
in connection with the Spring Financial Bill and with the normal Budget Bill in
September. Any proposal for new spending must be accommodated within the global
expenditure ceiling normally within the relevant expenditure area ceiling i.e. it
must be accompanied by a proposal for reductions in other areas. Nothing else would
be acceptable. The costing for any new proposals is done in the same way as for a
regular proposal presented in the Budget Bill. The number and volume of new
proposals is, however, quite limited. Rather than vehicles for new proposals,
supplementary estimates have, in recent years, been necessary to handle unexpected
volume increases in entitlement programs.
7. Problems and outstanding issues

7.1. Budget margin
Although the new process has transformed the dynamics of preparing and
deciding on the budget, it takes a while for bad old habits to die completely. And what
Swedish politics traditionally has been made of is promising the electorate additional
social benefits under the name of reforms. However, with the ceiling effectively set
three years earlier, the one seemingly remaining opportunity for politicians wanting to
squeeze in yet one more reform has been to eat away at the budget margin. For that
they have had to bet on being no unexpected cost increases in other areas.
Unfortunately this has invariably proven to be an overly optimistic assumption and the
price to be paid has been painful decisions and maneuverings at year-end to squeeze
under the ceiling and a concomitant loss of credibility.
7.2. Overall sector priorities
The top-down budgeting process, together with the increased focus on results, has
created an opportunity for pondering, in a more thorough and systematic way, the
overall composition of the budget. This opportunity has not really been seized.
It must, of course, be acknowledged that it is impossible to scientifically calculate
and in any definitive way decide what would be the optimal sector allocation of
government resources. There are too many imponderables, potential synergies and
external effects for that ever to be conceivable.
It could still be argued that a lot more could be done in the way of analyzing the
effectiveness and efficiency of programs even at a more aggregate level and seeing
to it that the outcomes of those analytical endeavors are fed into the political process.
For that to happen, the roles of the different actors in the budget process would need to
be reviewed in order to achieve a clearer division of labor and responsibilities. One
obstacle has previously also been the limited IT support for the budget process. That is
now changing with the building up of an internet-based FMIS that will soon service all
ministries and agencies. However, better IT support is a necessary but not a sufficient
condition for a division of responsibilities better in keeping with the new top-down
approach. A cultural revolution and new capabilities will be needed at all levels within
the Cabinet Office.
7.3. Long-term structural problems
The new multi-year top-down budget process has extended the budget horizon
beyond the immediately coming year. It has not, however, automatically led to a
systematic factoring in of the longer-term beyond the three-year horizon of the
expenditure ceilings. For Sweden, as for many other European countries, the fiscal
perspective for the period beyond 2008 is quite problematic because of the aging
population. Sweden does have a target for the budget balance that is higher than that
of most other countries within the European Union. Still it is probably not ambitious
enough to build the buffers that will be needed if the next generation is not to bear the
entire burden of the aging population. The problem is a specific and acute one for
Sweden, but a generic one for all MTEF type efforts; in incorporating the medium-term
into the budget process, one must not loose sight of the long-term!
The challenge is to create a framework for political decision-making that, just as
the new top-down budgeting process in Sweden, conditions the outcome in favor of
long-term collective rationality at the expense of short-term vested interests and vote
maximization.
7.4. Technical issues
Global, hard budget constraints are essential to fiscal discipline; but it makes a
difference how they are defined. There are several different dimensions to this issue.
For example, what should be the accounting basis for the constraints? In Sweden,
the accounting basis for the volume constraint (to use a generic term for the
expenditure ceiling) is the budgetary accounts, and they are expenditure-based. An
alternative would be to base the volume constraint on the general accounts which are
accrual-based. One important difference between the two sets of accounts is the
treatment of investment. In the budgetary accounts, the full outlay for investments is
included while in the accrual-based accounts only depreciation appears. The treatment
of investment in the expenditure-based budgetary accounts means that, in relation to
the global constraint, outlays for investments compete, krona for krona, with recurrent
expenditure. One effect is that, in a tight fiscal situation, reductions in the budget are
likely to focus on investment, simply because the government has more discretionary
power over it.
A second important issue as regards the global volume ceiling is what should be
included: total expenditure/cost for the entire general government or only central
government? If the constraint should cover only central government, should it
constrain total expenditure/cost or only that financed by central funds?63
Furthermore, should all types of expenditure/cost be included, or should some be
excluded because the government has very little control over it in the short term? In the
case of Sweden, interest payments on the central government debt are excluded. They
63 A technical question in that respect is whether it is possible, in situations where central funds are pooled
with other types of revenue, to attribute an expenditure/cost to a particular source of finance at least in a
credible way. And if one can not credibly follow it in the accounts, should one design a constraint that
presumes just that?
were with good reason excluded during the turbulent years of the 1990s when interest
rates were very volatile and even the slightest suspicion that the Swedish Government
would default on its loans for example because of putting interest payments under
the ceiling would have pushed interest rates even higher. In a more stable situation,
like today, the choice is not quite as obvious. Interest payments are largely exogenous
in the short term but in the medium- and long-term the Government can certainly
influence them through its debt policy. Including interest payments under the ceiling
would increase the incentives to reduce debt for the benefit of future generations.
And then there are the funds outside the budget, like the pension fund. In Sweden
they are included because payments in nature are very similar to regular government
expenditure; they can be influenced by political decisions. In a funded system the
Government would have no such right and it would be incorrect to include it.
As regards the budget balance, it is defined in national accounts terms. Having the
two global constraints defined in different terms complicates matters technically and
pedagogically. An alternative, given the existence of accrual-based accounts in the
areas for which the balance is calculated, would be to define it in straight accrual
accounting terms. The balance thus calculated could then for comparative purposes
cf. the Maastricht criteria be translated into NR terms.
Sweden has, for good or bad, made a set of choices for how to define the aggregate
ceilings. For any other country, the definitions would have to be based on an
assessment what the ceilings should and could achieve and what the technical options
for defining the ceilings are.
CHAPTER 6
The Swedish Budget Process: Expenditure Targets and Top-down Decisions
by
Niklas Frank, Sweden
1. Background
In Sweden the Parliament is highly involved in the budget process. From a formal
point of view the Parliament can change anything it wants in the budget. All the
standing committees in Parliament, there are 16 all together, are involved in the
process, and the responsibility for expenditure is divided between the various
committees. The budget is split in to 27 expenditure areas designed to reflect different
areas of political interest, and the responsibility for the expenditure areas are divided
between the committees according to their specialities. One committee may be
responsible for one or more expenditure area but no expenditure area may be split
between committees.
When the Government submits a bill to the Parliament, each Member of
Parliament has the unrestricted right during 15 days to submit counter proposals, so-
called motions. The motions are then distributed to the same committee that has been
assigned the bill. Based on the bill and the motions, the committee then produces a
report to the Chamber. This report contains the formal proposal on which the
Parliament votes, as well as written argumentation about all the proposals. The
committees must respond to every bill and every motion, none of them may be put
aside. This applies also in the budget process. The Parliament subsequently has to
process the Governments budget documents as well as all the counter proposals that
have been made by individual members.
Figure 6-1. Annual GDP growth 1980-2002

5
Sweden European Union
4
0
80 82 84 86 88 90 92 94 96 98 00 02
-1
-2
The Swedish budget process was radically reformed in the middle of the 1990s,
from a bottom-up to a top-down process. The starting point of the reforms was the
deep economic recession that hit the country in the beginning of 1990s. The
environment in which the budget reforms took place was an environment of crisis.
In 1993 which was the worst year of the recession unemployment rose to 8 percent,
a historically record high level. Growth in GDP was negative by 1.8 percent and the
situation for the public finances was very serious.
The Central Government borrowing requirement amounted in 1993 to almost 17
percent of GDP and the overall deficit in the public sector was almost 12 percent of
GDP. The deterioration of the public finances was great and very rapid. In the late
1980s Sweden had one of the largest surpluses in the OECD and only a few years later
one of the largest deficits. By 1993 the national debt amounted to 76 percent of GDP,
and was rising rapidly. Between 1988 and 1993 the public sector expenditure rose from
55 to 68 percent of GDP.
Figure 6-2. Public sector financial balance, percent of GDP, 1980-2002
6
Sweden European Union
4
0
80 82 84 86 88 90 92 94 96 98 00 02
-2
-4
-6
-8
-10
-12
-14
At the same time there were serious problems in the financial sector as well. The
Government had to spend about 4 percent of GDP to save banks from going bankrupt.
In November 1992 the fixed exchange rate was abandoned for a freely floating rate,
after the Central Bank had unsuccessfully tried to defend the fixed exchange rate by
raising the short-term interest rate to 500 percent. The value of the currency quickly
dropped by about 25 percent.
To stabilize the public finances and lay foundations for an economic recovery,
broad and radical measures were taken both reductions in public expenditure and
increases in taxes. There was also a broad political understanding that besides dealing
with the immediate crises in the public finances, there was a need to find ways of
preventing a similar situation from reoccurring in the future. In order to improve the
long-run stability of the public finances and prevent a repetition of the dramatic
deterioration during the early 1990s, the institutional arrangements around the budget
were fundamentally changed.
Chapter 6. The Swedish Budget Process: Expenditure Targets and Top-down Decisions 73
The new budget process, which was in place in 1997, was stronger than the old
process in virtually all respects. Changes were made in how the Government prepares
the budget and in the way the Parliament considers and approves the budget as well as
in the control and steering mechanisms on the implementation of the budget by the
various Government agencies. The principles and rules for the budget were given a
stronger legal status in the form of a budget law that was approved by the Parliament.
According to a comparative study about the strength of the budget processes in a
number of European countries, the Swedish process went from being one of the
weakest in Europe to being one of the strongest after the budget reform.
The reformed budget process enabled the Parliament to take a more general and co-
ordinated approach to the budget, focusing less on details and more on aggregates. The
level of Central Government expenditure is determined three years ahead by setting an
expenditure ceiling, and decisions on the budget are taken in two stages, the so-called
framework model. The decision-making approach is top-down, starting with the whole
and ending with the details. A surplus target for the whole public finances is being used.
In this presentation, I will focus on two elements of the new budget process, the two
elements I regard as being the most important. First there is the use of two budgetary
targets the Central Government expenditure ceiling and the surplus target for the total
public sector, and second there is the two-stage decision process in the Parliament.
2. Two budgetary targets, the expenditure ceiling and the surplus target
A central element of the reformed budget process is the usage of quantitative
budgetary targets. The Government has set up two such targets. There is a target
stating that the expenditure of the Central Government budget must not exceed the
expenditure ceiling that Parliament has decided upon. There is also a target for the
financial balance in the total public sector, stating that as an average over a business
cycle, there should be a surplus of two percent of GDP.
The two targets are different in their construction and use, and they should be
regarded as complementary parts of an integrated budget system.
2.1. The structure of the budget
The expenditure side of the Central Government budget is divided into 27

expenditure areas, which are subdivided into around 500 appropriations all in all. The
revenues are divided into about 150 revenue items.
Before the budget reforms there were a number of extra budgetary funds such as
the unemployment benefit fund, the health insurance fund and the early retirement
fund, to mention some. Expenditure from these funds amounted to a substantial part of
the state expenditure but was formally not a part of the budget. Thus there was a
significant discrepancy between the budget balance and the borrowing requirement.
One of the changes in the budget reform was to abolish all such funds and bring all
expenditure in to the budget. In the new system, the central Government budget deficit
therefore equals the borrowing requirement.
Another change was to abolish the use of net budgeting. All expenditure and all
revenue should be accounted for separately. This is an important principle that
increases the transparency and openness of the budget, as well as it makes
manipulation of the budget figures difficult.
2.2. The expenditure ceiling
2.2.1. Construction of the expenditure ceiling

The expenditure ceiling is a unique system that has no link to external demands or
international standards. It is a voluntary restriction taken on by the Government. It
should be pointed out that the budget law requires neither the Government nor the
Parliament to use a ceiling for the expenditure, but it does stipulate rules on how such
a ceiling should be handled if the Government should choose to use one.
In 1996 the expenditure ceiling was introduced. The system is based on a three-
year budget cycle. In the autumn of year t, the Government proposes a level of the
expenditure ceiling for the year t+3 to the Parliament. At the same time, the
Government puts forward a proposal about the distribution of expenditure between
the 27 expenditure areas for the year t+1, and a preliminary distribution for the years
t+2 and t+3. The Government also presents the expenditure ceilings for the years t+1
and t+2, which have previously been decided by Parliament one and two years earlier.
The Government is free to propose new levels for the ceilings even for the years t+1
and t+2, but in practice the Government has never done so. The only changes in
previously decided ceilings have been of a technical nature due to changes that has not
altered the budget balance.
The expenditure ceiling covers 26 of the 27 expenditure areas. The only
expenditure area that is exempt is the interest on the national debt, which differs from
other forms of expenditure in that it is less subject to control in the short run. In
addition to the 26 expenditure areas, the ceiling covers the old-age pension scheme,
even though this technically is not a part of the Central Government budget. Thus, the
only public expenditure not included in the expenditure ceiling is the interest on the
national debt and the expenditure of the Local Governments. (Concerning the 290 Local
Governments in the country, there is a legal requirement that their budgets must
balance each year. This means that all public expenditure is either subject to the
expenditure ceiling or the Local Government balance requirement.)
Since it is difficult to assess expenditure changes three years ahead with a high
degree of certainty, the expenditure ceiling is set at a level slightly above the estimated
expenditure level. The difference between estimated expenditure and the ceiling is the
so-called budget margin. This margin provides scope for meeting unforeseen increases
in expenditure or for taking further spending decisions without exceeding the ceiling.
However, use of the budget margin means that when Central Government expenditure
increases, there must be matching increases in revenue for the budget balance to be
maintained. The expenditure ceiling is defined the following way:
Expenditure area 1
+ ...........
+ ...........
+ Expenditure area 27
= Sum of expenditure areas
- Interest on National Debt
+ Public pension system expenditure
= Expenditure capped by ceiling
+ Budget margin (residual)
= Global Expenditure Ceiling
The expenditure ceiling is stated in nominal terms, which means that the level set
for each year of the three-year budget cycle is neither adjusted for inflation, nor
indexed in any other way. The calculation of the ceiling is based on a macro-economic
forecast that is presented in great detail in the budget documents. The inflation target
of the central bank is to keep inflation at 2 percent (+/- 1 percentage point), and the
macro-economic forecast for the budget is done in accordance with this target,
assuming that the central bank meets its target.
The Parliaments decision on the expenditure ceiling is a so-called guideline
decision, which means that it is not legally binding, and the Parliament is also able to
review an earlier decision on an expenditure area. This could happen if there was a
change of Government, in which case the incoming Government has the opportunity of
submitting a proposal that differs from expenditure levels determined earlier. The
expenditure ceiling should be regarded as a very strong budgetary commitment on the
part of the sitting Government to implement a policy where aggregate Central
Government expenditure (excluding interest on national debt) is kept within levels
previously determined.
2.2.2. Expenditure ceilings and entitlements

The expenditure ceiling is, as mentioned before, set in nominal terms and includes
all expenditure except interest on the national debt. Thus there is no kind of indexation
in the ceiling and there are many expenditure items included even though they are
highly sensitive to cyclical changes in the economy. This obviously means that there
are potential tensions built in to the system. What if inflation increases so that the
expenditures threaten to exceed the ceiling, even though the financial balance in the
public sector is not worsened? What if the legal rights for individuals to receive
entitlements exceed the limits set for the expenditure? Which is the stronger, the legal
rights of the individual or the budget restrictions?
The Swedish attempt to solve these problems is strict on one hand and pragmatic
on the other. When the ceiling is set it is crucial that there are credible forecasts on the
economy as a whole, as well as on the individual expenditure items. In fact, one
consequence of the new budget system has been to increase focus on the quality and
reliability of the economic forecasts.
According to the budget law, if there is reason to believe that the ceiling might be
exceeded, the Government is obliged to take measures to avoid this. At the same time,
an individual may take a case to court to receive his or her legal entitlement.
In the short run, the right of the individual takes the upper hand. The Government
can not deny an individual his legal entitlement by referring to the expenditure ceiling.
The budget law does not state any legal consequences for the Government that violates
it. In practice, the expenditure ceiling has never been exceeded, even though the
margin to doing so has been very small a few times. The Government has taken
measures by the end of the year to prevent the ceiling from being exceeded. What the
consequences would be of exceeding the ceiling has therefore not yet been put to a real
test, but surely there would be a political price for doing so.
If the expenditure forecasts indicate that the ceiling will be exceeded, for whatever
reason, the Government has to address the problem. In order to keep the expenditure
from exceeding the ceiling the Government must reduce some other expenditure, or in
the medium term, consider the sustainability of the entitlement rules. The expenditure
ceiling therefore works as a trigger for political considerations at an early stage, before
expansion of expenditure has gone too far. The system thus makes it impossible for the
Government to ignore or overlook a situation where the expenditure is about to exceed
the limits.
2.3. The surplus target
The surplus target is defined to include the three sectors that make up the entire
public sector, the Central Government, the Local Governments and the Public Pension
system. The target is specified in accordance with the European accounting system
(European System of Accounts, ESA95) and numbers are comparable to the other EU-
countries. This is also the system used when evaluating the European stability pact to
make sure that no country has a public sector deficit that exceeds 3 percent of GDP at
any time.
The target stipulates that as an average over the business cycle, the over all public
sector should show a surplus of two percent of GDP. For an individual year the surplus
might, and should, deviate from the two percent depending on the current
macroeconomic situation. Twice every year, the Government reports to the Parliament
about the financial balance and presents three-year forecasts for the economy and the
public sector financial balance.
The surplus target and the expenditure ceiling should be regarded as
complementary targets. When the expenditure ceiling is set, it must be set in such a
way that the surplus target is met, provided the estimated revenues. If revenues exceed
the forecast those extra funds may not be used to increase expenditure more than is
permitted by the expenditure ceiling. If revenues fall below the estimates then revenue
either will have to be raised or expenditure cut in order to maintain the surplus target.
Together the two targets provide a strong framework for the budget formulation.
3. The framework model

The decision process in the Parliament is called the framework model. It is a
process designed for top-down decision-making. First, limits are set on an aggregate
level and then the details are decided, under the aggregated restrictions. The process
revolves around two annual bills, the Spring Fiscal Policy Bill and the Budget Bill. The
Parliamentary decisions on the Budget Bill are taken in two distinct stages during the
autumn, leading up to the start of the fiscal year.
3.1. The Spring Bill
The budget process in the Parliament begins when the Government submits its
Spring Fiscal Policy Bill. The bill, which must be submitted no later than April 15,
contains an assessment of the current economic situation with perspectives for the next
three years and proposals for the orientation of the economic policy and the budget
policy.
In the Spring Bill there are no formal proposals from the Government involving
precise numbers. The bill contains detailed medium-term forecasts for the macro
economy, the Central Government expenditure and revenue, and the overall public
finances. In the bill the Government presents the expenditure for the coming three
years on expenditure area level, divided in to the 27 different areas. At this stage the
Government presents a calculation of the year t+3 to the Parliament for the first time.
Calculations of revenue are also presented in a relatively high level of detail.
The Spring Bill is thus used to give Parliament a rather detailed picture of the
economy as well as the budgetary situation. There must however be neither tax
proposals, nor proposals concerning the expenditure at this stage. The detailed
presentation of the budget gives the Parliament information about what the
Government intends to propose later in the Budget Bill, due in September.
As with all Government bills, the members of Parliament have 15 days to write
their counter proposals, their motions. In the case of the Spring Bill, the opposition
parties, currently four, present their alternative views on the economic situation and
their alternative suggestions regarding guidelines for the economic policy and the
budget policy. In this process the special committees in Parliament may submit their
opinions about the Government bill, as well as about the oppositions proposals, to the
Committee on Finance.
The Spring Bill, the proposals from the opposition and the opinions from other
committees, are all submitted to the Committee on Finance, which writes a report to
the Chamber where the political alternatives are presented. In the middle of June a
formal decision is taken about the guidelines for the economic policy and the budget
policy.
3.2. The Budget Bill
3.2.1. Overview
The Government submits its Budget Bill for the following year by September 20 at
the latest, a few days after the opening of the parliamentary session. During election
years the submission period may be extended, but no later than November 15. The
Budget Bill covers all expenditure and revenue, since no further proposals with
budgetary consequences may be submitted before the beginning of the fiscal year. The
gross principle is applied in that all revenue and expenditure are accounted for
separately.
All intended changes for the coming year in revenues and expenditure must be
presented to the Parliament in the Budget Bill. The estimates of expenditure as well as
revenue must include all the proposed changes. There must be no last minute changes
of taxes or expenditure after the presentation of the Budget Bill.
The expenditure is presented with its 27 expenditure areas and 500 appropriations,
and the revenue is presented with its 150 items. The opposition parties present their
alternative proposals to those of the Government. Usually the opposition parties
provide an overall presentation of their budget alternatives in special motions on
economic policy. In this context they present their proposals for the expenditure ceiling
and for the allocation of expenditure between the expenditure areas. In their motions
on economic policy the parties also present an overview of the changes in taxes and
charges they wish to implement, and how proposals for different reforms should be
financed.
A more detailed specification of the various parties budgetary proposals is set out
in special tax and appropriations motions, where the parties usually present their
proposals in the form of one motion per expenditure area.
3.2.2. The Budget decision, stage one

The decision-making procedure that the Parliament uses when considering the
Budget Bill is known as the framework model. This means that the Parliament first
takes a decision on the overall size of the budget, i.e. total Central Government
expenditure and how this should be allocated between the 27 expenditure areas. By
means of this decision, expenditure is fixed and the level determined for each
expenditure area, the so-called expenditure area frames. Thereafter these frames
function as limits which may not be exceeded in the second stage, when the Parliament
finalizes the budget.
In the first decision the Parliament also determines changes in the taxes and
charges for the next fiscal year. This enables the level of Central Government revenue
for the following year to be estimated. For obvious reasons, it is not possible to
determine limits for revenue in the same way as for expenditure. This is why the
Parliament does not make a decision on revenue limits. Revenues are determined by
how the tax base changes, which itself is related to changes in the economy, and the
rules applicable to taxes and charges. These rules apply until they are changed or
removed and are not considered by the Parliament each year when determining the
budget. The budget proposal can, however, contain proposals for changes in the
regulatory framework covering e.g. tax changes. The Parliament approves an estimate
of the Central Government revenues during the coming fiscal year, and at the same
time decides which changes in taxes and charges are to be implemented. Since the
levels of revenue and expenditure are thus determined, the level of borrowing
requirements can also be stated.
The Committee on Finance handles proposals on the frames for all the expenditure
areas. The other committees may submit their views to the Committee on Finance on
the expenditure areas for which they are responsible. The committees are free to
propose changes in the level of the different expenditure areas and/or recommend a
reallocation of funds between expenditure areas. In the same way, they can make
proposals that affect the level of Central Government revenue.
The Committee on Finance then examines these proposals and uses them as a basis
for submitting an overall proposal that will be considered by the Chamber around
November 20. The debate in the Chamber concludes with the Parliament making one
single decision determining several important elements simultaneously. The main
elements of the decision are:
z The overall expenditure ceiling for the year t+3.

z Expenditure ceilings for the years t+1 and t+2, in case they are being changed.
z The allocation of expenditure between the 27 areas for the year t+1.
z All proposed changes in taxes and charges for the year t+1.
z Approval of an estimation of total revenue for the year t+1.
The fact that all these points are decided by Parliament in one single decision, is
not a mere technicality, but is of crucial importance to the top-down character of the
budget process. The Committee on Finance decides the distribution of expenditure
between the expenditure areas, in a single decision. Thus any Parliamentary changes in
the Governments suggested distribution can only be done by the Committee on
Finance in this first stage of the decision process. This single decision system has at
least two important implications for the budget process.
Firstly, it sets a restriction on what the special committees can do in the second
stage of the process. The special committees are responsible for the distribution of
expenditure appropriations within their respective expenditure areas, but they have to
keep expenditure inside the expenditure area frame set in the first stage. A special
committee can therefore not expand the expenditure in its own area even if there is a
majority in that committee to do so. The room for maneuver in the special committees
is therefore highly restricted.
Secondly, it forces the opposition parties to present complete budget alternatives
instead of only picking the pieces they prefer to discuss. To present a feasible
alternative to the Governments budget proposal, the opposition has to present an
alternative decision including all its tax changes and its distribution of expenditure
between the 27 expenditure areas. All expenditure limits and tax changes are treated as
a single package, and alternatives are considered against each other as single
packages. The Parliament then takes a position on the entire package. This means
that even if the opposition parties agree on an issue, they cannot form a majority on
this issue alone and use this to reject a part of the Government proposal. To defeat the
Government in the parliamentary vote, the opposition must agree on an entire budget
package.
This budget decision is probably the single most important decision that the
Parliament takes during the year. If the opposition would actually win this vote there
would be a Government crisis. Without support for its budget policy the Government
might have to resign, or there would most likely be at least a vote of no confidence.
With this package decision, six weeks before the start of the year t+1, the
Parliament sets the expenditure limits and concludes its revenue deliberations and
determines the tax regulations that will be in force in the coming year. Still remaining
is the more detailed allocation of Central Government expenditure. This is dealt with
during the second stage of the budget process.
3.2.3. The Budget decision, stage two

In the second stage of the budget process, the 15 various special committees handle
their respective expenditure areas. The special committees put forward proposals on
the goals of the activities and the allocation of funds in their expenditure areas. The
whole expenditure area is treated in a single report and the Parliament makes a
decision on the appropriations and their distribution within the total expenditure area
by means of one single decision.
This is parallel to the decision of the distribution between expenditure areas in the
first stage, and the implication for the opposition parties is the same. It is not possible
to single out individual items within an expenditure area, and increase expenditure for
those items. In order to change the distribution within the expenditure area the
opposition parties must agree on all appropriations in that area. An increase in one
expenditure appropriation must therefore be financed by a decrease of another
appropriation in the same area. Any other way of financing the new expenditure is
closed. A special committee that is responsible for more than one expenditure area may
not reallocate expenditure between those areas. The decision on the frames in stage one
applies for each expenditure area regardless of which special committee has the
responsibility in the second stage.
The committee may not put forward any proposals that would lead to the
previously determined limits from stage one being exceeded. If a committee wishes to
take an initiative of its own, or act on a private members motion, the financial effects
must be taken into account. Such proposals must be financed by an equivalent
reduction elsewhere within the same expenditure area. It is not even permissible for
dissenting opinions to propose a limit higher than previously decided on by the
Parliament in the first stage.
Should a committee consider that an expenditure frame ought to be expanded, for
instance, on the grounds that there has been a change in some underlying variable, this
can only take place if the Parliament on a proposal from the Committee on Finance
decides so. In such a situation, the Committee on Finance may consider whether the
limit for other expenditure areas should be reduced. The aim of this procedure is that
the Committee on Finance and the Parliament should also consider how the proposed
increase in expenditure is to be financed.
Each special committee writes a report on their respective expenditure areas and
these are finalized in the beginning of December. After the debates in the Chamber,
decisions are taken on the various appropriations for the coming year. When this has
been done, the Government issues budget letters to its agencies with instructions for
how the appropriations are to be used.
When all the reports have been dealt with, the Parliament compiles the now
finalized Central Government budget for the coming year and hands it over to the
Government a few days before Christmas.
4. Effects of the reformed budget process

The changes in the budget process have had a number of consequences and I will
mention the most important ones.
4.1. It has moved focus towards the medium-term perspective
One important element in the budget process is the setting of a nominal

expenditure ceiling for the year t+3. This has made it necessary for the Government to
make a detailed macroeconomic forecast as well as detailed expenditure and revenue
forecasts for three years to come. This in turn has put pressure on the Government to
improve its forecasting methods. The requirement for the Government to act in case the
expenditure ceiling is threatened has moved increasing public expenditure higher on
the political agenda. With the reformed budget system, budgetary problems need to be
identified and dealt with in a relatively early stage.
4.2. It has put a minority Government in a strong position
Sweden has a proportional electoral system. The representation of the different

political parties in Parliament is strictly proportional to that of the popular vote,
limited only by a 4 percent minimum requirement of the votes to get seats in
Parliament. Without going in to details about different electoral systems, it is a well-
established conclusion that proportional systems have a tendency to produce weaker
Governments than systems based on majority voting. Sweden has a long history of
either Minority Governments depending on the additional support in Parliament of at
least one more party, or coalition Governments with up to four co-operating parties.
Strengthening the position of a Minority Government was therefore a desired effect of
the budget reform.
In order for the opposition to overthrow the Governments budget proposal, it has to
agree on a complete alternative. It is not possible for the opposition to agree on certain
items and change the Governments bill on those items only. Obviously, agreeing on the
entire budget is politically much more difficult than agreeing on individual issues. With the
reformed system, a Government that has a relative majority has a strong position in the
parliamentary budget process. In this way you might say that the reformation of the
budget process has moved power from the Parliament to the Government.
4.3. It has strengthened the Committee on Finance
The two-stage decision model has given the Committee on Finance the overall
responsibility for the budget in the Parliament. The Committee on Finance has been
given the power to restrict the other committees possibilities of spending resources in
their own areas. Giving this power to the Committee on Finance has been an important
element in creating a top-down structure of the process. The overall expenditure ceiling
and the distribution between expenditure areas are decided first, setting the limits for
the appropriations in the different expenditure areas. Provided that the expenditure
area frames are not exceeded, the special committees however still have great freedom
to redistribute the expenditure within their respective expenditure areas.
4.4. It has increased the transparency of the budget process
Transparency has improved in several ways. The budget has been made complete
and extra-budgetary funds abolished. Revenue and expenditure are accounted for
separately and net budgeting is not permitted. The need for the opposition to present
complete budget proposals has made the political alternatives more open and
comparable. It is no longer possible for an opposition party to present for instance large
increases in benefits without declaring how it will be financed.
5. Criticism and problems in the budget process

Even though there has been a high degree of political consensus surrounding the
reformation of the budget process, there has been criticism and some changes have been made.
5.1. The budget process takes all year
One criticism from the Parliament has been that the budget process takes all year.
Producing a complete budget alternative is a great deal of work for the opposition
parties, particularly for a small party with limited resources. This aspect led to a
change in the system in 2002. Before then the formal proposal on the overall
expenditure ceiling was a part of the Spring Bill. The Spring Bill then contained
guidelines for the economic policy and for the budget policy as well as an expenditure
ceiling for the year t+3. In 2002 this was changed so that the ceiling for t+3 is proposed
in the Budget Bill in September. This was a controversial change decided by a united
opposition against the will of the Government, who argued that the change would
weaken the budget process.
5.2. The budget margin is too small
When the system was launched, there was a guideline about the size of the budget
margin saying that the margin should be approximately 1.5, 2.0 and 2.5 percent of the
total expenditure for the years t+1 to t+3. As uncertainty about the numbers grew over
time, so would the budget margin. In practice, however, this margin has not been
upheld. Instead there has been a tendency to use the margin for new expenditure,
leaving little or no room to accommodate increased expenditure due to changing
underlying variables. Several times the Government has had to take short-term
measures in the end of the year to avoid the expenditure ceiling being exceeded. This
has had a tendency to shift focus back again from the medium term to the short term,
at least in the end of the fiscal year.
This problem has been discussed extensively as the Government has been
criticized for its handling of the budget margin. The Government has been requested
by the Parliament to present new guidelines about the size and use of the budget
margin in the Spring Bill in April this year.
5.3. Technical measures to avoid the expenditure ceiling
The lack of margin to the ceiling at the end of the year has led to short-term
measures to avoid exceeding the ceiling. At times such measures have been dubious
from the perspective of the budget law. The Government has for instance decided to
delay expenditure from December to January and there are examples where net
budgeting has been applied even though the budget law prohibits it. These short-term
measures are obviously closely related to the problem of having a too small budget
margin.
6. Concluding remarks
Both the Government and the Parliament have made evaluations of the budget
process, and both have concluded that the reformation has been very successful. The
two evaluations have pointed out the problems mentioned above, but put in
perspective of all the changes that were made, these problems must be considered
limited. There can be no doubt that the new process is far stricter than the old, since it
was strengthened in practically every respect and transformed from one of the weakest
to one of the strongest in Europe.
Figure 6-3. Public sector expenditure, percent of GDP, 1980-2002

70
60
50
40 Sweden European Union
30
80 82 84 86 88 90 92 94 96 98 00 02
How important is the budget process to the outcome of the budget policy? Would
the budgetary situation look different if the budget reform had never taken place?
Concerning this, a few observations can be made. The far-reaching decisions that were
taken in the middle of the 1990s to strengthen the public finances were all taken before
the reformation of the budget process. In the rather critical situation of 1993 there was a
great need to strengthen the budget and no time to wait. When the political will was
strong enough, drastic measures could be taken even without a strong budget process.
After the immediate crises, during the years in the late 1990s the economy was
performing relatively well. Between 1998 to 2000, the economy grew faster than the
European average. Tax revenues were exceeding the forecasts and there was a
substantial surplus in the total public sector. At the same time the expenditure ceiling
was set so that expenditure were reduced as a share of GDP. Between 1997 and 2000
the expenditure ceiling was reduced from 38 to 34 percent of GDP. Since then the
ceiling has been stable.
Figure 6-4. The expenditure ceiling, percent of GDP, 1997-2006
39,0
38,0
37,0
36,0
35,0
34,0
33,0
32,0
31,0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
I believe there is a rather strong case for the claim that the reformation of the
budget process has contributed to the stable development of the public finances since
the middle of the 1990s. A lot of political prestige has been invested in the new system,
both by the Government and the opposition, and the potential political risks of ruining
or undermining the system are considerable. During this period there has been a Social
Democratic minority Government that has negotiated the budget with either one or
two other parties, and the respect for the expenditure ceiling has been a cornerstone all
through those negotiations. As mentioned above, the budget margins have become
very small at times, but measures have been taken to stop the ceiling from being
exceeded. It is not a very brave guess that without the ceiling, the negotiations would
have generated significantly higher expenditures.
It should be added though, that the real test for the system is yet to come. During
the boom years of the late 1990s the surplus target for the public sector could be upheld
with relative ease thanks to the high levels of tax revenue. There was even room to
fulfil the surplus target and cut taxes at the same time. The last few years, however,
have shown a less favorable macroeconomic development. This has led to weaker
growth in tax revenues and increases in expenditure mostly related to health insurance
and unemployment. For the first time since the new budget system was put in place,
there is now simultaneous pressure on both budgetary targets, and the macroeconomic
situation is not helping to ease the pressure.
It is crucial to keep in mind, even though it may seem self-evident, that in the long
run no budget process is stronger than the political will to maintain it. There is a
delicate balance between a system strict enough to have substantial influence on the
budget policy, and loose enough to provide the necessary room for political maneuver.
CHAPTER 7
Medium-term Financial Planning in the Federal Republic of Germany
by
Astrid Lbke, Germany
1. Introduction
This paper provides an overview of the federal government's medium-term
financial planning in the Federal Republic of Germany.
Section 2 describes the historical development of medium-term financial planning
since the 1969 budget reform as well as the need for and the objectives of such financial
planning. The basic features of the content and of the planning period are outlined.
Section 3 shows in detail which parties offer advice and help shape the medium-term
financial planning and sketches the planning procedure. Section 4 takes as a specific
example the current budget preparation procedure to demonstrate how financial
planning can be realized in the same cycle.
2. Principles of medium-term financial planning

2.1. Historical development
The 1969 budget reform set up a mandatory, standardized framework for the
budget law of the central and regional governments. For one thing, the budget
provisions in the Basic Law of the Federal Republic of Germany (Sections 109 to 115)
were revised. For another, the Law on Budget Principles was passed. The Law on
Budget Principles contains the common principles which the federal government and
the Lnder must follow in managing their budgets. The Law on Budget Principles
provides the substructure upon which is built both the Federal Budget Code of the
German federal government, and the Land budget codes of all the federal States
(Lnder) of the Federal Republic of Germany.
The principal innovations introduced by the budget reform in conjunction with the
Law to Promote Economic Stability and Growth (StWG), which had already entered
into force in 1967, were the following:
z In their budget management, the Federation and the Lnder are enjoined to
take account of the requirements of macroeconomic equilibrium (stable prices,
a high level of employment, and external balance accompanied by steady and
adequate economic growth) (Article 109 (2) GG).
z Revenue from borrowing may not exceed the total expenditure for investment
provided for in the budget estimates (Article 115 GG). Exceptions to this rule
are permissible only to remedy a serious and sustained disturbance of
macroeconomic equilibrium.
z The Federation and the Lnder each base their respective budget management
on a financial plan covering a five-year period (Section 50 of the Law on
Budget Principles - HGrG).
z A financial planning council chaired by the Federal Ministry of Finance (FMF)

was set up to submit recommendations for co-ordinating the financial plans of
the Federation, the Lnder, the communes and associations of communes
(Section 51 HGrG).
z In order to ensure that the budgets of the federal government, the Lnder, and
the local authorities can be judged by uniform criteria, a standardized budget
system was developed with the following characteristics:
- system of classification by object: revenue and expenditure are classified by

categories (for example tax revenue, staff expenditure, interest payments,
and investment expenditure).
- system of classification by function: expenditure and revenue are classified
according to tasks of public administration (e.g. defense, environmental
protection and conservation, health).
Since this reform, adopting an approach geared to the medium term has become a
matter of course in all sectors of budget and fiscal policy. As employed by the FMF, the
financial plan has become a useful tool for fending off excessive demands on the
budget. The obligation to draw up a financial plan each year reinforces awareness
within the federal government of the need for fiscal discipline and responsible budget
policy and helps to ensure a balanced fiscal policy. In this way any need for fiscal
policy action can be identified at an early stage.
Though financial planning is a valuable instrument, it cannot guarantee the
implementation of a sound fiscal policy.
2.2. Need for medium-term financial planning
Medium-term financial planning is necessary in order to be able
z to take account of likely developments in public finance in the medium term

when making the decisions incorporated in the yearly budget;
z to identify at an early stage the financial impact of new measures, reform
projects, statutes and the like;
z to assess in good time the scope available for fiscal policy action in future
years; and
z to identify in good time undesirable fiscal policy trends with longer-term
impact and to counter them at an early stage.
2.3. Object of financial planning and planning period
"Financial planning" is taken to mean the actual process of planning, whilst the
"financial plan" is the outcome of this process.
Financial planning is detailed: Financial planning comprises the detailed planning by
budget title of all individual revenue and expenditure positions included in the budget
in the planning period, that is, not just aggregate figures for total revenue and
expenditure. Currently these comprise about 1,100 revenue titles and about 5,300
expenditure titles, grouped in chapters and departmental budgets.
Financial planning is comprehensive: It includes expenditure which can already
definitely be earmarked, and provision for risks which loom in the medium-term
planning period. Sound financial planning includes reserves as well.
Chapter 7. Medium-term Financial Planning in the Federal Republic of Germany 87
Financial planning has a macroeconomic orientation: It is based on the medium-term

projection of aggregate economic development produced by the federal government in
April of each year. From this, calculations can be made in particular of labor market
expenditure on the spending side and of tax revenue on the receipts side. In addition,
for revenue planning the figures produced by the Working Party on Tax Estimates are
also consulted. This working party consists of experts from the federal and Land
governments as well as independent experts.
The financial plan is not a fixed five-year plan of the type used in planned economies,
rather, it is a rolling plan: Each time the budget is prepared, the federal government's
financial planning is also brought up to date by the Federal Ministry of Finance and
rolled forward by one year. In concrete terms this means that at the same time as
preparation of the 2005 federal budget is now getting underway, work is also starting
on converting the currently valid 2003-2007 financial plan into the 2004-2008 financial
plan.
The planning period covers five years:
z the current fiscal year (now: 2004)

z the draft budget for the next fiscal year (now: 2005)
z three more genuine planning years (now: 2006-2008)
A five-year period has shown itself to be most practicable. A longer planning

period would involve too many uncertainties.
For comparison purposes, the published financial plan ought to be drawn up according to
the same system as the budget: In the German financial plan as published, expenditure is
presented in about 40 categories put together in accordance with functional
considerations. However, all medium-term planning data on roughly 5,300 expenditure
and some 1,100 revenue titles in the budget are available within the administration.
The financial plan is a government plan: It is negotiated and revised title by title with
the ministries. Then it is summarized into main categories and submitted to the
legislative bodies together with the government's draft federal budget for the next
fiscal year. There is no need for the financial plan to be adopted by the legislature
(unlike the budget). It is submitted to parliament for information only.
3. Participants and procedures in the medium-term financial planning of

the German federal government
3.1. Participants
3.1.1. Federal Ministry of Finance

The federal government's medium-term financial planning is under the overall
control of the Budget Directorate-General of the Federal Ministry of Finance. All
relevant information is brought together in the Policy Division for the Federal Budget.
In order to identify undesirable trends at an early stage and react appropriately, the
financial planning is continuously being updated within the government. But only once
a year in the course of the budget preparation for the next calendar year is a new
financial plan drawn up and presented to the legislative bodies.
The relevant divisions in the Budget Directorate-General are responsible for
individual departmental sections and chapters of the budget. Early in the year they
start re-evaluating the individual budget titles and estimating the risks connected with
them. Later in the budget preparation procedure they will have to negotiate their
estimates at working level with the departments for which they are responsible.
3.1.2. Financial Planning Council

The necessary co-ordination of the budgets and financial plans of the different
levels of government is undertaken in the Financial Planning Council. This is a co-
ordinating body at ministerial level with the FMF in the lead, comprising in particular
the finance ministers of the Lnder and representatives of the associations of
communes. Representatives of the Deutsche Bundesbank may participate in the
consultations. The Financial Planning Council meets twice a year (spring / autumn).
These meetings strive to reach agreement on future spending policy, which can also
find expression in agreements on expenditure ceilings.
3.1.3. Working Party on Tax Estimates

Estimates of future revenue are based on the estimates (spring/autumn) drawn up
by the "Working Party on Tax Estimates", meeting each year in May. This body
includes representatives of all the financial and economic experts in Germany:
z Federal Ministry of Finance,

z Finance Ministries of the Lnder,
z Six leading economic research institutes,
z Federal Statistical Office,
z Deutsche Bundesbank,
z Board of Experts for the Assessment of General Economic Trends, and
z Federal Union of Central Associations of Local Authorities.
3.2. Procedure
At the start of the calendar year, which is also the start of the budget preparation
year, the financial plan is updated based on the actual figures of the fiscal year which
has just ended. As the new budget is prepared, the individual revenue and expenditure
titles must be re-evaluated. At the same time, during this process account is also taken
of the risks and of any changes in the law. In this way it is then possible to roll the
individual titles forward for another three years and to evaluate them on this basis.
Once a first compilation has been made, the financial plan is first analyzed
internally by the Federal Ministry of Finance in order to identify any need for action in
areas in which particularly strong increases or undesirably high expenditures are
expected. This information may be input into the development of the policy guidelines
of the federal government.
At issue in the budget negotiations between the Federal Ministry of Finance and
the departments are the budget estimates for the coming fiscal year and also
simultaneously the financial plan. Thus the planned figures for next fiscal years are
debated within the government between the Federal Ministry of Finance and the other
federal ministries. The results of this process of consultation result in a further
adjustment of the financial plan.
It is essential for expenditure and revenue to be brought into line with the
aggregate productive capacity of the national economy. To this end, planning must be
based on a medium-term projection of key macroeconomic data, which can enable a
forecast to be made in particular of tax revenue and expenditure for the labor market.
In the German system of financial planning, a forecast of aggregate economic
development is made for the first two years of the planning period, taking account of
discernible trends in economic activity. The medium-term projection of the envisaged

development of the economy in the last three years of the planning period deliberately
makes no allowance for fluctuations in the level of economic activity.
Based on this macroeconomic projection, the Working Party on Tax Estimates
produces a forecast of tax revenues in the financial planning period.
3.3. Links between the medium-term financial planning of the federal

government and of the Lnder
All Lnder and communes in the Federal Republic of Germany are obliged to draw
up a financial plan.
In a country consisting of a federation of states, the fiscal policies pursued by the
different levels of government must be co-ordinated in the interest of achieving the
envisaged development of the economy as a whole. To this end, the financial planning
data of all levels must be combined in an aggregate public-sector budget. This means
that the budgets and financial plans of public authorities must be comparable and must
in particular employ a standardized system of budget classification.
As a member of the European monetary union, the Federal Republic of Germany
must fulfil certain criteria concerning the general government deficit (no more than 3
percent of gross domestic product) and the total amount of public-sector indebtedness
(no more than 60 percent of gross domestic product).
Since in Germany the Lnder and the federal government are autonomous and
independent of each other in their fiscal administration (Article 109 Basic Law), they
need to reach mutual agreement on their spending policies if the "Maastricht criteria"
are to be fulfilled. For this reason the Law on Budget Principles had added to it the new
Section 51 a, which sets up the so-called "national stability pact."
This development has lent increased importance to the Financial Planning Council,
as this has become the body in which the budgetary developments of the federal
government and of the Lnder are discussed and in which recommendations for the
observance of fiscal discipline are drawn up. Spending policies are discussed twice a
year in the meetings of the Financial Planning Council, and if possible agreement
is reached.
The comparability of the data is based on the arrangement of the budget system, a
standard which is equally binding for the federal government and the Lnder; it
formed part of the 1969 budget reform and is laid out in the Law on Budget Principles.
4. Medium-term financial planning in the budget preparation procedure

In view of the close links between the budget and the financial plan it is expedient
for the latter to be drawn up at the same time as the yearly budget. The financial plan
then serves a year later as starting-point and guideline for preparing the new draft
budget and the new financial plan.
4.1. Introduction
The Federal Ministry of Finance (FMF) is responsible for the budget preparation
procedure. The procedure begins one year before the passage of the budget law with a
circular to all government departments to submit their bids/estimates to the FMF.
In accordance with the arrangement of the budget system, as set forth in the
Federal Budget Code (BHO), the budget is subdivided into departmental budgets,
chapters and titles. The departmental budgets contain the funds appropriated for the
ministries. They are subdivided into chapters for the ministry, general appropriations,
subordinate authorities, etc. Each chapter is subdivided into titles, which are the
smallest subdivisions in the budget.
At present the federal budget comprises approximately 1,100 revenue titles and
5,300 expenditure titles, grouped together in chapters and departmental budgets. These
titles are subject to ex-ante control in medium-term financial planning and in the
budget preparation procedure. After the close of the fiscal year, they are audited by the
Federal Court of Audit (FCA) to see if they were managed effectively.
4.2. Schedule
The budget preparation procedure may be seen from the following diagram:
Table 7-1. Preparation of the budget 2005
2003 C ir c u la r t o t h e d e p a r tm e n t s f o r t h e p r e p a r a t io n o f t h e
D ecem ber b u d g e t ( 2 0 0 5 ) a n d t h e fi n a n c ia l p la n ( 2 0 0 4 - 2 0 0 8 )
2004
P r e p a r a tio n o f th e e s tim a t e s o f t h e b u d g e t a n d t h e
Janu ary
f in a n c ia l p la n
F e b ru a ry
M a rc h
D e p a r t m e n ta l b id s s e n t t o F M F ;
S t a r t o f n e g o t i a t io n s F M F / d e p a r t m e n t s a t t e c h n ic a l le v e l
A p r il
M e d i u m - t e r m p r o je c t io n o f e c o n o m ic d e v e lo p m e n t a n d
M ay m e d iu m - t e r m ta x r e v e n u e e s t im a t e ;
F in a n c ia l P la n n in g C o u n c il
June S t a r t o f b u d g e t n e g o tia ti o n s a t m in is t e r ia l l e v e l
C a b in e t r e s o lu ti o n o n d r a f t b u d g e t a n d f in a n c i a l p l a n
J u ly
D ra ft b u d g e t s e n t to B u n d e s ta g (B T ) a n d B u n d e s r a t (B R )
August
f o r d e lib e r a t io n a n d f in a n c i a l p l a n f o r in f o r m a t i o n
S e p te m b e r 1 s t r e a d in g in t h e B T
1 s t r e a d in g in t h e B R
S t a r t o f d is c u s s io n s in B T c o m m it t e e s
O c to b e r
S h o r t - t e r m p r o je c t io n o f e c o n o m ic d e v e lo p m e n t a n d
s h o r t - t e r m t a x r e v e n u e e s t im a te ;
N ovem ber F in a l d is c u s s i o n i n B T b u d g e t c o m m it te e ;
F in a n c ia l P la n n in g C o u n c il
2 n d a n d 3 rd r e a d i n g i n t h e B T
2 n d d e b a te in th e B R
D ecem ber P r o m u lg a t io n o f th e b u d g e t s t a t u t e
2005 S t a r t o f b u d g e t e x e c u tio n
2006 R e n d e rin g o f a c c o u n ts
4.3. Negotiation of the medium-term financial plan
After the circular on the preparation of the budget has seen sent to the
departments, the preparations for the medium-term financial planning begin in the
FMF. This includes a stocktaking of the finished titles of the fiscal year which has just
come to an end and the forecast of future expenditure based on the individual titles.
The tax estimate of the previous autumn as well as discussions of the Financial
Planning Council form the first foundation for the macroeconomic benchmark data.
The bids/estimates of government departments are submitted to the FMF in the
spring and are checked against the principles of need, conformity with regulations, and
economy.
These principles also guide the negotiations which follow at working level with
the departments on the draft budget and the financial plan. The divisions assigned to
examine the titles ("twinned" divisions) are independent of the departments and are
subordinated through the Budget Directorate-General to the executive level of the FMF
(Finance Minister and State Secretaries).
Any issues that cannot be settled at this level are the subject of further negotiations
at directorate or ministerial level. This process is aimed not only at bringing together
the positions of the FMF and of the departments. At the same time, the macroeconomic
benchmark data must be reconciled with the detailed, title by title planning in the draft
budget and with the financial plan. This is an iterative process in which each and every
individual title is negotiated between the FMF and the department concerned and
rolled forward for the financial planning period, so that within the administration
medium-term planning data is available for all the expenditure and revenue titles. The
law the Federal Budget Code requires that a realistic estimate be made for every
budget title, so the key data of the starting position may have to be corrected. In this
process, more recent data is provided by the macroeconomic projection in April and
the latest tax estimate in May.
After having been adopted by the cabinet (June), the draft budget is submitted to
the Bundestag and the Bundesrat for consultation. In addition to the draft budget, the
federal government's financial plan (scope and nature of expected revenue and
expenditure over a five-year period; Section 9 StWG) is also presented. However, the
Bundestag and the Bundesrat merely take notice of the financial plan.
At this point in the budget preparation procedure, the medium-term financial
planning has been completed. The parliamentary procedure concerns only the draft
budget for the coming fiscal year.
4.4. Consideration of the draft budget in the Bundestag
After the first parliamentary reading in the Bundestag, the draft budget is referred
to the Budget Committee of the Bundestag, which takes charge of the subsequent
deliberations.
The Budget Committee scrutinizes all the estimates and, where necessary, submits
proposals for amendment. The decisions of the Budget Committee are prepared by a
number of committee members. The rapporteurs/co-rapporteurs of the Budget
Committee discuss the draft budget with representatives of the supreme federal
authorities concerned and of the FMF and the Federal Court of Audit (FCA). Proposals
submitted by the rapporteurs form the basis of deliberations in the Budget Committee
and are adopted in most instances. A detailed debate is generally held in the Budget
Committee if the rapporteurs and co-rapporteurs are unable to reach agreement on
specific points or if an issue is of such fundamental significance as to require to be dealt
with by the Budget Committee itself.
The comments of the specialized committees and of the Bundesrat are taken up in
the deliberations of the Budget Committee. Issues that cannot be finally disposed of
when the departmental budgets are considered by the Budget Committee are shelved
until the so-called "settlement" session. This is usually held in November, and marks
the conclusion of the Budget Committee's deliberations on the draft budget. The FMF
submits documentation for decisions to be taken in the settlement sessions combining
all the deferred issues and other matters on which it considers a decision necessary (the
so-called settlement items).
This is followed by the second and third readings of the budget law in parliament,
during which minor amendments are made, and the final debate in the Bundesrat.
5. Conclusion
Medium-term financial planning in the Federal Republic of Germany has proved
to be an effective instrument in the quest for a stable fiscal policy.
It is based essentially on the fact that each of the approximately 6,400 revenue and
expenditure titles are re-evaluated at the start of the budget preparation procedure and
planned for the next four years, taking into account the expected risks. At the same
time the Federal Ministry of Finance evaluates the overall economic situation of the
country on the basis of the macroeconomic data. In this way the German federal
government gains an exact picture which is determined by the macroeconomic climate
and which also takes detailed account of future revenue and expenditure.
On the one hand the analysis of these data makes it possible to identify any fiscal
policy action which needs to be taken by the government. On the other hand from the
negotiation of the financial plan with the departments in the budget preparation
procedure there emerges a target specification on the expenditure ceilings for the
coming years. The financial plan forms the basis for the budget preparation and
facilitates the budget negotiations in the following years.
CHAPTER 8
Medium-term Perspectives in Budgeting: The U.S. Case
by
1. Introduction
In the United States, medium-term aspects of budgeting are an integral part of the
annual budget process, and as such are somewhat difficult to disentangle from the
annual budget process. This paper will outline the major rules and process in multi-
year budgeting, covering both the executive and legislative budget processes.
After presenting the necessary institutional background for understanding the U.S.
public sector, the evolution of budgeting, and key actors, the paper turns to multi-year
macrofiscal forecasting in the annual budget process. The paper then turns to sectoral
allocation of resources, including conflict resolution mechanisms. The paper then
briefly addresses the question of policy content and ownership of the executive budget,
the role of the legislature, and in-year spending ceilings. The paper concludes with a
summary assessment of multi-year budgeting in the U.S. and some lessons learned.
2. Background
2.1. Governmental structure and public finance roles.
The United States has a presidential system of government. Power is separated

between three branches of government (executive, legislative, judicial).
z The legislature (Congress) has the ultimate authority over spending, granted
to it under the Constitution.64 Providing authority to spend is exclusively a
legislative power. The executive branch may not spend more than the amount
appropriated, and it may use available funds only for the purposes approved
by Congress. In theory, the Congress could develop a budget without any
presidential input, although this has not been the case in modern history. It is
also important to point out early that while the President does propose one
comprehensive budget, the Congress enacts the budget through 13 separate
appropriations bills or votes, roughly corresponding to function.
z The President and executive branch have important roles in the budget
process in that (i) they propose the budget first, thus setting the agenda; (ii)
they have the detailed spending information needed to prepare good budgets;
and, (iii) the President has Constitutional power to approve or veto legislation
(including the 13 appropriations bills).
64 No money shall be drawn from the Treasury, but in Consequence of Appropriations made by Law.
2.2. Evolution of Budgeting
The relationship between the executive and legislative branches over budgeting
and control of public finances has developed gradually over the 200 years of the
countrys history, with the balance of influence over spending shifting between the two
branches. Until 1921, spending agencies submitted budget requests directly to the
Congress. While Congress had preeminent authority in spending decisions, it was very
dependent on the budget estimates and requests submitted by spending agencies. As
the complexity of the public finances expanded, this uncoordinated approach placed
too high a burden on Congress to prepare a coherent fiscal position, and also provided
too little formal room for Presidential policy in advancing a coherent executive fiscal
position.
The Budget and Accounting Act of 1921 established the executive budget process,
requiring the President to prepare and submit a comprehensive federal budget to
Congress annually. The Act also created the Bureau of the Budget (renamed the Office
of Management and Budget (OMB) in 1974) to support the President in carrying out
coordinating the executive branch budget. OMB assists the President in overseeing the
preparation of the Federal budget and also supervises budget execution
/implementation. OMB evaluates the effectiveness of agency programs, policies, and
procedures, assesses competing funding demands among agencies, and sets funding
priorities. OMB ensures that agency reports, rules, testimony, and proposed legislation
are consistent with the President's budget and with Administration policies. In
addition, OMB oversees and coordinates the Administration's procurement, financial
management, information, and regulatory policies. And, not a small task, OMB is in
charge of general management oversight and reforms in the public sector.
Along with OMB, the Council of Economic Advisors and the Treasury Department
assist the President by providing economic projections and revenue estimates.
With the passage of the Budget and Accounting Act, influence over budgeting
shifted to the executive branch. This tendency generally continued until the early 1970s,
when Congress began to reassert its role. By the mid-1960s, the size and complexity of
the Federal budget had grown significantly. Some programs from the Depression
(Social Security) and War on Poverty (such as Medicaid and Medicare) were large and
growing, and were entitlement programs not governed by the annual budget. To
manage these programs required more than an annual focus in the budget.
In 1970, Congress passed the Legislative Reorganization Act, requiring more
budgetary information from the President, including that the budget cover the budget
year and four additional years.
In 1974, Congress passed the Congressional Budget and Impoundment Control
Act. The Act further reinforced the congressional budget role establishing important
tools for managing its handling of the budget. The House and Senate Budget
Committees were created to coordinate congressional consideration of the budget, and
the Congressional Budget Office was established as a source of nonpartisan analysis
and information, including providing an alternate economic forecast for the Congress.
The law also established a specific timetable for action on the budget and a concurrent
budget resolution on budget aggregates (a legislative device that does not require
Presidential approval, but guides legislative action on the budget).
CBO's role is to provide the Congress with objective, nonpartisan analyses on
economic and budget decisions, and with information and estimates for the budget
process. CBO prepares analyses and estimates relating to the budget and the economy,
and presents options and alternatives for the Congress, but does not make policy
Chapter 8. Medium-term Perspectives in Budgeting: The U.S. Case 95
recommendations. CBO helps the Congress formulate a budget plan, stay within that
plan, and assess the impact of federal programs.
Congress also has the Congressional Research Service, which provides research on
specific topics for Members of Congress, as well as the General Accounting Office
the Supreme Audit Institution or external audit body that provides evaluations of
programs in addition to audit reports.
To understand the U.S. budget process generally, including multi-year budgeting,
it is important to understand both the executive and legislative budget processes.
3. Foundations of budget preparation: macrofiscal

3.1. Executive branch macroforecasting, expenditure forecasting, and fiscal
envelope
Formal econometric models are not used in preparing economic forecasts of key
economic variables for use in the President's Budget. Professional economists in OMB,
Treasury, and the Council of Economic Advisors (the 'troika') focus on trends in six key
variables:
(1) real GDP growth;

(2) CPI (higher CPI lowers indexed tax revenues and increases budget spending
via automatic cost-of-living allowances, etc.);
(3) GDP price index (higher GDP growth increases tax revenues);
(4) unemployment;
(5) short-term interest rates (90-day Treasury Bill rate); and,
(6) long-term interest rates (10 year Treasury bonds yield).
Some attention is also paid to projected income distributions, which will effect tax
revenue projections (profits versus wages). Policy discussion focuses more on the
wedge between the GDP price index and CPI than on each measure itself. The CPI
generally out-paces the annual rate of GDP growth. The rule of thumb: the wider the
gap between the two measures, the worse for the Federal Budget.
In debating the appropriate level of key economic variables for use in the
President's Budget, participants do receive other forecasts as benchmark levels (several
academic private sector estimates, as well as estimates by international organizations).
The existence of these alternate estimates is thought by some to place constraints
on the extent to which political officials might manipulate economic estimates beyond
reason in order to support their political agenda via the budget. Regardless, alternate
estimates do provide substantive grist for the debate.
The process of forecasting is really an incremental one. Analysts begin with the
most recent official forecast, and then proceed to review the key variables in light of
trends in the economy. The current forecast is modified where a consensus is reached
among OMB, Treasury, and CEA on necessary changes.
Generally, only first-order effects on the economy are considered. Professional
economists in the executive branch are loath to attempt second-order estimation (e.g.
the effect on tax revenues of a change in tax laws which change behavior).
Forecasts are updated every six months. Executive branch economists constantly
monitor key statistical releases on economic activity, and are always thinking of the
implications of the news for the official forecast.
3.2. Process: from policy approval to budget production
Once staff of the troika have finished their review and developed opinions on
necessary revisions, senior policy officials meet to discuss the forecast. Eventually, the
Chairman of the CEA, the Director of OMB, and the Secretary of Treasury must agree
on the final numbers before they can be used to develop the Budget.
Policy consensus must be reached for economic forecasts. The key policy officials
involved in fiscal policy must reach some consensus on broad policy directions as they
consider the economic forecasts. (In the 1980s, this policy review often resulted in more
optimistic estimates, which led to the appellation 'rosy scenario' forecasting based on
economic theory dubbed 'voodoo economics.')
Generally, the key policy officials will include the troika, as well as senior
administration officials (possibly the President, Vice-President, Chief of Staff, and other
White House political staff). These arrangements have varied by Administration. Under
President G. H. W. Bush, domestic and economic policy were the domain of two very
influential officials, the Chief of Staff John Sununu and the Director of OMB Richard
Darman.
Under President Clinton, a formal advisory body called the National Economic
Council (NEC) was established in 1993 to provide advice to the President and build
consensus among key policy officials. The Council was composed of the President,
Vice-President, seven domestic Cabinet Secretaries, the Secretary of State, Director of
OMB, Chair of the Council of Economic Advisors, National Security Advisor, Assistant
to the President for Domestic Policy, and Secretary of Defense.
The important point to bear in mind is that there were vehicles for achieving policy
consensus within Government. As of 1993, these have become more formalized in the
NEC. This would be roughly analogous to Cabinet consensus in Parliamentary systems.
3.3. From forecast to budget preparation
Once approved, CEA staff use a private sector econometric model (dubbed the
'black box') to produce somewhat greater detail for the forecast, especially
consumption, investment, and export components of the national accounts. Also,
greater detail in terms of income distribution is sought for use by Treasury in
forecasting revenues. While based on an econometric model, the black box really is an
accounting device or reverse model, where key variables are entered, and consistent
details are generated. The black box assures that basic national accounting identities are
observed to produce a generally consistent set of supporting detail for the approved
policy variables. This black box output detail is then given to OMB and Treasury for
production of further detail.
The U.S. Treasury is responsible for tax-based revenue forecasts, while OMB
generally estimates user fees and similar non-tax revenues. Treasury has formula based
on historical tax collection records of estimated revenues given the basic national
account and income distribution data.
OMB has simple formulae based on recent historical trends and simple regression
models which provide greater detail on inflation, a wider array of interest rate
estimates, etc. (e.g. Medical inflation is, on average, 2 percent greater than the CPI
estimates). These more detailed estimates are then provided to the executive branch
agencies for use in estimating future year or budget year costs. Agencies use estimates
of inflation and possible pay increases in calculating their budget requests (policy) and
current services estimates.
The set of variables which the Federal Executive Branch has determined are
necessary for producing the budget are listed in Box 8-1.
Box 8-1. Economic variables used in developing the President's Budget
Basic Budget Data

1. Pay and personnel-related costs for civilian employees
2. Pay and personnel-related costs for military personnel
3. Non-pay costs for discretionary programs
Alternate Price Measures

4. GDP Deflator (chain-weighted);
5. CPI-W (quarterly and annual);
6. States-local expenditure Index;
7. Federal non-pay, non-defense expenditure deflator;
8. Federal non-defense non-residential structure construction;
9. Private construction non-residential structures.
Revenue-related estimates
10. GDP (quarterly and annually)
11. GDP Price Index (quarterly and annually)
12. Civilian unemployment (quarterly and annually)
13. Corporate profits
14. Goods imports
15. Personal income, broken into wage and non-wage income (interest income, dividend
income, rental income, proprietors' income) (quarterly and annually)
16. Civilian labor force (quarterly and annually)
17. Treasury bill rates (91, 182, and 364-day rates (quarterly and annual)
18. Treasury notes and bonds (2, 3, 5, 7, 10, and 30-year rates) (quarterly and annually)
19. Prime rate (quarterly and annual)
20. LIBOR (quarterly and annual)
21. AAA state and local bond rate (quarterly and annual)
22. Mortgage rate (quarterly and annual)
23. Energy prices (imported and domestic oil, coal, natural gas, and a hybrid gas and oil
index) (quarterly and annual)
24. Housing starts, mortgage originations (existing and new homes), average single-family home
3.4. Expenditure trends
Many of the routine budget tasks, such as applying a pay inflator to base year pay
and projecting into future years, are automated in the OMB budget database system.
The appropriate economic estimate is fed into the OMB computer, and the relevant
accounts are directly updated.
In early budget discussions at the macro level, OMB uses the most recent approved
budget information (congressionally-approved budget), adjusted for information on
implementation of the current budget, to make projections of spending trends. Several
concepts unique to the U.S. budget system need to be understood; baselines,
discretionary spending, and mandatory spending.
Baselines provide the President and Congress with a framework for making and
enforcing budgetary decisions, serving as a benchmark. OMB and CBO baselines
project federal spending, revenue, and budget surplus or deficit amounts that would
occur if existing budget policies were left unchanged. The rules for calculating the
baseline are set in law.
Discretionary spending applies to programs whose funding must be provided
annually in the budget. Mandatory spending encompasses entitlements to individuals
and other programs whose spending is controlled via separate legislation, and not
subject to the annual budget.
Mandatory spending and receipts are generally assumed to continue at the level
prescribed by existing law. Discretionary spending is assumed to continue at the level
of the current year's spending level, adjusted to reflect certain assumptions. The
projections for mandatory spending, receipts, and discretionary spending are based
upon economic assumptions (e.g., economic growth, inflation, and unemployment) and
technical assumptions (e.g., demographic and workload changes).
There is no formal or informal discussion between executive branch staff and staff
at the Federal Reserve on economic forecasts. However, Washington, D.C. based
economic professionals do have a general idea of what others in their field are thinking.
The Federal Reserve has their own model, and do not share their forecasts. The
Federal Reserve MPS model does, however, have some close similarity to the private
sector economic model the executive branch uses for obtaining more detailed estimates.
However, since the President's Budget is released publicly, including the basic
economic assumptions underlying the budget, the Federal Reserve has the basic
information on Executive Branch economic thinking to use in their own forecasts and
decision-making.
3.5. Congressional Budget Office Forecasting
Once the President's Budget economic forecasts are released with the President's
Budget, the Congressional Budget Office makes their own set of estimates. CBO is
thought to follow a similar consensus and incremental approach to forecasting.
However, CBO does invite private and academic forecasters to present their views of
the economy. Also, CBO, being a single agency, has a somewhat easier time in arriving
at consensus estimates, versus the executive branch with three separate agencies
involved. CBO does subscribe to some private sector forecasting services, though which
ones, and how these private forecasts are used in developing the official CBO forecast,
is not known.
OMB and CBO do have some different assumptions and conclusions on income
distribution and some key variables. Particularly with regards to income distribution,
OMB has tended to estimate greater GDP growth and greater corporate profits
which are highly taxed and hence greater Federal tax revenues, than CBO has
projected.
Because of lingering suspicion over OMB economic estimates after President
Clinton's inauguration, the President declared that CBO economic assumptions would
be used instead of OMB's. While only lasting for a short while, the President's decision
did serve to remove different economic estimates as a source of budget debate, and also
chastened the troika to produce more pragmatic estimates.
Publication of the economic estimates in the Presidents Budget also allows close
scrutiny of the soundness of the estimates.
3.6. Revenue forecasting
As noted above, within the executive branch, the Treasury has general
responsibility for tax policy, including estimating the effect on revenues of changes in
tax policy.
Most of the laws establishing the federal government's revenue sources are
permanent and continue year after year without any additional legislative action.
Revenue legislation may be considered in the congressional budget process when
President proposes changes in taxes.
3.7. Debt
Within the executive branch, the Treasury also has responsibility for debt policy,
Because of the size of the economy, debt policy has generally not been a major focal
point in budget debates, except for the 1990s under President Reagan, and again now
with the rapid increase in deficits. Most of the focus has been on taxes (revenues) and
spending, with deficits and debt a residual. (Note that for smaller or less open
economies with more limited access to capital markets, debt policy is a far greater
issue.)
Perhaps the more immediate administrative issue for the Treasury in debt policy is
Congressional limits on debt. The gross federal debt consists of the debt held by the
public plus the debt held by government accounts. Almost all of the gross federal debt
is subject to a public debt limit. The Treasury can borrow funds only up to statutory
debt limit. Most gross federal debt is held by the public, though a significant amount
(roughly 30 percent) is held by government accounts (primarily trust funds such as
Social Security, whose surpluses must be invested in special federal government
securities by law).
Congress occasionally will enact increases to the debt limit. Within the
congressional budget process, the annual congressional budget resolution specifies the
appropriate level of the public debt for each fiscal year covered by the resolution. In
order to actually change the public debt limit, however, legislation must be passed by
both Houses of Congress and signed by the President. The House of Representatives
Ways and Means Committee and the Senate Finance Committee are the responsible
committees of Congress for debt limit issues.
In addition to the annual budget process, debt limits are another means for
Congress to control the executive branch, and force the executive to discuss debt issues
directly.
A Mid-session Review is also required by the law to be sent to Congress by July 15
of each year. The Mid-session updates the economic estimates and Presidents Budget
proposal.
4. Strategic Allocation
4.1. Sectoral ceilings
In the U.S., sectors are termed functional classification. The President's budget
classifies federal budgetary activities into functional and subfunctional categories that
represent the major purposes of the federal government. Each budgetary activity of the
federal government, including budget authority, outlays, tax expenditures, and credit
authority, is classified into a subfunction based on the primary purpose it serves
without regard to the agency or other unit responsible for it. There are currently 21
functional and 77 subfunctional categories. Figure 8-1 presents the current functional
and subfunctional classification.
The functional categories provide a broad statement of budget priorities and
facilitate the analysis of trends in related. They are not used for setting ceilings.
Figure 8-1. Functional Classification

Indicative ceilings for the annual budget process in the U.S. are set according to the
ministry (department) classification set for the budget year, and four subsequent years.
Generally, OMB begins with the second year of the prior budgets multi-year forecast,
and makes adjustments for total spending trends and limits, economic or legislative
changes, and also to reflect known policy preferences. A new policy reserve has
sometimes been set at this stage, representing a pool of unallocated resources that will
be used to fund new initiatives (whether arising from the President or from ministries).
In some cases these are provided to the President for information, and in other
instances not, depending on the degree of interest of the President in budget issues and
the degree of delegation to staff and the Director of OMB.65
The ministry ceilings are indicative. OMB requests that agencies stay within the
ceiling, but there are no formal sanctions for not doing so. Because the ceilings are not
Presidentially-approved or sanctioned, they have few teeth. Even if there were, it is
highly unlikely a cabinet member would be reprimanded or dismissed for not meeting
the ceiling. However, there are more informal pressures for compliance that operate
more at a career staff level. In advising the Director of OMB and President, OMB staff
do make note of whether the ministry complied, and by how much they were off
ceiling. Further, even if the ministry does not specify where they would reduce to meet
the ceiling, OMB staff will recommend where reductions should take place. So, it is in
the interest of ministries to themselves identify low priorities.
Further, implicitly, there is always some question in ministry budget staff minds as
to whether OMB scrutinizes their budget more closely if ceilings are ignored. (For small
agencies, over which OMB has more direct influence, there is greater likelihood of
compliance with the ceilings.) Regardless, ministry requests are generally within 5-10
percent of ceilings.
4.2. Conflict resolution and consensus-building
After receiving ministry requests, OMB prepares policy papers for the Presidents
decision. There is a well established process for decision-making, with only the most
significant issues ever making it to the Presidents desk. Many will be made by the
OMB director, and for smaller issues, by political appointees within the OMB.
After decisions are taken, the decisions are passed-back to the ministries, usually
around Thanksgiving (November). Ministries are given a tight deadline for reviewing
and formally registering any appeal to the passback levels. This process of passback
and appeals can be very contentious between OMB and ministries.
The process of resolving appeals involves direct meetings between OMB staff and
ministry budget staff and focused debate. Efforts are made to resolve as many issues as
possible at a staff level, before raising the issues to a political level. Eventually, some
issues may need to be resolved by the Director of OMB and affected cabinet secretary.
Generally, few issues require appeal all the way to the President. Some Presidents
(Ford, Carter, and Clinton) have taken a direct role in deciding appeals. Others (Nixon,
Reagan, George H.W. Bush and George W. Bush.) have delegated appeal decisions to
presidential aides or the OMB Director.
65 While the formal Presidents Budget document includes the budget year and four additional year, down
to the program level, over the last decade or so the Senate has included language in the budget resolution
requiring ten-year estimates. The executive has generally complied with this request, submitting these estimates
at a more aggregate level.
Throughout this process, it is the better prepared, better argued, and better
documented case that tends to win out. For OMB and agency staff, the premium tends
to be on data and analysis. For highly political issues, of course the political officials
will take the decision and weigh many factors. Even the best analysis will not win out
against politics. But for the majority of issues OMB would deal with during this time,
data, documentation, and analysis are the key factors.
Out-year estimates can sometimes also be used in resolving disputes with agencies,
at least for discretionary spending. There have been instances where an agency agreed
to a lower increase in its proposed budget in exchange for increases in its budget for the
budget year +1 or beyond. Of course, the subsequent years ceiling will be revisited in
light of changing circumstances during the next budget cycle, and the agreed increase
may not in fact happen.
With this in mind, OMB puts a premium on hiring the best staff, and having access
to as much data as possible. It is not uncommon for, OMB staff outnumbered by
budget staff in ministries armed with agencies own data, to provide critical
information or data at a key moment in negotiation that wins out when the agency
may not have itself reviewed its own data.
Once agreements are reached and appeals resolved, the ministries are responsible
for preparing the budget documentation and justification. In addition to the formal
Presidents Budget, each agency will prepare a separate budget justification for
transmittal to Congress, as well as congressional testimony for their executives. OMB
reviews and approves all these documents before they are sent to Congress, to assure
they are all in alignment with Presidential/administration decisions and policies.
Agencies do have direct relationships with Congressional committees and staff,
and have been known to provide information informally which sometimes undermines
the budget. However, OMB staff closely monitor these discussions, and do report back
to the Director of OMB on ministry behavior. Egregious breaches might solicit a call
from the OMB Director to a Minister, but these are not common.
4.3. Final budget allocation
Final budget allocations (passback and appeals) are set relative to the ceilings. This
has been especially true with the enactment of legally binding ceilings on discretionary
and mandatory spending through the Budget Enforcement Act (BEA). While the BEA
addresses aggregate spending levels, these are translated into ministry ceilings, and the
sum of ceilings must of course match the legal limits.
The actual ministry budget allocation, while taken relative to the ceiling, is really
determined by the accumulation of individual decisions on programs and policies. The
policy papers prepared for decision (referred to above) incorporate extensive analysis
by OMB staff of agency budget requests, as well as other research or findings (e.g. audit
findings) regarding programs and policies. The analysis results in OMB staff
recommendations for specific changes in funding, programs, policies, employment, etc.
It also reflects the findings of OMB staff from budget hearings held with each agency
(at which each agency can make their case for their request). As each decision is taken,
its effect relative to the ceilings is monitored to assure the totals are not exceeded.
While only some appeals will be addressed directly by the President, in the end it
is the Presidents Budget that is transmitted to Congress, under his signature. The
President is ultimately responsible for its content. The Cabinet does not formally
approve the budget.
4.4. Roles
The U.S. does not have a formal MTEF. The budget process has evolved,
particularly after 1974, to incorporate medium-term projections in the process, as well
as ministry strategic plans. Generally, the relationship between OMB and ministries has
not changed dramatically due to the introduction of medium-term forecasts. Other
factors influence relations more. For example, during periods of deficits, Presidents
with a policy of spending reduction, and OMB Directors with high influence over
domestic policy, OMB tends to be stronger relative to the ministries, and during these
times OMB may be less collegial in relations with ministries. At other times, OMB
influence may be less, and relations with Ministries become far more collegial.
That said, there was some movement under President Clinton towards a more
collegial approach by OMB. Early in the Clinton Administration, a survey was
conducted of spending ministries as to what they thought of OMB, and barriers to
agency efficiency. Not surprisingly, the results were rather negative, with ministries
feeling OMB was too much into rule-mongering and second-guessing, and too high-
handed. The results were at first surprising to OMB staff, and certainly not accepted.
After a series of brown-bag discussions between staff and the OMB Director, some
acceptance of the results seemed to sink in, with staff understanding that to really be
effective, the ministries needed to be partners and understand the overall constraints.
No subsequent surveys have been taken, so it is not clear if the ministries would report
differently now than a decade ago.
5. Policy and budgets

Generally, because of the requirement for the President to submit a budget, and
that the budget embodies the President and Administration policies, it is owned.
Presidents can and do issue veto threats to Congress over exclusion (inclusion) of
issues that they support (oppose).
Much of the Presidents involvement in the budget takes place through the OMB.
When macroforecasts are developed and the budget outlook prepared, OBM advises
the President and presents policy recommendations. The President makes policy
decisions, which are passed to agencies through OMB. These initial policy decisions
may set out broad budgetary goals, or may include specific policy proposals and
spending ceilings to guide agencies in preparing their budget requests. Policy guidance
on budget development often varies from year to year, depending on current economic
and political conditions.
As noted above, ministries develop their budget requests. Direct Presidential
involvement in many details is limited by the nature of the process. Presidential
involvement tends more to policies and some specific proposals.
With respect to multi-year budget estimates, it is important to note that the
Presidents Budget includes two sets of estimates for Congress: current services
baseline and Presidents policy (showing the multi-year impact of the Presidents policy
choices). These aid in transparency, and allow quick identification of those areas where
major budget policy changes are proposed.
6. Legislature
As noted in the introduction, the Congress has the authority to change the
executive budget proposal completely, including total revenue, total expenditure,
deficit, debt, and allocations to ministries. This authority is in practice counterbalanced
by the executive authority to veto the budget bills.
Over time the congressional budget process has been improved to give Congress
greater ability to manage its own affairs. General examples of these from the early
1970s were mentioned in the introduction. This section will go into the congressional
budget process in more detail.
The Congressional Budget and Impoundment Control Act of 1974 established the
basic structure for the congressional budget process, creating the House and Senate
Budget Committees and CBO. The Congress budget process starts in February with the
submission of the Presidents budget, after which the House and Senate Budget
Committees begin developing their budget resolutions.
The annual budget resolution establishes total spending and revenue for five or
more years, allocates spending by functional classification, and includes CBO
baseline estimates of revenues. The spending amounts are then allocated to the
House and Senate Appropriations Committees for discretionary spending, to the
authorizing committees for mandatory spending, and to the House Ways and
Means Committee for revenue targets. The Appropriations Committee of each
house subdivides the discretionary allocation among its 13 subcommittees (broadly,
the major functions). The committee allocations are one means of enforcing the
aggregate spending levels in the budget resolution by holding committees
accountable for meeting their targets.
After extensive hearings, each of the subcommittees reports one of the 13 regular
appropriations bills to its respective full appropriations committee, which in turn
reports the bills to their Chamber. A cost estimate of each bill is prepared and
compared to the amount allocated or subdivided to the relevant subcommittee. The
Congressional Budget Act prohibits bills from exceeding the budget resolution
committee allocations. However, to be enforced, a Member of either Chamber must
raise a point of order when the bill is being considered on the floor of the respective
Chamber.
For mandatory spending, revenues, and debt limits, the process of reconciliation
(also established under the Congressional Budget and Impoundment Control Act) is
used to enforce the budget resolution allocations. Congress includes reconciliation
instructions in the budget resolution directing the authorizing committees to
recommend changes in law to achieve the direct spending, revenue, and debt limit
targets in the budget resolution. The Budgets Committees of each Chamber then
compile the various proposed legislative language from the authorizing committees
into a reconciliation bill. Unlike the process for discretionary spending, there are no
points of order for a reconciliation bill that does not meet the budget resolution
mandatory spending targets. Amendments on the floor of each Chamber may seek to
bring the reconciliation bill closer to the budget resolution targets, but that is entirely at
the discretion of a Member, and subject of course to a successful vote for the
amendment. For revenues, a point of order may be raised for amendments that would
reduce revenues below the budget resolution.
The Congressional Budget and Impoundment Control Act does prohibit either
Chamber from considering amendments to a reconciliation bill that would increase the
deficit. The law also provides a congressional budget process timetable; the budget
resolution should be adopted by April 15, and the reconciliation bill by June 15. (These
are frequently not met.)
The Budget Enforcement Act of 1990 also required that any measure reducing
revenues below the budget resolution or increasing mandatory spending be offset with
equivalent revenue increases or mandatory spending reductions (commonly referred to
as the pay-as-you-go rule). This is also enforced by a sequestration process, whereby
any violation will trigger an offsetting sequestration of nonexempt mandatory
spending programs.
7. In-year Effect
Generally, OMB has been responsible for costing new legislation. As a practical
matter, not all legislation is costed. Only executive branch proposals, and legislation in
Congress likely to be passed by Committee and proceed to the floor.
During the budget process, scorekeeping (measuring the budgetary effects of
pending and enacted legislation against the baseline) is used to compare proposed
budget policy changes to existing law and to enforce spending and revenue levels
agreed upon in the budget resolution and other budget legislation, such as the
discretionary spending caps and pay-as-you-go rules.
During the fiscal year, scorekeeping is also used for new legislation. As OMB must
clear all legislation before transmittal to Congress, no uncosted or unfunded legislation
is transmitted from the executive. All proposals must fit within the ceilings.
8. Discussion and overall assessment

This presentation of medium-term budgeting in the U.S. has focused on the
process, technical details, rules and roles. It has tried to provide both the hard facts,
and some of information on how it works in practice. But to the reader it may appear a
dizzying blizzard of rules and procedures. This summary tries to draw out the key
features of the process, and also the key strengths and weaknesses of the process.
One aspect of note is in fact the interplay between executive and legislature. The
competition between the two branches has led to innovations in budgeting and a fairly
high degree of accountability in public finances. That is not to say all spending is
efficient. But there are pressures to become more efficient.
For the medium term aspects of budgeting, the important points are:
(1) In considering policy and programs, thinking beyond one year to the outyear
effects of proposals and existing spending is common throughout the system.
Over a quarter century of requirements for multi-year estimates, and
progressively greater attention to these estimates, have made multi-year
thinking a normal part of public finance.
(2) The baseline estimates serve as a useful reference point for considering changes
in policy and programs over several years, and better inform decision-makers.
This has likely led to different choices in programs, and in reallocation from
lower to higher priority programs, early in the budget process but it is not
possible to systematically measure this effect.
(3) Beyond the information effect noted above, including use of indicative ceilings
in the budget process, the multi-year estimates had limited bite. Not until the
large deficits of the 1980s and general political and public consensus that
deficits were bad did multi-year estimates and ceilings gain real teeth. And
further, not until the Gramm-Rudman-Hollings and Budget Enforcement Act
legislated ceilings and pay-go requirements did the real effect of ceilings and
estimates come to be felt.
(4) Some means of bringing consensus on multi-year policy is critical. It is this
consensus that shapes behavior and changes direction the most. For the U.S.,
consensus within the executive branch is achieved through the NEC. Within the
legislative branch, it is achieved via the negotiations that take place in the
Budget Committees in crafting the budget resolution (frequently including
negotiation with the executive branch). And finally, the Budget Enforcement
Act represented the executive-legislative consensus necessary in the U.S.
political system to really begin to reign in deficit spending.
(5) Despite the presence of multi-year estimates for a quarter century, this has not
prevented growth in spending. It has influenced choices, and might even have
contributed to a shift towards more mandatory spending programs not directly
under annual budget control. For the U.S., information on future impact has not
been enough to effectively limit spending growth. This might be part of the
curse of riches, where deficits and postponement of decisive choices on fiscal
matters can be financed given the easy access to international capital markets
enjoyed by the U.S. The impact of multi-year information might be greater for
countries with less easy access to capital markets. For the U.S., hard, legally
binding fiscal targets were required to make a difference.
(6) The Gramm-Rudman hard targets proved too demanding, and the targets could
not be reached through the political process. The BEA proved more successful,
by limiting discretionary spending and requiring off-sets for deficit-increasing
changes to entitlement programs. These procedural rules changed the
incentives in the budget system to limit spending growth, and this proved more
successful than hard deficit reduction targets.
(7) With the short-lived budget surpluses in the late 1990s and early 2000s, the
consensus on budget discipline broke down, and the BEA lost emphasis. The
multi-year estimates are firmly entrenched and still prepared, and at least
provide the political officials with clear information on the direction they are
heading. The availability of this information does likely serve as some break on
excess, but again the teeth are lacking. However, as the outyear picture
worsens with continuous provision of multi-year estimates, the pressure to
contain spending will likely reassert itself a self-correcting mechanism, albeit
with lags in effect. Even in the absence of hard legal constraints, multi-year
information is of value.
While the deficits are large, and there is increasing suggestions of the imprudence
of current policy, some of the expansionary fiscal policy is intended to boost economic
growth. The policy is intended as counter-cyclical just as large surpluses were
incurred during years of strong growth and used to pay down the debt, so now deficits
are incurred. One can debate whether it is really needed, or needed in the size and form
in which it is incurred (tax cuts). It does not mean the underlying budget system is
entirely broken. Where a policy decision is taken to operate an expansionary fiscal
policy, it does not mean the underlying process is dysfunctional. It would be if the
expansionary policy were taken absent information on likely outcome, but that is not
the case.
Part II
Performance Management
CHAPTER 9
An Overall Framework for Considering Performance Management
by
Joel Turkewitz, The World Bank
1. Overview
Moving This paper provides a brief overview of major issues in the design of
performance-enhancing reforms in the public sector. As this paper is written as an
input into a dialogue between the Government of Korea and the World Bank on
improving public expenditure management, specific particular attention has been
given to experience and learning in achieving greater performance orientation through
alterations in the allocation and spending of public financial resources. The first section
of the paper reviews some current thinking about the fundamentals of improving
performance in the public sector. The second section presents a discussion of different
approaches to generating enhanced performance orientation in the budget process, and
the final section examines the evidence to date on the impact of performance-oriented
reforms and the implications of that evidence on the design of performance
management programs.
2. Performance in the public sector

Performance is not a new idea in the public sector. The earliest state bureaucracy,
established in ancient China, was founded upon the importance placed on a highly
trained cadre of people working in the interests of the state. As the public sector has
evolved over time, it has consistently retained a commitment to performing work in the
public interest in an effective and efficient manner a concept enshrined in the notion
of a public service ethic.
While the general concept of performance has a long history, the idea of
performance in the public sector has been reinvigorated over the course of the past two
decades. As governments expanded in the post-World War II era, pointed questions
started to be asked by the voting public as to what was being accomplished. The
greater the external pressure on politicians to demonstrate that money was being well
spent, the greater the internal need on public servants to identify and quantify public
sector products and positive societal outcomes. Performance in the public sector
became increasingly associated with measuring, monitoring, and evaluating the
outputs of projects and the outcomes of policy initiatives. 66 The last 20 years has
witnessed the introduction of reforms throughout the developed and developing
worlds designed to orient public sectors towards the more effective and efficient
delivery of services.
66 For a discussion of the idea of performance in the public sector, see Schick (2003).
110 Part II. Performance Management
The drive to improve public sector performance has given a new look to the public
sector as it entered the 21st Century. Processes and organizations have been remade
through innovations like program budgeting, medium-term expenditure frameworks,
accrual accounting, and performance auditing. Privatization and decentralization have
reshaped the state, while contracting-out and performance contracting has modified
public sector employment. Enhanced monitoring and evaluation, program reviews,
and increased civil society participation in setting policy and reviewing results have
changed how managers and politicians are held accountable for results.
Decades of reform efforts have achieved mixed results. Initial expectations in
countries like the United States and Sweden that improved performance could be
rapidly achieved through changes in the budget process have failed to materialize,
while experiences in such places as New Zealand, South Africa and many smaller
public sector organizations has been much more positive. The inability to state
definitively whether performance management works is not a data problem but instead
reflects the fact that reforms vary greatly in their objectives, the traditions and histories
of the countries undertaking reform, and the variations in the reforms that have been
tried. Some countries that have undertaken performance-oriented reforms have
succeeded in improving their overall public finance performance, reallocating
resources across sectors, and improving efficiency within programs. In other countries,
reforms appear to have done little more than reshuffle arrangements within the public
sector with no discernable impact on public finance or service delivery.
Reforms that have succeeded in creating well-performing public sector have
benefited from inputs that can neither be planned nor designed such as good timing,
effective leadership, and governmental stability. Successful reforms have drawn upon a
generally accepted group of ideas regarding the organization and operation of high-
performance public sector environments. As stated by Allen Schick (1999), those ideas
include:
z Performance improves when managers are told what is expected of them, and
results are measured against these expectations.
z Performance improves when managers are given flexibility in using resources

to carry out assigned responsibilities.
z Performance improves when operational authority is devolved from central

agencies and departmental headquarters to operating levels and units.
z Performance improves when government decisions and controls focus on

outputs and outcomes rather than on inputs and procedures.
z Performance improves when managers are held accountable for their use of
resources and the results that they produce.
It is interesting to note that ideas about performance improvement look not to

specific processes or rules but rather at the role of managers and the alignment of
managerial incentives. Performance-enhancing efforts that have focused on adopting
new techniques, even those as far ranging as the implementation of a medium-term
expenditure framework, have succeeded in impacting on performance only to the
extent that they have generated a change in the informal rules and culture that defines
how managers operate.
Chapter 9. An Overall Framework for Considering Performance Management 111
Convergence on the factors that generate performance improvements, does not

equate to convergence in reform design. Efforts to increase performance in the public
sector are as varied as the countries that have designed them. At the same time, the
individual reform programs that have been developed have at their core a limited
number of drivers of change. Some reforms seek to engender improvement through
increased competition, some by greater managerial authority. Others look to enhanced
program oversight and review, or deregulation and simplification to generate change.
A particular reform might draw upon multiple drivers of change, although not all
combinations are equally compatible.67 For example, New Zealand has worked hard to
enhance competition in the public sector, but has combined this with increased
attention to program review. Canada, on the other hand, has repeatedly worked to
introduce program review, while the U.S. has tried a combination of enhanced
managerial autonomy and program review.
There is no evidence to suggest that any single driver of change dominates other
models. Successful reforms are crafted around the strengths of existing organizations
and institutions and are specific to their environment. The importance of categorizing
reforms based upon their central change strategy is that each strategy raises a specific
set of challenges. Reforms designed around enhancing competition need to find ways
to minimize transaction costs and ensure that clients have sufficient information to
make informed choices. Alternatively, freeing up managerial authority is likely to
succeed only if accountability system are strong and managers have the ability to
manage for results.
Up until now, I have not included in my discussion of performance-oriented
reform a discussion of the role of monitoring and measurement. Performance and
measurement are undoubtedly linked in the minds of many people as is evident by oft-
repeated phrases, such as If you do not measure results, you do not know if you are
succeeding or the stronger That which gets measured gets done. Much writing on
the subject often assumes that monitoring performance leads directly to improving
performance. Work in the area has focused disproportionately on the monitoring side
with little attention given to the pathway by which performance monitoring gets
translated into performance improvement.
There is little empirical reason to be sanguine about the impact that performance
information per se has on organizations and individuals since both frequently have a
variety of ways to deflect or reinterpret performance data in ways that allow them to
retain comfortable patterns. Indeed, it may very well be that performance information
is valuable only after an organization has been transformed into one that is oriented to
continuous improvement in its effectiveness and efficiency. Britains Next Steps
Initiative, widely regarded as a successful performance innovation, is an example of
this pathway the impact of robust performance information was predicated on
redesigned service organizations solely focused on the efficient delivery of services.68
While monitoring and results information have a vital function within performance
reforms, they are not the foundation for change.
67 See Schick (1999) for a discussion of strategies for reform, and the associated case studies for analysis of
specific reform efforts. Efforts to increase managerial autonomy may have a hard time co-existing with efforts
to enhance program oversight, especially when those efforts are introduced simultaneously.
68 See Schick (2003) for a discussion of the fallacy of performance as measurement.
3. Improving performance through the budget process

Much of the work designed to enhance performance in the public sector has
focused on altering budgetary processes in order to more closely link spending and
performance. Attention to the budget process is completely sensible given the role of
the budget as the primary instrument for setting government policy. The link between
budgetary reform and performance-oriented change is so strong that sometimes the
entire performance movement is grouped together under the banner of performance
budgeting. This is a mistake since there are many efforts to enhance public sector
outcomes that do center on the budget.
While it is useful to distinguish between broader performance management and
performance budgeting, it is also useful to define what is meant by performance
budgeting. For the purposes of this paper, we will utilize Professor Marc Robinsons
definition: performance budgeting refers to procedures or mechanisms intended to
strengthen links between the funds provided to public sector entities and their
outcomes and/or outputs through the use of formal performance information.69 This
definition does not limit performance budgeting to government-wide budgeting or
exclusively relate the concept to central allocation processes.70
Like performance management, efforts at performance budgeting have existed
for some time. Attempts in the early 1960s in the United States and later in Sweden
were undertaken with the expectation that it would be possible to create a system
that relatively mechanically linked allocations and performance. It is perhaps not
surprising that these efforts to reform the budget process into a technocratic
exercise divorced from politics did not succeed. The U.S. Planning-Programming-
Budgeting System (PPBS) attempted in the 1960s and the famous performance
budgeting effort of its time has been described by one noted observer as producing
a vast amount of inchoate information characterized by premature quantification
of irrelevant items.71
Although early efforts were not promising, performance budgeting has grown
to be a standard feature of the budgeting process in most developed countries. An
OECD 2001 survey found that 70 percent of member countries included
performance information in their budgets. Perhaps a better measure of the
penetration of performance ideas into the budget process is that 40 percent of
survey respondents answered that their measures distinguished between outputs
and outcomes.72
As we will see in the case studies that accompany this piece, performance
budgeting comes in many different forms. All of the reforms share a movement away
from traditional concentration on inputs and an attempt to counter-balance the forces
of incrementalism in budget setting.73 The following table from Professor Philip Joyce
expresses some of the major differences between performance and traditional
budgeting.74
See Robinson (2004).

69
An alternative term performance-informed budgeting has been suggested recently in the United States.
70
See Philip Joyce (2003).

71 See Wildavsky (1969).
72 Kristensen (2002). For a useful discussion of performance budgeting, see Diamond (2003).
73 Incrementalism refers to the tendency for budgets to be determined by slight modifications in allocations
based upon a determination of what is needed this year to do what was done last year.
74 Joyce (2003), p.15.
Table 9-1. Difference between performance and traditional budgeting
Traditional Budgeting Performance Budgeting
Inputs as ends in themselves Relationship between inputs and outputs
Change in inputs and results for the entire

Changes in inputs as the margin program (for example, how much more
money for how much more results)
Divorced from planning and management in Budgeting integrated with planning and
agencies management
Budgeted resources Costs
The different approaches to performance budgeting can be largely divided into

efforts that focus on tighter ex ante linkages between results and budgets (through
targeting) and those that concentrate drawing tighter ex post linkages between
performance and allocations through the use of budgetary incentives.75
The strongest example of ex ante targeting is provided by the U.K. Public Service
Agreements (PSAs). The PSAs, signed by each agency, set out key objectives and
targets. Headline targets focus on broad outcomes such as improving literacy in
children. These targets are translated into operational targets at the agency of program
level which concentrate on outputs, activities, and processes. The PSA approach was
adopted at the start of beginning of the Labour Government in order to ensure that
planned increases in funding for education and health were tied to concrete
improvement commitments.
Output-purchase budgeting represents an alternative approach whereby
government purchases a specified output based upon detailed unit cost calculations.
This radical version of budgeting for results has been pursued most vigorously in New
Zealand starting in the late 1990s. The creation of a budget model based upon a
purchaser-provider relationship is the extreme form of budget reform that focuses on
outputs. In such a system, each agency is viewed as a business and failures within the
production function can cause governmental organizations to register losses.76
The two approaches discussed above work by linking budgeting with ex ante
expectations of performance. Efforts of this sort have frequently experienced incentive
problems since the consequences of not meeting targets has remained relatively
uncertain. Poor performance can be a justification for increasing budgets (since an
objective may be critical and poor performance may indicate insufficient funding) just
as good performance can justify holding a budget allocation steady or even reducing it
(program goals have been achieved or the program has demonstrated its ability to
produce results with existing funding). For this reason, some commentators have
questioned the extent to which budgeting based upon expectations of future
performance provide adequate incentives for managers to truly improve agency and
program performance.
75 The following discussion draws heavily on Robinson (2004a) and Robinson (2004b).
76 For a discussion and analysis of the reforms, see Schick (1996) and Robinson (2002).
Performance approaches that link ex post evaluations with future allocations seek
to increase the pressure on managers to perform. The Bush administrations Presidents
Management Agenda (PMA) and the complementary Program Assessment Rating Tool
(PART) in the United States are leading examples of a strategy dedicated to
incorporating past results into future allocations. The PMA requires the OMB to assess
the management of agencies, with a focus on the degree to which the agencys senior
management utilizes performance information in their strategic and day-to-day
decision-making. The PART program requires the Office of Management and Budget
(OMB) to determine whether a program is well functioning, with over 50 percent of the
PART score determined by an analysis of results and accountability. Both PMA and
PART are integrated into the budget process by the OMB, as well as through the
unusual step of making PART scores public and using a traffic light scale (green for go,
yellow for caution, and red for halt) for its PMA evaluation in order to ensure
maximum accessibility of agency rating to politicians and their constituencies.77
The current U.S. approach is fundamentally similar to a series of reforms
undertaken by the Government of Canada. An extensive program review was initiated
in 1994 on the heels of efforts to rationalize programs that was begun in the 1980s.
While the initial efforts focused on building up evaluative capacity in the Comptroller
Generals Office, the later, more successful efforts have required explicit reallocation of
resources within fixed budgets and have been combined with other reforms designed
to enhance managerial attention to service delivery and citizen needs. Detailed
program evaluations have also been a feature of the Australian governments approach
to improving performance, especially in the period from 1987 to 1997, but have been
abandoned in the period since 1997.78 The challenge faced by all efforts to conduct
systematic program evaluations is the sheer time and cost required to perform rigorous
evaluations. In many cases, program evaluations appear to have been forced to
compromise on quality, and have led to enormous conflicts within public
administration with relatively minor gains.
As with performance reforms in general, there does not appear to be a compelling
case for promoting one performance budgeting reform over the other. The case studies
that accompany this report will provide greater detail as to the mechanics of each of
these strategies information that will be essential for Korean policy-makers in their
determination of a suitable approach for the country. Instead of looking at their
differences, I want to turn now to a brief consideration of two features that are common
to all performance budgeting efforts the restructuring of budgets around programs
and the selection of performance indicators. It is the authors belief that the former, the
creation of programs, has generally received insufficient attention while the latter has
received too much focus.
If one were to read through recent discussions of performance budgeting, it would
be easy to get the impression that the core of the effort was defining the right set of
indicators and ensuring their rigorous monitoring. It might come as a surprise that
such efforts would likely be useless if the operative budget retained its traditional line-
item structure. In a traditional line-item model, it is hard if not impossible to determine
the policy objectives that are driving spending or to identify the amount of money
77 The Bush Administration has combined these enhanced agency and program reviews with a significant
increase in the use of monetary performance bonuses to reward mangers for high performance. The
effectiveness of such performance rewards in generating effective and efficient public sectors has been
challenged.
78 For a discussion of the role of evaluation in the Australian reforms, see Mackay (2004).
spent on achieving a particular goal. Enhanced expenditure classification systems can

sometimes allow for expenses to be more closely allocated to individual projects but
inherently fail to capture all relevant costs. Restructuring the budget into a program
structure creates the foundation for any serious performance budgeting effort.
We choose the term restructuring the budget rather than reconfiguring it because it
requires ministries and agencies to rethink their activities as well as recalculate their
spending. The initial stage in the process is for a government body to define its
programs a group of activities and projects, usually under a single manager, which
contributes to a specific program objective. This is anything but a trivial exercise, as the
initial PART review revealed. In the 2003 exercise, the OMB determined that fully 50
percent of programs could not be evaluated because they failed to have defined
objectives or measurable results an astounding conclusion given that the
Government Performance Results Act (GPRA) had been enacted in 1993 precisely to get
agencies to define programs and objectives.
Program budgeting is now a common feature within the OECD. A set of generally
accepted principles exist that may help guide the transition from line-items to
programs that provide a constructive basis for performance budgeting. As stated by
Jack Diamond (2003, p.17), these principles include:
z Programs are mono-functional so that each program is linked to only one

function.
z Programs are hierarchically constructed, so that each program has a number of

subprograms and each subprogram can be decomposed into a number of activities
and projects. Each subprogram is related to only one program, likewise each
activity and project is related to only one subprogram.
z Each program has an appropriate size for efficient management.
z Programs and subprograms should be defined in a way to support political

decision-making and prioritization, by making clear the relationship between the
resources used and the envisaged outputs and envisaged policy results
(outcomes).
z Programs must consider all related activities (including regulatory ones) and
projects, which together assist in achieving their objectives. This means that capital
and recurrent spending should be considered together in judging program
performance relative to its objectives.
z Accountability of subprograms should have clear managerial responsibility,

usually and preferably, within a single organizational unit.
z Responsibility for implementing each particular program should almost always

align by administrative unit to one chapter in the budget. Where it is not possible
to maintain this, it is important to assign lead roles to a particular chapter.
Defining a program is an exercise in translating policy into action. It involves

policy-makers working alongside managers in order to ensure an alignment between
agency functions, managerial responsibilities, and oversight arrangements. Moreover,
it also requires that financial management arrangements be configured in order to
ensure the accurate assignment of costs, both direct and indirect, as well as the creation
of system to supply relevant and timely data on finances and outputs. The integration
of financial management with budgeting is an enormous task, frequently requiring new
cost coding as well as new accounting practices, such as accrual accounting.
Establishing a program structure to budgets is, therefore, improperly understood
as a procedural step on the way to introducing performance measures. When it is
properly done, establishing a program structure requires changes within agencies, as
well as the creation of a completely new relationship between budgeting and financial
management. The importance of the changes that can be introduced in the
programming exercise have already been mentioned in the discussion of overall
performance reform the process of organizational and institutional restructuring
precedes the capacity to utilize data to improve performance.
Against this backdrop, the selection of performance indicators can appear to be a
relatively minor component of reform. This is incorrect since it is clear that selecting
inappropriate indicators or establishing unreasonable targets for performance can
destroy the momentum of reform. Indeed, establishing useful indicators and utilizing
data in the budgetary process is critical to achieving the objectives defined in the
programming process.79
Struggles to define useful indicators have been reported almost universally among
the countries in the vanguard of reform. This reflects the complexity of the task as well
as the unfamiliarity of most public sector managers in thinking in monitoring terms.
Undoubtedly, problems will be encountered by nations or agencies embarking on their
own performance budgeting effort, especially in the tricky area of identifying realistic
outcomes and ascribing causality in areas effected by multiple programs. At the macro-
level, monitoring of performance is likely to be a frighteningly inexact undertaking.
Countries are fortunate in that they now have the ability to examine a growing
glossary of indicators defined in the course of previous reforms as they develop
measures that are suitable for them. Experience also suggests that the development of
indicators is a continuous process, as measures evolve over time due to dialogues
between managers and policy-makers and the modification of programs.
4. Results of performance reforms and implications for the design of

performance-enhancing reforms
The author has repeatedly stated in the sections one and two that evidence on the
impact of performance reforms in general, or specific strategies, is hard to come by and
generally inconclusive. This result obtains from a variety of factors already discussed.
For the most part, performance management remains, in Allen Schicks elegant turn of
phrase, an idea whose time has come but whose implementation has not.
The above statement should not be understood to mean that performance reforms
have generated no improvement in performance since this clearly is not the case. The
implementation of performance-orienting reforms in countries like New Zealand, the
U.K., and others is associated with improvements in overall public finance
performance, as well as more efficient service delivery. The case studies that
79 As Diamond (2003) writes, experience suggests that the benefits of a budget program format will soon
be lost unless departments (and most central agencies) continue the momentum and purpose of reform by
moving to develop standards of service delivery and search for ways by which these standards can be
continuously improved and services delivered more effectively.
accompany this piece, along with other studies referenced in this paper, suggest some
of the successes along with the challenges of demonstrating impact.
One of the problems in coming to empirically based conclusions on effective
program design is a result of the restricted nature of existing analysis. As one
commentator has recently noted, Regrettably, the empirical literature on government-
wide performance budgeting is disappointingly limited in scope and methodology,
and does not provide for strong conclusions about the efficacy of these systems. 80
The absence of a robust empirical literature may be due in part to the complexities of
reform but also may be related to somewhat mistaken preconceptions as to where
performance management and performance budgeting would have greatest impact.81
Early efforts at performance management and performance budgeting were
invigorated by an interest in centralized management of economies and social
programs. The early attempts at government-wide performance programming,
borrowed from the U.S. Defense Department, envisioned the ability to distribute
budgetary resources across the government in order to maximize societal welfare. Since
that time, the performance movement has evolved greatly but has retained a
overwhelming interest in improving the allocative efficiency across sectors through
centralized processes involving a limited number of organizations. In the U.S., this has
meant that attention has focused on the creation and approval of federal budgets, with
a specific concentration on those processes involving the OMB.
The initial and continuing focus on centralized allocative processes has perhaps
decreased appreciation of the other processes that determine how public money is
allocated and spent. For example, agencies provisionally allocate money through
their budget submissions, and managers further reallocate money once budgets have
been determined in the operation of their programs. Money also gets reallocated
within the execution process through delays in payment or via payment to third
parties.82
It may well be that performance budgeting and performance management is most
likely to have impact at levels that have been obscured by our preconceptions. There is
significant evidence to suggest that managers in a large variety of programs within the
United States have made use of performance information to direct resources to greater
effectiveness and inefficiency. This has included defining inter-agency allocation
processes that are influenced by performance as well as mid-year reallocations based
on interim program results.83
Ultimately, improvements in public sector performance come about through some
combination of better calibrated allocative processes distributing resources to better
designed programs that are managed in an environment that demands performance. In
this brief review, we have suggested some of the choices to be made in program design,
and some of the factors associated with changing managerial cultures and improving
decision-making. I conclude this brief piece with a quote from Allen Schick, At the
80 Robinson (2004a), p.42.

81 Robinson suggests that the paucity of data is due to a lack of interest on the part of economists and
financial management analysts in the issue, as well as the preponderance on performance budgeting reform
that has taken place in the United States.
82 See Joyce (2003) for an excellent discussion of the variety of allocation processes within the federal
budget process, as well as an examination of how performance information can be utilized across the entire
budgeting process.
83 See Joyce (2003). For a discussion of how outcome information is being used at the federal level in the
U.S., see Hatry, et al (2003).

end of the day, improving performance is essentially a matter of getting organizations

and the people who work in them to behave differentlyTo make a difference, an
organization must be obsessed with purpose, so that the critical actions it takes in
allocating staff and resources, designing and delivering services, assessing results and
feeding back the findings to new decisions are all done through the lens of
performanceGovernments that dont manage for results dont budget for results,
even if they install the outward trappings of performance budgetingRather than
being the locomotive that drives government to change, performance budgeting is the
caboose that confirms the transformations that have been made.84
84 Schick (2003), p. 8, 19, 22.

References
Diamond, Jack, From Program to Performance Budgeting: The Challenge for Emerging Market
Economies, IMF Working Paper 2003, WP/03/169.
Hatry, Harry, et al., How Federal Programs Use Outcome Information: Opportunities for
Federal Managers, Managing for Performance and Results Series, IBM Center for the
Business of Government, May 2003.
Joyce, Philip, Linking Performance and Budgeting: Opportunities in the Federal Budget
Process, Managing for Performance and Results Series, IBM Center for the Business of
Government, 2003.
Kristensen, K., Overview of Results Focused Management and Budgeting in OECD Member
Countries, OECD, Paris, 2002.
Mackay, Keith, Two Generations of Performance Evaluation and Management Systems in
Australia, draft paper, 2004 (on file with the author).
Robinson, Marc, Does Performance Budgeting Work? An Analytical Review of the Empirical
Literature, draft IMF paper, 2004a (on file with the author).
______, Performance Budgeting: Lessons from a Comparative Overview, manuscript, 2004b
(on file with the author).
______, Output-Purchase Funding and Budgeting Systems in the Public Sector, Public
Budgeting and Finance, 22 (4), 2002.
Schick, Allen, The Performing State: Reflection on an Idea Whose Time has Come but Whose
Implementation has Not, GOV/PUMA/SBO(2003) 17, June 2003.
______, Opportunity, Strategy, and Tactics in Reforming Public Management, paper presented
at the OECD Symposium Government of the Future: Getting from Here to There, PUMA,
SGF(99)4, September 1999
______, The Spirit of Reform: Managing the New Zealand State Sector in a Time of Change, a
report prepared for the State Services Commission and the Treasury, Wellington, 1996.
Wildavsky, A., Planning-Programming-Budgeting: Rescuing Policy Analysis From PPBS,
Washington, D.C., 1969.
CHAPTER 10
Introducing Performance Management
by
Youngsun Koh, Korea Development Institute
1. Performance management framework

1.1. Structure of performance management framework
Performance management typically consists of the following four steps:
(1) Setting performance targets

- Delineating the mission and strategic objectives of the organization
- Defining quantitative and/or non-quantitative performance indicators
- Choosing targets for indicators
(2) Designing the program

- Designating who is in charge of the program
- Designing the service delivery system
- Planning human and financial resources
- Drawing up evaluation plans
(3) Implementing the program

- Delivering service with inputs of human and financial resources
- Measuring performance indicators
(4) Assessing the performance

- Performance monitoring
- Program evaluation
As the last step indicates, there are various tools to assess performance. The first is
performance monitoring. In performance monitoring, performance indicators are
defined and then measured on a regular basis. The measured performance is in turn
compared to the targets that were chosen at the start of the period. In Korea, examples
can be found in the Performance Budget Pilot Project that was carried out in 1999-2002
and the Performance Management System that was introduced in 2003.
Performance monitoring can produce information on outputs and outcomes in a
frequent and timely manner at relatively low costs. Based on this information, the
organization in charge of the program can modify its business plans and service
delivery system. On the other hand, monitoring by itself can rarely explain the
causality between inputs and outputs/outcomes. It only describes the measured
performance of a program, and says nothing about the extent to which the program
contributed to the observed performance. In addition, performance monitoring does
not question the relevance and appropriateness of the chosen performance indicators,
and takes as given the objectives and design of the program. Moreover, it is often
difficult to measure with only a limited number of indicators the performance of a
program that has diverse stakeholders and multiple objectives. In some cases, it is even
impossible to define performance indicators. Despite these limitations, however,
performance monitoring plays an important role in the overall performance
management process by providing the basic information on program performance.
A second tool for performance assessment is program evaluation. According to the
OECD (1999), performance evaluation can be defined as a systematic and analytical
assessment addressing important aspects of a program and its value and seeking
reliability and usability of results. Program evaluation addresses the question of why
and how the program produced certain outputs and outcomes. It purports to analyze
scientifically the impact of a program on the expected and unexpected results, describe
objectively the factors that contributed to the success or failure of the program, and
provide information that can be used to redesign the program and for other purposes.
Program evaluation can also question the relevance and appropriateness of the chosen
performance indicators and reassess the objectives and design of the program. On the
other hand, program evaluation usually requires large amounts of money and time,
and cannot be performed on all programs.
Of course, there are other tools for performance assessment than these two. Audits
have traditionally been the most important instrument to ascertain the regularity and
integrity of public sector operations. More recently, performance audits have been
employed by the supreme audit institutions of many countries to analyze the efficiency
of government programs. In addition, many countries have also been utilizing
performance appraisal of individual government employees to attribute the outputs
and outcomes of an organization to those of individuals, and to reflect the information
on their promotion and salary determination.
Table 10-1. Key Differences between evaluation and monitoring
Monitoring Evaluation
z periodic; z Usually episodic;

z assumes appropriateness of program, z Can address a wide-range of potential
activities, and indicators; questions about a policy, program, or project;
z tracks progress against small number z Can identify what has happened as a result of
of indicators; an intervention and provide guidance for
z usually quantitative; future directions;
z use data routinely gathered or readily z Can address how and why questions;
available; z wide-range of quantitative and qualitative
z cannot indicate causality; research methods possible;
z difficult to use for impact assessment; z can use data from different sources;
z usually internal. z can identify unintended as well as planned
impacts and effects;
z can involve internal, external, or self evaluation.
Source: Perrin (2000).
Among these tools, performance monitoring and program evaluation are deemed
the most important, and accordingly they take a central stage in performance
Chapter 10. Introducing Performance Management in Korea 123
management. Some, such as Rist (2003), even equate performance management with
performance monitoring and evaluation (PM&E). Table 10-1 contrasts the main
characteristics of monitoring and evaluation.
1.2. Performance assessment: an example
An example is provided below of the performance monitoring and evaluation.

The Korean Government has made a lot of efforts to reduce pollution in rivers.
Figure 10-1 shows the BOD (Bio-chemical Oxygen Demand) level of four major rivers
in Korea. The graph indicates a weak declining trend in the BOD level since the mid-
1990s, but the trend is not clear.
Figure 10-1. BOD levels of four major rivers
Source: Kang and others (2003).
Kang and others (2003) tried to estimate the impact of pollution abatement
investment in the following way. First, they calculated EBOD, which is defined as the
difference between the actual BOD and the assumed level of BOD when no investment
has been made since 1990:
EBODt = 3.46 (w1 Dt + w2 Wt + w3 Lt) BODt ,
where
3.46 = BOD in 1990 (ppm),

Dt = pollution by households (1990 = 1),
Wt = pollution by firms (1990 = 1),
Lt = pollution by livestock farming (1990 = 1),
w1, w2, w3 = weight of each pollution source, and
BODt = actual level of pollution (ppm).
Because they did not have data on w1, w2, and w3, Kang and others (2003) assumed
various levels for these weights. In the next step, they estimated the following equation
with the standard regression method:
EBODt = + Rt-1 + Pt + ut ,
where
Rt-1 = accumulated investment after depreciation up until year t-1 (100 million
won), and
Pt = annual precipitation (100mm).
The estimated value of ranges between 2 and 2.5, and indicates that the pollution
abatement investment has not been very cost-effective. According to Table 10-2, each
investment of 100 million won (around 83 thousand USD) reduces pollution by 0.00002
ppm. In this case, it costs 500 billion won (around 420 million USD) to reduce pollution
by 0.1 ppm.
Table 10-2. Estimation results
Weights
R2 DW
w1 w2 w3 (x 0.00001) (x 0.01)
0.4 0.1 0.5 -1.02 2.14 (3.47) 6.16 (1.67) 0.64 1.80
0.4 0.2 0.4 -1.05 2.28 (2.75) 7.50 (1.37) 0.53 1.56
0.3 0.2 0.5 -0.84 2.19 (2.67) 6.71 (1.24) 0.51 1.59
0.3 0.4 0.3 -0.90 2.47 (1.93) 8.20 (0.97) 0.36 1.33
0.2 0.3 0.4 -0.87 2.33 (2.22) 7.46 (1.08) 0.43 1.43
0.2 0.4 0.4 -0.07 2.38 (1.86) 7.41 (0.88) 0.34 1.35
0.1 0.4 0.5 -0.49 2.28 (1.78) 6.61 (0.78) 0.31 1.36
0.1 0.5 0.4 -0.52 2.43 (1.60) 7.36 (0.74) 0.27 1.29
Note: In parentheses are t-statistics.
Source: Kang and others (2003).
In this example, the annual measurement of BOD in Figure 10-1 corresponds to

performance monitoring, while the estimation of the impact of pollution abatement
investment in Table 10-2 corresponds to program evaluation.
It should be noted that the above program evaluation has a couple of deficiencies.
First, Kang and others (2003) took account of only point-sources of pollution, and did
not include pollution from non-point sources such as farm land and roads in their
estimation. Presently, there are no data on the latter kind of pollution. Second, they
estimated the impact of pollution abatement investment on the average level of BOD
across the four major rivers, rather than on those of individual rivers. This is again due
to the lack of data.
Such a lack of adequate data can be observed in other areas as well, and poses a
serious threat to the successful introduction of a performance management framework
in Korea.
1.3. Uses of Performance Information
Performance information produced in the performance assessment process can be

utilized in many ways. The OECD (1999) defines three objectives of evaluation, which
can be readily extended to the case of performance management in general:
z improving decision-making;
z assisting in resource allocation; and
z enhancing accountability.
According to Schick (2003), the big problem is not in measurement, but in

application. Much attention has been paid to the former, but not enough to the latter.
With some notable exceptions, governments that invest in measuring performance
rarely use the results in managing programs. They do not base civil service salaries on
performance, nor do they hold managers accountable for performance or allocate
resources on this basis. Efforts to budget on the basis of performance almost always
fail, as do reforms that aim to link pay and performance.
This observation indicates that producing performance information is a separate
issue from utilizing it. Below we will discuss various problems surrounding the uses of
performance information.
1.3.1. Improving decision-making

Most importantly, performance information should be used by those who are
directly responsible for the success of the program. For example, the program manager
can identify more efficient and effective ways to implement the program, change the
design, and utilize the lessons learned in planning future programs. At a higher level of
decision-making, the ministry in charge can modify current policies or propose new
ones based on performance information. In short, performance information provides an
opportunity to learn and improve.
A performance management framework cannot be established on a firm ground
unless it is used as a tool for learning. As Perrin (2002) notes, there appears to be
general agreement that performance management can only be effective if people
working within the system actually believe in the value of a results-orientation. With a
true results-focus, a results-focus permeates everything that one does, for example,
with managers and staff constantly asking themselves what the benefits of their
activities are supposed to be, how they know, and what they can do to be more
effective. Otherwise, what takes place would likely be mainly a paper exercise, with
people going through the motions rather than really acting in a results-oriented
manner.
A great danger emerges when managers and staff do not believe in the value of a
results-orientation, which is distorted performance data. Performance data often rely
on official statistics produced, for example, by national statistical offices. But more
often they are produced by managers and staff themselves. When they suspect that
performance information will work against their own interests, they have an incentive
to distort the data in their favor. At the least, when they perceive performance
management as an unnecessary and onerous exercise, they will take little care in
recording and reporting the data, and the data quality will suffer.
In this regard, improving decision-making is the most important use of

performance information. The information produced for this purpose also forms the
basis for information produced for the purposes of assisting in resource allocation and
enhancing accountability. Managers and staff should check whether the expected
outputs and outcomes have been obtained at each stage of the program
implementation. They should also scan all major aspects of the program. They
therefore need to produce as detailed and diverse information as possible on program
performance. In contrast, higher-level decision-makers or the budget authorities need
far less information on individual programs unless they want to overwhelm
themselves with an excessive amount of data. The information supplied to them is
therefore necessarily a subset of that produced by the program.
1.3.2. Assisting in resource allocation
1.3.2.1. Two ways to link performance to budgeting

Performance information can be used in the budgeting process. In particular, when
the government is faced with severe budget constraint, performance information can
provide valuable guidance in setting up priorities across different expenditure areas
and reallocating resources from less efficient and effective programs toward more
efficient and effective ones.
Resource allocation proceeds in multiple steps (Politt, 2001). In the first step, the
budget is allocated among various sectors such as defense, education, and
infrastructure investment. In the next step, sectoral budgets are further divided among
major programs. For example, within education, the budget is allocated among
elementary, secondary, and higher education. In the third step, program budgets are
allocated to individual projects (free lunch for elementary school students, for example)
and/or institutions (universities, for example). In the final step, the budget is allocated
to specific activities (teachers salaries, research support, library maintenance, etc.).
The budget authorities usually have the final say in the first and the second steps.
Line ministries often take responsibility for the next two steps, as they correspond
roughly to the decision-making process within ministries. Below we will confine our
attention to the first and the second steps, namely the macro-budgeting carried out by
the budget authorities.
There are two ways to link performance to budgeting. One is pursuing a direct
linkage between performance and budgeting, and the other an indirect one. In a direct
linkage, a functional relationship between performance and budget is assumed. For
example, the Government Performance and Results Act (GPRA) of 1993 in the U.S
originally defined performance budget as that which presents varying levels of
performance resulting from different budget amounts.
In an indirect linkage, no such one-to-one relationship is assumed between
performance and budgeting. Performance information serves only as one of many
references. The budget authorities take other social, economical, and political factors
into account when deciding on budget allocation. Subsequently, poorly performing
programs are not necessarily cut back or discontinued, and are sometimes assigned
more resources.
1.3.2.2. Direct linkage between performance and budgeting

Most commentators agree that a direct linkage entails many problems (Politt, 2001,
Perrin, 2002, and GAO, 2002).
First, one should be able to measure performance precisely in order to directly link
it to budgeting. But many government programs do not lend themselves to precise
measurement. Basic research programs, for example, aim to accumulate and diffuse
scientific knowledge, which in itself is not quantifiable. Even when quantification is
possible, it may be difficult to establish the causality between the intervention and final
outcomes. The educational achievement of middle school students, for example, can
depend on the current education as much as on the previous ones they received in
kindergartens and elementary schools.
Second, program costs should also be measured precisely. To this end, overhead
costs from support services should all be allocated to programs. In addition, economic,
rather than historic, costs should be the basis for cost comparison, and this requires the
introduction of accrual accounting in budgeting. For example, annual depreciation of
buildings and facilities should be added to program costs. While many countries
currently employ accrual accounting for their financial statements, only a handful of
them, including the U.K, New Zealand, and Australia, have introduced accrual
accounting in budgeting. This indicates the degree of difficulty involved in the task.
Third, even if it is possible to measure both performance and costs precisely, it will
be difficult to link them directly when their relationship is not stable. And even if their
relationship is stable, past stability does not guarantee the same stability in the future.
For example, a large expansion of a certain program can result in substantial changes in
average and marginal costs. When this is expected, the government cannot rely on past
data to budget for expansion.
The experiment by the U.S government shows that these concerns are real. The
GPRA required the Office of Management and Budget (OMB) to report on the
feasibility and advisability of including a performance budget as part of the Presidents
budget (GAO, 2002). In 2001, OMB reported the results of five performance budgeting
pilots that explored agencies capabilities of more formally assessing the effects of
different funding levels on performance goals. Overall, OMB concluded that the pilots
raised key challenges regarding performance budgeting at the federal level including,
for example, the following:
z In many instances, measuring the effects of marginal, annual budget changes

on performance is not precise or meaningful.
z While continuing to change from a near total reliance on output measures to
outcome measures, it will be much more difficult to associate specific levels
with those outcomes, particularly over short periods of time.
z Establishing clear linkages between funding and outcome will vary by the
nature of the program and the number of external factors.
z Delays in the availability of performance data, sometimes caused by agencies
reliance on non-federal program partners for data collection, will continue to
present synchronization problems during budget formulation.
Based on these findings, OMB is currently aiming at an indirect linkage between

performance and budgeting, as indicated by the following passage from the budget
circular for FY 2005: A performance budget consists of a performance-oriented
framework, in which strategic goals are paired with related long-term performance
goals (outcomes) and annual performance goals (mainly outputs). At a minimum,
resources are aligned at the program level within this framework, and agencies are
encouraged to align resources at the performance goal level (OMB, 2003).
1.3.2.3. Indirect linkage between performance and budgeting

Even pursuing an indirect linkage between performance and budgeting is not
easy. Many of the problems discussed above still apply. When we cannot quantify and
measure the performance or cannot identify external factors that affect the outcomes,
even an indirect linkage is difficult to achieve. In addition, resistance from stakeholders
can make the task even more difficult.
First, in order to reach the complex and sensitive distributional deals which
budget-making entails, and in order to create or maintain sufficiently broad coalitions
of support (or at least, acceptance) for continuing this program or cutting that one,
politicians (both in the executive and the legislature) need to appeal to vague and
general values (Politt, 2001). The last thing they are interested in, during this delicate
and frequently adversarial process, is careful comparative evaluations of rival
programs or the specification of precise operational priorities and targets. Such
exercises would show only too clearly who gains and who loses and what is the
relative cost-effectiveness of different programs, and would thus make coalition-
management all the more difficult. The progress of performance management, by
contrast, requires the participants to discuss and agree on realistic, measurable, dated
goals, targets, and standards, with a highly specific identification of client groups and
their preferences.
Moreover, politicians are under constant pressure to propose solutions to present
political, social, and economic problems, and once their proposals have been accepted
and implemented, they need to prepare themselves for the attacks from opponents.
When somebody comes along with performance data and says that the proposed
program has been ineffective in producing the promised result, only embarrassment
will follow.
For these reasons, politicians usually do not welcome performance management.
Second, line ministries may not want performance management for similar
reasons. For many civil servants, the most serious threat is the condemnation of poor
performance by the public and politicians, which should be avoided at all costs.
Third, the budget authorities may not want to link performance to budgeting even
in an indirect way. In any country, the primary responsibility of budget authorities lies
in collecting budget requests from line ministries, setting priorities, allocating the
budget, and containing the total within a certain limit. In carrying out this function,
budget authorities may fear that performance management would reduce their room
for discretion. For example, they would sometimes feel obliged, against their will, to
increase the funding level for those programs that showed good performance.
There is also what might be termed a cultural divergence between financial and
performance management (Politt, 2001). Financial management by the budget
authorities emphasizes the central control of spending ministries demands for
resources. In contrast, performance management encourages continual and
spontaneous learning and improvement in a spirit of devolution and empowerment.
This tension between control and spontaneity can hinder a whole-hearted acceptance of
performance management by the budget authorities.
1.3.2.4. A prudent and gradual approach

The above discussion suggests that a one-to-one direct link between performance
indicators and budget discussions generally is not possible or appropriate (Perrin,
2002). Nevertheless, integration does not have to be viewed as being all or nothing. It
would not be responsible for a government to forsake completely any hope of linking
resources and performance in some way at least. Information about results clearly
should represent one source of information for use in planning and decision-making,
including decisions about future allocations. One should also remember that perfect
information is not essential. Decisions are always made on the basis of imperfect
information. Making available at least some information on performance may be able
to add some greater degree of confidence to the decision-making process. At the least,
it can be used to raise questions that should be considered or explored through other
means.
At the same time, efforts should be made to minimize the resistance by various
stakeholders to performance management. A better understanding of their incentives
and interests would help in this regard. Further, a realistic expectation on what
performance management can bring about will also help. Success is guaranteed only by
a prudent and gradual approach; a firm belief in the value of performance-orientation,
but with a clear recognition of the gap between hope and reality.
1.3.3. Enhancing accountability

Performance information can also be used for enhancing accountability. Figure 10-
2 illustrates the accountability relationship between citizens, the legislature, the
executive, and others. The legislature and executive are held directly responsible to
citizens, and their accountability is ascertained through the general election (and the
presidential election in a presidential system). The executive is held responsible to the
legislature through parliamentary hearings, budget approvals, and so on. Within the
executive, staff are held responsible to managers, and agencies are held responsible to
parent ministries. When the central government provides funds for or imposes
regulations on local governments, an accountability relationship can be established
between them. Lastly, when ministries contract out part of their operations to outside
contractors, the latter are held responsible to the former for the delivery of goods and
services.
Figure 10-2. Accountability relationship
Citizens
The The Executive

Legislature
Ministries
Local governments
(Managers)
Outside contractors
Agencies
(Staff)
Note: The arrows denote accountability relationships.
In this universe of relationships, performance information can be used to assure

accountability. In a line ministry, for example, managers can review performance of
their staff when considering a promotion or a salary increase for the latter. Citizens can
refer to the performance budgets and other documents when assessing the
effectiveness of government programs. In addition, a formal contract on performance
can be agreed on between those involved in the accountability relationship. For

example, line ministries may stop paying for the service when an outside contractor
does not fulfil the terms of performance contract. But even when such formal contracts
do not exist, each of the above accountability relationship assumes an informal and
implicit contract between the concerned parties.
One should remember, however, that in making efforts to meet the numerical
targets specified in performance contracts, government employees may forget their
broad and non-quantifiable responsibility to work for the public good as civil servants.
Performance contracts can also foster silo-views in the public service and hamper
cooperation between individuals and organizations. When this happens, the overall
performance of the government can be poorer than before the introduction of contracts.
Performance contracts should therefore be devised in such a way as to minimize these
kinds of risk. Setting a cross-departmental target for interdependent programs is an
example.
Trust is another issue to consider in utilizing performance information for
enhancing accountability. Schick (2003) emphasizes that performance contracts within
government are better characterized as relational contracts. They are not real
contracts in the sense that they rarely have the same effect as those between private
parties. For one thing, within the government, politicians and managers rarely have an
arms-length relationship; for another, politicians often have little or no effective
recourse and have to turn the other cheek when managers fail to perform. They cannot
close down the Department of Education and in some cases are pressured to reward
the poor performance by increasing the resources provided. In contrast to formal
contracts, relational contracts derive their force from the incentives of the parties to the
contract to behave in a cooperative, trustworthy manner because they are
interdependent. Ministers need managers to produce public services and managers
need ministers to get them resources and political support. When the relationship
works well, the parties may periodically review progress under the contract and
discuss problems with a strong result-orientation.
In contrast, when the public sector lacks trust between its members, relational
contracts cannot hold. The contracts cannot be expected to enhance performance by
spurring the parties to focus on results. Even worse, they can generate rancor among
the employees and lower the morale if the employees find the contracts as unfair and
unjust. It is therefore important to build trust in the public sector in general and in the
performance management system in particular before utilizing performance
information for accountability purpose.
2. Current performance Management Practices in Korea

2.1. Overview
Currently, several systems of performance management are simultaneously at

work within the Korean government. They are; (i) the Performance Management
System, (ii) the Government Operations Assessment System, (iii) Management by
Objectives, and (iv) performance audits by the Board of Audit and Inspection. Table 10-
3 explains the main features of these systems. Below we will take a closer look at the
first and second systems.
Table 10-3. Systems of performance management

System
Main features
(Agency in charge)
Performance - Objectives: To shift the focus of expenditure management from inputs to

Management outputs/outcomes by setting up performance goals, measuring the
System (Ministry of indicators, and feeding back the information to resource allocation.
Planning and - Based on the pilot project on performance budgeting carried out in 1999-
Budget) 2002.
- Objectives: To enhance the efficiency, credibility, and accountability of

government operations by assessing the performance and
Government
organizational capacity of ministries and citizens satisfaction with
Operations
them, and by utilizing the results in improving the current operations
Assessment System
and planning the future ones.
(Office for
- Composed of the assessments of (i) the central government ministries,
Government Policy
(ii) local governments, (iii) specific programs, etc.
Coordination)
- The Basic Law on Government Operations Assessment was introduced
in 2001 to provide a legal basis for the system.
- Objectives: To foster competition within public service by providing

Management by
performance-based rewards to individual civil servants based on a strict
Objectives
performance assessment of policies and programs.
(Ministry of
- Applies to civil servants of degrees 14. Sets up individual performance
Government and
targets, assesses their achievement, and reflects the results on their
Home Affairs)
annual salaries (for degrees 1- 3) and bonuses (for degree 4).
- Objectives: To analyze and evaluate the efficiency and equitableness of

major government activities and programs from a holistic and
Performance comprehensive perspective and to propose effective alternatives, rather
Audits (Board of than focusing on the disclosure of irregularities.
Audit and - Performance audits carried out in 1999 corresponded to 3.7 percent of all
Inspection) on-site audits.
- Performance audits still mostly address the issue of regularity, and do
not meet the standards for program evaluation.
2.2. Performance Management System
2.2.1. Overall design and implementation schedule
The Performance Management System (PMS), which started in the spring of 2003, is
led by the Ministry of Planning and Budget (MPB). PMS corresponds to performance
monitoring and does not include program evaluation as its component. PMS requires
line ministries to set up performance goals and indicators, prepare annual performance
plans and performance reports, and submit them to the MPB at the start of the annual
budget cycle. The MPB is in charge of making the system in place by examining the
status of PMS in line ministries, coordinating the implementation, and feeding back the
results to resource allocation.
PMS covers only part of ministries activities, as those activities not involving large
sums of expenditure (such as pure policy-making) are excluded from performance
monitoring. Also, activities for which the benefits of performance monitoring are
expected to be small (such as wages and salaries, basic program expenditures,85 and
general administrative expenses) are excluded as well.
For now, PMS applies to 22 leading ministries but will be eventually extended to
all ministries in the central government. The leading ministries were asked to list the
programs (PMS programs) that will be subject to PMS, pick 30 percent of them for
2004, set up strategic and performance goals for these programs, and develop
indicators. In 2004, they begin to measure these indicators, and set up goals for the rest
of their PMS programs. Non-leading ministries will start developing their indicators in
2004.
Table 10-4. Timetable for the Performance Management System
Year Leading ministries Non-leading ministries

z Develop performance goals and
2003
indicators.
z Prepare performance plans for FY z Develop performance goals and
2005 (At least 30 percent of PMS indicators.
programs should be covered).
2004
indicators for remaining PMS
programs.
z Prepare performance plans for FY z Prepare performance plans for FY
2006 (All PMS programs should be 2006 (A least 30 percent of PMS
covered). programs should be covered).
2005
indicators for remaining PMS
programs
2007. 2007 (All PMS programs should be
2006
z Prepare performance reports for FY covered).
2005.
2008. 2008.
2007
z Prepare performance reports for FY z Prepare performance reports for FY
2006. 2006.
z For the first time, all ministries will prepare performance reports for FY 2007 that
2008
cover all PMS programs.
Source: MPB, Guidance on the PMS, 2003.
85 In Korea, programs with small and recurrent costs are designated as basic programs after negotiation
between line ministries and the MPB. Others are designated as major programs.
2.2.2. Coverage of PMS

Programs in the general and special accounts that consist entirely of local shared
taxes, formula-based grants, financing items, internal transactions between accounts,
and contingencies are excluded. Basic programs are excluded except when it is deemed
necessary to subject them to PMS. Those programs for which the benefits of
performance monitoring are expected to be small are also excluded.
Similar standards are applied to the programs in public funds. In 2003, public
funds totaled 59 in number. Among them, 20 are excluded from PMS. Examples of
those excluded are the Foreign Exchange Stabilization Fund, various financial funds
(which reside outside the scrutiny of MPB and parliament), and those soon to expire..
2.2.3. Performance plans and reports

A performance plan sets out in a systematic way the mission, strategic and
performance goals, performance indicators, and the budgetary resources of the
ministry. It should also contain explanations on how the indicators will be measured
and how it can be verified. Optionally, it can describe external factors that affect the
performance.
Figure 10-3. Performance structure
Strategic goal Performance goal Performance indicator Programs

Mission
Source: MPB (2003).
A performance report describes in a systematic way the degree to which goals

have been met. It should list performance goals and indicators as originally set out in
the performance plan; describe the performance in terms of goals and indicators;
explain the reasons for any poor performances; summarize the assessment; and
describe future plans. Optionally, it can include audit results by the Board of Audit and
Inspection and summary findings of program evaluations.
2.2.4. Utilizing the performance information

According to the MPB, the performance reports should tell us not only whether
goals have been met or not but also why. When the poor performance is not due to
uncontrollable factors (such as the unexpected change in economic circumstances or in
laws and regulations), the program will receive an extensive review by the budget
authorities. When it is difficult to identify the reasons for poor performance, a program
evaluation will be planned.
PMS will provide one source of information for budget allocation across sectors
and programs. Line ministries should present performance information when
requesting resources, and the MPB will utilize it when preparing the National Fiscal
Management Plan and drafting the budget.
2.2.5. PMS in a performance management framework

To the extent that PMS is composed of strategic goals, performance goals, and
performance indicators, it is very similar to the performance management framework
employed in the U.S, the U.K, Australia, and others. But a significant difference exists
in its coverage. PMS is confined to a subset of activities carried out by ministries. In the
case of the Ministry of Education, for example, PMS programs take up only 30 percent
of total expenditure in 2004. In contrast, the countries mentioned above monitor the
performance of all major programs.
Table 10-5. Budget of the Ministry of Education

(Unit: Billion won)
2003 2004
Total expenditure 4,112.3 4,535.2
Major programs (A+B) 3,905.7 4,323.4

PMS programs (a+b) 1,319.3 1,339.9
The general and special accounts 2,490.3 2,639.0

Major programs (A) 2,348.0 2,485.9
PMS programs (a) 1,319.3 1,339.9
Public funds 1,622.0 1,896.2

Major programs (B) 1,557.7 1,837.5
PMS programs (b)
Source: MPB.
The pilots on performance budgeting that were carried out in 1999-2002 covered
all expenditures of line ministries. When significant difficulty was encountered in
developing performance indicators, the MPB decided to change the design of the
system and cover only those programs for which clear performance targets can be
established. But the current design can result in the exclusion of important activities
and programs from performance management. In particular, programs whose
expenditure consists mainly of wages and salaries or recurrent running costs will be
excluded regardless of their importance, as well as the grant programs for local
governments. If these programs take a large part in carrying out the mission of the
ministry, performance management will lose its substantial meaning.
Of course, there exist cases where performance indicators are difficult to build. But
indicators are only a part of the whole exercise. A more important aspect of
performance management lies in clarifying the mission of the ministry and classifying
its activities in terms of strategic and performance goals. In this process, managers and
staff are reminded of the ultimate objectives that they want to achieve, and a cultural
change toward a results-orientation emerges within the organization.
2.3. Government Operations Assessment System
2.3.1. Overall design

The Government Operations Assessment System (GOAS) aims to enhance the
efficiency, credibility, and accountability of government operations by assessing the
performance and organizational capacity of ministries and the citizens satisfaction
with them, and by utilizing the results to improve current operations and to plan
future ones (OGPC, 2000). GOAS is composed of five types of assessments; (i)
assessments of the central government ministries (aka organizational assessments), (ii)
assessments of local governments, (iii) self-assessments of the central government
ministries and local governments, (iv) assessments of specific programs, and (v)
assessments of agencies by their parent ministries.
Table 10-6. Components of Government Operations Assessment System
Types of assessments Main features
z Assesses ministries in terms of (i) the implementation and

Assessment of the central performance of their major programs, (ii) their
government ministries (aka organizational capacities, and (iii) citizens satisfaction
organizational assessment) with them.
z Covers 43 ministries, agencies, and commissions.
z The central government ministries assess the operations

delegated to and performed by local governments and
Assessment of local their grant-based programs and implementation of the
governments national agenda.
z Covers 16 provinces and metro-cities. Municipalities can
also be subjected to the assessment.
Self-assessments by the
z The central government ministries and local governments
central government ministries
assess their own operations.
and local governments
z Evaluates those programs that have drawn large public

attention or that have surfaced as national and social
Assessments of specific
issues.
programs
z In some cases, assesses the implementation of major
policies and examines urgent problems.
z Each ministry assesses their agencies in terms of the

Assessments of agencies by implementation and performance of their major
their parent ministries. programs, their organizational capacities, and citizens
satisfaction with them.
Source: OGPC (2003).

GOAS is led by the Office for Government Policy Coordination (OGPC) within the
Office of the Prime Minister. The OGPC performs organizational assessments and
oversees other assessments with the help of the Committee for Policy Assessment
(CPA). The latter is an advisory committee serving the prime minister86 and has private
sector experts as its members.
Below we will explain the organizational assessment, self-assessment, and
assessment of specific programs in greater detail.
2.3.2. Organizational assessment

Organizational assessment examines ministries in terms of (i) the implementation
and performance of their major operations, (ii) their organizational capacities, and (iii)
citizens satisfaction with them. Assessment of the implementation and performance of
major operations is further divided into (i) assessment of major polices and (ii)
indicator-based assessment of operations.
Table 10-7. Organizational assessment
Types of assessments Main features Frequency
Assessment of z Assessment of major policies z Biannual

implementation and
performance of major
operations z Indicator-based assessment of operations z Annual
z Assesses the management capacities of

Assessment of
knowledge-base, organization and z Annual
organizational capacities
personnel, and policies.
z Assesses citizens satisfaction with major

Assessment of citizens
policies, customer services, and z Annual
satisfaction
administration.
The most important one of these is the assessment of major policies. It is carried
out over the core programs of the ministries and/or those programs that embody the
major policy agenda of the government. When selecting the programs to be subjected
to the assessment, the OGPC selects them at the level of bureaus. Once selected, they
become part of the self-assessment programs. That is, ministries should first provide
their own assessment of the programs selected by the OGPC. The OGPC then reviews
these self-assessments and provide their own views with the help of the CPA. The
results are reported to the prime minister in a ministerial meeting twice a year.
The criteria for assessment as suggested by the OGPC are presented in Table 10-8.
86 The prime minister in Korea is appointed by the president just like the ministers and does not have a
political affiliation with a particular party.

2.3.3. Self-assessment
Self-assessment is carried out over the central government ministries and local
governments. Self-assessment of the central government ministries is implemented in
much the same way as the organizational assessment, except that ministries themselves
select the programs to be subjected to the assessment. Each ministry also has a
Committee for Self-assessment in place, which provides assistance in making self-
assessment. The criteria for self-assessment are the same as those for the organizational
assessment (Table 10-8).
Table 10-8. Criteria for assessment
Aspects Criteria Questions
z Does the objective conform to the policy agenda of the

Appropriateness
government and respond to the changes in policy
of the policy
environment?
objectives
z Is the objective clearly presented?
Policy
formul- z Have operational targets and tools to achieve the objective
ation been set out?
Adequacy of the
z Have due steps (such as a survey of the public opinion)
implementation
been taken in the planning process?
plans
z Has cooperation been sought with other ministries that
have related roles in achieving the objective?
Efficiency in z Is the operation being carried out on schedule?

implementation z Are the resources being spent efficiently?
Policy
implemen- z Are the changes in policy environment adequately
Appropriateness
tation accounted for during the implementation process?
of the
z Are adequate efforts being made for public relations?
implementation
z Is a cooperation network in place with other ministries
process
that have related roles?
Achievement of
z Have the objectives been achieved?
Perfor- the objectives
mance Effectiveness of
z Does the policy have real impacts on the public?
the policy
2.3.4. Assessment of specific programs

In the assessment of specific programs, those programs that have drawn large
public attention or that have surfaced as national and social issues are evaluated. The
evaluations are carried out by the OGPC. The CPA first selects the programs to be
evaluated and for each evaluation, designates one of its members as a supervisor. The
supervisors then organize evaluation teams that include outside experts and the OGPC
officials. Ministries in charge of the programs also participate in the evaluations.
Evaluation results are reviewed by the CPA and reported to the prime minister.
Line ministries can also initiate evaluations. In this case, they need to submit their
evaluation plans by mid-March to the OGPC and report the evaluation results to the
prime minister.
2.3.5. GOAS in a performance management framework

GOAS has many features that can be found in a performance management
framework. The OGPC (2003) requires ministries to state their policy objectives and the
expected results of their programs in terms of quantifiable targets similar to those
employed in performance monitoring. In addition, the assessment of specific programs
corresponds to program evaluation.
But the similarity ends there. There are significant differences between GOAS and
a genuine performance management framework. First, GOAS does not require
ministries to set up missions and strategic and performance goals and classify their
programs in a systematic way. The focus of the assessment is placed on a limited
number of programs that are deemed important. No clear explanation is given on the
relation between the selected programs and the overall mission and strategies of the
ministry. Second, despite the OGPCs call for quantifiability, performance targets are
often stated in abstract terms and lack concreteness and output- and outcome-
orientations. That is, one can see what ministries are doing but not the ultimate benefits
they bestow on the public. Third, enough use has not been made of the assessment of
specific programs. Since 1998, only about 5 programs have been evaluated each year.
The quality of evaluations is deemed low in most cases.
3. Developing a performance management framework

3.1. Need for expanding the coverage of PMS
Our goal should be developing PMS into a full-fledged performance management

framework. That is, the coverage of PMS should be expanded to all major activities of
the ministries. In addition, GOAS should be eventually incorporated into the new
framework, making it possible to manage ministerial performance in a comprehensive
and coherent way. In my view, there are several obstacles to these efforts.
First, there is a general misunderstanding within the MPB about the purpose of
performance management. Many officials in the MPB view it only as a tool for resource
allocation and disregard its role in improved decision-making by line ministries. For
these officials, setting targets, measuring performance, and then reflecting the results
on budget formulation is all that counts. As a result, the MPB decided to confine
performance management to those programs whose performance can be easily defined
and measured in numerical terms. Even then, suspicions on the practicability of
performance management and worries about the constraint it can impose on the
budget authorities discretion are holding back the support for PMS within the MPB.
Second, most line ministries in Korea do not have a sound system of planning and
reporting established within them. All ministries report their annual plans to the
president at the start of the year, but they are usually 20-30 pages long and contain
main policy directions only. Some ministries sporadically publish 5-year or longer-term
plans, but these plans mostly lack a close examination of resource availability in the
future. They stop at expressing the hopes and expectations of the ministries.
As discussed previously, performance management can have the greatest impact
when it is accepted and utilized in the decision-making process of line ministries and
when it is part of their planning and reporting process. That is, a system of planning
and reporting is a precondition for good performance management. When it is not well
established in line ministries, and when they do not feel the need for such a system, it
is difficult to instill a performance-orientation in them.
Moreover, when the primary use of performance management as envisaged by the
budget authorities lies in linking performance and budgeting, line ministries will take
it as another source of external interference in ministerial affairs and resist its
introduction.
Third, the diverse system of performance management as explained in the
previous section poses another obstacle to transitioning toward a new framework.
PMS, GOAS, MBO, and performance audits are likely to overlap each other and be a
source of turf-wars between the sponsoring agencies. To avoid this possibility, the MPB
openly announced that it will differentiate PMS from other forms of performance
management by focusing on the programs with large expenditures and measurable
performances.
But the current fragmented efforts by various central agencies are only second-best
to a consolidated drive toward a comprehensive and coherent framework for
performance management. Such fragmentation can increase confusion in line
ministries, overburden them with paperwork, and generate skepticism. It should
therefore be minimized in coming years.
3.2. Major issues in developing a performance management framework
3.2.1. Avoiding a direct linkage between performance and budgeting

As repeatedly stressed above, we should avoid directly linking performance to
budgeting. Performance data obtained in performance monitoring and evaluation can
and should be utilized as one source of information for budget decisions. But it is
neither desirable nor possible to set up a one-to-one relationship between performance
and resource allocation for most programs. The MPB should be aware of this fact.
Otherwise, performance management will be forever stuck in the current approach that
focuses on those programs whose performance can be easily defined and measured in
numerical terms.
In this regard, it is worth noting that the performance budget as currently
prepared in the U.S federal government does not aim for a direct linkage between
performance and budgeting. Similarly, no direct linkage is being pursued in the U.K
and Australia, which represent another group of countries that have systematically
introduced a performance management framework in their government.
3.2.2. Establishing a sound system of planning and reporting in line ministries

As discussed above, a system of planning and reporting is a precondition for good
performance management. A first step to help build the system is requiring line
ministries to prepare annual business plans and performance reports. The business
plan would describe in detail the mission, the strategic and performance goals, the
activities to attain the goals, and the associated resources. The performance report
would review the performance results and discuss the future course of actions.
Table 10-9. Documents required in GPRA

Documents Requirements
z Should cover not less than 5 years, and should be updated at least 3 years.
z Must contain:
- A comprehensive mission statement for major functions and operations
of the agency;
- General and outcome-related goals;
- A description of how the agency will achieve the goals and the
Strategic plan
operational process and resources required;
- A description of how the goals relate to annual performance plan goals;
- An identification of key factors external to, and beyond the control of,
the agency that could significantly affect the achievement of goals; and
- A description of program evaluations, with a schedule for future
program evaluations.
z Should cover each program activity in the agencys budget.
z Must:
- Establish goals that define the level of performance to be achieved by a
program activity;
- Express goals in an objective, quantifiable, and measurable form unless
an alternative form is approved by OMB;
Performance
- Describe the operational processes and resources required to achieve
plan
goals;
- Establish performance indicators to be used in measuring or assessing
the relevant outputs, service levels, and outcomes of each program
activity;
- Provide a basis for comparing actual results with the established goals; and
- Describe the means to be used to verify and validate measured values.
z Should include actual program performance results for the 3 preceding
fiscal years.
z Must;
- Review how successfully performance goals were achieved;
- Evaluate the performance for the current year relative to the
performance goals achieved during the fiscal year(s) covered by the
reports;
Performance
- Where goals are not met, explain and describe (a) why the goals were
reports
not met, (b) plans and schedules for achieving the goals, and (c) if the
goals are impractical or infeasible, why that is the case and what action
is recommended;
- Describe the use and assess the effectiveness in achieving performance
goals of any waiver under 31 U.S.C section 9703; and
- Include the summary findings of program evaluation completed
during the fiscal year.
Source: GAO (1998).
In the U.S, federal agencies should prepare strategic plans, performance plans, and
performance reports under GPRA (Table 10-9). In Korea, line ministries are expected to
prepare performance plans and reports under PMS. More detailed requirements like
those in GPRA would help enriching the plans and reports. The MPB can also compare
them across different ministries and propose best practices. It can also rate their quality
and publish the result on a web-site. Of course, line ministries should be given the
freedom to determine the document formats and encouraged to experiment with
different modes of presentation as long as they satisfy basic requirements.
A second step to build a system of planning and budget is introducing a medium-
term expenditure framework. The reason for the weak planning capacity in many line
ministries can be traced back to the uncertainty surrounding the amount of available
resources in coming years. It is very difficult and often meaningless to plan for the
future when uncertainty is large. I believe this explains why most ministries currently
live from hand to mouth, without a serious effort to prepare a realistic long-term plan
and design their annual operations within this plan. A medium-term expenditure
framework will certainly be of much help in this regard.
Third, when reviewing the budget requests from line ministries, the MPB should
pay attention not only to the outputs and outcomes of individual programs but also to
their design and implementation process. This point is illustrated in the Program
Assessment Rating Tool (PART) as employed in the U.S federal government. The PART
purports to provide a consistent approach to rating programs. It consists of
approximately 30 questions, which are divided into four sections (OMB, 2003). The first
three sections address program purpose, strategic planning, and management,
respectively. The fourth hones in on results and asks whether performance can be
demonstrated. Thus a substantial part of the PART questions deal with the issues of
program design and implementation under the premise that a good performance
cannot follow from a poor design and/or implementation.
3.2.3. Consolidating the diverse systems of performance management

As noted above, PMS and GOAS should eventually be merged into a new
framework for performance management. MBO should also be subsumed under the
new framework. For this purpose, once a system of planning and reporting is
established, all documents required under different systems should be consolidated
into the performance plans and reports, greatly reducing the workload for line
ministries.
But the question still remains who will take charge in leading this effort; i.e., who
will review the plans and reports of individual ministries, discipline the laggards, and
coordinate the whole process. At face value, the OGPC has these responsibilities within
the government. But with only limited manpower and without strong leverage against
line ministries, the OGPC is not in a position to pursue this effort without help from the
MPB. On the other hand, the MPBs role is confined to budgetary matters within the
government, and many believe that pursuing a government-wide administrative
reform is beyond its ambit.
Therefore a close cooperation between the OGPC and the MPB looks essential,
with the OGPC taking the lead and the MPB providing logistics. In particular, officials
in the MPB should remember that a performance management framework was initiated
and led by the ministry of finance in most countries. In these countries ministry of
finance, with the best talent within the bureaucracy, had the capacity to perceive the
necessity for change and to carry out the government-wide reforms.
3.2.4. Emphasis on program evaluation

There appears to be increasing awareness among many countries that performance
monitoring and evaluation can provide different types of complementary information
(Perrin, 2002). An effective performance management system, in particular one that can
provide a focus on outcomes, requires both monitoring and evaluation.
Data obtained through monitoring is rarely sufficient to get at outcomes, and used
alone, is not capable of identifying attribution or causality (e.g. to what extent can one
say that a program activity is directly responsible or not for the given results?).
Monitoring information often can best be used to direct management attention and to
raise questions that should be explored in greater depth through evaluation.
Evaluation has the potential to identify what actually has happened as a result of a
program or initiative, whether planned or not, including unintended outcomes and
effects that often may be more significant than the stated objectives. In particular,
evaluation has the potential to identify why and how outcomes have come about. This
is essential information in order to be able to attribute effects to program activities. It is
also critical information in order to make decisions about future policies and programs,
to assist in program improvement, and to be able to generalize what has happened
from a single setting to elsewhere.
In a country like Korea which has little experience in evaluation, a formal strategy
to introduce the evaluation practice seems desirable. The Australian Ministry of
Finance, for example, imposed the following four requirements on line ministries
(Mackay, 2003):
z Every program should be evaluated every 3-5 years.

z Each portfolio (i.e., comprising a line ministry plus other agencies) should prepare
an annual portfolio evaluation plan, with a 3-5 year forward coverage, and submit
it to the Department of Finance these plans comprise major program evaluations
with substantial resource or policy implications.
z Ministers new policy proposals should include a statement of proposed
arrangements for future evaluations; and
z Completed evaluation reports should normally be published, unless there exist
important policy sensitivity, national security, or commercial-in-confidence
considerations, and the budget documents which ministries table in parliament
each year should also report major devaluation findings.
This strategy had significant influence in spreading the evaluation practice among
line ministries in subsequent years. The MPB should consider adopting a similar
strategy. In addition, attention should be paid to the accumulation of data, which is
essential for good evaluation. As illustrated at the beginning of this paper, almost all
sectors are in dire need of basic data. In particular, an income panel data set, none of
which exists at the present, would provide a valuable basis to assess the various
welfare programs that have been expanding rapidly in recent years.
3.2.5. Restructuring budget accounts

Under PMS, line ministries are asked to reclassify their programs following the
structure of strategic and performance goals. This requires in many cases aggregation,
disaggregation, or consolidation of existing accounts. The restructuring should
continue in the future as PMS expands its coverage over all activities of line ministries.
The primary responsibility for restructuring as well as for developing strategic and
performance goals would lie with line ministries themselves. Of course, they should
consult the MPB and parliamentary committees and get their approval when necessary.
A closely related issue is the allocation of overhead costs across programs. Ideally,
program costs should reflect not only the direct costs of running the program but also
the indirect costs associated with central services such as payroll management and
ministry-wide IT investment. This will facilitate the comparison of costs between
programs. But this task can encounter serious practical difficulty and impose undue
burden on line ministries. It should therefore proceed to the extent that it is feasible
and meaningful.
Figure 10-4 illustrates how the Immigration and Naturalization Service (INS) of the
U.S government realigned its programs and accounts to link them to performance
goals. In the new account structure, the support and administration account is
temporary, capturing the overhead and support costs that could not be easily spread in
2003 (the first years of restructuring). INS spread these costs in the 2004 budget.
Figure 10-4. Restructuring of accounts in INS
Existing Program Structure New Program Structure New Account Structure Performance Objectives
Border Patrol
Inspections Border Enforcement
International Affairs
Enforcement
Objective: Secure the ports
of entry, land border, and
Intelligence coasts of the U.S. against
Immigration Enforcement
Interior Enforcement unlawful entry.
Account
Investigations Objective: Facilitate lawful
travel and commerce across
the borders in the U.S.
Detention &
Detention & Removals
Removals
Objective: Deliver services
to the public in a
Adjudication &
professional and courteous
Naturalization
Immigration Services manner and ensure that
Immigration Services
Account correct immigration benefit
International Affairs decisions are made in a
Benefits timely and consistent
fashion.
Information and
Records Management
Information Resource
Management
Data and Objective: Strengthen
Communications human resource
recruitment and retention
Support and
Construction and efforts and provide for a
Administration
Engineering workforce that is skilled,
diverse, and committed to
excellence.
Training Infrastructure and
Administrative Support
Legal Proceedings
Management and
Administration
Source: OMB (2002).
3.2.6. Introducing accrual accounting in budgeting
At a later stage, we should also consider introducing accrual accounting in

budgeting. The restructured budget accounts mentioned above will enable us to better
assess and compare the performance and costs of individual programs. However, as
long as the programs are costed on a historic, rather than economic, basis, the
information on costs has limited value. The accrual accounting would reveal the true
economic costs of the programs.
Here again a pragmatic approach is needed, focusing our efforts on those areas
that make significant differences in costing. For example, the U.S government has
introduced accrual accounting in federal credit programs direct loans and loan
guarantees in 1990. Currently, efforts are being made to extend it to three additional
areas; government employees pension programs, the removal of hazardous materials,
and capital assets. According to OMB (2002),
z Pensions and retiree health benefits are earned as Federal employees work; they
are paid much later, after the employees retire. The legislation transmitted [to the
congress] would require program and other employer accounts to pay the
employer share of the accruing costs of these benefits to retiree benefit accounts.
These accounts would pay the benefits when they come due.
z Similarly, programs that generate hazardous substances would be required to pay
the accruing cost to clean up contaminated assets at the end their useful life. These
payments would go to funds responsible for the cleanup.
z In contrast, capital assets are bought before they are used. In this case, an agency
Capital Acquisition Fund (CAF) would be created. Following good budget
practice, the CAF would request budget authority (BA) up front to acquire assets
that are included in the budget, and outlays would be recorded when payment is
made. However, this BA would be in the form of borrowing from Treasury
authority. The CAF would then borrow for the period of the assets useful life;
collect annual capital user charges in proportion to asset use, and make the
mortgage payments to Treasury.
Other countries have already introduced accounting in budgeting in full scale;

examples include the Resource Accounting and Budgeting in the U.K and the Outputs
and Outcomes Framework in Australia.
3.2.7. Greater use of performance agreements

When a performance-orientation has been reasonably established in the
government, a greater use of performance agreements can be encouraged. A minister
and staff members would agree on a set of performance targets, review the progress,
and discuss problems. Similar practices can be introduced between managers at all
levels and their staff and between ministries and their agencies.
Under the current system in Korea, performance cannot significantly influence the
annual salaries of individual employees. Salaries are determined in most part by the
years of service and rank of the individual. But performance agreements do not require
performance-based monetary rewards to be effective. As discussed earlier, they
provide for relational contracts and help enhance performance by spurring the
parties to the contract to focus on results.
But it should be remembered that the introduction of performance agreements
should be preceded by establishing a reasonably working performance management
framework. Otherwise, they can generate rancor and lower the morale among the
employees. It is important to build trust in the public sector in general and in the
performance management system in particular before utilizing performance
information for accountability purposes.
References
GAO (General Accounting Office), Managing for Results: The Statutory Framework for
Performance-based Management and Accountability, GAO/GGD/AIMD-98-52, January
1998.
________, Managing for Results: Agency Progress in Linking Performance Plans with Budgets
and Financial Statements, GAO-02-236, January 2002.
Kang, Man-ok, Byung-kook Lee, and Im-ho Lee, Environmental Budgets in Korea, Research
Monograph, Korea Environment Institute, 2003 (in Korean).
Mackay, Keith, The Two Generations of Performance Evaluation and Management System in
Australia, in Performance Evaluation System and Guidelines with Application to Large-Scale
Construction, R&D, and Job Training Investments, World Bank and Korea Development
Institute Conference Proceedings, Korea Development Institute, July 2003, pp.45-60.
MPB (Ministry of Planning and Budget), Guidance on the Performance Management System, 2003 (in
Korean).
OECD, Improving Evaluation Practices: Best Practice Guidelines for Evaluation and
Background Paper, PUMA/PAC(99)1, January 1999.
OGPC (Office for Government Policy Coordination), Government Operations Assessment: E-
Community, 2000. 5 (in Korean).
________, Guidance on the Government Operations Assessment, 2003. 1 (in Korean).
OMB (Office of Management and Budget), Analytical Perspectives: Budget of the United States
GovernmentFiscal Year 2003, U.S. Government Printing Office: Washington, 2002.
________, Performance and Management Assessments: Budget of the United States GovernmentFiscal
Year 2004, U.S. Government Printing Office: Washington, 2003.
Perrin, Burt, Implementing the Vision: Addressing Challenges to Results-focused Management
and Budgeting, OECD, 2002.
Politt, Christopher, Integrating Financial Management and Performance Management, OECD
Journal on Budgeting, Vol. 1, No. 2, 2002, pp.7-37.
Rist, Ray C., Designing and Building a Performance-Based Monitoring and Evaluation System:
Structures and Strategies, in Performance Evaluation System and Guidelines with Application
to Large-scale Construction, R&D, and Job Training Investments, World Bank and Korea
Development Institute Conference Proceedings, Korea Development Institute, July 2003,
pp.31-43.
Schick, Allan, The Performing State: Reflection on an Idea Whose Time Has Come But Whose
Implementation Has Not, GOV/PUMA/SBO(2003)17, OECD, May 2, 2003.
CHAPTER 11
Performance Management in the Health Sector
by
Heesuk Yun, Korea Development Institute
1. Need for performance evaluation in the health field

Most public interventions in the healthcare field intend to increase access to
medical care or provide better treatment for patients, and usually result in increased
expenditures. In Korea, the Ministry of Health and Welfares budget accounted for
7.38 percent of the gross national product in 2003 up from 3.82 percent in 1995. Also,
total expenditure for public health shows a steeper increase over the same period. The
budget size for public health and medical care, and medical security increased from 1.3
trillion won to 5.2 trillion won, an increase of nearly 300 percent (Ministry of Health
and Welfare, 2003).
In the case of the United States, unprecedented growth in social and health
programs for the disadvantaged began during the Lyndon Johnson presidency in 1964,
when Medicare, Medicaid, and other health care programs were launched, and the
Congress mandated and funded performance evaluations. Considering this, it is
natural that there is increasing awareness for cost containment and cost-effectiveness in
the health field in Korea nowadays.
Further, Korea is becoming a more advanced country in terms of public healthcare.
Last year the secretary of Ministry of Health and Welfare declared the War against
Cancer, which suggests that the targets of public health actions have now expanded
beyond infectious diseases to include chronic diseases. This is a common trend
indicating a certain degree of development in public health. As shown in the
experiences of more advanced countries, the area of public health actions will keep
expanding into violence, emerging pathogens, and the social contexts that influence
health disparities. With the continuing rise in healthcare costs expected to coincide
with this trend, attempts to investigate new proposals for various public interventions
will be needed while examining ongoing interventions carefully for more effective
management of public health care.
In practice, in advanced countries, large expenditures generally increase the
importance of accountability of public and private resources invested in health
programs and medical services, and the parliament is concerned about holding
programs accountable for the funds. Also, citizens, employers, the government, and
other purchasers of health services are concerned with clinical and fiscal accountability,
or evidence that healthcare actions deliver effective and quality services in an efficient
manner. Pressure of cost containment is being strengthened in Korea as well. The
Ministry of Health and Welfare, faced with rising costs, is already considering
abandoning the traditional Fee For Service (FFS) payment system and expanding
Diagnosis Related Groups (DRG) to a wider set of treatments.
Increasing emphasis on prevention is another factor stimulating interest in
performance evaluation in the health field. Many people and health professionals
believe that prevention may be a better alternative to a cure. As a society enters into a
more developed stage, more preventive programs and technologies emerge to improve
the nations health. Immunizations to prevent disease, mammography screening to
detect breast cancer, and campaigns promoting smoking cessation are such. However,
since the potential efficacy of preventive care is hard to prove with certainty, the need
for rigorous evaluation will increase in the near future in Korea.
Against this background, this paper attempts to summarize the essential elements
and special characteristics of performance evaluation in the health field in relation to a
general evaluation framework. Also, emphasizing the need to build an infrastructure
for evaluating health interventions, the paper introduces the U.S. experience of
devising indicators for measuring health performance. In doing so, standards for
effective evaluation of health interventions will be briefly reviewed, and implications
for establishing a performance evaluation in the health field in Korea will be extracted.
2. Types of performance evaluation in the health field

Any organized public health action can be subject to a performance evaluation.
Public health actions include direct service interventions, community mobilization
efforts, research initiatives, surveillance systems, policy development activities,
outbreak investigations, laboratory diagnostics, communication campaigns,
infrastructure-building projects, training and educational services, and administrative
systems. For convenience, these can be divided into three categories; the evaluation of
health programs, the evaluation of the health care system, the evaluation of health
services (Grembowski, 2001).
2.1. Evaluation of health programs
This type of evaluation targets programs created to reduce or eliminate a health

problem or achieve a specific objective. Comprehensive inventory of the categories of
health programs that can be implemented to achieve specific health objectives is the
subject of evaluation. Examples include exercise programs created to increase physical
activity and prenatal care programs to prevent medical complications associated with
low birth weight.
2.2. Evaluation of the health care system
This evaluates the performance of the health care system. A health care system has
a structure defined by laws and regulations, and has as its components personnel,
facilities, organization, and financing. The characteristics of the population that the
system serves and the physical, social, and economic environment where they live are
also included in the structure components. The structure of the system influences the
process or delivery of health services, which in turn produces outcomes. The system
performance evaluation typically examines the influence of the structure and structure
components on the outcomes. For example, if whether the patents payment method to
primary care physicians has any influence on patients utilization of health services.
Chapter 11. Performance Management in the Health Sector 149
2.3. Evaluation of health services
This examines the efficiency and effectiveness of specific health services provided
to patients in the system. Evaluation is focused on the measurement of benefits, health
outcomes, or a specific health service associated with particular medical technology
compared to the costs of producing them. Technologies that produce large benefits at
low costs have greater worth than technologies that offer few benefits at high costs.
3. Objectives of public health intervention

Most interventions in the health field have multiple objectives, and the objectives
consist of a mix of different dimensions such as time, place, method and generality.
This multiplicity of objectives is often a source of unproductive disagreement among
evaluators. Confusion about objectives could be eliminated by recognizing that these
objectives can be classified in a number of ways, and the most common way is ordering
objectives by generality (Suchman, 1967).
In this case, objectives may range from the most general such as reducing
mortality to the very specific such as reading a health pamphlet. Furthermore,
objectives and sub-objectives can be hypothesized corresponding to the various steps
or actions that make up a program. While these are a continuous series of events,
dividing them into a hierarchy of sub-goals is essential for evaluation purposes, each of
which may be the result of successfully achieving the preceding goal, and a
precondition to the next higher goal. Figure 11-1 illustrates a cumulative chain of
objectives progressing from the most immediate practical objectives toward the
ultimate ideal goal in a step-wise chart with different levels of evaluation for health
educational literature.
After determining the degree to which an objective has been achieved, this finding
becomes a step toward achieving the next higher objective.
This logic is based on the assumption that mortality will eventually be affected if a
chain of prior accomplishments is fulfilled. In terms of the ultimate objective, this kind
of evaluation assumes that the literature will reach a large proportion of the population
for whom it was intended and be read, which in turn will have some effect in
motivating the reader to go through the recommended procedures, and hence, reduce
mortality as a result of it.
An assumption must be made whenever one moves from a higher-order objective
to a lower one. Thus, every lower-level objective must assume all of the assumptions
made for all of the objectives above it. Any program based on a false set of assumptions
cannot be justified, even if sound evaluations are completed for each of the other
objectives individually. In order to perform a successful evaluation, one needs to prove
the intervening assumptions through research.
Figure 11-1. Chain of objectives in the health field
Reduction in morbidity or mortality
Number or proportion of persons who are

meeting prescribed and accepted standards.
Number or proportion of persons who change

their patterns of behavior in accordance with
the new knowledge. This may be verbalized,
but is more accurate when observed in action.
Number or proportion of persons who change

opinions or attitudes from the new knowledge.
Number or proportion of persons who learn the facts

contained therein.
Number or proportion of persons who glance at or read it.
Number or proportion of persons who see the material.
Number or proportion of persons who receive the material.
Number of requests received for the material, or number distributed.
Number of pieces of literature available for distribution.
Pre-testing of literature by special readability formulas.
Source: Suchman (1967).

4. Special characteristics of performance evaluation in the health field
Performance evaluation in every field follows general steps. First, questions calling
for the evaluations, which are often too general and vague, need to be transformed into
questions which are precise enough to be measured and evaluated. This is followed by
evaluation designing, analysis with credible evidence and presenting conclusions, and
lastly, sharing lessons and ensuring use.
Keeping in mind this commonality, however, particularity of evaluation in the
health field should also be emphasized, which brings special features in each step of
the evaluation process.
4.1. Scientific rigor required
Since the beginning of modern medicine, rigorous scientific inquiry has been the
most important factor that characterizes the adoption of new procedures and practices
in the health field. This in turn characterizes performance evaluation in the area of
health in a way that all causal links, explicit or implicit, in the intervention logic should
be certified with scientific rigor.
Intervention logic is a hypothesis on the sequence of events for bringing about
change by synthesizing the main program elements into a picture of how the program
is supposed to work. As a conceptual link from a programs inputs, or resources
devoted to it, to its output, and to the achievement of results and outcome, intervention
logic summarizes the program's overall mechanism of change by linking processes to
eventual effects (CDC, 1999). Elements in a logic generally include inputs (e.g., trained
staff), activities (e.g., identification of cases), outputs (e.g., persons completing
treatment), and results ranging from immediate (e.g., curing affected persons) to
intermediate (e.g., reduction in tuberculosis rate) to long-term effects (e.g.,
improvement of population health status).
A main purpose of evaluation is to assess effects, and intervention logic, an
explanation of what the public action is supposed to achieve and how it is supposed to
achieve it, is the basis of the assessment. Since there are usually causal assumptions
hidden beneath the surface relating to how the program is supposed to generate its
supposed effects, and the state of the external environment such as other policy
interventions and other external factors, the logic needs to be investigated carefully. In
other words, the main job of the evaluator is to examine the relationship between
activities of interest and observed consequences. This can be accomplished by
identifying and critically assessing these hidden assumptions concerning conditions for
the intervention to be effective. In the health field, intervention logic often involves
scientific theory in laboratory diagnosis of disease or pharmaceutical knowledge,
which makes it more complicated to identify the hidden assumptions and to
investigate their uncertainties. Inevitably, scientific specialty is heavily involved.
A noticeable example is the growing interest in preventive care. While advocates
of preventive care assert that diagnosing and treating illness at an early stage can save
a significant amount of money and successfully restore the patient to good health, the
efficacy of preventive care is difficult to prove. It requires conceptualization of those
specific conditions and socioeconomic characteristics of patients for whom preventive
care might have some impact. For example, little is known about the circumstances
under which various screening programs, such as Pap smears to detect cervical cancer,
are effective.
Figure 11-2 shows an example of a community-based program to reduce lead

poisoning among children living in low-income, dilapidated housing areas in New
York City in 1970. The lead poisoning problem mostly affected young children who ate
paint chips based with lead that peeled from house walls. In order to reduce childrens
exposure to lead-based paint, the city started a program to screen 120,000 children for
lead poisoning. If the lead level in a childs blood was high, the childs apartment was
inspected, and if evidence of lead exposure was detected, wallboards were erected to
prevent further exposure. The intervention logic illustrated in Figure 11-2 assumes the
housing repairs would have direct influence in reducing lead poisoning among
children.
The result was that the incidence of lead poisoning decreased by 80% in 3 years.
However, it was not the installation of wallboards that caused the decrease of
incidences. Instead, due to the wide-scale screening and publicity about the program,
parents monitored their childrens behavior more diligently, preventing them from
eating paint chips. This example shows that a causal assumption in the intervention
logic, which is that childrens exposure to lead paint would be reduced by house
repairs, had missed important links. In this case, the program worked as intended but
not because of the reasons specified in its intervention logic. It might have had a better
result if the resources allocated for housing repairs had been invested differently such
as in expanding the number of houses being screened and campaign on children being
exposed to lead poisoning.
Figure 11-2. Lead poisoning program
Children with Medical

Screening
Lead Poisoning Treatment
Reduction in
Lead Poisoning
of Children
Identification of
Housing Repair
Hazardous
Housing
Source: Grembowski (2001).

Figure 11-3 shows an example of a smoking cessation program to reduce low birth
weight among low-income, pregnant women. Sheehan (1998) examined the association
between stress, addictive behaviors, and low birth weight among 5,295 low-income,
inner-city women in Hartford, Connecticut. The arrows in Figure 11-3 indicate direct
and indirect relationships between various factors and low birth weight. Addictive
behaviors such as smoking and alcohol consumption have direct effect on low birth
weight in the causal relationships. If a pregnant woman smokes or drinks alcohol more,
the childs birth weight would be lowered. The standardized regression coefficients
indicate the size of the change in low birth weight produced by a standardized change
in various factors. Economic stress has direct effects on family stress and social
support, and family stress and social support, in turn, have direct effects on addictive
behavior. Hence, the three stress variables have indirect effects on low birth weight
through their direct effects on smoking and drinking behavior.
Figure 11-3. Smoking and low birth weight
Older Than .42

17 Years
.78
-.13 No Social
Support
.55
.41
.75 .90
.35 Economic
Stress
Addictive .28 Low Birth
Behavior Weight
.89
.30
Family .32
.35
Stress
Medical
.19
Risks
Source: Sheehan, Stress on Low Birth Weight, reproduced from Grembowski (2001).
This causal relationship has some implications about the programs expected
outcome. First, the program alone is likely to reduce but will not eliminate the low
birth weight problem. In the cause and effect structure, 90 percent of the variation in
low birth weight is not explained by the variables in the structure. Also, the figure
indicates that both medical risks and addictive behavior have direct effects on low birth
weight. Thus, because the unexplained variance is large and other risk factors exist
beyond the programs control, the program will at most reduce the low birth weight
problem.
The size of reduction can also be projected. It is suggested that the programs
indirect effect on low birth weight will be smaller than 0.28. If the program reduces
smoking by 0.25, then the programs indirect effect on low birth weight would be 0.25
0.28 = 0.07.
Second, the causal relationship in the figure also suggests that the programs
performance would be increased if the program was modified. As shown, because
stress has a direct effect on addiction and indirect effects on low birth weight, the
program would perform better if it also includes a part for stress reduction.
This is an example showing that scientific and statistical reasoning offers an
understanding of how the program is supposed to work, and how efficient the
program will be in achieving its objectives. Evaluation research in the health field is
normally based on theory from laboratory experiments, economic as reasoning and
statistical projection, and it should be constantly adjusted with scientific rigor
whenever new information about the evaluation logic is added to the information set
or the contextual situation is changed.
4.2. Contextual complexity
In most cases, health programs lie in complicated contexts including historical and
geographical settings, political, social, and economic conditions, and influences of
related or competing organizations. The most important are the socio-economic
conditions. The evaluator should consider the needs of the target population and the
particular problems which the program is designed to address. It is difficult to identify
what effects are genuinely caused by a program and to separate these effects from
other influences on the socio-economic problems. Hence, the evaluator needs to
understand these intertwined factors in order to design a context-sensitive evaluation,
to interpret findings accurately, and to assess the generalizability of the findings.
It should also be noted that because health programs are often created to reduce
health disparities across socio-economic groups, evaluators also must consider political
aspects among various income groups and politicians, and interest groups surrounding
a health care program (Shortell, 1978). Since politics is about how we attach values to
policy issues, politics and values are inseparable from the evaluation of health
programs in nature. Consequently, evaluations are always conducted in a political
context where a variety of interest groups compete in their own interests.
5. Methods of performance evaluation for health care intervention
5.1. Estimating counterfactual situation
When we say that certain effects were produced or caused by an intervention, this
means that if the intervention had not taken place or had taken place in a different
degree, those effects would not have occurred, or would not have occurred in the same
degree. Simply put, evaluations are carried out to determine what would have
happened without the intervention. This is called the counterfactual situation
(European Commission, 1997), and ideally, deriving the counterfactual situation with
certainty is desired.
Evaluation would be easy when comparing two groups that are identical in every
respect except that one group is exposed to the intervention while the other group is
not. In the real world, however, this is not the case since we can never find two groups
that are identical in every respect except for being exposed to a program. The non-
equivalence of the two groups means that the counterfactual situation has to be
estimated rather than derived, and this emphasizes the importance of evaluation
designs scientific accuracy and data credibility.
The methods for evaluation design are drawn from social, behavioral, statistical
and health sciences. Design types include experimental, quasi-experimental, and
observational designs, and according to the specific evaluation case, the best fitting
design should be selected to provide the appropriate information to address the case
specific questions (Fink, 1993). Experimental designs use random assignment to
compare the effect of an intervention with otherwise equivalent groups. Quasi-
experimental methods compare nonequivalent groups (e.g., health program
participants with those on a waiting list) or use multiple waves of data to set up a
comparison. Observational methods use comparisons within a group to explain unique
features of its members (CDC, 1999). The choice of design has implications for what
will count as evidence, how that evidence will be gathered, and what kind of claims
can be made. In particular, health problems are usually entangled with other living
conditions or other health problems where the origin of the cause is hard to trace.
Hence, special attention must be given when selecting a design. At the same time, data
availability has enormous importance in performing evaluation in all types of design.
5.2. Customizing tools for evaluation
Though it is true that more rigorous methodological tools and designs focusing on
the impact or outcome of evaluations are essential parts to ensuring the quality of
healthcare evaluation research, what is more pressing is the need to understand
specific program components and the process by which they are implemented since
every health program has its special features due to scientific hypotheses or the
contextual particularity in which it lies. The interdependency of the health program
with other political and environmental factors stated above should also be considered
carefully.
For this reason, the observed golden rule is that there is no single evaluation
methodology which is universally applicable for the entire health field. Instead, the
choice of method should be determined by the particular evaluation problems at hand.
Actually, arbitrary choice of method at the beginning without careful examination of
the case often results in a poor evaluation. The evaluation methodology should be
carefully chosen after examining the targeted health evaluation case, and more
importantly, should be customized and modified to reflect the particularity of the case.
Also, if a program's activities are aligned with those of other programs operating
in the same setting, or with change of social, economic conditions, certain effects cannot
be attributed solely to one program or another. In this case, the goal of the evaluation is
to gather credible evidence that describes each program's contribution in the combined
change effort.
6. Prerequisites for starting-up performance evaluation in the health field:

evaluation infrastructure
6.1. Construction of representative frameworks for evaluating public healthcare

intervention
Although it was previously emphasized that each evaluation uses a different and
customized methodology for the targeted program, it should also be highlighted that
there exists a strong need for a recommended general framework for health field
evaluation that is both a synthesis of existing evaluation practices and a standard for
further improvement.
The existence of a standardized framework supports a practical approach to
evaluation that is based on steps and standards applicable in public health settings.
Because the framework is purposefully general, it provides a guide for designing and
conducting specific evaluation projects across many different areas of public health
intervention. In addition, the framework can be used as a template to create or enhance
program-specific evaluation guidelines that further operationalize the steps and
standards in ways that are appropriate for each program (CDC, 1999).
By integrating the principles of the general framework into all program operations,
it is possible to stimulate innovation toward the improvement of outcomes and better
detection of program effects. Consequently, detecting these effects in a more efficient
and timely manner will enhance the evaluators ability to translate findings into
practice. Furthermore, guided by the steps and standards in the framework, a basic
approach to program planning will also evolve, and integrated information systems
will support a more systematic measurement. Also, lessons learned from evaluations
can be used more effectively to guide changes in public health strategies.
It should be noted that the standard framework should be considered as a
practical, non-prescriptive tool for individual evaluation work; in that it only acts as a
guide for public health professionals in their use of program evaluation. However,
designed to summarize and organize essential elements of program evaluation, the
framework comprises steps in program evaluation practices and standards for effective
program evaluation. Adhering to the steps and standards of this framework will allow
a better understanding of each program's context and will improve the way program
evaluations are conceived and conducted.
It is also worth noting that the general framework can encourage evaluation
research to be integrated with routine program operations. Because performance
evaluation is in nature a case-specific analysis, it is not easy to practice it consistently
across program areas, or sufficiently integrate it into the day-to-day management of
most public health interventions. Hence, there needs to be a standard framework that
offers operating principles for guiding public health activities.
These principles of evaluation work as guidelines to improve how public health
activities are planned and managed. It would be possible to ensure that new and
existing programs follow these principles by requiring each program to conduct
routine, practical evaluations that provide information for management and improve
program effectiveness. In that way, public health professionals will recognize that the
basic steps of the evaluations framework are parts of their routine work. In day-to-day
public health practice, program goals are defined; core questions are stated; data are
collected, analyzed, and interpreted; judgments are formed; and lessons are shared.
6.2. Constructing indicators for major healthcare evaluation types
Evaluation is about revealing the value of a program. This involves making value
judgements on the degree to which a programs performance has been good or bad, so
predetermined and transparent indicators are needed to ensure that value judgements
do not become arbitrary. For this purpose, a prerequisite for evaluation research is to
set up a system of criteria by which to measure the observed effects of a program and
the benchmarks on the proper function or success of a program. However, setting the
system of indicators may be difficult for a number of reasons. For one, objectives can
sometimes be expressed in very vague terms, and a single program may have multiple
objectives, either in terms of results of outcomes, some of which may carry relatively
more weight, or even be incompatible with others. Also, objectives may also evolve
over time, as the programs environment evolves. Furthermore, indicators should
ideally allow evaluators to compare the performance with other policy instruments in
the same field of action because even if a program falls short of achieving its objectives,
its performance may not necessarily be unsatisfactory.
In the health field, where intervention logic is grounded on scientific knowledge
and all the political, social, and economic interests, indicators should function as the
guiding rules that pertain to the evaluation's focus and questions. The role of indicators
should translate general concepts regarding the health program, its context, and its
expected effects into specific measures that can be interpreted, while providing a basis
for collecting evidence that is valid and reliable for the evaluation's intended uses. For
this reason, setting up a system of indicators in the area of health is more complicated
than in other fields, because indicators need to present even the puzzling aspects of the
laboratory hypothesis in a meaningful form for monitoring.
Indicators include measures of the program's capacity to deliver services, the
participation rate, levels of client satisfaction, the efficiency of resource use, and the
amount of intervention exposure. Also included are outcome measures such as changes
in participant behavior, community norms, policies or practices, health status, quality
of life, and the settings or environment around the program. Furthermore, multiple
indicators are needed for tracking the implementation and effects of a program.
Developing multiple indicators should be based on intervention logic, and in this
sense, intervention logic can be used as a template to define indicators leading from
program activities to expected effects. For each step of the model, qualitative and
quantitative indicators could be developed to suit the concept in question, linking
assumption, the information available, and the planned usage of data. Relating
indicators to intervention logic also allows detection of small changes in performance
faster than when relying on a single outcome as the only performance measure. In
addition, lines of responsibility and accountability are clarified in this approach as the
measures are aligned with each step of the program strategy. Consequently, combining
intervention logic and indicators results in a set of measures that detect the
consequences of intermediate effects on health outcomes of the program.
Even during evaluation, indicators might be modified or new indicators can be
adopted. According to changes in the environment or context, indicators need to
adjust. It should also be noted that measuring performance by tracking indicators is
only part of an evaluation and must not be the only basis for judging whether an
intervention is a success or failure. Serious problems can result from using performance
indicators as a substitute for the entire evaluation process and reaching final
conclusions. For example, an indicator such as a rising rate of disease might be
assumed to reflect a failing program when, in reality, the indicator is influenced by
changing conditions that are beyond the program's control.
6.3. Preparing data sources
When an indicator is proposed for use in performance monitoring, appropriate

data must be available to support its use. Ideally, data would be collected from the
specific population of interest, within the relevant time frame, using valid, reliable, and
responsive measures. However, collecting data is expensive work, and realistically,
efforts to expand data collection can reduce the resources available for programs. As a
result, performance evaluations often have to rely on data collected for other purposes,
so evaluators must understand limitations on the applicability of the data. Whether or
not specific populations of interest are included in samples from which data are drawn
and whether data are collected sufficiently often, or are made available soon enough,
need to be considered in the monitoring process.
Another problem arises when only data sources that provide information on a
particular group are available, and comparisons need to be made among groups. In this
case, understanding the limits of available data has critical importance in drawing
appropriate inferences (Drummond et al., 1997).
7. Examining indicators for healthcare evaluation: experience of U.S.

At the request of the U.S. Department of Health and Human Services (DHHS), the
Panel on Performance Measures and Data for Public Health Performance Partnership
Grants was established. The panel was given the task of examining the state of
performance measurement for public health and to recommend performance measures
in ten areas including: chronic disease; sexually transmitted disease (STD), human
immunodeficiency virus (HIV) infection, and tuberculosis; mental health;
immunization; substance abuse; and three areas of prevention of special interest to
DHHSsexual assault, disabilities, and emergency medical services. The panel focused
on measures the government would be able to use over the next 3-5 years to negotiate
agreements and monitor performance in these areas.
More than 3200 measures were proposed to the panel through various outreach
efforts. The panel used four guidelines for assessing them, and measures that scored
the highest were recommended for use in performance monitoring. The objective of
recommending performance measures is to provide technically sound methods for
assessing progress in meeting public health objectives and to provide the government
practical and useful tools to advance their public health objectives (National Research
Council, 1997).
7.1. The panels framework for assessing suggested indicators
Although health outcome measures are widely used and have intuitive appeal,
they are insufficient by themselves for monitoring the efforts of a given program to
reduce complex public health problems. Many measures of health outcomes are
affected by various factors that are not under the health interventions control,
subsequently, compromising the validity of measures of program effect. Hence,

changes in outcomes cannot be attributed only to specific program effectiveness.
Another constraint is that many important public health objectives, such as lowering
the incidence of cancer and HIV infection, cannot be achieved over short periods of
time to derive an outcome measure.
For example, if the policy interventions goal is to reduce the mortality rate from
breast cancer, it can reduce the risk of such adverse health outcomes by increasing the
number of mammograms it provides to women aged 50 and over. However, the
effectiveness of the intervention cannot be shown in a short period of time.
Furthermore, there are a series of process activities (health education programs,
surgical and nonsurgical treatment, and postoperative follow-up care) and capacity
indicators (number of trained staff and facilities offering mammography screening)
that are believed to be related to the level of mortality from breast cancer.
In this case, the evaluator needs to consider whether desired health outcomes are
achieved, whether specific agency commitments are carried out, and whether the
agency has the capacity to conduct all the necessary processes. Outcomes are
fundamental, and any process or capacity measure used to assess performance should
be widely accepted as being closely related to them.
For this reason, the panel recommends using intermediate outcome measures,
such as risk status, for which there is general consensus that the result being measured
is related to the health status outcome, and that performance monitoring must make
use of process and capacity measures to complement available measures of outcomes.
Each process and capacity measure should be accompanied by reference to published
clinical guidelines or other professional standards that describe the relationship
between the process measure or capacity measure and the desired health outcome.
Health outcome: change in the health of a defined population related to an

intervention.
Risk status (intermediate outcome): change in the risk demonstrated or assumed
to be associated with health status.
Process: what is done for defined individuals or groups as part of the delivery of
services, such as performing a test or procedure or offering an educational service.
Capacity: the ability to provide specific services, such as clinical screening and
disease surveillance, made possible by the maintenance of the basic infrastructure of
the public health system, as well as specific program resources.
To sum up, the effectiveness of an intervention in using resources can most

appropriately be evaluated by assessing the degree to which desired changes in health
outcomes are achieved, together with a judgement of the degree to which those
changes can be attributed to the intervention. When a firm causal link between the
resources and processes used and the health outcome sought has not been established,
as is often the case, or when the intervention resources are a small part of all the
resources that contribute to the outcome, the panel believes that performance
assessment must necessarily depend on a combination of health outcome, process and
program capacity measures.
And performance measures should be understood and adopted as the product of
an evolutionary process, to be revised as additional empirical evidence is obtained and
better methods of data collection are implemented.
7.2. Guidelines for the assessment of proposed indicators
(1) Indicators should be aimed at a specific objective and be result oriented. They must
clearly specify a desired public health result, including identifying the population
affected and the time frame involved. Process and capacity measures should clearly
specify the health outcome, or long-term objective, to which they are thought to be
related.
(2) Indicators should be meaningful and understandable. They must be seen as important
to both the general public and policy makers at all levels of government and should
be stated in non-technical terms.
(3) Data should be adequate to support the indicator. Adequate data on the populations
of interest must be available for the use of indicators and have the following
characteristics:
Data to track any objective must meet reasonable statistical standards for accuracy
and completeness;
Data to track any objective must be available in a timely fashion, at appropriate
periodicity, and at reasonable cost; and
Data applied to a specific indicator must be collected using similar methods and with
a common definition throughout the population of interest. Comparisons of an
indicator across regions are valid only if the definition and collection methodology are
consistent across regions.
(4) Indicators should be valid, reliable, and responsive. Indicators should as much as
possible, capture the essence of what they purport to measure, be reproducible
(reliable), and be able to detect movement toward a desired objective (responsive).
7.3. Example: Chronic disease
Prevention of chronic disease is the primary goal of many health programs, but
chronic disease incidence and mortality data are not useful because the expected time
period between most prevention activities and the effect of those activities on disease
incidence or mortality tends to exceed several years, which also exceeds the time that
health departments are generally willing to wait to assess the interventions
effectiveness. Therefore, potential chronic disease measures are focused on risk
reduction and screening, and process measures.
Potential Risk Status Measures

Risk status measures represent intermediate health outcomes.
Tobacco
Individual adult Percentage of (a) persons aged 18-24 and (b) persons aged 25 and older
currently smoking tobacco
Individual youth Percentage of persons aged 14-17 (grades 9-12) currently smoking tobacco
Individual Percentage of women who gave birth in the past year and reported
pregnant woman smoking tobacco during pregnancy
Individual Percentage of employed adults whose workplace has an official policy
working adult that bans smoking
Nutrition
Content Percentage of persons aged 18 and older who eat five or more servings of
fruits and vegetables per day
Content Percentage of persons aged 14-17 (grades 9-12) who eat five or more
servings of fruits and vegetables per day
Total calories Percentage of persons aged 18 and older who are 20 percent or more
above optimal body mass index
Exercise
Individual adult Percentage of persons aged 18 and older who do not engage in physical
activity or exercise
Individual youth Percentage of persons aged 14-17 (grades 9-12) who do not engage in
physical activity or exercise
Screenings and
Tests
Hypertension Percentage of persons aged 18 and older who had their blood pressure
checked within past 2 years
Cholesterol Percentage of women aged 45 and older and men aged 35 and older who
had their cholesterol checked within past 5 years
Breast Cancer Percentage of women aged 50 and older who received a mammogram
within past 2 years
Colon Cancer Percentage of adults aged 50 and older who had a fecal occult blood test
within past 12 months or flexible sigmoidoscopy within past 5 years
Cervical Cancer Percentage of women aged 18 and older who received a Pap smear within
past 3 years
Diabetes
HbA1C Percentage of persons with diabetes who had HbA1C checked within past
12 months
Foot exam Percentage of persons with diabetes who had a health professional
examine their feet at least once within past 12 months
Eye exam Percentage of persons with diabetes who received a dilated eye exam
within past 12 months
Examples of Process Measures

Nutrition Program Strategy Enable children to learn healthy dietary habits
Process Measure Percentage of schools with menus that meet dietary guidelines
for fat content and five or more servings of fruits and vegetable
daily
Physical Activity Program Increase opportunities for sedentary working adult to exercise
Strategy
Process Measure Percentage of worksites with worksite wellness programs that
include physical exercise
Smoking Program Strategy Limit illegal youth purchase of smoking tobacco; Increase the
price of tobacco products; Restrict smoking tobacco advertising;
Restrict indoor tobacco smoking; Educate children about
hazards of smoking tobacco; Increase access or availability of
smoking cessation programs; Market effective antismoking
message to the general public
Process Measure Percentage of vendors who illegally sell smoking tobacco to
minors; Amount of excise tax per pack of cigarettes; Percentage
of communities with ordinance or regulations restricting
smoking tobacco advertising; Percentage of worksites that are
smoke free; Proportion of schools with age-appropriate
smoking prevention activities and comprehensive curricula;
Proportion of current tobacco smokers visiting a health care
provider during the past 12 months who received advice to
quit; Percentage of adults who can recall seeing an antismoking
message during the 12 months following a media campaign; etc
Screening Program Strategy Educating patients regarding need for and appropriate timing
of screening tests
Process Measure Percentage of persons with diabetes receiving diabetes health
education
Screening Program Strategy Improving access to screening services
Process Measure Percentage of managed care organizations in which patients
can schedule mammograms at convenient times for them
Screening Program Strategy Implementing tracking and recall systems
Process Measure Proportion of providers with chart-based or other real-time
system for identifying women in need of mammography
Examples of Capacity Measures

Resources Number of full-time health department employees for chronic
disease prevention
Number of public service messages prepared by state agency
shown annually for chronic disease prevention
Proficiencies Number of key surveillance systems and data sets (i.e., death
certificates, cancer registry data, birth certificates, Behavioral
Risk Factor Surveillance System (BRFSS), Youth Risk Behavior
Surveillance System (YRBSS), hospital discharge data, Medicare
and Medicare encounter information and other relevant local
data sets) that are establish and maintained
Percentage of local health departments receiving technical
assistance and training
Percentage of labs that meet quality standards
Planning Percentage of population served by systematic community
planning process, with leadership provided by the official
health agency and participation of all relevant groups (e.g.,
consumers, providers, advocators)
Percentage of population covered by written comprehensive
chronic disease prevention plan(s) containing priorities and
objectives based on needs, resources, and local demands
Community Involvement Percentage of health care providers working under agreements
established with public health departments to provide
population-based prevention programming to reduce major
risk factors for premature morbidity and mortality
Proportion of health department programs that operate within
the framework of a community coalition or have a community
advisory group
7.4. Example: STDS, HIV, and tuberculosis
HIV infection, tuberculosis, and many of the STDs resemble noninfectious chronic
disease. The interval between the acquisition of infection and the development of
serious consequences may be years long. Monitoring the long-term consequences of
these infections does not provide a useful short-term indication of the performance of
prevention efforts, so measuring in this public health area is more complex than in
others. Three general strategies for preventing these diseases are: reducing the risk of
exposure, reducing the probability that an exposed person becomes infected, and
reducing the duration of the infectious state among persons who became infected.
Potential Health Status Outcome Measures

Incidence rates of selected Rate of reported gonococcal urethritis in men
STDs Rate of reported chlamydial urethritis in men
Rate of reported cases of primary and secondary syphilis
Rate of reported cases of congenital syphilis
Incidence rates of HIV Rate of reported newly diagnosed cases of HIV infection
infection Rate of perinatally acquired HIV infection of infants
Prevalence rates of selected Prevalence rate of gonococcal infection in women in defined
STDs populations
Prevalence rate of chlamydial infection in defined populations
Prevalence rate of syphilis in defined risk groups, e.g., pregnant
women
Prevalence rate of rectal gonococcal infection in men
Prevalence rate of HIV Seroprevalence of HIV infection in defined populations at high
infection risk of the infection, e.g., pregnant women who abuse drugs
Potential Consumer Satisfaction Outcome Measure
Rates of consumer satisfaction with STD, HIV, and tuberculosis treatment programs
Potential Risk Status Measures

Rates of sexual activity among adolescents aged 14-17
Rates of sexual activity with multiple sex partners among people aged 18 and older
Rates of condom use during last episode of sexual intercourse among sexually active adolescents
aged 14-17
Rates of condom use by persons aged 18 and older with multiple sex partners during last episode
of sexual intercourse
Rates of condom use during last episode of sexual intercourse among men having sex with men
Rates of injection drug use among adolescents and adults
Completion rates of treatment for STDs, HIV infection, and tuberculosis
Example of process Measures
Program Strategy Reduce barriers to receiving treatment from specific providers
Process Measure Percentage of patients with insurance coverage for specific
treatments
Process Measure Percentage of patients reporting no transportation barriers to
obtain necessary services
Process Measure Percentage of physicians and other care providers receiving cultural
competency training
Program Strategy Improve quality of services provided
Process Measure Percentage of cases followed up after most recent contact
Process Measure Percentage of known intravenous drug users with access to needle
exchange program
Examples of Capacity Measures

Resources Percentage of high-risk communities with nearby testing and
screening services from multiple types of health providers and
public health organizations
Planning Percentage of the states population who reside in communities
that are engaged in formal community processes for assessing
and planning for HIV/STD/TB prevention and treatment
services
7.5. Example: Mental health
Although there is little agreement on linkages between specific process and mental
health outcomes, there is some agreement on the dimension that are important in
evaluating mental health services; quality assurance, access to services and utilization
of services, consumer satisfaction with services, and psychological and social outcomes.
Potential Health Status Outcome Measure

Percentage of persons aged 18 and older receiving mental health services who experience
reduced psychological distress
Potential Social Functioning Outcome Measure
Percentage of persons aged 18 and older receiving mental health services who experience
increased level of functioning
Percentage of persons aged 18 and older receiving mental health services who report increased
employment (including volunteer time)
Percentage of persons aged 18 and older with serious and persistent mental illness receiving
mental health services who live in integrated, independent living situations or with family
members
Percentage of children aged 17 and younger with serious emotional disorders receiving mental
health services who live in noncustodial living situations
Percentage of persons aged 18 and older with serious mental illness who are in prisons and jails
Percentage of children aged 17 and younger with serious emotional disorders who are in
juvenile justice facilities
Percentage of homeless persons aged 18 and older who have a serious mental illness
Potential Consumer Satisfaction Outcome Measures
Percentage of adolescents aged 14-17 or family members of children and adolescents or both
who are satisfied with: (a) access to services, (b) appropriateness of services, and (c) perceptions
of gain in personal outcomes
Percentage of persons (aged 18 and older) or their family members or both who are satisfied
with: (a) access to mental health services, (b) appropriateness of services, and (c) perceptions of
gain in personal outcomes
Examples of Process and Capacity Measures

Program Strategy Improve access to services/utilization of services
Process Measure Percentage of adults with serious and persistent mental illness
who use health services
Process Measure Percentage of youth with serious emotional disorders who use
mental health services
Process Measure Percentage of those who use services that are voluntary
Process Measure Percentage of people requesting services who begin receiving
services within 2 week of the initial request
Capacity Measure Percentage of primary care providers who receive supplemental
training in mental health services
Program strategy Improve quality assurance
Process Measure Percentage of service plans that include input from consumers
and family members
Process Measure Percentage of children less than 5 years old who are screened
and assessed for mental health intervention
Capacity Measure Percentage of primary care providers who use standardized
screening tools for assessing the mental health status of primary
care clients
7.6. Example: Immunization
The vaccine preventable diseases are well understood. Cases are easily identified,
and health outcomes are well defined. A strong, causal relationship exists between
immunization and disease prevention. Clearly defined process (vaccination rate) and
outcome (disease incidence rate) measures are preferred, and they should be
responsive to effective interventions within a short period of time.
Potential Health Outcome Measures

Reported incidence rate of representative vaccine-preventable diseases
Potential Health Outcome Measures
Age-appropriate vaccination rates for target age groups for each major vaccine group:
Children aged 2 years; children entering school at approximately 5 years of age
Mumps, measles, and rubella
Diphtheria-tetanus-pertussis
Polio
hemophilus influenza B
varicella
Adults aged 65 years and older
diphtheria-tetanus
hepatitis B
influenza
pneumococcal pneumonia
Examples of Process Measures

Program Strategy Improve access to immunization services
Process Measure Percentage of population who do not cite financial resources as a
barrier to immunization
Program Strategy Increase parent education and awareness
Process Measure Percentage of parents with children under 18 who believe that
the benefits of immunization outweight the risks
Process Measure Percentage of parents with children under 18 who report
receiving immunization reminders from their immunization
providers
8. Summary
As stated in the beginning, performance evaluation in the health field is an
increasingly important area of study for future government officials, health
professionals, and service providers. The essence of evaluation is to isolate causes of
particular events or outcomes. If a particular intervention appears to be associated with
a beneficial effect, one must find out whether the effects can really be attributable to the
intervention or whether they might be the result of some other factors. With this
commonality, what distinguishes evaluation of public health action from other fields is
the use of rigorous scientific methods involving laboratory frontier knowledge and
contextual complexity requiring socio-economic and political awareness.
In order to set up an evaluation system in the field, building infrastructure is
required more than anything. The framework for evaluating public health intervention
needs to be devised, indicators need to be collected and assessed for future use. Also, to
utilize indicators, appropriate data for each indicator should be gathered and
organized. Furthermore, since indicators are designed to detect changes in a limited
time period, following up data collection should be made within a reasonable time.
Although it will take a tremendous amount of resources to accomplish these
prerequisite tasks, now is the time for Korea to tackle them in order to remain
accountable and to manage surging healthcare costs.
References
Centers for Disease Control and Prevention, Framework for Program Evaluation in Public Health,
1999.
Drummond, Michael F., Bernie J. OBrien, Greg L. Stoddart, and George W. Torrance, Methods for
the Economic Evaluation of Health Care Programmes, Oxford University Press, 1997.
European Commission, Evaluating EU Expenditure Programmes: A Guide, 1997.
Fink Arlene, Evaluation Fundamentals: Guiding Health Programs, Research, and Policy, SAGE
Publications, 1993.
Grembowski, David, The Practice of health Program Evaluation, SAGE Publications, 2002.
Ministry of Health and Welfare, Yearbook of Health and Welfare Statistics, 2003.
National Research Council, Assessment of Performance Measures for Public Health, Substance Abuse,
and Mental Health, (eds.) Edward B. Perrin and Jeffrey J. Koshel, 1997.
Shortell, Stephen M. and William C. Richardson, Health Program Evaluation, The C. V. Mosby
Company, 1978.
Suchman, Edward A., Evaluative Research: Principles and Practice in Public Service and Social Action
Programs, Russel Sage Foundation, 1967.
CHAPTER 12
Performance Management in Education: A Call for Paradigm
Reform in Korea
by
Ju Ho Lee, KDI School of Public Policy and Management
1. Introduction
Foreign institutions have affirmatively evaluated the performance of primary and
secondary education in Korea. Such is the case when these assessments are based on
outstanding math and science scores achieved during international competitions.
However, underlying these superficial results are factors that significantly influence
parents and students such as school satisfaction and teacher evaluations, which ranks
in the bottom tier among OECD countries.
This paper sheds light on student performance, which rests on the current college-
admission system, not the performance-management system. In short, the overall
increase in test scores can be attributed to the severe competition characterized by
students will to enter better universities, rather than the notion that a performance-
driven system in schools are increasing the quality of students.
Under the college-admission system, performance-management system is not
operating effectively. Schools are protesting the sole reliance of college admission-
based studies, emphasizing the need for character development. At the same time,
teachers and administrators at the Ministry of Education and Human Resource
Development are actively dissuading the implementation of scholastic-based
performance-management system. However, the absence of such system leads to a
vicious cycle, where students continue to seek private tutors through the inadequacy of
schools, rooted by skyrocketing demands for college preparation.
As the struggle to provide performance-management system for primary and
secondary schools emerge, surveys such as the National-Standard of Education
Achievement, City and Provincial Office of Education Evaluation (henceforth office
evaluation), and the Primary and Secondary School Evaluation (henceforth school
evaluation) are beginning to broach. Hence, this paper will assess the efforts of the
current performance-management system in primary and secondary schools and
propose a new scheme to improve the current performance-management system in
South Korea.
The structure of this paper is as follows. In section 2, a general overview of the
performance in schools and the relationship between the results of such performance
and the characteristics of performance-management system will be discussed. In
section 3, a general overview of the performance-management system in schools will be
examined and assessed. Finally, section 4 will propose reforms to improve the
performance-management system for primary and secondary schools in South Korea.
2. Performance in education
Korean scholastic attainment based on international assessment test is clearly high
among OECD countries as depicted in Table 12-1. For each category, Korea scored first
for science, second for math, and sixth for reading. However, when taking their
distribution into consideration, students below level 1 (1%) and students at level 5 (6%)
are quite low, meaning a large chunk of the students are positioned between levels 3
and 4 (39% and 31%, respectively). Although the size of students below level 1 is ideal,
the size of students at level 5 only resembles the 20th ranking country among the 27
countries surveyed in this test. Hence, although Koreas small proportion of low-
performance students is ideal, the incapability to educate outstanding students poses a
serious problem.
Table 12-1. Performance test

(Unit: Ranking out of 31, Percentage)
Korea Japan U. S. U. K. France Germany Sweden Canada Italy Mexico
Scientific literacy 1 2 14 4 12 20 10 5 23 30
Math literacy 2 1 19 8 10 20 15 6 26 30
Reading literacy 6 8 15 7 14 21 9 2 20 30
Below level 1 1% 3% 6% 4% 4% 10% 3% 2% 5% 16%
At level 1 5% 7% 12% 9% 11% 13% 9% 7% 14% 28%
At level 2 19% 18% 21% 20% 22% 22% 20% 18% 26% 30%
At level 3 39% 33% 27% 27% 31% 27% 30% 28% 31% 19%
At level 4 31% 29% 21% 24% 24% 19% 26% 28% 19% 6%
At level 5 6% 10% 12% 16% 8% 9% 11% 17% 5% 1%
Source: OECD, Knowledge and Skill for Life, PISA 2000.
In the results from PISA 2000, student engagement shows us very interesting
point.
In Table 12-2, student engagement is illustrated by asking students whether they
strongly agree, agree, disagree or strongly disagree to the following statements about
their school: I feel like an outsider (or left out of things), I make friends easily, I feel like
I belong, I feel awkward and out of place, other students seem to like me, and I feel
lonely. Forty-one percent of Korean students reported to demonstrate a low sense of
belonging, the lowest out of the 27 OECD participating countries. This is a striking
result, which challenges the academic achievement demonstrated in Table 12-1.
Furthermore, low participation, which measures how many times in the past week
students say they missed school, skipped classes, and arrived late, shows that only 8
percent of students feel that they have low participation. Perhaps pressures from
parents to attend classes and competition from colleges that require class participation
have contributed to this ironical result.
Chapter 12. Performance Management in Education: A Call for Paradigm Reform in Korea 171
Table 12-2. Student engagement

(Unit: Percentage)
Korea Japan U. S. U. K. France Germany Sweden Canada Italy Mexico
Low sense of Belonging1) 41% 38% 25% 17% 30% 23% 18% 21% 23% 22%
Low participation2) 8% 4% 20% 15% 15% 13% 24% 26% 22% 21%
Note: 1) Percent of students in each OECD country who have a low sense of belonging.
2) Percent of students in each OECD country with low participation at school.
Source: OECD, Student Engagement at School, PISA 2000.
However, why is it the case that high scholastic achievement and low sense of
belonging coexist in South Korea? One plausible explanation would be the prevalence
of privately funded private tutors and remedial classes outside school. The data from
Table 12-3 shows that Korea and Japan have remarkably high participation of
educational courses outside school. Interestingly, the inclusion of private tutors
reduced Japans participation to 17.3 percent while maintained the Korean numbers to
rank first at 57.5 percent.
The low assessment of Korean teachers as shown in Table 12-4, adds to the
possibility that high scholastic achievement scores received by students should be
attributed to private tutors rather than their schools. In table 12-4, students were asked
to indicate the frequency with which teachers in the language of assessment show an
interest in every student's learning, give students an opportunity to express opinions,
help students with their work, and continue to teach until students understand. Values
above the OECD average, set to 0, indicate higher than average student perceptions
that teachers are supportive, while negative values indicate below-average student
perceptions of teachers' supportiveness. Korean students report the lowest support
(-0.67) from their teacher.
Table 12-3. Student participation in educational courses outside school

(Unit: Ranking out of 26, Percentage)
Korea Japan U. S. U. K. France Germany Sweden Canada Italy Mexico OECD
Average ranking 2 1 8 13 N. A 20 21 18 24 3
Participation1) 63.6% 70.7% 24.7% 20.1% N. A. 10.4% 7.8% 14.4% 5.6% 51.4% 24.9%
Average ranking 1 21 18 19 N. A. 11 16 14 4 6
Participation2) 57.5% 17.3% 28.8% 24.4% N. A. 36.2% 7.8% 31.6% 48.0% 47.1% 32.3%
Note: 1) Percent of students who have sometimes or regularly attended courses in the language of assessment,
courses in other subjects, or extension or additional courses in the last three years.
2) Percent of students who have sometimes or regularly attend courses in the language of assessment,
remedial courses in other subjects, training to improve study skills or private tutoring in the last
three years.

Table 12-4. Teacher evaluation
(Unit: Ranking out of 27, WARM estimator87)
Korea Japan U. S. U. K. France Germany Sweden Canada Italy Mexico OECD
By Ranking 27 18 4 1 19 24 7 6 21 13
Student1)
Mean -0.67 -0.17 0.34 0.50 -0.20 -0.34 0.21 0.31 -0.28 0.07 0.00
Ranking 26 10 18 13 N. A. 16 5 11 25 3
By
Principle2)
Mean -0.72 0.14 -0.04 0.02 N. A. -0.01 0.34 0.08 -0.69 0.39 0.00
Note: 1) Index of teacher support and performance on the combined reading literacy scale, by national
quarters of index.
2) Index of principals' perception of teachers' morale and commitment and performance on the
combined reading literacy scale
In addition, school principals were asked to provide their views on teachers'

morale and commitment. To do so, they were asked to indicate how strongly they
agreed or disagreed with statement such as teachers work with enthusiasm, teachers
take pride in this school, and the morale of teachers in this school is high. In the
case of Korea, teachers comparatively have the lowest levels of morale and
commitment by the standards of school principals.
3. Performance-management in education
3.1 Performance monitoring88
The Central Education Research Center conducted two nationwide scholastic

achievement assessments in 1959 and 1963, evaluating 5 percent of 5th and 6th grade
elementary students. This helped identify and solve problems associated with teaching,
87
The index was scaled using a weighted maximum likelihood estimate, using a one-parameter item
response model (referred to as a WARM estimator; see Warm, 1985 with three stages: A). The question
parameters were estimated from equal-sized sub-samples of students from each OECD country. B) The
estimates were computed for all students and all schools by anchoring the question parameters obtained in the
proceeding step. C) The indices were then standardized so that the mean of the index value for the OECD
student population was zero and the standard deviation was one(countries being given equal weight in the
standardization process). It is important to not that negative values in an index do not necessarily imply that
students responded negatively to the underlying questions. A negative value merely indicates that a group of
students (or all students, collectively, in a single country) or principles responded less positively than all
students or principles did on average across OECD countries. Likewise, a positive value on an index indicates
that a group of students or principals responded more favorably, or more positively, than students or principals
did, on average, in OECD countries.
88
This section is heavily drawn from Lee, Myung-Hee (2000).
improve the curriculum, and collect fundamental resources necessary to encourage

learning. Consequently, in preparation for the middle school admission without
examination policy of 1969, the Central Education Research Center conducted the
basic scholastic assessment for primary school to test changes in the scholastic ability
of elementary students. In this test, 300 classes composed of elementary students from
1st to 6th grades were tested in Korean, Math, Social Science, and Science. The results
of the assessment were compared and analyzed by region, grade, sex, and scholastic
ability by calculating their average and standard deviation. Furthermore, areas that
were affected by the admission without examination policy Seoul, Pusan, and
Daegu were analyzed separately.
In 1974, The Korea Education Development Institute began the Elementary and
Middle School Scholastic Assessment intended for 1 percent of elementary and middle
school students nationwide to evaluate the efficiency in education efforts such as the
curriculum and the education methodology and to collect data for future plans. In
1977, in preparation for the nine-year mandatory education program, they conducted
the Elementary and Middle School Scholastic Assessment to further understand
students family conditions, school life, and scholastic ability. In 1980, to assess the
effectiveness of the new academic system fifth academic curriculum primary
education assessments were held. Again in 1983, 1984, and in 1985, multiple
evaluations of the new academic system were administered to compare and analyze for
future development of academic curriculum.
Moreover, in 1973 and in 1980, the Korean Institute for Research in the Behavioral
Sciences (KIRBS) oversaw the Nationwide Evaluation of Korean Education and the
Evaluation of Korean Primary Education to survey the academic ability of
elementary students. By evaluating Koreas mandatory education, they were able to
produce an analysis of education funding, which was intended to improve the quality
of education in the forthcoming academic systems.
With the establishment of the National Institute for Education Evaluation,
scholastic achievement surveys were administered in greater scope and at a national
level. From 1986 to 1992, 1 percent of all Korean primary and secondary students per
grade were assessed; from 1993 to 1997, all students were tested. The assessments
administered by the National Institute for Education Evaluation sought to collect and
analyze the effects, quality, and level of primary and secondary education from a
national level, understand the students, teachers, and other factors that influence
scholastic achievement, and finally develop top-notch academic materials to be
distributed and utilized for teaching at schools. Although the same institution
conducted these evaluations, variations in academic years, curriculum, scope, and
content have made it impossible to obtain long-term trends from these surveys.
Korean scholastic achievement tests were administered to understand the
academic level and to assess the curriculum and pedagogical methodology of primary
and secondary school students. However, it is striking that such national-level test does
not directly influence the accountability of primary and secondary education.
In the case of England, tests such as the Standard Assessment Test (SAT) and the
General Certificates of Secondary Education (GCSE) have functioned to survey national
goals and check for accountability by school or by region. In the U.S., the National
Assessment of Education Progress (NAEP) has been analyzed to evaluate student
performance by state.
However, Korea still does not possess a concrete scholastic achievement system
that is operated at a national level. Furthermore, surveys conducted by the Central
Education Research Center, KIRBS, and the National Institute for Education Evaluation
have not functioned to receive substantial feedback and to maintain a nationwide
quality. Hence, future assessments must focus on the implementation of education

policies from the level of the city and provincial office of education. In addition,
evaluations must be modified to objectively assess and monitor education performance
of individual schools to understand the scholastic achievement from a national level.
3.2. Project evaluations
Project evaluations in the Korean education sector refer to the progress in

implementing education policies and their assessments, such as the City and
provincial Office of Education Evaluation and Primary and Secondary School
Evaluation. Although these assessments are not exactly performance evaluations, the
goals outlined by these surveys show that they are, in general, project evaluations.
The background behind the introduction of the City and provincial Office of
Education Evaluation lies in the progress of introducing the Institution Evaluation
System, enforcing the Autonomy on Regional Education, and promoting the May
31st Education Reforms (Kong, Eun-Bae et al., 2002). According to the manual of the
City and provincial Office of Education Evaluation, the goal of these assessments
was to strengthen the accountability of the city and provincial offices of education
given by their autonomy. In other words, the implementation of the Autonomy on
Regional Education, the increase in autonomy and decentralization from the ministry
of education, and the transition from the Regional Education Subsidy Distribution
System to the Total Distribution System have raised the issue of accountability.
Moreover, the necessity of increasing competition among suppliers of education to
meet the demands of the education system as outlined in the reform agenda of May 31,
1995 have resulted in connecting the outcomes of the City and provincial Office of
Education Evaluations to the amount of funding. The idea behind this was to increase
the quality of education and to promote education reforms.
The City and provincial Office of Education Evaluation (henceforth office
evaluations) and Primary and Secondary School Evaluation (henceforth school
evaluations) have their legal basis from Article IX Clause 2 of the Primary and
Secondary Education Law and intended for regional-administrative institutions
(usually city and provincial offices of education) and national, public, and private
primary and secondary schools. The standard stated in the Enforcement Ordinance on
Primary and Secondary Education Law includes the composition and application of
funds, construction of schools and educational facilities, organization of the
curriculum, and personnel maintenance and welfare. Article IX Clause 2 of the
Primary and Secondary Education Law explicitly states the execution of school and
institution evaluations for an efficient accomplishment of education administration.
More specifically, office evaluations were introduced to bring about favorable
competition between city and provincial offices of education, promoting accountability,
increase in the quality of education, and raise voluntary efforts caused by the financial-
support system of the government (Kong, Eun-Bae, 2002). Hence, office evaluations
have the following goals: evaluating the process of implementing policies,
nationalizing regional differences, increasing education quality through autonomous
competition, providing information regarding the distribution of resources, and
proliferating excellent cases of administrative methods (Kong, Eun-Bae, 2002).
89
There are two goals for school evaluations. First, it provokes accountability
generated from the increase in autonomy for school management. Second, it provides
89
This paragraph is based heavily from Baek, Sun-Geun (1997) and KEDI (2003).
feedback and guidance for various educational activities and maintenance facilities.
Thus, it functions to boost efficiency and effectiveness. Finally, the disclosed
information provided by the assessment gives consumers of education a better idea
about each schools pedagogical activities and management programs to select schools,
teachers, and courses. However, as indicated by Baek, Sun-Gun (2003), most office and
school evaluations focus on achieving only the second goal outlined above.
3.2.1. City and provincial Office of Education Evaluation (office evaluation)

As of 2003, the office evaluation has been conducted 6 times since its introduction
in 1996 (administered every year from 1996 to 1999 and biennially afterwards). Office
evaluations have been conducted to expedite efficient education administration,
increasing school accountability, and raising issues relating to the quality of education.
As demonstrated in Table 12-5, the areas of evaluation are quite comprehensive,
but different each time the test is administered. The vast scope of evaluating standards
dictated in the Enforcement Ordinance of Primary and Secondary Law and the
ability of the Minister of Education and Human Resource Development to include
items at his will could lead to differences in testing areas and distribution. As a result,
areas of assessment, content, and distribution have in fact changed over the years and
have lead to problems in its consistency. Furthermore, taking regional differences into
consideration, evaluations are conducted by city and province. In addition, as
described in Table 12-6, evaluation results are disclosed and categorized according to
score and linked to government funding depending on their performance.
Table 12-5. Areas of assessment, number of subjects and items, and distribution of the
office evaluation
Round Number of Number Distribution Total

Areas of Assessment
(Year) Subjects of Items (City/Province) Points
1st Round z Education Policy 7 25 100
200
(1996) z Education Reform 15 50 100
z Class Reform 9 40 200
z Demand-Based School System 4 19 110
2nd Round z Embodiment of Welfare 5 24 130
700
(1997) z Efficiency in Information and Distribution of 4 19 120
z Administrative and Financial Support System
z Will to Promote Education Reform 5 17 140
z Classroom Innovation 3 43 130
3rd Round z Efficiency in Operating Finances 4 16 130
500
(1998) z Demand-Based School System 3 12 120
z Strengthening the Support System 5 18 120
z New School Culture to Improve Classes 3 39 160
z Performance-Based Teachers 4 18 60
4th Round
z Student and Parent-Based School 4 20 140/115 600
(1999)
z Efficiency in Operating Finances 4 19 100
z Strengthening the Support System City:4/Province:5 33/40 140/165
z Major National Policy Project 2 13 165
5th Round z Autonomy and Special Projects of the City or 3 6 125
500
(2001) z Regional Office of Education
z General Policy Project 10 17 210
Source: Kong, Eun-Bae (2002), p.30.
Table 12-6. Budget distributed according to the results of the office evaluation
(Units: Hundred Million Won)

Total Amount
1996 1997 1998 1999 2001
Subsidized
Total Subsidized
500 800 1,500 1,500 1,000 5,300
Budget
Basic/Difference 250/250 100/700 150/1,350 150/1,350 200/800 -
City Average 33.63 52.76 82.35 82.35 67.09 305.83

Provincial
33.13 53.71 102.62 102.62 58.93 351.02
Average
Total Average 33.33 53.33 93.75 93.75 62.5 331.25
Maximum 57.36 99.53 171.54 140.436 89.43 496.81
Minimum 17.04 24.49 27.56 41.71 46.87 202.59

- Basic Allotment
- Basic allotment (50%) - Basic allotment - Basic allotment - Basic
(12.5%) + Allotment (13.3%) + Allotment by
+ Allotment by Project (10%) + Allotment Allotment
by Total Score
(40%) + allotment to the Total Score Obtained by Total Score (20%) +
Obtained (40%) +
best Office of Education (50.7%) + Allotment by Obtained (40%) + Allotment
(10%) allotment to an Project (33.3%) + Allotment to an by Total
- 1 billion Won outstanding Office of Allotment to the overall outstanding Office Score
Education (45%) +
distributed to each office outstanding Office of of Education Obtained
Education (2.7%) (50%) (80%)
allotment to the
of education as basic
Basis for - Index of Basic - Index of Basic - Index of
overall outstanding
allotment and remaining
Distribution Allotment = Student Allotment = Basic
Office of Education (50%) + Number of
10 billion Won according Student (50%) + Allotment =
(2.5%)
to the Index of Basic Students (50%) Number of Student
Allotment = Student - Index of Basic - 50 billion Won subsidy Students (50%) (80%)
(50%) + Number of Allotment = Student increase + special
Students (50%) (50%) + Number of subsidy of 100 billion
Students (50%)
- 5 billion Won subsidy Won
increase - 80 billion Won
subsidy increase
Source: Office Evaluation Committee (for each year), and KEDI (2002).
3.2.2. Primary and Secondary Education Evaluation (School Evaluation)

There are three types of school evaluations: surveys conducted by the Korea
Education Development Institute under the request of the Ministry of Education and
Human Resource Development, tests conducted by the city and provincial offices of
education, and assessments as part of office evaluations. Surveys supervised by the
Ministry of Education and Human Resource Development are conducted by visiting
individual schools to identify the conditions geared toward improving school quality,
ameliorating the quality intended by school evaluations, identifying problematic
conducts of education, and providing a measure of scholastic achievement and school
satisfaction. In other words, this assessment identifies the current status of the
individual school from a national level, examines the effects of policies implemented by
the government at a grass-roots level, and fosters the exchange of information between
teachers, schools, and research centers with the results produced by the ministry (Baek,
Sun-Gun, 2003).
Table 12-7. Areas of assessment and contents of the school evaluation conducted by the
ministry of education and human resource development
Area Assessed Research Content
- Instruction through an emphasis on educational goals

Curriculum /
- Pedagogical activity that reflects student characteristic
Pedagogical Activity
- Pedagogical activity that develops voluntary study skills
Pedagogical Activity - Searching and discovering ones specialties

Outside the Curriculum - Livelihood guidance/basic attitude toward democracy
- Disguising duty, communication, animosity among coworkers

- Assisting the development of a faithful environment. Elevating
the specialties of the personnel
Activities that Support
- Resources necessary to conduct educational activities
Education
- Amicable communication between the school and parents and
parent participation at school. Supporting and Utilizing regional
resources
- Studying methods
- Communication skills
- Logic and creativity
Scholastic Achievement
- Relationship with adults
- Ego-identity
- Responsibility
- Students
Satisfaction - Parents
- Teachers and administrators
- Degree of difficulty involved when administering the 7th
Curriculum
- Class utilizing ICT
Education Policy - Achievement Test
- Special ability/aptitude education
- Reducing students per class
- School Administration Committee
Source: Ministry of Education and Human Resource Development and Korea Education Development
Institute (2003), pp.14-17.
Table 12-8. Contents of the field evaluation conducted by the ministry of education and
human resource development to evaluate schools
Areas of Assessment Major Questions
What are the basic characteristics of the school (composition of

Introduction of School
personnel and surrounding environment)?
Do the personnel of the school agree with the goals and plans?
Do the curricular-educational activities reflect each students
Curricular-Educational
specialty (student level, interests, abilities)?
Activities
Are curricular-educational activities developing voluntary studying
skills?
Are pedagogical activities outside the curriculum adequately
Pedagogical Activities exhibiting student skill and potential?
Outside the Curriculum Do pedagogical activities outside the curriculum help develop
qualities possessed by democratic citizens?
To provide better educational activities, is the school operated and
administered efficiently?
Is the school working to improve its teaching and quality?
Activities that Support
Does the school possess necessary facilities in conducting educational
Education
activities?
Are the parents, school, and the local community cooperating to
improve the quality of education?
While supervising the 7th Curriculum, to what degree has the school
faced difficulties?
Are ICT-utilizing classes aiding student comprehension?
Education Policy Are special-skill classes helping the development of student quality?
Is reducing class size helping to improve the school quality?
Is the Student Administration Committee contributing to the
democratic management of the school?
Source: Ministry of Education and Human Resource Development and Korea Education Development
institute (2003), p.22.
School evaluations held at the city and provincial offices of education level are
conducted periodically every two to three years, using the Primary and Secondary
School Evaluation Manual, of 1997 to inspect the progress of implementing education
policies and projects. In addition, these tests are acclimated to the trends and directions
of the individual city and provincial offices of education. More specifically, the Seoul
Metropolitan Office of Education has the following school-evaluation goals: first, it
should trigger the efficiency and the accountability of pedagogical activities, second, it
must promote autonomous and creative educational activities to alter the education
environment, and third, it should elevate the quality of education by discovering and
proliferating an excellent case of educational activity (Seoul Metropolitan Office of
Education, 2004). School evaluations administered by the Seoul Metropolitan Office of
Education are written and site surveys, composed not of qualitative tests but of
effort assessments, measuring the progress of educational activities geared toward
students and improvements in lecturing methods.
Table 12-9. Areas of assessment and evaluation tasks for general secondary schools of
Seoul metropolitan office of education
Area of Assessment Distribution Evaluating Project
Revitalizing exchange and experience learning.

Internalizing Character
Formulating an autonomous living environment.
Education based on 160
Creating a school without violence.
Family-Related (16%)
Promoting thrifty spending and environmental education.
Experiences
Internalizing unification education.
Activating enlightening activities.
Education to develop Invigorating special and unique skills.
150
special skills to fulfill Internalizing science education.
(15%)
dreams Systemizing career guidance.
Strengthening physical education.
Improvement of class to increase creativity.
Sustaining the innovation 290 Reforming evaluation to elevate logic and problem-solving skills.
of classroom and evaluation (29%) Establishment of guidance for responsibility to increase ability.
Promoting effective reading.
Spreading English education.
Increasing computer skills and educating information-
Developing knowledge 120
technology ethics.
and information skills (12%)
Provoke information-utilizing skills for teachers.
Use and developing a school homepage.
Promoting an autonomous specialty project for each school.
Increasing opportunities for parents to join school activities.
Constructing a school Internalizing the management of the schools administration
280
community for all to committee.
(28%)
participate Improving student welfare.
Developing teachers specialization and working conditions.
Enforce an appropriate and efficient budgeting of school funds.
1,000
5 Areas of Assessment 25 Projects and 25 Items.
(100%)
Source: Seoul Metropolitan Office of Education (2004), pp. 10-72.
The problems associated with the current office and school evaluations are their
ambiguity: the vagueness of their secondary goals (Kong, Eun-Bae, 2002). More
specifically, the extravagance of evaluating items and the lack of consistency has also
been pointed (Baek, Sun-Geun and Baek Sung-Joon). Moreover, there has been negative
feedback regarding the inadequacy in actively using the test results. However, in this
case, the problem lies in examining the survey as a source of spreading excellent cases
and correcting bottlenecks as a reflux mechanism. Thus, rather than liming the
assessment to disclosing outstanding cases, there must be successive measures under
the auspices of the Ministry of Education and Human Resource Development or by a
regional office of education. Therefore, through rigorous project evaluations,
modifications must be made to the project and its affected budgeting system. For
evaluation results to influence the performance-management system, achievement goal
and indices must first be developed, while redefining the content and procedure of
assessment.
3.3 Performance-based budget
Education budget falls under the jurisdiction of the Ministry of Education and
Human Resource Development. The annual revenue for 2004 increased by 6.1 percent
(compared to the annual revenue of 2003) to 4.6 trillion won of which general-
accounting revenue increased by 5.7 percent to 252.7 billion won and special-
accounting revenue increased by 6.1 percent to 4.4 billion won (both figures compared
to 2003) (Ministry of Education and Human Resource Development, 2004). Annual
expenditures were 4.0 trillion won for central-education finances and 22.4 trillion won
for provincial-education finances (84.9 percent of total budget).
In the process of producing the education budget, the city and provincial offices of
education present their annual revenue and expenditure budgets to the Minister of
Education and Human Resource Development. Consequently, the minister analyzes
the budget; for cities or provinces that have conspicuous declines in efficiency and
financial health, the minister is authorized to order inspections of the respective office
of education, disclose all results, and provide guidance and maintenance to retain
financial stability (Ministry of Education and Human Resource Development, 2002).
Furthermore, to determine whether funds were used according to their respective
proposals, evaluations are conducted annually before the distribution of resources to
increase transparency, performance, and administrative skills. However, a large
portion of the budget is spent on personnel expenses and on school administration,
making performance management only possible for policy projects.
The long-term plan of the government is to sustain a sound financial basis,
construct performance-based financing, and maximize autonomy and participation
(Ministry of Education and Human Resource Development, 2003a). Much attention has
been given to the construction of performance-based financing that the government
proposed, since it will foster performance analysis and evaluation of financial projects,
which will be reflected in the distributed budget.
The forecast for provincial finances look rigid due to projects improving education
environment, constructing facilities for the 7th Curriculum, executing the overall policy
to improve teaching professions, and promoting information and technology.
Moreover, with the implementation of new policies from the current administrating
government, difficulties are expected to be forthcoming in the finances of provincial
education (Ministry of Education and Human Resource Development, 2003a). Making
efficiency and performance-lead finance system the main priority of the provincial
education finance, the introduction of performance-based budgeting and its simulation
will be conducted (Ministry of Education and Human Resource Development, 2003a).
Furthermore, an analytical index of efficient investment on provincial education
finances will be developed and applied to promote soundness, efficiency, fairness,
investment return, and autonomy for individual schools.
According to the Special-Accounting Education Budgeting Proposal (Ministry of
Education and Human Resource Development, 2003b) to maximize financial
management and trigger efficient financial investment of performance-based financing
simulation, ten to twenty projects will be selected by each city and provincial offices of
education, compose performance index and proposal, which will be used for the
following year. In preparation, a budget-assessing committee will be formed to
evaluated and analyze financial priorities.
Performance-lead budgeting links projects and resources to performance. By
giving autonomy to each divisions expenditure and evaluating their performance, they
can assess the amount of money that can be distributed. Moreover, by disclosing
performance goals and actual results, this method generates transparency and
accountability.
At the moment, there are 27 performance-based budgeting projects administered
by the Ministry of Education and Human Resource Development via performance
index developed by the Ministry of Budget and Planning. For instance, the goals of the
curricular reform and continued support project have two goals: implementing a
national curriculum and developing specific performance goals and indices to support
research related to curricular improvement. Although the Ministry of Education and
Human Resource Development planned partial implementation of performance-based
budgeting, no specific strategic and performance goals and indices have been
developed.
To successfully introduce performance-based budgeting, transparent budgeting
and execution, complete disclosure of the entire process, and performance goals and
indices must first be developed. However, even with the planned simulation of
performance-based budgeting in 2004, the Ministry of Education and Human Resource
Development has not even developed a performance index. To include performance-
based budgeting to individual schools, an assessing body and evaluating methodology
must also be constructed.
Since 1999, the Ministry of Education and Human Resource Development has
simulated performance-based budgeting for the life-long education project, but did
not reap the results it wanted. The difficulty in assimilating an econometric and
performance index, unclear achievement results within a fiscal year, and the
unwillingness to cooperate to assess school performance by scholastic achievement,
academic level per school, and the rates of matriculation and employment, have
contributed to its failure. In any case, with high fixed cost associated with the budget of
city and provincial offices of education and the Ministry of Education and Human
Resource Development there are hopes that performance-based budgeting will be
implemented partially for national policy projects and projects at a city and provincial
offices of education level.
4. A call for paradigm reform

Primary and secondary school education policies in Korea have focused their
efforts to unify the school curriculum and evenly distribute funds rather than
managing and improving the performance of education. Their root causes could be
found in the equalization policy, which stresses the equal distribution of funds, and the
difficulty that follows the assessment of performance in education. However,
90
education, like other public and private sectors, needs to maximize its goals and
performance via an efficient allocation of resources.
Hence, this paper proposes the implementation of an Accountability-
Strengthening Program, which links funds to the achievement of pre-selected goals
designed to foster performance in primary and secondary education that will increase
student selectivity, maximize efficient fund allocation based on education demands,
and diversify the school system and its curriculum. Furthermore, improving the
90
Since the year 2000, the Ministry of Planning and Budget is overseeing the performance-based funding
system for institutions capable of implementing such system and is increasing their targets (Ministry of
Planning and Budget, 2001).
diversity of the education system, increasing the scholastic ability of low-income and
low-performing students, and financially rewarding education institutions that have
attained the goals set by the government and objectives of the Accountability-
Strengthening Program are intended for all public and private schools. However,
before such programs can be implemented, research pertaining to measuring student
achievement and school performance should be conducted to set a standard for
assessing school accountability.
4.1. Elevating the Education Performance Evaluation
Prior to implementing a performance-managing policy that fosters the

performance in primary and secondary school education, a methodology that
systematically measures student achievement and education performance for each
individual school needs to be researched and developed. Hence, creating a standard
and methodology to quantitatively assess student achievement, which could be used to
calculate individual school performance must first be made. Moreover, the construction
of such system must consider the following.
First, student achievement must be assessed via scholastic achievement and
character development, because these two are the functions of school education.
Furthermore, this paper proposes the implementation of a national or a regional
achievement evaluation held periodically. In addition, such evaluation must disclose
all results as a way of producing an unbiased and objective survey of school
performance. Likewise, a quantitative indicator of character development must be
developed and used to assess students periodically; the regional office of education
should use the results of such findings to construct new programs to foster character
development and support their implementation.
Second, there are three ways of evaluating student achievement: using the absolute
achievement, the change in achievement, and the success in improving student
achievement, which fell previously below the lowest standard. The first method uses a
students absolute achievement as a basis for evaluating performance for an individual
schools education. The second method uses the amount of change in achievement,
91
rather than their absolute values to determine educating performance. The third
method uses the success or failure of improvement for students whose achievement
92
falls below the lowest standard to assess performance. As to which program a school
chooses to participate depends on the type of education program for each school. In the
case of primary and secondary public schools in Florida, it incorporates the three
methods outlined above.
4.2. Introducing an Accountability-Strengthening Program
This paper proposes the implementation of an Accountability-Strengthening

Program for all public and private primary and secondary schools receiving
government aid. The Accountability-Strengthening Program calls for diversifying
primary and secondary education, evaluating accountability tests for the diversified
Charter Schools in the US use this method to evaluate school performance.

91
In the case of the No Child Left Behind Program in the US, which began in January 2002, success or
92
failure to improve student achievement make the basis for determining performance, intended for students in
the bottom 30 percent.
education program, and systemizing funds to be rewarded for education programs that
are more successful than others, which could be classified into three categories:
z Strengthening accountability through diversifying the school system,

z Strengthening accountability through the improvement of student
achievement for students of low-income and low-ability, and
z Strengthening accountability by rewarding institutions that have fulfilled
predetermined goals
Diversifying the school system via strengthening accountability has been restricted
as a supplementary issue to other reform agendas, pursued by education reform
committees. Thus, the central idea underlying the strengthening of accountability
through school-system diversification lies in the increased selectivity of schools for
students and the competition among schools, which ultimately leads to improvement
of performance of education in schools. The Charter School system in the US and the
93
system outlined in the FKI-CEPRI report are specific examples of this program. By
providing individual schools comprehensive jurisdiction over school management and
curriculum, thereby diversifying education, accountability of education is strengthened
through enabling student selection. More specifically, by implementing a formula-
94
funding system concurrently with the accountability-strengthening program, funds
will be efficiently allocated to schools where students choose attend, increasing
primary and secondary education performance.
Furthermore, this paper proposes the implementation of accountability-
strengthening program via increasing student-achievement performance for students
of low-income and low-ability. This proposal is different from the first, because it is
intended for all public and private schools, including charter schools. The focus for this
95
program lies in remedial classes, which will increase the scholastic performance of
low-income and low-performing students. Policies for students of low-income are
geared toward financial assistance (tuition grants and the provision of meals at school)
and only directed toward the extreme poor. Moreover, remedial classes for low-
performing students were focused on assisting the extremely poor-performing
minority. In other words, programs intended to increase performance for low-income
and low-performing students are almost nonexistent.
This paper proposes government provision of assistance for programs geared
towards increasing scholastic performance for low-income and low-performing
students. In doing so, a systematic research pertaining to improve student performance
must be conducted beforehand. The No Child Left Behind Program of the US that
93
The FKI-CEPRI report proposed the implementation of a charter school operating along the following
principles: increased student selectivity of schools, provided comprehensive authority over school operation
and funds, and issued discriminated government funding contingent upon student size.
94
Formula Funding is the specification of the government-discriminated funding system based on student
size often used by charter schools. Formula funding can be divided into four categories basic student costs,
curriculum costs, remedial-educational costs, and school-operation costs and the calculation involved for each
category involve detailed conditions.
95
Remedial education refers to remedial classes aside from regular courses and livelihood guidance,
excluding curriculum guidance.
began in 2002 could be a good case to model. In addition, the government must also
provide assistance for remedial programs in each school, geared towards students
whose household income and scholastic performance is below a substantial level.
Furthermore, schools receiving aid from the government must provide remedial
96
education for low-income and low-performing students; continued government aid
97
must be contingent on the performance of the remedial program. Active student and
parent participation must also be considered, and research must also be conducted to
98
attract participants.
Finally, this paper proposes an accountability-strengthening program that utilizes
a rewarding system. In this program, individual schools that have accomplished goals
outlined by the government to improve education performance will be rewarded to
promote school effort to ameliorate their system. More specifically, the government
99
will set achievement goals and recruit schools that want to compete within this
condition. Among the participating schools, the government will financially reward the
school and its participating members. To strengthen accountability through a
rewarding mechanism, this paper proposes the following.
First, regulations pertaining to school affairs should be temporarily exempt for
participating schools so that they could conduct various experiments to achieve the
goals outlined by the government. If schools are restricted by regulations, their
attempts for experimenting will be limited. In other words, the director of a regional
education office may function as an exogenous variable, affecting education
performance. Thus, a temporary exception for regulations related to school affairs
could be implemented in the same manner as charter schools.
In regards to the reward bestowed by the government, there must also be benefits
for the members of the school, including teachers and students. Without the active
participation of teachers and students, it is difficult to conduct experiments to attain
100
performance goals. Hence, providing direct rewards (incentives) will promote the
active participation to achieve these goals.
96
Because remedial education could be operated separately from regular classes, accountability-
strengthening program must be intended for all public and private schools receiving aid from the government.
97
In the case of the American No Child Left Behind Program, direct selection of teachers and curriculum
by students and assessment of scholastic improvement have contributed to the strengthening of school
accountability.
98
In a society that stresses academic attainment, students may feel embarrassed to participate in a program
for low-income and low-performing students. (Stigma effect) Thus, to encourage active student participation,
direct and indirect incentives must be provided.
99
The government can choose diverse goals such as improving of scholastic achievement and curriculum,
innovating class lectures, and developing character.
100
Rewards for teachers could be a bonus in their salary or additional benefits; for students, scholarship could be
considered.
Table 12-10. Accountability-Strengthening Program
Diversifying the School System to Assess Accountability (for Charter Schools)

z Diversifying the school system through the introduction of charter schools. (Guarantee
school selectivity, provide autonomous operation rights, and financial assistance through
formula funding)
z Test the accountability of the charter school by evaluating its education performance and
assessing the possibility of charter extension.
Improving Scholastic Achievement for Low-Income and Low-Performing Students (for all
public and private schools)
z Implement a remedial education program to increase the performance of low-income and
low-performing students.
z Implement an achievement-evaluating system to assess the quality of remedial education.
z Test for accountability based on the improvement of student achievement.
Promoting Competition Among Schools via a Rewarding System (for Schools that desire
competition)
z Government will outline performance goals and recruit schools that want to participate in
this competition.
z Bestow autonomous operation rights for schools participating in the competition,
comparable to the level provided to charter schools.
z Direct rewards for the school, teacher, and students involved.
References
(In Korean)
Baek, Sun-Geun, Proposal for connecting School Evaluations and the City and Provincial
Offices of Education Evaluation at the National Level, Prospects and Tasks for the Creation of
an Educational Society, presented at the Fall Academic Seminar of the Academic Association
of Korean Education Evaluation, Korea Education Evaluation Institute and Korea
Education Development Institute, 2003.
Chun, Sae-Young and Lee, Sun-Ho, Research on the Basic School Fee Distribution Model of
Formula Funding, Journal of Education Finance, Vol. 11, No.2, pp. 149-173, 2002.
Kim, Mynug-Suk, Goals and Uses of a National-Level Scholastic Achievement Test, Korea
Institute of Curriculum and Evaluation (KICE), 1998.
Kim, Sung-In, Problems, Analysis, and Improvement Proposals for College-Admission Policy,
Education Development: January/February, pp. 52-55, 2001.
Kong, Eun-Bae et al., Research to Propose an Improved Way of Conducting City and Provincial
Offices of Education Evaluations, Korea Education Development Institute, 2002.
Lee, Myung-Hee, Problems and Directions for Improvement for Koreas Scholastic
Achievement Evaluation System, Search for Koreas Curriculum and Education Evaluation for
the 21st Century, Education Policy Seminar, 2002.
Ministry of Education and Human Resource Development, Budget for 2004 Self-Governing
Bodies Special-Accounting for Education, Ministry of Education and Human Resource
Development website (http://www.moe.go.kr), 2003a.
, Summary of the Budget for 2004 Special-Accounting for Education, 2003b.
Ministry of Education and Human Resource Development and Korea Education Development
Institute, Survey of the Overall School Evaluation Results, 2003.
(In English)
Evers, Willianmson M. and Herbert J. Walberg, School Accountability, Hoover Institution Press,
2002.
Florida Information Resource Network, http://www.firn.edu.
Hassel, Bryan C., The Charter School Challenge, Brookings Institution Press, 2003.
Hill, Paul T., Robin J. Lake, and Mary Beth Celio, Charter School and Accountability in Public
Education, Brookings Institution Press, 2002.
Korea Education Development Institute, Research on the City and provincial Office of
Education Evaluation System and Methods for Improvement, 2002.
Lombardi, Joan, Time to Care, Temple University Press, 2003.
Ministry of Education and Human Resource Development and Korea Education Development
institute, Reports on the Results of the Overall School Evaluation, 2003b.
No Child Left Behind Program, http://www.nochildleftbehind.org
OECD, Student Engagement at School, www.oecd.org, 2003.
, Knowledge and Skills for Life, www.oecd.org, 2001.
Office Evaluation Committee, Report on City and provincial Office of Education Evaluation,
each year.
Seoul Metropolitan Office of Education, Report of School Evaluation for the 2003 Academic
Year, 2004.
CHAPTER 13
Establishing a Performance Management Strategy: Governing for Results
by
Harry P. Hatry and Ritu Nayyar-Stone, The Urban Insitute
1. Introduction
1.1. Rationale for performance management
Government agencies can achieve several important benefits if they regularly

(such as quarterly or monthly) track the performance of their programs. Program
managers and other public officials need information on both the outputs (the physical
products they produce, such as meters of road repaired and number of clients treated)
and the outcomes (the effects the program has on their customers, generally, citizens)
of their work. This is needed to help managers and other public officials identify
problem areas, identify needed improvements, and later assess the results of the
actions taken.
Principal benefits of implementing a performance management system are as
follows:
z Performance information can be used to improve service quality and results.

Over the long run, this is likely to be the most important benefit of
performance management.
z Performance information can be used by program managers and other public
officials to help them develop and justify their budgets.
z Performance information, particularly outcome information, can help
motivate public employees to make improvements and can also make their
jobs more interesting, especially if the outcome information (and not only
workload information) is regularly made available and discussed with them.
z Performance information is also an important part of a government's
accountability and transparency processes. Performance information adds
accountability, not only for financial probity but also accountability for using
public funds to accomplish important results.
Despite these benefits few countries have implemented a comprehensive

performance management system across most, if not all, all agencies at the national
level. Some possible reasons are as follows. Implementing a performance management
process will usually be quite new and perhaps strange and threatening to public
employees. It requires some changes in approach, and some new data collection
procedures (such as regular surveys of program customers). Making the transition is
not easy and takes time. The results, however, over the long run are likely to be well
worth the effort, especially in terms of improved services to the nations citizens.
Performance management appears to be more prevalent in local government,

where the connection between government agencies and citizens is closer and more
easily seen. The City of Baltimore (Maryland in the United States), is a recent example
of successful implementation of performance management - see the text box insert.
Baltimores Citi Stat, U.S.A.
Over the last several years, the Mayor of Baltimore has adopted the CitiStat approach as
a primary management tool for several of the citys agencies. Supervisors from different
agencies appear before the mayors cabinet every two weeks to discuss their agencies
performance data. A number of improvements were noticed very quickly.
- In the first three-month period there was a 25 percent decrease in overtime in the
Water and Waste Division.
- During the regular meetings, department performance graphs are displayed for all to
see, and trends of either good or bad performance are usually easy to see.
- City officials note that there is no longer any stigma of having to visit and respond to
questions from city leadership.
- There are more opportunities for interagency collaboration and for identifying and
sharing better practices.
Once it was decided to implement CitiStat it took three months to collect and analyze the
performance information and make a presentation at the Mayors office. Currently 14 agencies
are involved in the CitiStat process. The first agency (solid waste collection) made its
presentation on June 2000. Within four months seven agencies were included in the process
regularly reporting data to CitiStat and having bi-monthly meetings to use the information in
improving service quality.
The mayors office estimates a start-up cost of $20,000 that mostly includes infrastructure and
overhead cost and not staff salaries. Agencies can collect data with any software and in any
format. However, they all use a similar format in reporting the information to the Mayors
office.
1.2. Categories of performance information
Table 13-1 presents the categories of performance information that are relatively
familiar to agency managers. Data on the amount of resources expended for particular
programs (inputs), and the products and services that a program has completed
(outputs), are regularly used by governments to track program expenditure and service
provided. Results-based information (outcomes) is rarer, even though it is considerably
more helpful in determining the results and accomplishments of the service. Finally,
information that relates the amount of input to the amount of output or outcomes
measures efficiency (or its reciprocal) productivity.
Chapter 13. Establishing a Performance Management Strategy: Governing for Results 189
Table 13-1. Performance information definitions and sample indicators
Inputs: Resources used to produce outputs and outcomes.

z Number of full-time equivalent person-years were spent on the program
z Program Expenditures
Outputs: Products and services delivered.

z Number of people served
z Number of applications processed
z Number of inspections made
Outcomes: Results and accomplishments of the service provided.

z Employment rate for persons that completed the program
z Percent of youth graduating from high schools in schools supported by the program
z Infant mortality rate in areas targeted by the program
Efficiency and Productivity: The amount of input to the amount of output (or outcome).
The ratio of the amount of input to the amount of output (or outcome) is labeled
efficiency. The inverse, the ratio of the amount of output (or outcome) to the amount of
input, is labeled productivity.
z Cost per gallons of drinking water delivered to customers (output based)
z Cost per number of roads that were improved from poor to good condition
(outcome based)
Sources: Adapted from Hatry (1999), and Nayyar-Stone, Mark, Cowen, and Hatry (2002).
1.3. Content of report
This report provides recommendations for a multi-year strategy and specific steps
for providing and using performance information to produce a truly citizen outcome-
oriented government. It provides a framework for the introduction of a performance
measurement/performance management system in governments throughout the
world. The basic assumption is that the purpose of government is to support its citizens
in achieving the highest quality of life for the funds that the government is spending.
Performance measurement/performance management systems basically serve two
purposes: providing accountability and improving programs and services so as to
make them more efficient and effective. The key question for accountability is: What
are citizens getting for their money? The key question for improving services is: How
can we make programs and services better?
Throughout this report the term performance measurement is used to cover
measurement of both outputs and outcomes. However, the report focuses primarily on
the need to measure and manage outcomes, the most neglected, and most important
category of measurement. Outputs are important to measure since they provide basic
information as to the governments physical products. The relation of outputs to
expenditures is much easier to assess than the relationship of outcomes to either
outputs or expenditures. However, outputs tell little about the success of government
programs and services in achieving results important to citizens.
Performance management occurs when managers use the information from the
performance measurement process to help them manage and improve programs and
services. Performance management, it can be argued, is the most important reason for
performance measurement, even more important in the long run than use of
performance measurement to make government accountable. Fortunately, a good
performance measurement system can simultaneously provide information useful for
both purposes: achieving accountability and improving services.
Government also involves elected officials. They play a major role, as will be
discussed, in moving a government to performance measurement and performance
management. They should be considered to be an integral part of Governing For
Resultsboth by their support for the effort and by themselves using the information
from the process.
This report does not discuss fiscal or financial planning. These, of course, are of
major importance for successful governing but are beyond the scope of this paper.
Section 2 provides a sample of a time-phased set of actions needed for
implementing a performance management process. Section 3 identifies the roles and
responsibilities of the major participants, such as: (a) national offices such as the
Ministry of Planning and Budget; (b) the line ministries and agencies; (c) the
supporting ministries and agencies; and (d) the legislature. Section 4 discusses a
number of key elements relating to designing a performance management system.
Section 5 discusses a number of important elements in implementing the design.
As of this time, many national governments have begun introducing some form of
performance measurement. However, few appear to have made major progress in
introducing performance management. Even these efforts have tended to focus on
outputs and not on the results of those outputs. In addition, the focus in almost all
national governments of which we are aware, has been to move as soon as possible, if
not immediately, into performance budgeting.
Much less attention has been focused on helping managers use the information to
improve their services to citizens. This, unfortunately, means that examples of good
performance management practice by national governments are rare. We have, thus,
had to draw from the experiences of sub-national governments, such as local
governments, where more active progress has been made. Overall, however, the reader
needs to recognize that while lip-service on performance management has been
present for years, the actual practice at any level of government, while growing
considerably, is still rare.
2. Time-phasing of basic steps in implementing performance management

This section provides a model schedule for the first three years of implementing
a performance management process for a national government. A three-year period is
likely to be required for initiating such a process. However, full implementation is
likely to take many more years.
Countries such as Australia, Canada, Sweden, United Kingdom, and the United
States have each taken a number of years including periodic shifts in direction of their
processes. As noted at the beginning of this report, country initial efforts have focused
on performance measurement, performance budgeting, and sometimes strategic
planning. The use by agencies and their programs for improving services, and not
merely for budget preparation and justification, appears to be rare and a much more
recent development (It is not clear to us that this has yet been accomplished in a
comprehensive way). Local governments, particularly a number in the United States
have made more progress in using performance information for internal management
uses, but this also has not happened overnight.
2.1. During the first year or so:
Step 1: Initial implementation steps Establish an oversight committee,

develop a multi-year plan, and identify the resources needed
Establish a government-wide Performance Measurement/Performance Management

Steering Committee to provide comprehensive support, guidance, and advice on
performance management and measurement efforts throughout the government.
Agencies face many common issues implementing performance management
successfully. Each agency should not have to start from scratch. Table 13-2 provides
suggestions on the committees composition. This will also help to build support for
the performance management process. Each ministry and each agency might form a
similar steering committee.
Table 13-2. Possible composition of a national oversight steering committee
The Steering Committee might include representatives from the following types of
organizations:
z Budget Office, e.g., ministry of planning and budget
z Finance Office.
z National Planning office.
z Ministerial offices, perhaps up to three
z Operating agencies (at least three)
z Civil Service/Human Resource office.
z Office of the Auditor/Inspector General
z Outside expert
The Ministry of Planning and Budget in Korea is planning to establish a Central

Performance Management System Advisory Group one at each ministry. Advisory
groups are to consist of officials from the MPB and line ministries and private sector
professionals.101 High-level officials need to be fully supportive of this performance
measurement/management-by-results approach for it to be fully implemented and
sustainable over time.
Develop a multi-year implementation plan for performance measurement/performance
management (perhaps covering the next 3-5 years). The plan should identify which
organizations should do what and by when (for example identification of indicators, data
101 Kim, Jay-hyung (2003), p. 8.

collection, and analysis). Figure 13-1 provides a sample three-year schedule that focuses
on the technical steps for any single agency or program within the agency.
Figure 13-1. Sample Performance Management System schedule for each program/service
36 month schedule
Project Steps Month

0 3 6 9 12 15 18 21 24 27 30 33 36
First Year
Step 1:
Early
Implementation steps
Step 2-3:
Market and conduct
training on PM
Steps 4:
Identify data sources and data
Collection plan
Step 5:
Develop & implement process for
Citizen input
Second and Third Year
Step 6-7:
Establish baseline, set targets, &
establish data quality control
Procedures
Steps 8:
Report performance
Information regularly
Step 9-11: EarlyUses

Use performance information LaterUses
on a regular basis
Step 12:
Begin performance review and
recognition awards
Estimate the start-up costs and subsequent on-going implementation costs for the
performance management system. Seek funding for any significant added start-up
costs. Over the long run, performance management is expected to become a basic part
of the work of public managers and be routinely financed by regular operational

funding. (Note: this step may need to be delayed until development of the data
collection plan described in step 4 below.)
Step 2: Market performance management among employees
Market performance measurement/performance management among government

employees. Disseminate the reasons for regular outcome monitoring throughout the
government. Make it clear that this has high level and widespread support. Seek
support from elected officials, the business community, and the general public.
A key marketing approach is for upper- and middle-level managers to require
regular performance reports and to use them as an opportunity to hold program
performance review meetings. If program outcomes are a major subject of those
sessions, this will indicate to other government personnel that results are of key
importance. This type of meeting is discussed in more detail in Step 12.
Step 3: Develop guidance materials and a training plan for performance management
Develop and disseminate guidance materials on performance measurement/performance

management for use by both management and no-management personnel. Include definitions
of key terms relating to performance measurement, such as
outcome, result, output, etc. needed to avoid confusion in terminology among
agencies and their programs. This includes developing a manual on performance
measurement and on the constructive use of performance information (transforming
performance measurement into performance management) for use by individual
agencies and their programs.
Develop a plan for performance measurement/performance management training
programs and technical assistance. Estimate who will need training (e.g., what levels
and what functions) and when, and how many persons need to be included. Assign
responsibility for this step.
The training on performance measurement might include information on:
z What performance measurement is (focusing on measuring outcomes).

z How performance measurement can be used for managing performance.
z The types of data collection procedures (especially customer surveys, trained
observer ratings, and use of agency records).
Trainers in practical outcome-based performance measurement are likely to be

quite scarce. An option is to consider participants in the government's early
performance measurement programs efforts as candidates for trainers, even if the
training is for other programs or other agencies.
If the country has chosen to adopt a pilot approach to implementing performance
management (see section 4), ask personnel who participated in successful early trials of
outcome-based performance measurement systems to help with performance
measurement training. When feasible, also ask those personnel to provide technical
assistance to other agencies in their performance measurement development efforts.
Step 4: Identify current data availability and develop a data collection plan
Identify the current performance data available and the major performance data
gaps in what performance information managers need. Ask agencies to identify
performance indicators for which regular data are desirable for each of their major
programs. The indicators should include both outputs (the amount of work
accomplished) and outcomes (the quality and effects of the programs work on
citizens). The indicators should include both those for which data are currently
available and those for which data are not currently available but can be obtained in
the future. (Many programs will probably not currently be systematically obtaining
outcome data.) Customer surveys and "trained observer" procedures are both likely to
be useful to many agencies for collecting outcome information for their performance
measurement systems.
Request that each agency develop a data collection plan for the performance
indicators, particularly the performance indicators for which data are not currently
being collected. The data collection plan should: define thoroughly each performance
indicator; identify how the data will be collected (the data collection procedure that is
expected to use); the timing and frequency with which the data for each indicator will
be collected and reported; who will be responsible for obtaining the data on each
indicator, and any special resources needed for collecting the data.
Given the considerable potential usefulness of customer surveys for many
agencies, the government should consider providing some national technical assistance
in customer survey procedures to agencies and their programs. Any agency, and any
program within an agency, that plans to use regular customer surveys will face a set of
common issues.
Step 5: Develop a process for citizen input
Develop and implement a process for the government and its individual programs
for obtaining input from citizens on the quality and effects of the agencies services.
First, develop and implement a process for obtaining input from citizens for selecting
the outcome indicators, such as through the use of focus groups or hearings. Second,
consider using citizens to help with data collection, such as doing trained observer
ratings of educational or health facilities, roads, or landfills in their city.
2.2. During the second and third years:
Step 6: Establish a baseline and set targets for improvements
First obtain data for performance indicators to establish a baseline. This is vital for
future target setting by program managers. Such targets, of course, need to be related
to the resources expected to be available to the program.
Begin holding regular (perhaps quarterly) performance reviews whereby a
manager, at any level, meets with staff to discuss the findings and what needs to be
improved and how shortly after the performance report for the latest reporting
period has become available. In subsequent meetings assess the extent to which
changes have led to desired improvements.
Step 7: Develop procedures for data quality control
Develop procedures for data quality control. Such offices as those of an Auditor
General and any Agency Comptrollers might be encouraged to incorporate
performance data collection procedure reviews in their work. The purpose here is to
encourage, and ensure, that the outcome data are of satisfactory quality.
Step 8: Report performance information regularly
After the performance measurement process has been underway for a year or so,
ask each operating agency to develop annual reports for the public, ones that include
outcome information. Whenever possible, break out (disaggregate) outcome
information to show outcomes for different customer groups. Some examples of
candidate breakouts are income, age, gender, and location.
Involve citizens in group meetings to discuss the findings from the outcome
reports. These meetings would include discussion of customer/citizen responsibilities
in producing successful outcomes, as well as the role and improvement actions needed
by government programs.
Step 9: Use performance information in budget submissions
Begin calling for agencies to provide performance information in budget

submissions at the beginning of the second year of implementing a performance
management strategy. This is an early use of performance information. The
Government and its Budget Office can begin to request (and, at a later stage, require)
outcome information from agencies as part of each agencys budget submissions.
Budget-year targets for each performance indicator should also be requested for
agencies for each program. The budget requested for each program is that expected to
be needed to help produce the outcomes targets that the program proposes for the
budget year for each of its performance indicators.
Budget reviews by the Budget Office and, internally, by the agencies themselves,
should include reviews of past achievement of outcomes and reviews of the outcomes
that agencies hope to achieve in the new budget year (with the resources requested by
the programs).
If the government reviews outcomes as part of its review of agency budget
proposals, this will be a major motivator to encourage agencies and their programs to
obtain and use outcome information.
Step 10: Use performance information to make service improvements
Only by using performance information can each agency get the full benefit of
implementing a performance management strategy at the national level. Governments
initiating such a process often spend considerable time collecting data and then stop at
reporting the performance information the use of such information does not get
sufficient attention. One approach is to develop and implement service improvement
action plans. These focus attention on priority improvements by using performance
indicators to establish improvement targets and then track progress towards those
targets. This can be done towards the end of the first year after implementing a
performance management system.
A service improvement action plan is usually developed and implemented
through a working group created within the agency. This group can include senior
management, technical, and line staff. Researchers or external experts (from the
business or academic community) can also be members. The working group starts by
assessing the critical issues facing their agency. It then selects outcomes and indicators,
establishes baseline data, and sets improvement targets, develops and implements an
action plan with clear roles and responsibilities to achieve the targets, and collects data
again in 3, 6, or 9 months to identify the extent to which the desired
improvements in service outcomes have occurred.
As shown in the text box insert, a municipality in Albania has used these action
plans.
Improving Roads and Footbridges in Balduskhu, Albania
In 1999, Baldushku a rural commune consisting of 14 villages (average population of 400)

decided to prioritize the improvement and building of new roads and footbridges that connect
the villages to the main road. They did the following:
- Developed outcome performance indicators. For example, (1) increase the percent that
rate the condition of access roads as adequate or better from 37 percent to 45 percent,
and (2) decrease the number of footbridges rated in bad or worse condition.
- Undertook a citizen survey to obtain a satisfaction rating on roads and footbridges in
different zones of the commune. Survey results were distributed to all 14 villages via a
newsletter.
- Launched a public awareness campaign for citizen contributions to the building of
three footbridges.
- Obtained donor funding (World Bank, Soros, and International Organization for
Migration) to purchase road and footbridge maintenance equipment.
Step 11: Include performance measures in contracts
Consider using performance contracting when contracting for important

program services whether to non-government organizations (NGOs), for-profit
businesses, or other public agencies. Performance contracts include outcome indicators
and outcome targets in the contracts. Thus, performance contracts identify what
outcomes are required and how much outcomes are expected, rather than specifying in
the contract how to do the service. Performance contracts sometimes detail specific
monetary rewards and/or penalties for particular levels of outcome. (For example,
contracts for employment programs might pay the contractor a specified number of
money for each client for whom a job was found.)
The use of performance information in contracts can be initiated after the agency
and its program have sufficient performance data to establish performance targets for
the contract.
Agencies should encourage government contractors to use outcome measurement
procedures. Outcome data will help contractors regularly assess how successful their
activities are and help them identify elements of their services that need improvement.
Subsequently, the outcome information can help them assess the extent to which those
improvement actions have been successful.
Step 12: Begin annual performance reviews and recognition awards to motivate
staff
Begin providing good-performance recognition rewards for agencies, or programs

within agencies, for excellence in helping produce high outcomes. Such awards would
reward agencies or programs that have achieved high levels of outcomes and those
whose outcomes have improved significantly in the past year.
Begin including in annual performance reviews of program manager, and other
key staff, a review of the most recent available findings on program outcomes. If the
program has not yet implemented an outcome monitoring process, the performance
reviews of the manager and key staff would include, as one of the elements being
appraised, the extent to which they are making progress in implementing such a
process.
3. Roles and responsibilities
3.1. The role of the national oversight office (such as a ministry of planning
and budget) and its relationship to the line ministries
This is a key, if not the key, relationship needed to develop an effective

performance management process. The national ministry should provide basic
guidelines and ground rules for the overall process. At the same time, the challenge is
to provide enough flexibility to the individual line ministries and their agencies to
develop performance indicators, data collection systems, and reports appropriate for
their own needs. Providing common definitions of key performance measurement
terms and basic performance reporting formats that all agencies need to use in
reporting to the center is desirable to avoid confusion and inconsistent reporting.
See the Philippines example in the text box insert as a sample national government
initiative.
Piloting Local Government Administrative Reform in the Philippines
Within the context of the Medium-term Philippine Development Plan 1999 to 2004, the national
government has implemented the Local Development Administrative Measurement System
(LDMAS). The LDMAS was developed in consultation with national and local government
agencies, civil society groups, donors and academics and is organized around 11 key local
government performance domains (e.g., fiscal management, service delivery, land use,
infrastructure, etc.). These domains are measured through 67 input, process, output, and
outcome indicators, and have been recently pilot tested in 6 municipalities with the following
impact:
- Each of the six municipalities reported that the indicators provided administrators
with important measures of performance that were previously not accessible.
- In the small GoA municipality, the indicators directed attention to poor fiscal
performance of the city administration, leading to the initiation of a series of measures
to improve the fiscal performance.
The national office would call for ministries and their agencies to provide data on
their major outputs, outcomes, and efficiency indicators. This would be a requirement
in budget submissions. The national office probably would also call for reporting on
these indicators at regular intervals throughout the year, such as quarterly. The latter is
particularly important for moving from a purely performance measurement system to
performance management. These regular reports provide a basis for program reviews,
permitting more timely mid-course modifications throughout the year.
The national office should establish basic common reporting formats that it needs
to review performance. For budget submissions, it might call for separate reports on
each agencys performance indicators. Or more commonly done it might call for
identifying the indicators and their values along with the funding information for each
program. The national office might request that indicators be reported by category of
indicator (such as whether the indicator measures an output, an intermediate outcome,
an end outcome, or efficiency). The form might contain separate columns for reporting
indicator values for some number of past years, columns for the budget-year
performance targets for each performance indicator. (The form might also call for
reporting anticipated values for performance indicators for programs whose budgets
are expected to produce products in years beyond the budget year. This appears to be
rare in current performance budgeting systems, at least as a formal requirement.) The
national office should also include requirements for agencies to provide explanations
when the indicator values are considerably worse, or better, than the targets. See the
Texas and Louisiana example in the text box insert.
Requiring Explanatory Information: Texas and Louisiana, US
The states of Texas and Louisiana in the United States have required formal explanations for any
performance indicator values that are not within either plus or minus five percent of the target
that the agency had set for the performance indicator for the reporting period.
Both states require quarterly reporting on performance to the governor and legislature, and
have established web-based reporting systems, so that agencies can easily enter their results
every quarter. For example, see the Louisiana Performance Accountability System (LaPAS) at
http://www.state.la.us/opb/lapas/lapas.htm
A key element for governments that have performance measurement systems is

that providing the data is likely to be perceived by many managers as being
threatening. This is not an unreasonable concern by managers. This fear can lead to
various forms of non-cooperation, playing games with the numbers, reporting only on
indicators that have had favorable values. The national office needs to take some steps
to alleviate this problem, for example, by establishing a constructive rather than
punitive framework in using performance information.
The national office would review the indicators provided to it by ministries and
their programs to make sure they are sufficiently comprehensive and appropriate.
When not satisfied, the office would work with the ministry and agency to work out a
more satisfactory set of indicators. The national office also will be concerned about the
quality of the data. Suggestions for alleviating this concern are discussed later in this
report.
The national office also must be careful not to over-react to worse than expected
performance figures. It should avoid automatically blaming agency or program
management. Before acting, any reviewing office needs to first seek to determine the
reason for the poor results. For example, weak outcomes may have occurred not
because of any inherent problem with the program concept management but because
of external factors (such as worsening economic conditions that hampered the ability of
government employment programs to find jobs for unemployed workers).
Preferably the national office would act as a helper to line ministries and their
agencies in developing their own performance management processes. This is likely to
be a different perspective than is typical for national budget offices that consider their
role as solely to review agency budgets to keep them as low as possible.
The government is likely to need a separate unit from the unit that examines
budgets to develop the performance management guidance and provide assistance to
agencies (such as the management part of the Office of Management and Budget).
Each ministry and each operating agency will have its own special performance
measurement problems. The use of an across-the-government oversight group, as
discussed elsewhere in this report, can be used to help work out conflicts when they
arise between agency flexibility and the oversight offices needs for common reporting.
It is likely highly desirable for the national ministry to select the terms for the various
types of indicators, define them carefully, and provide examples. Unfortunately,
considerable confusion has resulted in many governments because of lack of at least
some attempt at standardization.
As noted in section 2, implementation is likely to be eased considerably, and
ministry and agency resistance reduced, if the government establishes a government-
wide oversight group that involves ministries and agencies in guiding implementation.
Table 13-3 summarizes the key responsibilities of the national oversight office.
Table 13-3. The responsibilities of the national oversight offices
1. Establish ground-rules and guidelines for required regular reports and for budget
submissions (such as: definitions of required types of performance indicators; guidelines
for presenting budget period targets for individual performance indicators; number of
prior years for which data are to be reported, requirements for providing explanations
for unexpected numbers; and the format for presenting this information).
2. Provide training to budget examiners in performance analysis.
3. Establish requirements, if any, for long-range strategic planning (including the number of
years to be covered and the content see Table 13-6 for suggestions on content).
4. Provide support and at least some resources for initial agency training and technical
assistance in performance measurement and performance budgeting.
5. Design incentives to agencies for performance measurement and performance
management (with an emphasis during at least the initial years in non-monetary rewards
for good performance, such as recognition awards for agencies and programs with
sustained good outcomes).
102
3.2. Role of ministries and agencies
The work of ministries and their agencies are, of course, essential to the successful
implementation of a performance management system. Each Ministry will have to
make a number of decisions specific to its own implementation of the system. These
decisions are listed in Table 13-4. Each decision is discussed below.
Table 13-4. Decisions at the ministerial and agency levels
Who will take the lead in establishing a performance management system in the ministry?
Who will take the lead in establishing performance management in the agencies?
Who is responsible for what (strategic missions and policy objectives; agency objectives,
indicators and targets)?
Who will select the performance indicators?
Will there be a strategic plan and when?
Who will be trained, in what, and by whom?
What resources are needed and available to implement the process?
What reporting requirements will be assigned?
How will performance information be used in preparing and justifying budgets?
How will internal performance targets be established?
What kind of IT support system will be necessary?
What kind of incentives will be used to encourage improved performance?
What is the role of performance information in contracting for services?
What will be the role of local governments?
What will be the role of customers, businesses, and nongovernmental organizations in
performance management?
What quality control procedures will be used?
When and in what form should performance information be shared with the public?
3.2.1. Who will take the lead in establishing performance management in the Ministry?
Each minister will need to designate a senior person to lead in establishing a
performance management system at the ministry. The day-to-day management and
supervision of this system will usually not be feasible for the minister. If possible, the
designated individual would be a non-political appointee, to assure continuity to the
system even with a change in government. This person should have sufficient authority
to make some key decisions on behalf of the minister. (Management and supervision of
full-scale performance management within a ministry might require, perhaps, 40
percent of this persons time over a one-to-two year period.)
Each ministry will likely find it very useful to establish a ministerial oversight
committee of, say, 8-12 individuals to oversee development and implementation of
performance management. In addition to the person appointed by the minister to guide
this process, the committee might consist of:
102 Some of the issues discussed in this section are drawn from Hatry (2002) and Nayyar-Stone, Mark,
Cowen, and Hatry (2002).

z High-level officials of a few agencies within the ministry

z Representatives of the planning and budgeting office in the ministry
z Representatives of the ministrys human resources office, such as the official
in charge of training
z A leading expert in evaluation or performance measurement, whether inside
or outside the ministry and government
z A representative from the national ministry of planning and budgeting (to
provide a broader perspective)
z A representative from the information technology office
z Possibly, one or two persons from outside the government who represent the
sector that the ministry serves.
Korea, in addition to establishing a national performance measurement system

advisory group, also is planning to establish a performance management advisory
group in each ministry.
3.2.2. Who will take the lead in establishing performance management in the agencies?
Each agency will need to identify personnel who will lead the agencys effort. The
agency head will need to designate that individual. Probably the agency should also
form an overall coordinating/oversight group consisting of representatives from a
number of programs, if not all major programs, in the agency that would work
together. The group would recommend agency strategies for such performance
management elements as developing an overall schedule, training of staff, common
reporting formats that supplement those required by the ministry, and obtaining
resources where needed.
3.2.3. Who is responsible for what?

Clear responsibilities and timelines will need to be developed for each of the
following:
z Developing an overall mission for the ministry and specific missions for the
103
individual agencies and their individual programs.
z Identifying the outcomes sought by each agency and each program,
developing the specific performance indicators needed to track progress of
each desired outcome, and who and how will targets be developed for each
performance indicator.
z Prioritizing and defining the agencies and programs that will be the initial
pilots adopting performance management if the government decides to
start with a pilot rather than full-scale adoption of performance management.
(It is probably preferable to cover the entire ministry from the start, rather
than specifying a few agencies as a pilot. This issue is discussed further in
Section 5.)
103 It is very difficult to define the word program very specifically. In general, a program is a group of
government activities that has a particular mission and is sufficiently substantial to warrant being identified as
a specific entity with its own budget. Each government needs to establish its own set of programs. Some
examples of what might be separate programs are: road maintenance, road construction, communicable disease
prevention (but the health agency might want to have separate programs aimed at particular diseases, such as
AIDS-prevention), water pollution control, and foreign trade development.
The overall mission for the ministry and its strategic/policy objectives are a major
basis for agency goals, objectives, performance indicators and targets. Therefore, it is
important that they are clear and specific enough to provide guidance to the agencies.
3.2.4. Who will select the performance indicators?

Generally, the individual agency programs are asked to develop the set of
performance indicators for the program. However, the ministry and agency need to
determine what reviews will be made of the programs choices (A national ministry,
such as the Ministry of Planning and Budget, should also later review the performance
indicators, at least the key performance indicators that it believes it needs for its
reviews and should be provided to elected officials).
Programs tend to focus on outputs (and to a lesser extent, unit-cost), indicators
over which they have most control, with little focus on outcome indicators. However, a
key principle of performance management is that a major focus is to assess and attempt
to improve outcomes important to citizens. Thus, upper-level reviews of the programs
choices of performance indicators will almost certainly be needed to assure that the
program includes an adequate balance of outcome, as well as output, indicators.
The identification of a clear set of clear mission/objectives and valid indicators
that can assess results is a critical foundation for a performance management process.
Shortcomings in these have led to many problems in implementing useful performance
104
management.
3.2.5. Will there be a strategic plan and when?

The concept of strategic planning is excellent. It is highly desirable that each
government agency periodically undertake an extensive examination of what its
missions is, what the agency wants to accomplish in the future (its objectives), what the
future is likely to look like, and what steps the agency needs to undertake to reach its
future objectives.
However, it is not necessary to carry out a full blown strategic planning exercise in
order to start implementing performance management. It likely would be better for the
agency to focus initially on determining what the agencys mission is, what type of
outcomes it seeks, and what outcome indicators should be tracked to assess progress in
achieving those outcomes. It probably should wait until it has gained experience in
performance measurement, after which each ministry may want to prepare strategic
105
plan.
Once a ministry has an established performance management system and a
strategic plan, they will have very strong synergies and complement each other very
well. For example:
z The mission statement, objectives, key outcome indicators included in the

strategic plan can also be an integral part of the performance measurement
system.
z The performance measurement system can be constructed to track the key
outcome indicators for assessing progress towards the objectives included in
104 This was noted by Gustafsson in this conference volume in describing the Swedish system. It appears to
apply equally to most, if not all, the other major country efforts to date.
105
One form of strategic planning is appraisal or situational analysis (also known as a strength,
weaknesses, opportunity, and threat analysis (SWOT)).
the strategic planas well as providing other, more detailed, indicators

needed for program reviews and to help in program improvement.
z The first year of the strategic plan should be the basis for establishing the
budget for that year.
The strategic plan provides an appropriate opportunity to stress that service to

customers is a key issue for the ministry, even for line staff. It can also serve as a public
information document for the public. The main concern with strategic planning is the
time and effort needed to develop a thorough and useful strategic plan. Many
government organizations that have prepared strategic plans only prepare complete
new plans every few years. They do, however, up date them perhaps every other year.
If this is not done, plans can quickly become obsolete and irrelevant.
Probably the main value is not the strategic plan itself but rather the strategic
planning exercise, forcing the government to think about and evaluate new strategies,
new approaches, especially in the face of changing environmental and technological
conditions. Many, if not most, strategic planning exercises, and the strategic plan that
results, have fallen far sort of being thorough examinations of the agencys mission, the
future, and alternative strategies. This issue is discussed further in Section 5.
3.2.6. Who will be trained, in what, and by whom?

Adoption of performance management requires adequate staff training to ensure
they have the skills and understanding of performance measurement and performance
management concepts and procedures.
Several levels of training are required for senior managers in the ministry
headquarters, for managers and supervisors in individual line agencies, and for many,
if not most, line and support staffs. It can be argued that every one in the government
should have exposure to the concepts of citizen-focused services, a central concept of
performance management. Probably the most efficient way is to first train senior
managers at the ministry and agency level using external trainers and any training
materials developed by the Ministry of Planning and Budget.
Following this, the ministerial staff should adjust the material so that it is specific
to the ministry and its agencies. Training for managers and professional staff in the
agencies might be provided initially by staff at the ministry. However, in the long run,
such training should be considered to be the responsibility of each agency.
Performance management should be considered to be just part of good management.
At the beginning of training for line staff, a one-hour presentation by the Minister and
head of the agency, indicating their support for performance management, can be very
effective in clearly indicating that these high-level officials take performance
management seriously.
It is important that training is done not only at the start of the Performance
Management Strategy, but also on an ongoing regular basis to reconfirm techniques
and learn new approaches. Training will also be required for new staff joining the
ministry. The content of the training is discussed in Section 5.
3.2.7. What resources are needed and available to implement the process?
The ministry with its agencies and programs will need to consider what added
financial costs and time will likely be needed. Then it will need to consider where that
cost and time will come from. Some initial assistance preferably would come from the
national ministry sponsoring the performance management effort, such as for initial
training. Some might come from outside funding sources such as the multi-lateral
banks and bilateral organizations.
Over the long run, however, the ministry and its agencies will need to be
responsible for their own performance management efforts. Performance management
should be considered to be just a part of good management. The potential costs of a
performance management process are discussed further in Section 5.
3.2.8. What reporting requirements will be assigned?

The ministry will want to specify reporting requirements for its agencies and in
a manner compatible with the requirements specified by the national budget ministry.
The agencies will want to specify reporting requirements for its programs. Reporting
requirements include:
z What types of performance information is to be reported;

z What explanatory information should be provided if actual performance
levels miss or considerably exceed, the targets for the performance indicators
and steps being planned or undertaken to correct the problem;
z In what formats;
z When they will be required.
While each agency and each program may choose to track a relatively large
number of indicators for internal use, each will need to select a shorter list for
external reporting.
In addition, the ministry and/or agencies, with their programs, need to develop
guidelines for analysis of the performance data, especially outcome data. The raw data
are rarely useful to managers and staff. Data on individual events need to be
aggregated and then grouped in ways that can be really useful. For example, outcome
data will usually be considerably more useful to managers if the performance data are
grouped by such characteristics as the following:
z By demographic groups, such as: age group, gender, race/ethnicity, and income group
z By geographical areas, such as: regions, districts, and municipality
z By the particular office or facility providing the service (it is a basic principle
of performance management that each manager should be provided
performance data related to each manager's own area of responsibility).
A major use of such groupings of the outcome data is to provide benchmarks

against which managers and others can assess progress. (Other benchmarks include
comparisons to outcomes in other similar countries. This is already commonly done for
a number of basic countrywide indicators, such as infant mortality rates. This requires
that reasonably similar data are being collected.)
3.2.9. How will performance information be used in preparing and justifying budgets?
Budget preparation and justification is of major concern to ministries and their
agencies and programs. Some budget instructions will come from the national budget
ministry. In a performance management system, a major new budgeting element is
performance budgeting the inclusion and consideration of output and, particularly,
outcome information as part of budget submissions.
However, each ministry and its agencies will likely want to determine the way
budgets will be prepared and submitted in more detail than that included in the
guidance provided to it by the national budget ministry. The ministry and agencies will
want to select those performance indicators that are most appropriate for inclusion in
budget submissions. Agencies and their programs will inevitably need for internal
management purposes considerably more output and outcome information than is

likely to be needed by the national budget ministry.
A major requirement in performance budgeting is to set targets for each reported
performance indicator for the budget year. The ministry and its agencies will need to
develop guidelines rules as to how these might be chosen. We note that choosing
targets for outcome indicators is likely to be more of an art than a science. While often
it will be possible to relate outputs to budget size, this is seldom, if ever, true for
outcomes. Seldom, if ever, will formulas become available that permit precise forecasts
of outcomes from budget size.
3.2.10. How will internal performance targets be established?

Setting performance targets for each output and outcome indicator at the
beginning of each year, and then reviewing performance against those targets
periodically throughout the year, are basic elements of performance management. As
noted above, establishing annual targets is usually required as part of budget
submissions. However, for internal management purposes, it is good practice to also
set targets for shorter reporting periods, such as for each quarter of the year or for each
month. Following are some ideas for selecting targets:
z First, consider previous performance.

z Consider new internal developments such as new technologies likely to affect
future performance.
z Consider external factors that can be expected in the future, such as
population and economic changes
z Consider benchmarking against the best, such as the outcomes achieved by
the programs best performing regional offices.
z If benchmarking against the best is considered too great a challenge, use the
average performance of all units.
z Make sure the targets chosen are feasible, given the agencys budget and
staffing plan for the year.
z Consider seasonal factors.
3.2.11. What kind of information technology (IT) support system will be necessary?
Computerized data-processing capability and skills in the ministry and agencies
will make it much easier to collect and analyze performance information and in a
timely manner. The extent of technical assistance from IT offices (whether coming from
the national level of government and/or from the ministry) will likely be important to
each program attempting to provide accurate and timely performance data on many
performance indicators.
If possible, information on each agency can also be placed on the world-wide-web
for easy accessibility and update.
3.2.12. What kind of incentives will be used to encourage improved performance?

Performance data can be used to help motivate agencies, programs, and their
staffs. A list of incentives that include performance as a major criterion, and have been
used in the past by governments, is presented in Section 4.
The incentives available will likely be largely determined by national, not the line,
ministries, especially monetary incentives. However, line ministries and their agencies
will likely have some flexibility in their application. Ministries and their agencies are
likely to have authority to decide for themselves whether to use non-monetary
incentives. These include:
z Recognition awards, whether awards are to be made to whole agencies, whole

programs, teams of workers, or individuals.
z Regular program performance reviews, perhaps held with staff shortly after
each performance report has been issued and disseminated to staff. At these
sessions, staff would discuss where performance was very high and where
performance was disappointing, discuss why, and suggest actions to improve
performance in the future. Note that this motivational approach does not use
awards to motivate. Rather, this approach motivates by establishing that
management is taking performance information seriously and wants to use
the information in a constructive way to improve public services. Regular
performance reviews may be the most effective motivator, non-monetary or
monetary! (Some large cities in the United States, especially New York City
and Baltimore, have begun using this approach in recent years with what
appears to be considerable success.)
z Allocating discretionary funds to agencies (or to individual agency programs)
for programs with high performance (such as providing extra resources for
training, for other professional development opportunities, or for new
equipment.
3.2.13. What is the role of performance information in contracting for services?

If an agency contracts or provides grants to other organizations for services to
customers, it can include performance targets in the agreements, especially outcome
targets, and then compare actual performance to those targets. This is called
performance contracting. The agreements can also include specific rewards and
penalties for meeting or exceeding targets and for failing to meet the targets. (One of
the first uses for performance contracting has been in road construction where bonuses
are provided for completing construction projects ahead of schedule and for reduced
fees for missing the schedules completion dates. In recent years such contracts have
been used for a wide variety of services, such as employment and adoption services).
An additional motivator for contractors to perform well is to include past performance
history as a criterion for future awards. Outcome-based performance contracts will be
made attractive for contractors, and reduce their resistance to them, by exchanging this
greater accountability for more flexibility in how the work is performed.
Government agencies need to consider a number of important issues when using
performance contracting:
z The agency needs to have capacity to establish such contracts so they are fair
to both the public and the contractor.
z The agency needs to have the capacity to monitor contractors carefully to
obtain the info it needs to determine the extent to which contractors have met
their targets and be able to judge the validity of claimed reasons outside
their control preventing them from meeting their targets.
z Performance targets should be compatible with, and support, the performance
indicators in the programs performance measurement system.
z The program will probably need to encourage, and help, contractors maintain
their own performance measurement processes but require them to allow the
program to undertake periodic quality control audits of the contractor's data
systems.
3.2.14. What will be the role of local governments?
Ministries and agencies that provide resources, or regulate, services provided

by regional or local governments, need to consider what performance information they
need for their oversight work. The ministries and agencies will need to work with those
governments to determine the reporting requirements. This includes identifying what
data on what performance indicators will be needed and the reporting timing. (Local
governments are likely to find the performance data useful as well.)
3.2.15. What will be the role of customers, businesses, and nongovernmental organizations?
These stakeholders have important role in at least three stages of implementing
a performance management system. Here are some potential roles:
z They are important sources of ideas for determining what outcome indicators
the program should track. This information can be obtained by holding focus
group meetings with customers, businesses, and NGOs.
z They can be important source of data for measuring outcomes. Reliable
feedback can be obtained from citizens and businesses by sponsoring quality
surveys.
z They are important potential customers for performance reports. They can
become supporters of the activities of the ministry if the reports focus on
outcomes (and not outputs) and provide clear, understandable, information of
interest to them.
NGOs can also join together in identifying the principal interests of their
constituents, and prepare report cards (see text box insert) to monitor the governments
performance.
Using Citizen Report Cards in Bangalore, India
In 1993 and 1994 the Public Affairs Center (an NGO), concerned about the deteriorating quality of
public services, developed and implemented a citizen satisfaction survey that measured user
perception on the quality, efficiency and adequacy of basic services extended by 12 municipal
agencies. The results of the survey were translated into a quantitative measure of citizen
satisfaction and presented in various media in the form of a report card. The 1994 survey was
followed up in 1999.
- Eight of the 12 agencies covered in the 1994 report card made attempts to respond to
public dissatisfaction. The worst rated agency the Bangalore Development Authority
reviewed its internal systems for service delivery, trained junior staff and began to co-
host a forum for NGOs and public agencies to consult on solving high priority civic
problems such as waste management.
- The report cards were also successful in generating political momentum for reforms.
Popular local and regional media carried regular stories on the report card findings.
Citizens have also been inspired to take initiative towards improving services and have
subsequently engaged in the state-citizen Swabhimana partnership in Bangalore a
program to arrive at sustainable solutions to longstanding city-level problems. The Chief
Minister of Karnataka has also established a Bangalore Agenda Task Force of prominent
citizens to make recommendations for the improvement of basic services.
Report cards have also been used as an effective tool in Ukraine, US, the Philippines and
elsewhere.
3.2.16. What quality control procedures will be used?

When implementing performance measurement, ministries and their agencies and
programs need to consider ways to ensure reasonable quality of the performance
measurement information. Data quality may not become a major issue until the
performance measurement information has begun to be used for high stakes purposes,
such as in making major decisions. However, attention should be given to data quality
from the beginning when important new and modified data collection procedures
are being introduced. Data quality should also be a topic included in performance
measurement and performance management training programs.
Periodic external, more independent, quality control reviews are highly desirable.
Audits may include random checks of data at 2 or 3 randomly selected agencies.
However, ultimately the responsibility for data quality belongs to the level where the
data originates. Responsibility will usually reside with the individual programs in
agencies. (An example of an exception would be citizen surveys conducted by higher
level organizations and seek performance information from multiple programs in the
survey.)
3.2.17. When and in what form should performance information be shared with the public?
Depending on the agency (and type of service), performance information can be
communicated to citizens on a quarterly, half-yearly, or annual basis. Reports should
not focus only on agency missions and activities, but also on service quality and
outcomes. Performance information is also most useful when adequately broken out
into categories such as gender, income, age, education levels, and geographical location
(rural or urban for example, or district/zone within a city). For external reporting
such as to the public it is important to keep in mind the following:
z Be selective as to how many and which indicators are included. Ministries and
agencies will internally track several more performance indicators than are
reported to the public. However, when selecting indicators for external
reporting, include the data on pre-selected performance indicators whether
performance has been good or bad. This will help establish the credibility of
the information.
z Make the reports easily understandable. As appropriate, use charts, graphs,
and diagrams to supplement tables and descriptive information.
z Obtain feedback from the public about the usefulness and readability of the
reports.
z If performance has been poor, consider attaching information identifying
plans for improvement.
3.3. The role of the legislature
The legislature is both a major motivator for, and a major user of, performance
information. Having legislation that requires the government to report performance
information provides one basic building block for performance measurement and
performance management.
Whether, and how, the legislature uses the performance information that it
receives can also have significant effects on the success of a performance management
system. Legislatures in some countries may have little real power so they cannot make
much use of such information even if they had good relevant results-based
information. In a parliamentary government the executive and legislative functions
tend to blend together. Nevertheless, most of these same issues apply.
Whether or not the government has a parliamentary system, line agencies usually
are reluctant and apprehensive about providing performance information to the
legislature. And many legislators may not be interested in, or may not trust, the
quantitative data provided by line agencies. Agencies may provide too much data on
too many indicators, many of which may not be very important (usually being output,
not outcome, data). For all of these reasons, thus far, use of performance information
by legislatures to make policy or even appropriation decisions seems quite sparse. The
lack of such use, while perhaps acceptable to the line agencies, also means that the
legislature is not motivating the agencies to strengthen their performance measurement
and performance management processes.
At some point, we expect that legislatures will become more interested in
performance information, especially information which represents important outcomes
for their citizens. Legislatures are likely to be interested, when good data on these
become available to them, in such data as that on educational achievement, on
indicators of health, on crime statistics, on measurements of air and water quality, on
transportation safety indicators.
4. Designing a performance management system

While designing a performance management system at the national level several
issues should be addressed. These are discussed in some detail below.
4.1. Need for cultural change
The concepts of outcome management are quite straightforward and common

sense. Nevertheless, outcome management does appear to require a shift in thinking
for government employees. Government managers have traditionally focused on their
budgets and expenditures that is, inputs. When they think of their products,
government managers have typically focused on the physical outputs they produce.
This is understandable since this is what managers can most readily count and perceive
as being their product. To switch to a focus on a customer/citizen orientation that is,
outcomes agencies need to change the way they think. Such an orientation means
that public services need to focus much more on what good they can do for citizens,
information much less currently available. The shift towards a performance-oriented
culture requires the following:
z That staff adopt the values, beliefs, and behavior that enable the government
to achieve its goals.
z That performance improvement becomes a way of life.
z That the various parts of the government begin to use the performance
information to help them operate their agencies and programs (for example,
by using regular program reviews and recognition awards to motivate
employees and underpin the performance culture).
Another major barrier to altering public manager culture is managers (legitimate)

concern that performance information will be used against them. If the government
emphasizes the use of performance information to make service improvements, to
make the organizations into learning organizations, this problem will be alleviated.
See the text box insert of the use of performance information by North Carolina, US.
Performance information can be used not only to identify problem areas, but also to
provide appropriate commendations or rewards to staff for high performance in the
face of adverse external factors. Problem areas can also be viewed as a signal for a need
for greater training and/or technical assistance. This, rather than the punitive element
of performance information, needs to be stressed to make managers less wary about
collecting performance information.
Using Outcome Information to Identify Technical Assistance Needs, North

Carolina, US
The North Carolina Womens and Childrens Health Division reports on performance to all
100 county health directors. Outcomes are compared, and used to identify those that need
most technical assistance (provided through regional consultants). Indicators include
adolescent pregnancy rate and percent women with short birth interval.
Another, important, way to alleviate such concerns is to explicitly include in any

performance management system the opportunity for agency and program managers
to provide explanations as to why particular performance levels have not met
expectations. This can be quite important because of the many external factors over
which managers have little or no control. They do not want to be blamed for poor
performance if it occurred because of factors outside their control.
In some instances, national service providers may be monopolies and not face
competition. They may be fully comfortable about their outputs and unwilling, or
not see the need, to focus on outcomes. In this situation the Steering Committee could
stress that (even though the agency does not need to worry about competition)
performance information can help agencies use limited resources more efficiently,
improve service delivery, and develop a better relationship with citizens.
4.2. Take advantage of existing public sector strengths
National governments, of course, differ in many ways. They can have different
cultures, different levels of resources, different degrees to which managers have
authority to take various actions, different degrees to which staffs have had training or
other background in analytical processes, different levels of IT capacity, different
histories as to the collection of performance data, different degrees of red tape that
reduces managers ability to implement performance measurement and use the
performance information to make changes, and so on. In addition, agencies within a
national government can also differ considerably in all these elements.
The features of performance management described throughout this report apply
in general to all governments and agencies (and programs). However, the degree to
which these elements can be accomplished, and when, can differ considerably. For
example, some countries, some programs, will need considerably more time before
they can obtain important outcome data. In such instance, these programs will need, at
least initially, to use less sophisticated data collection procedures, with less detailed
coverage, and less frequent data collection than are preferable. Manual data processing
procedures may need to be used until adequate computer technology and the training
to use it are obtained.
In some countries, or some agencies, managers may have little flexibility to use the
data to change service procedures or even suggest such changes. This can significantly
reduce the value of the information collected.
Some agencies in some countries already are collecting data of reasonable quality
useful for performance reporting. For example, agencies may be collecting data on
infant mortality, unemployment, incidence of various diseases, school dropouts, school
attendance rates, etc. (Even if such data are being collected, countries may not be using
this information for management purposes. Also, data collected are usually not broken
out by age, gender, location, income, education, etc. information likely to be
especially useful for identifying needed improvements.)
To the extent that outcome data are being collected on a regular basis, countries
can use this as a model for initiating performance management in the collecting agency
and other agencies. A national government might start its performance management
effort by identifying programs that already collect data that can be readily adapted to
performance information. Similarly, some programs, some agencies, might have
expertise in particular data collection procedures. For example, some agencies, and
programs, might be undertaking at least periodic surveys of citizens. The persons
knowledgeable about the procedures might be able to help other agencies, and
programs, use surveys, say of their own customers, to obtain outcome information.
(For example, if the National Statistics Offices already undertakes periodic national
surveys of the population, the staff of these offices might be able to provide training or
technical assistance to other agencies, or might be able to undertake the survey work
for the other agency.)
Other public sector strengths and skills that can be taken into consideration while
designing a performance management system are:
z Accounting system. A good accounting system that can be flexible enough to

provide cost accounting by program (as a primary basis for performance
budgeting and measurement of program efficiency) would put that
government far ahead for these components of performance management.
z Working relations across various departments and agencies.
z Business community. In countries in which the national government has
good, constructive, relations with the business community, they might be able
to provide whether voluntarily or under contract help in some of the
techniques and procedures of performance measurement.
4.3. Incentives to encourage performance management
A major issue in performance measurement and performance management is the

need to encourage government personnel to take the effort seriously. A number of
incentives can be used to help motivate public employees to focus on results and
service quality that citizens receive. These include:
4.3.1. Non-monetary incentives
z Offering more flexibility to ministries, agencies, and their programs in

exchange for achieving results. Managers might be given more control over
procurements, personnel changes, or more flexibility in use of their funds
if, for example, they had met their performance targets for some previous
number of reporting periods. This can have major appeal to agency and
program officials.
z Providing recognition awards to programs or agencies that, for example, had

met their performance targets for some previous number of reporting periods.
z Providing added opportunities for outside training
z Distributing the performance reports to program personnel so they are kept
aware of the progress their programs are making.
z Calling for agencies and their programs to report the outcomes of their
programs to the public.
z Holding regular program review meetings with program staff in which
performance is a major topic. Such How-Are-We-Doing? sessions can be the
most effective way to motivate public employees. They can demonstrate to
staffs that management is taking the results-oriented approach very seriously.
Such sessions would be held shortly after each regular performance report
was prepared. At the meetings, the group would assess where the programs
results were good and where not so good. The group would attempt to
identify why unexpectedly good or poor results had occurred and then come
up with an action plan for improving future program results. If the group
could not identify the reasons, the program might assign individuals to seek
such reasons and report back on their findings. In later meetings, after
changes had been made, the group would review the latest performance
information to assess whether the changes had achieved the results expected.
z Including performance criteria as part of annual individual performance
appraisals (if the government has such an annual performance appraisal
process).
z Recognizing good performance in one part of the program by asking staff to
provide technical assistance to another part of the program.
4.3.2. Monetary rewards
z Providing monetary awards to programs or agencies that, for example, had

met their performance targets for some previous number of reporting periods.
The use of such additional funds would might be limited to such activities as
special staff training, attending conferences, computer upgrades, and the like.
z Linking individual salary to performance. However, such attempts in the past
have often back-fired, causing problems and resentment by those not
receiving the added pay. This approach requires very careful design.
z Providing one-time bonuses to individuals or to teams who have achieved
high levels of performance.
4.4. Use of pilots or complete coverage
When starting up a performance measurement or performance management

system, a government may choose to pilot the process in a few agencies and programs
or call for coverage across all, or at least most agencies, especially the line agencies.
In our experience, across-the-board coverage is likely to be the better option. Pilots
suggest that the government does not have a full commitment to performance
management and represents only a half-hearted commitment to a performance
oriented culture. Using a pilot process will also likely stretch out full government
implementation for many years and the momentum from the initial support may
dissipate by the time the pilots are concluded. In addition, some agencies and
programs not covered are likely to take the position that their programs are different,
and more difficult, than those in the pilot. Thus, they will argue that the pilot was not
helpful to them, even if the pilot has been successful.
The Government of Thailand in the mid-1990s experimented with plots in three of
its agencies, but without a wider level of support the initiative waned by the time the
plots were completed, even though the pilots appeared to be successful.
The United States in 1993 passed performance measurement/performance
budgeting/performance management legislation that was also agreeable to the
executive branch with reporting by all agencies to begin in fiscal year 1999. A three-
year pilot effort was called for by the legislation, involving a small number of federal
programs. Most other agencies and programs stood by and waited, probably expecting
the new requirements to go away. The pilot effort appears to have delayed many
agencies from taking the effort seriously until it became clear that the government was
serious about the requirement. The pilot did have the advantage of providing time for
getting across the basic concepts. However, the initial reports from most federal
agencies were quite weak, probably because the non-pilot agencies and programs
delayed full-scale effort.
Clearly, different agencies, and different programs within agencies, will have
different levels of implementation success. It is very hard to tell in advance which
agencies and which programs will readily take to performance management and
develop more sound processes. Starting with wide coverage will at least get each
agency moving towards performance management even though, as expected, wide
differences in implementation quality occur.
One possibility that might be considered is while starting with a government-
wide implementation to select 3 or 4 agencies with whom to work intensively so as
to accelerate the process. This can have several beneficial results: more quickly yield
results in order to show other agencies the promise of the system; reveal problems with
the system that could then more swiftly be fixed; serve as a source for examples and
trainers for the rest of the government.
4.5. Projects vs. programs
Some governmental activities can be labeled projects. Projects have a fixed time
period and, usually, a somewhat narrow focus (such as individual road or wastewater
treatment construction projects). Many donor-funded activities fall into this category.
But much of what government and its agencies do are continuing activities, such as
maintaining roads, educating children, providing health care, regularly inspecting
business and natural resources for environmental problems, etc.
Performance management is an important approach for both programs and
projects.
4.6. Relation to in-depth program evaluations
In-depth program evaluations can be very useful in providing evidence as to the

success of a program and the factors that have contributed to the success or lack of
success of the program. Outcome measurement by itself will seldom provide sufficient
information as to the extent to which the program contributed to the outcomes or the
factors that have affected the outcomes. The manager of any sports team, such as
soccer, needs to know the running score, whether the team is winning or losing and by
how much. This is information that is provided by an agencys performance
measurement process. But the score does not tell why the team is winning or losing. To
determine that, the manager, whether of a sports team or of a government program,

needs to undertake more of an in-depth evaluation.
However, in-depth program evaluations require special resources and require
considerable time and effort. Governments that have resources may want to introduce
in-depth program evaluations from the start. In any case, few governments, and few
ministries, will be able to undertake more than a very small number of such
evaluations in any given year. Thus, only a small proportion of programs can be the
subject of in-depth program evaluations in any year.
For governments with limited resources, or just getting into the area of
"evaluation", it is probably better to learn to walk first before trying to run. That is,
such governments and their agencies are likely to be much better off by focusing on
obtaining outcome information from a variety of their programs and services in their
line agencies. Such regularly-collected outcome data will often be quite useful for later
in-depth program evaluations. Most evaluations are done after the fact, and thus may
depend on the extent to which outcome data were earlier available so that the current
accomplishments can be compared to the outcomes before the program was initiated or
in past time periods.
Over the long run, however, it is desirable that government ministries develop
capacity to undertake in-depth program evaluations on a select number of particularly
important programs. Table 13-5 provides a set of criteria that a ministry might use to
help it select which programs and services that will be the subject of an in-depth
program evaluation in any particular year.
A good step is for each ministry annually to prepare a program evaluation plan, or
at least update its most recent plan. Because such evaluations often will take more than
one year, such plans might cover at least two, if not three years into the future.
Ministries are likely to use outside contractors (such as universities or other non-profit
organizations) to undertake at least some of these in-depth evaluations so the
ministry does not have to maintain, and pay for a continuing in-depth large-scale
program evaluation capability. However, even if the Ministry contracts out all its in-
depth program evaluations, it will need to have staff to properly write and monitor
these contracts.
Table 13-5. Criteria for selecting issues for program evaluations
1. Can the results of the evaluation influence decisions about the program?
- Programs for which a decision regarding continuation, modification, or termination needs
to be made are obvious candidates.
- Poor candidates are those about which decision makers have strong preconceptions of
program value or for which there is considerable support by influential interest groups if
these circumstances make it very unlikely that the program would be altered. In some
cases, however, the program may be of such great importance to a community that
government officials may proceed with the evaluation and be prepared to seek changes
despite political risks.
- Can the evaluation be completed in time to be helpful to decision makers? Evaluations
completed after public officials become committed to a decision are useless.
2. Can the evaluation be done?

- Are sufficient data obtainable on important effects of the program? Program evaluations
can never cover all effects, but before an evaluation is begun it should be clear that it will
be possible to gather meaningful data on significant aspects of the program. For example,
the long-term impacts of job training, health, or education programs may not appear for
many years. In most cases, governments are unable to wait this long. Nevertheless,
intermediate results can often be made available and thus provide significant information for
interim decisions.
- Can sufficient resources be obtained to meet the time schedule and technical
requirements of the evaluation?
- Is the program actually in place, and is it being implemented as originally planned? If
not, an evaluation may not be appropriate.
- Has the program been stable enough so that an evaluation will provide relevant
information? A program that is constantly changing or is about to change in significant
ways is not a good candidate for evaluation.
3. Is the program significant enough to merit evaluation?

- Programs that involve a large amount of government resources or that have important
potential benefits or possible negative consequences to the public should be given higher
priority. Thus, the likely cost of the evaluation should be compared to the possible
decreased cost or improved effectiveness that may result.
- Is the program suspected of being marginal in its performance? If so, there may be opportunities for
making major improvements or cost reductions. Relatively old programs that have not received an in-
depth evaluation in a long time may fit this criterion.
- Is the program a new one whose potential benefits and costs are particularly uncertain?
The program should be reviewed after operating long enough to demonstrate its
effectiveness. This probably means a minimum of six months to one year of operation for
most programs, though even these time periods are often too short for evaluating long-
term effects. New programs are often more open to alteration because they have not yet
gained a strong constituency.
- Programs that are candidates for expansion are particularly suitable candidates for an
evaluation.
Source: Hatry, Winnie, and Fisk (1981).
5. Implementing a performance management system

This section discusses several issues that emerge and are relevant while
implementing a performance management system.
5.1. Capacity building training and technical assistance needs
As noted above, despite the common-sense focus of outcome management,

managers and professional staff throughout the government are likely to need a
meaningful amount of training. This includes both training in how to undertake
performance measurement (the technical side) and how to use the information
obtained (managerial training).
5.1.1. Technical training

Managers and professional staffs within both national and line agencies (including
those in administrative offices such as finance, procurement, and personnel offices) will
need some degree of technical training. This includes exposure to such subjects as:
z The distinctions between indicators of inputs, outputs, efficiency, intermediate

outcomes, and end outcomes including use of such tools as outcome sequence
charts (logic models) and the use of focus groups to obtain information on how to
identify appropriate performance indicators.
z Background in the concepts of analysis, with at least a basic understanding of basic
measurement approaches and the use of data comparisons to help the programs assess
how well the program is doing and the progress being made.
z Exposure to the basics of surveying program clients (likely to be a major source of
data for many public programs) to maximize the likelihood that surveys will be valid
and unbiased. (This training includes some exposure to sampling concepts to enable
agencies to avoid the need and added costs for surveying large numbers of clients.)
z The elements, and applications, of trained observer approaches in which personnel
are trained to make systematic ratings of physical conditions as a way to measure
various aspects of service quality
5.1.2.Training in using the performance data

Often badly overlooked in many, if not most, governments throughout the world
is the lack of attention to training managers, especially on how they can use the
performance information, especially to help them make program improvements and
not only using the performance measurement process to satisfy higher level
requirements. This badly neglected topic has meant that agency managers and ministry
officials tend to look on performance measurement as primarily something to get done
in order to satisfy some higher-level authority. The important need is to transition
managers from performance measurement to performance management. This means
training in such elements as the following, each of which can considerably increase the
value of the data to managers and their staffs:
z The need to obtain and report performance data in sufficient detail to enable
managers to identify and compare performance achieved for the various key
client groups the program serves, such as client demographic characteristics
(such as for various age, gender, income, and ethnicity groups).
z The need to obtain and report performance data in sufficient detail to enable
managers to identify and compare performance achieved for key service
characteristics (such as individual organizational units or regions and for

different service delivery characteristics where different amounts and/or
types of a service are delivered to clients).
z The need to obtain and report performance data on many specific service
characteristics, which while they may not be of interest to higher-level
officials, are important to citizens (such as obtaining information on the
timeliness, helpfulness, courteousness, and accessibility of the services).
z The need for regular and reasonably frequent collection and reporting of the
performance data. Typically higher-level officials examine performance data
on an annual basis, typically during the budget/appropriation process.
However, annual reporting is not likely to be adequate for managerial
purposes. Reporting performance data at least quarterly is likely to be
desirable for managers and their staff to track what is happening. This added
frequency will make it more feasible for managers to make mid-course
corrections and then have the opportunity to determine in later quarters
whether or not changes that were introduced have led to desired
improvements in performance.
z The utility of regular managerial performance reviews with staff after each
performance report has been discussed. Such reviews can be used to identify
where performance is on target and where not, to identify possible
explanations for unusually good or unusually disappointing per levels; and to
suggest ways to improve future performance (such as by identifying practices
that have been particularly successful for some organization units or for same
client groups that might be transferred to other units or other client groups).
z The need to undertake thorough searches for explanations for performance
that is unexpectedly for or good. A search for explanations should be an
explicit part of the performance management process. This topic is discussed
in more detail in a later section.
z The need to have data that becomes available and is processed in a reasonably
and timely way, such as within no more than one month after the end of a
reporting period. Data that takes years, or even months, to obtain will lessen
the ability of managers to react to that data and to make improvements.
The training, such as that described above, preferably will encompass all levels of
management and supervision. Professionals who are involved in some way in program
and policy issues also are likely to need at least some training in performance
measurement and its use. Provision also needs to be made, not only for initial training
in the initial years of the introduction of the performance management process, but also
for training for new managers and new professional staff, whether they come from
inside or outside the government.
Such wide coverage, while highly desirable, will likely put a strain on the
resources of a government and its ministries. Of course, not all the training needs to be
done at once but rather over a period of years. Much training will come internally,
through the on-the-job training of staff by persons who have had obtained some
previous experience and knowledge in performance measurement and performance
management issues. Some agencies in many countries will have staff that have
experience in performance data collection, analysis, and reporting, such as staff in
health education, and transportation agencies. These persons and their experience
might be drawn upon for training and technical assistance.
One strategy, that has been used, is to use persons that have completed the
training to help train others. Those persons might come from outside the agency or
even outside the ministry. In the long run, providing adequate training is the
responsibility of the agency itself. An agency can choose to use outside consultants or
inside the government personnel, or some combination. As far as the cost of such
training (and technical assistance), a good argument can be made that such training is
just a basic part of management, and the costs should be borne by the agency itself.
5.2. Performance information in the budget/appropriation process: performance

budgeting
In most countries the desire for information on performance by a legislative body

or by a national office, such as the Ministry of Planning and Budget, has been the cause
of the performance measurement activity. The usual reason for the high level interest is
to use the performance information in the governments budgeting and appropriation
process. Ultimately, it is assumed that budget decisions will be improved if they can be
made at least partly on the basis of data on recent accomplishments of programs and
program estimates of future performance if a budget of a particular size is approved.
The key question for public officials in the budgeting/appropriation process is: What
are we getting for the money that the government is spending?
Budgeting will become much more effective if the agencies and their programs
provide performance data of reasonable quality.
What is not widely recognized is that budgeting requires estimating the future.
Performance measurement (and program evaluation) is about measuring what has
happened in the past. Measuring results is difficult. Projecting into the future, even one
year into the future, with all the uncertainties and potential contributing factors is even
more difficult. Thus, budgeting will always likely be as much of an art as a science.
Estimates of how much particular physical output a particular amount of funds will
produce in the way of outputs often can be estimated reasonably accurately. However,
estimating the relationship between funding and outcomes is much more of a problem.
Many external and internal factors can affect the outcomes, such as changes in the
national economic situation, unusual weather conditions, international issues that
arise, loss of key personnel, unexpected behavior of citizens relating to a particular
service, etc.
The first need is for the government and its agencies to provide quality timely data
on performance to provide a basis for forecast performance for the budget year and
beyond.
A second need in performance budgeting is for agencies and their programs to be
asked to provide targets as to the likely outcomes that will occur with the amount of
funds requested. When funding levels are changed from those proposed, whether the
changes are made by a national office or by the legislature, the agency needs to be
given the opportunity to revise its targets accordingly. Target setting is even more of an
art than other parts of the budget process. In the initial year, or two, in which an
agency collects performance data for new indicators, targets should be considered as
particularly questionable until the agency has gained experience with the data for
those particular outcome indicators. Agencies should not be criticized for substantial
differences from targets for such new performance indicators.
Another complication for national governments, one not often discussed, is that
the funds for a particular budget year often produce outcomes in future years beyond
the budget year. It takes time for funds to be expended and then for outcomes to be
achieved. The outcomes from budgets from a particular budget year may not occur for
one, two, or say three years after the funds have been appropriated. Preferably, this
means that agencies are asked, as part of the formal budget process, to provide out-
year estimates of out-year outcomes for those programs for which significant outcomes
are expected to occur in later years. The authors of this report have seldom seen this
done, at least not in a systematic way.
There is a timing issues that are should consider. Full scale performance budgeting
cannot be done until agencies have a performance measurement process that provides
performance data. However, governments just beginning to develop performance
measurement systems can at least qualitatively address the question of what they are
getting for the money. Programs and agencies preparing budgets but lacking
performance data need to consider at least qualitatively the outcomes achievable from
requested amounts. Similarly, those examining budget requests will need to seek
qualitative information from requestors as to what outcomes they expect even if no data
are available. This at least will begin to encourage agencies and their programs to begin
thinking about, and giving consideration to, outcomes.
In the initial year, before any historical performance data are available, so that the
agency has had little experience with the data, relating budget amounts to the
performance indicators and target setting are not likely to be meaningful. The initially
available past data might be reported, but any targets set for those performance
indicators for the budget period probably should be considered to be questionable.
After experience is gained with the data, the performance indicator targets estimated
for the budget period, based on the budget amounts requested, will have a better basis.
Performance budgeting fully occurs when budget amounts are linked to outputs
and outcomes. This may mean simply displaying the amount of output or outcomes
expected for each budget program line. This is currently what is most often initially
done by governments that are introducing some form of performance budgeting.
However, real performance budgeting does not occur until agencies and their program
use the performance information to help make and justify their budget choices.
A special problem arises for performance budgeting for governments that only use
some form of line item budgeting. To be meaningful, linking performance to the
budget requires that the government have program/service categories to which
performance information can be related. Thus, this is a prerequisite for performance
budgeting.
A word of caution is in order at this point. Governments and other stakeholders
need to recognize that linking future outcomes to budget amounts will probably
seldom, if ever, be precise. Too many outside factors can affect outcomes to enable
agencies and their programs to have precise formulas that link funds to outcomes.
5.3. Long-range/strategic planning
Here we discuss issues involved with program and policy planning that contains a
major focus on results (and not only fiscal planning) beyond the budget year. These
plans represent an opportunity for a government, agency, and/or program to review
thoroughly what it is doing, how it went about it, and to identify what it should do in
the future, including possible major changes. To do this properly can take considerable
time and effort, including: making projections as to the nature and amount of the
future demand for the governments service; predicting the environment in which the
service will find itself; and analyzing various options as to what the government/
agency/program should do.
A first issue is whether strategic plans need to precede the development of a
performance measurement system. Performance monitoring does not appear to require
a strategic plan. Agency and program mission statements and general statements of the
agencys, and programs, objectives do need to be developed as a startling point for

both purposes. It can be argued that strategic planning will be better done if it is based
at least in part on the information coming from the performance measurement process.
Sweden initially required agencies to critically review what they were doing, but for a
variety of reasons has dropped this requirement.106 The United States requires five-year
strategic plans with interim updates. However, the development of these strategic
plans came after development of the initial performance measurement systems in the
agencies.
Nevertheless, long-range, strategic plans that consider both costs and outcomes are
conceptually quite desirable. The World Bank is encouraging developing medium-
term expenditure frameworks (MTEF) with a time horizon of about three years into
the future. (MTEFs have been prepared by many OECD countries. However, the
primary focus of MTEF has been on fiscal and financial policy, without much attention
to the role of information on policy and program outcomes for citizens.) The United
States government requires each federal agency to develop strategic plans, covering
five years into the future, and to periodically update them. Some planning can go out
much further into the future, ten or twenty years. The World Bank and other funders
have also encouraged multi-year Poverty Reduction Strategy Papers.
Any such long-term plan, regardless of whether it covers three, five, ten, or more
years into the future should focus not only on funding needs but also on what results
are expected and desired. (This vital element of long-range planning has often been
neglected.) Long-range plans should identify the key outcomes, provide key outcome
indicators for assessing progress, the latest data on those indicators, and provide
targets as to what the plan expects to achieve if the plan is successfully executed. This
provides the major basis for tracking progress against the plan.
Such plans can become obsolete quickly so governments, such as Sweden and the
United States call for their reviews periodically, perhaps every other year, if not
annually, to consider the latest developments. Strategic plans, if properly done, require
considerable time and effort by public officials to develop the plan. Thus, it is not likely
to be feasible to redo completely each year the plan. However, important changes and
unexpected events can occur that might significantly alter either the expenditure flow
or results expected. Thus, a government should probably at least review its plan on a
regular basis, such as annually or every other year, and modify it as appropriate.
Long-range, strategic planning is an opportunity for an agency to re-think what it
is doing. Full strategic planning would include such elements as those shown in Table
13-6.
106 See the paper by Gustafsson in this volume.

Table 13-6. Suggested elements in a states strategic planning process
z Identification of the population served and outcomes sought by the agency
z Identification of specific outcome indicators by which progress will be measured
z Examination of the future environment within which the agency and its programs operate
(such as demographic, economic, and social factors and trends)
z Identification of the latest available baseline values for each outcome indicator
z Examination of alternatives and practical options for achieving outcomes, including the
current service delivery approach
z Analysis of each options likely cost, feasibility, and effect on the outcomes, including
estimates of the out-year values for each outcome indicator included in the plan
z A process for obtaining input from the legislature, agency customers, agency employees,
and interest groups
z Selection of the strategic option to include in the plan.
Source: Adapted from Liner et. al., (2001).
Typically, most such long-range, strategic, plans have neglected many of the
elements listed in Table 13-6. They often provide little indication that the planners have
examined thoroughly future environmental characteristics (such as likely population
and technology changes). Seldom do they appear to have examined in a systematic way
alternative service delivery approaches to what they are currently doing. (Rather, the
plans merely present what agencies have been doing in the past and extend the current
process into the future.)
At which level or levels should long-range plans be prepared? Should each
Ministry, each agency, each program, have its own long-range plan? And should the
government as a whole have a go-wide strategic plan? Probably each ministry and each
agency within ministries should have its own plans. Probably major agency programs
should each have their own plans. Conceptually, a government-wide strategic plan is
also an excellent idea. A problem is that such plans can become quite political and be
revised every time there is a change in presidents and prime ministers. These plans,
thus, can become wish lists, and get-elected statements than thoughtful reviews of
what the government is about and what it should be aiming for in the future.
A key question is what are the purposes of long-range plans? Are they worth the
time and effort? One purpose often claimed for plans is that they provide a vision for
the future. As noted earlier, the planning process provides an opportunity for a
government, ministry, agency, or major program to examine from a fresh perspective
what it is supposed to be doing and how to do it. Such a plan, if it contains outcome
targets for the future, also provides a basis for assessing the governments progress
over time in achieving those objectives. And the plan should provide key input into
decision that need to be made in the next budget period.
To be of practical use, these plans should be linked to the organization's annual

plans, such as its budget and business plan. Thus, the first year of the plan should
provide a basis for the next budget plan.
5.4. Controlling data quality
Once performance data become available on a regular basis and become used for
high stakes purposes, public officials and key outsiders will become very concerned
about the quality and accuracy of the performance data. Government organizations
need to plan early how to maintain reasonable control and thus confidence in the data.
Inevitably, the primary responsibility for performance data quality will be that of
the organizations responsible for collecting the data and as a practical matter, this
responsibility will usually need to reside with the ministries and their agencies. Even in
those situations when an independent, outside organization is conducting an in-depth
program evaluation, the evaluation will likely require some data obtained by the
agency itself rather than new data obtained by the evaluating organization.
Some outside quality checks will need to be established, with outside-the-agency
review of at least a sample of performance indicators and performance data. The
sources might include periodic reviews by:
z A ministerial office, such as its office responsible for management,

performance, or budget oversight.
z A ministerial or agency inspector general office, if these exist.
z A national audit or inspector generals office perhaps located under the
legislature or a completely independent office.
The State Auditors Office of the State of Texas in the United States is required by
state legislation to review annually a sample of performance indicators in each major
agency to assess the quality of the data collection procedure for the performance
indicators. It has established a regular process for these reviews. The United States
federal government has a somewhat looser approach. Agency inspector generals
offices may undertake such reviews, but these are not formally required, nor does the
U.S. General Accounting Office undertake such reviews as an on-going activity. The
federal budget examiners who try to become familiar with the programs whose
budgets they review conceptually are in a position to make some appraisals of the
quality of the data that are provided to them. However, again this is not an officially-
required activity.
Performance data quality control appears throughout the world to be an
underdeveloped area. This is partly because the collection, reporting, and use of
performance data, particularly data on program outcomes, is still a new activity form
most governments and most agencies. A major problem is that agencies will be
generating large amounts of performance data, making it infeasible to review more
than a small portion of the data each year. A key ingredient for maintaining data
quality may be data transparency. When agencies report and disseminated
performance data widely, such as to the public, this is likely to place pressure on the
reporting agencies to provide data that can pass open review.
5.5. Costs of implementing and maintaining performance management
The additional financial cost of performance management is likely to be

surprisingly small. Most governments that have implemented a performance
measurement process consider the process to be a basic part of operations and have
expected agencies and their programs to absorb any cost without an added budget. To
our knowledge no one has provided cost estimates, either in monetary units or even as
to the amount of time, probably because of this. As discussed below, many factors will
affect how much additional financial costs and staff time will be needed.
However, a small amount of added cost appears to be a wise investment for
such activities as startup training in performance measurement and performance
management for managers and professional staff. Some legislatures have approved
hiring of small numbers of executive branch agency analysis staff. The legislature of the
state of Louisiana (in the United States) offered such funding support, but very few
agencies asked for the added staff.
Probably, the major cost for implementing, and subsequently maintaining,
performance management is the time of existing personnel. Some actual financial cost
will be required, such as for outside training and technical assistance.
Most costs, financial and staff time, will likely occur during the start-up year or
two. An important consideration is the likely need for individual ministries, agencies,
and programs, to establish new, or modified data collection procedures (including data
processing) for obtaining outcome data not previously available. Another need for at
least some ministries and agencies will be the cost of a small number of added staff to
oversee the performance measurement process, analyze the data, and provide reports
of the findings. This role, however, might be filled by existing staff, where appropriate
staff are available, and other roles, such as detailed oversight of line-item information
is diminished. For example, performance measurement has a citizen/customer focus.
Many programs will likely need to survey their clients on a periodic basis to obtain
outcome information. Many governments already undertake occasional national
surveys; however, for performance measurement, considerably more survey work is
likely to be needed and on a regular, at least annual, basis. For some surveys, agency
personnel probably will be able to use their own current resources or those of the
government office that does national survey work. However, if an outside contractor or
university is needed to administer the survey, this can entail added financial cost.
As noted above, the key cost is most likely to be staff time. Substantial staff time
is likely to be needed in the start-up phase of performance measurement and
performance management for such activities as training, selection of performance
indicators, and data analysis. As noted above, most governments introducing
performance measurement and performance management (such as Australia, Canada,
New Zealand, United Kingdom, and United States) have required the cost of these
activities to be absorbed by the agencies and their programs, as merely being part of
basic operations. A key issue for less industrialized countries is whether they are
willing to reduce other activities such as their attention to line-item budgeting and
instead ask staff to focus more on service outcomes.
Other cost issues relate to the current capacity of the government and its various
agencies. To the extent that agencies already have analytical staff and technology, such
as computers and copiers, effective and efficient implementation will be easier.
Governments, and agencies, with more limited resources will need more assistance and
have to settle for less, more selective, measurement and reporting.
Some ministry start-up costs should preferably be covered by the Ministry of
Planning and Budget, as an incentive for ministry participation. Bilateral or multi-
lateral funding organizations are also a source of funding for the start-up costs and
initial training associated with developing and implementing performance
management. Subsequent costs for the training of staff and on-going implementation
can be considered the responsibility of individual ministries and their agencies and be
funded from their own resources as has been done by all the governments that we
have examined.
References
Gustafsson, Allan, Performance Management and Budgeting: a Case Study of Sweden, World
Bank Working Paper, undated.
Hatry, P. Harry, Performance Measurement: Getting Results, 1999. Washington, D.C.: The Urban
Institute.
Hatry, P. Harry, Richard E. Winnie, and Donald M. Fisk., Practical Program Evaluation for State
and Local Governments, Second Edition, Washington, D.C.: The Urban Institute, 1981.
Kim, Dong Yeon, Key Issues for Introducing MTEF and Top-down Budgeting in Korea,
Washington, D.C.: The World Bank Working Paper, 2003.
Kim, Jay-Hyung, Performance Evaluation and Management System in Korea, prepared for a
conference on Performance Evaluation System and Guidelines with Application to Large-scale
Construction, R&D and Job Training Investments, jointly held by the World Bank and Korean
Development Institute, July 24-25, 2003.
Liner, Blaine, Harry P. Hatry, Elisa Vinson, Ryan Allen, Pat Dusenbury, Scott Bryant, Ron Snell,
and NCSL, Making Results-Based State Government Work, Washington, D.C.: The Urban
Institute, 2001.
Nayyar-Stone, Ritu, Katharine Mark, Jacob Cowen, and Harry P. Hatry, Developing a
Performance Management System for Local Governments: An Operational Guideline,
prepared for UN-HABITAT and the World Bank, June 2002.
CHAPTER 14
Results Based Budgeting in Australia
by
Richard Murray, Australia
1. Outline
z Overview of the fiscal framework
- Charter of budget honesty

- Medium term fiscal strategy
- Accrual budgeting
z The outcomes and outputs framework
- Outcomes, outputs and administered programs

- Case study: education
z What have we learned?
- Programs versus outcomes

- Measuring performance against outcomes
- The need for central oversight
- Evaluation
2. The fiscal framework
2.1. Charter of budget honesty
z Guides rather than mandates the conduct of fiscal policy. Requires the
Government to:
- outline its medium term fiscal strategy

- publish financial statements consistent with Australian and international
accounting standards
- specify fiscal policy actions that are temporary in nature, and the process for
their reversal
- publish an Intergenerational Report every 5 years
2.2. Medium-term fiscal strategy

z Strategy since 1996/97 is for budget balance on average over the cycle
z Supplementary objectives are:
- maintain surpluses while economic growth prospects remain sound

- no increase in the tax burden from 1996/97 levels
- halve net debt achieved in 2000-01
- improve the Commonwealths net worth position over the medium to longer
term
Figure 14-1. Improvement in net debt
Per cent of GDP Per cent of GDP

20 20
Projections
Estimates
Outcomes
15 15
10 10
5 5
0 0
-5 -5
1982-83 1986-87 1990-91 1994-95 1998-99 2002-03 2006-07
Chapter 14. Results Based Budgeting in Australia 229
Figure 14-2. OECD comparison of net debt
Per cent of GDP Per cent of GDP

90 90
80 Decade from 1995 to 2004 80
70 70
60 60
50 50
40 40
30 30
20 20
10 10
0 0
Germany Total OECD European Japan United States Australia
Union
2.3. Accrual budgets
Provides a more comprehensive picture of government operations
z Improves focus on fiscal sustainability (the balance sheet) and intergenerational

equity
- compares assets and liabilities (rather than just inflows and outflows of cash)
z for the Budget and forward estimates

z for individual agencies as well as whole of government
z Imposes more business-like discipline on agencies

z Ensures more focus on net worth
3. The outcomes and outputs framework

3.1. Outcomes, outputs and administered programs
z Outcomes are defined by Finance as:
The impact sought or expected by government in a given policy arena.

z The focus is on change and consequences: what effect can government have on the
community, economy and/or national interest?
z Outputs have been defined by Finance as:
The actual deliverables goods and services agencies produce to generate the desired
outcomes specified by government.
z Customers include members of the public, industries or sectors, ministers, members of

Parliament, other agencies or even in some instances, interests (eg the national interest).
z Administered Programs are defined by Finance as:
Those resources administered by the agency on behalf of the government (such as

transfer payments to the States, grants and benefits) to contribute to a specified
outcome.
z They are identified separately from outputs because they involve different
accountability requirements.
Figure 14-3. A model of the framework
Government Policy Platform and Priorities
Outcome A Outcome B Outcome C Outcome D Outcome E
Agency responsible for Agency responsible for

Outcomes A and B Outcomes C, D and E
Outputs and
Outputs and Administered
Administered
Items for Outcome E
Items for
Outcome B
3.2. Case study: education
3.2.1. Outcomes
Outcome 1 Outcome 2 Outcome 3
Individuals achieve Individuals achieve Australia has a strong

high quality relevant skills and science, research and
foundation skills and learning outcomes innovation capacity and is
learning outcomes from post school engaged internationally on
from schools and education and science, education and
other providers. training. training to advance our
social development and
economic growth.
Output 1.1: Funding Output 2.1: Funding for

for schools vocational education
and training Output 3.1: Research
Output 1.2: infrastructure
Assistance to Output 2.2: New
individuals apprenticeships Output 3.2: Assistance
including those with for science
special needs Output 2.3: Assistance collaboration and
for skills and career innovation
Output 1.3: development
Assistance for Output 3.3: Support
quality teaching and Output 2.4: Funding for for the Australian
learning higher education education and
training export
Output 1.4: Output 2.5: Assistance industry and
Assistance for to post school students international
transition through including those with relationships
and from school special needs
Outcomes should:
Outcome 1: Individuals achieve

high quality foundation skills
and learning outcomes from
schools and other providers.
Very Good
Focus on the end result,

not the means of Outcome 3: Australia has a
achieving it strong science, research and
innovation capacity and is
engaged internationally on
science, education and training
to advance our social
development and economic
growth.
Fair / Poor
Outcomes should:

high quality foundation skills and
learning outcomes from schools
and other providers.
Very Good
Be succinct
Outcome 3: Australia has a
strong science, research and
science, education and training to
advance our social development
and economic growth.
Fair
Outcomes should:
high quality foundation skills
and learning outcomes from
schools and other providers.
Very Good
Be specific to the area

being addressed strong science, research and
science, education and training
to advance our social
development and economic
growth.
Fair
Outcomes should:
learning outcomes from schools
and other providers.
Very Good
Be stated in a way that
allows the target group to be
identified
Fair
Outcomes should:

learning outcomes from schools and
other providers.
Good
Enable the formulation of
Outcome 3: Australia has a strong
effective indicators
science, research and innovation
capacity and is engaged
internationally on science,
education and training to advance
our social development and
economic growth.
Poor
Outcomes should:

other providers.
Very Good
Be free of value-laden,
generalized or aspirational Outcome 3: Australia has a
Good / Fair
Outcomes should:
Outcome 1: Individuals achieve high

quality foundation skills and learning
outcomes from schools and other
providers.
Good
Be amenable to
extension across
agency boundaries
internationally on science, education
and training to advance our social
development and economic growth.
Good
Outcomes should:

other providers.
Very Good
Be able to be related to
Budget documentation
internationally on science, education
and training to advance our social
development and economic growth.
Fair
3.2.2. Outputs
For Outputs it should be possible to:
z define outputs in terms of what is produced;
z measure both departmental outputs and administered programs in terms of

price, quantity and quality (PQQ); and
z relate outputs to an outcome
- if this cannot be done the agency should cease to produce the output or
identify an additional outcome that the agency should achieve.
Output Groups
Output 1.1: Funding Output 2.1: Funding for Output 3.1: Research
for schools vocational education and infrastructure
Output 1.2: Assistance training Output 3.2:
to individuals Output 2.2: New Assistance for science
including those with apprenticeships collaboration and
special needs Output 2.3: Assistance for innovation
Output 1.3: Assistance skills and career Output 3.3: Support
for quality teaching development for the Australian
and learning Output 2.4: Funding for education and
Output 1.4: Assistance higher education training export
for transition through Output 2.5: Assistance to industry and
and from school post school student international
including those with special relationships
needs
Defines what is being provided, price and quantity are measurable and Outputs
relate well to Outcomes.
3.2.3. Targets and performance indicators
Figure 14-4. Mean scores of Australian students in reading, mathematical and scientific
literacy compared with OECD average for 2000 (set to 500)
S c ie n tific lite r a c y 528
M a th e m a tic a l lite ra c y 533
R e a d in g lite ra c y 528
525 526 527 528 529 530 531 532 533 534
Figure 14-5. Graduate starting salaries as a proportion of Average Weekly Earnings
Bachelor VET
60.3
2002 82.7
60.2
2001 85.8
60.8
2000 84.2
59.9
1999 81.6
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
per cent
Figure 14-6. Number of patents issued to Australian residents in Australia
2001/02 1,763
2000/01 1,418
1999/00 1,295
1998/99 1,419
0 500 1,000 1,500 2,000
4. What have we learned?

4.1. Programs versus Outcomes
z Ministers often make policy decisions based on programs

z Budget documentation tends to provide information on Outcomes
z There has been a move towards providing more administered program
information to Government
4.2. Measuring performance against Outcomes
z Many outcomes are too broad. For example The Defense of Australia
- In this case the Government has moved to define 7 sub outcomes for reporting
purposes
z Outcome 1: Command of Operations in Defense of Australia and its interests

z Outcome 2: Navy Capability
z Outcome 3: Army Capability
z Outcome 4: Air force Capability
z Outcome 5: Strategic Policy
z Outcome 6: Intelligence
z Outcome 7: Superannuation and Housing Support Services for Current and Retired
Personnel
4.3. The need for central oversight
z A devolved management model and relative financial autonomy has led to a lack
of central knowledge on financial information
- Leads to a lack of scrutiny of agency performance by central agencies

- Directly affects the ability of central agencies to advise on Budget Strategy
z The Government has acted to provide the Department of Finance with additional
resources and mandated program and cash reporting
4.4. Evaluation and Audit
z Evaluation of performance information is essential to measuring the achievement

of Outcomes
- Senate Estimates Committees scrutinize performance information ex-ante

- The Audit Office scrutinize performance information ex-post
z Are Chief Executives held accountable for Outcomes?

CHAPTER 15
Performance Management and Budgeting: a Case Study of Sweden
by
Allan Gustafsson, Mapsec
1. Country Context
Sweden is a constitutional monarchy and the King is the formal Head of State. The
Constitution is based on the principles of popular sovereignty, representative
democracy, and parliamentarism. Sweden has a unicameral parliament, the Riksdag,
with 349 seats. All elections are by proportional representation. A party must gain at
least 4 percent of the national vote to qualify for representation. There are presently
seven parties represented in the Riksdag.
Swedish politics has, since the 1930s, been dominated by the Social Democratic
Party. The party has since then, except for two short periods, been in power as the
dominant party in a formal coalition, or as a minority government with the explicit
backing of one or more minor parties in the Riksdag. Presently the Social Democratic
Party is backed by the Greens and the Leftist Party but their collaboration is limited to
economic matters.
The ideological pillars of the Social Democrats can be said to be three: private
production, public redistribution, and public provision of welfare services. The
economic policy of the Social Democrats has generally been pragmatic, but the 1960s
and 1970s saw a rapid growth in the size of the share of national income that passed
through the government. Most of that growth was in the directly redistributive part of
government spending, i.e. in entitlement programs. Another basic tenet of the
Swedish model has been and continues to be that social benefits shall be general in
nature. That is they shall not focus uniquely on the most disadvantaged of society but
rather every tax payer should feel that he gets something in return for his tax
contribution. The trend towards an ever-growing public sector stopped abruptly in the
early 1990s when a fiscal crisis, in the form of a ballooning deficit, forced the
Government to dramatically curtail spending. As a result the share of General
Government in GNP was rapidly brought down from 70 percent during the height of
the crises to approximately 55 percent where it has now leveled off.
In the context of the topic of this paper it is noteworthy that the dominant share of
government spending more than 50 percent refers to individually consumed social
services. A little less than one third can be classified as individually consumed private-
type services while only one tenth refer to classic collective goods.107 In other words:
107 A service is classified as individual if it is associated with positive marginal costs for each person
consuming it. It is conventionally classified as being of a social type because they are traditionally provided by
the public sector and/or the public sector determines its price and availability for example health care,
education, child care, care for the elderly, culture, etc. The category individually consumed private-type
services refer, for example, to support to agriculture and housing allowances. Collective goods refer to defense,
central administration, etc.
there is an individual, identifiable consumer at the receiving end of the vast majority of
the activities of the Swedish public sector!
2. Origins of the reform and the historical context

The present system of public management is not the result of a one-time
comprehensive reform but rather of a drawn-out series of partial reforms the
aggregate result of which could be termed performance management. Whether or not it
can be claimed that performance budgeting has been introduced depends on how you
define it.
This long reform process dates back to the 1960s when in an Official Government
Report it was suggested that what was called program budgeting be introduced. The
inspiration came from the US military under McNamara. The fundamental idea was
that the budget should be divided into programs. It should be possible to identify the
outputs of programs as well as the quality of these outputs. It should also be possible
to calculate the full cost the outputs, including indirect costs. The proposal, however,
drew a line for what indirect costs should be apportioned to final output. Central
administration, for example, was to be treated as a separate auxiliary activity.
Somewhat contradictorily, it was suggested that separate programs should be set up
for investments. To complement the one-year perspective of the budget, it was also
suggested that every agency should elaborate multi-year plans showing planned
output and costs. In regard to appropriations, the Official Government Report
suggested two alternatives: either that they be expressed in terms of cost with a
separate set of accounts in expenditure terms (modified cash) or that they be expressed
in expenditure terms with a separate follow-up in cost terms.
The ideas put forth in the Official Government Report on Program Budgeting were
tested in a number of agencies over several years. The pilot tests were eventually
abandoned and the attempts have often been maligned as unrealistic and mechanistic.
It is nonetheless clear that the Official Government Report and the pilots had a
profound impact on the reforms implemented during the decades to come.
The Official Government Report on Program Budgeting was followed by a number
of new studies, proposals, and initiatives. In the early 1970s a new Official Government
Report termed the Budget Review presented a first proposal for a division of the
budget into a number of broad areas. The study proposed 23 sectors under which some
3-8 program groups would be defined. Every program group could contain 2-8
programs. Appropriations would apply to programs. A lot of emphasis was put on the
need for long-term planning and a number of instruments to bring in the long-term
into the management process were proposed. The Official Government Report had no
immediate practical impact on processes and procedures but some of its ideas have
later been picked up, for example the division of the budget into a number of broad
areas. The latter reform was introduced as part of the introduction of top-down
budgeting in the late 1990ies. (See Chapter 5)
Results-based management as we see it today was further shaped by a 1985
Official Government Report which resulted in concrete government proposals tagged
on to the supplementary estimates in 1986 and in 1988. A starting point for the
proposals was that they should permit a continuous review of government activities
with a view to reallocating resources from low-priority to high-priority areas.
Thus, the budget process should be changed so that:
Chapter 15. Performance Management and Budgeting: a Case Study of Sweden 243
z the Riksdag (Parliament) and the Government would focus less on the details
but rather provide general guidance and define the objectives of government
activities;
z decision that customarily had been taken outside the budget process, but
which had an impact on the budget, should be handled within the budget
process;
z a longer-term perspective was introduced in the budget process;
z productivity and efficiency would be enhanced allowing the provision of
more relevant services;
z agencies were given greater responsibility for the use of the appropriations
provided for the running of the agencies and for the organization of its
activities; and
z the analysis and the reporting on results achieved would be improved.
One important and concrete change of the budget process was the introduction of
three-year cycles for the presentation of the agencies budget proposals. Every third
year, the agency was to produce a more ambitious and soul-searching document in
which it was expected to critically review what it was doing and how it went about it.
The proposals for year 2 and 3 needed not go to the same depth but were rather to
serve as bridges between the more thorough reviews. The cycles of the agencies were
staggered so as not to overburden the Government Offices. On the basis of the more
thorough review carried out every three years, the agency would receive a firm three-
year budget frame.
The system was maintained for a few years but was dropped in the middle of the
nineties when the urgency to reduce spending left no room for elaborate and
thoughtful prioritization processes. The whole idea that the agencies would voluntarily
question their own activities at least beyond the margin was probably unrealistic
anyway and it is doubtful that the system would have survived even in calmer seas.
A reform ushered in by the 1988 bill a reform that has survived is the
requirement that each agency shall produce a yearly report on results. After the
introduction of accrual accounting in 1992, the report on results is normally combined
with a statement of accounts.
The 1988 bill furthermore announced a number of changes in the legal framework
for the management of the agencies as regarded:
z the management of the agency direct income;

z the setting of fees;
z the use of framework appropriations;
z the financing of agency investment;
z the accounting framework; and
z cash management.
These intentions were gradually translated into government ordinances and

instructions during the first part of the nineties.
Not specifically presaged by the bill of 1988, but of particular importance was the
introduction of accrual accounting in 1992/93.
The Generic Budget Law of 1996 gave a results-based budgeting a legal basis. The
Performance Budgeting Project, generally known by its Swedish acronym VESTA,
initiated in 1997 picked up a number of loose ends left by close to three decades of
piecemeal reforms with the objective of creating a more consistent and systematic
performance orientation to the entire management framework. Its work focused on the
coverage, basis, and structure of the budget, the definition of appropriations, the
control and management of investments in infrastructure and in the defense area, and
on the structure and flow of performance information. A first Ministry of Finance
White Paper was prepared in 2001. The work of the project has now been integrated
into the normal development work of the Budget Department of the Ministry of
Finance and continues. Some of the original ideas of the project are shelved for the
moment, while others are pursued. A new white paper resulting from the continued
work is expected during the first part of 2004. The most recent reform initiatives are
discussed below in the section on outstanding issues.
3. Elements of performance management and budgeting in Sweden

Rather than an outcome of a single comprehensive and consistent reform,
performance management in Sweden is thus, as shown above, the result of a long series
of piecemeal reforms strung out over a thirty-year period. Despite the mixed record of
the early experiments, the course was really set in the 1960s with the introduction of
the notion that the public sector should be managed rather than simply administered.
Many of the basic ideas of the early proposals have also survived, albeit often in a
different format, while others have been added. The initial years were, however,
largely experimental and the concrete Swedish performance management model as we
see it today is the outcome of reforms implemented during the last fifteen years.
What follows is an overview of the elements that in some way or another
constitute performance management la Sude as we see it today. Problems,
shortcomings and future areas of development are discussed in the final section.
3.1. The objective of performance management
The fundamental objective of performance management is to increase the

effectiveness and the efficiency of public action in general and of the application of the
resources managed by the public sector in particular.
The concrete content of performance management is set of rules, mechanisms,
processes and technical support systems that will enhance effectiveness and efficiency.
In principle the arena for performance management ought to span all actors involved
in shaping and executing public action from the legislative bodies to the civil servants
delivering services to the public. In other words it ought to involve: a) the
parliamentary process, b) the relationship between the legislature and the government,
c) the internal processes in the Government Offices, d) the role of and relationship with
audit institutions, committees, independent think-tanks, etc., e) the relationship
between the Government Offices and the agencies, f) the internal process within the
agencies, and g) the interaction between the agencies and the public.
In practice performance management has come to be associated very much with
the relationship between the Government and the agencies, with the internal
operations of the agencies, and their relationship with the public. Most of the elements
listed below fall into those three categories. The shortcomings and the areas for further
development relate mostly to categories a) d).
As the content of public activity and public spending (doing the right thing) is
determined primarily at least at the more aggregate level in the first four processes,
while efficiency (doing things right) is determined more in the latter three, the focus
of reforms have thus been on enhancing efficiency.
The driving force of the reforms has been the conviction that efficiency is a
function of authority and accountability. If institutions and individuals are given
authority to decide on how best execute a task they will behave responsibly and
efficiently, provided they are held accountable for the result.
The listing of the elements of the Swedish performance management model is thus
divided into elements that give authority to institutions and individuals, and those that
are put in place to achieve accountability.
3.2. Authority
3.2.1. Decentralization to local level

This paper will focus on central government. However, it must not be forgotten
that the perhaps most important delegation of authority within the Swedish public
sector is that to local government. Of the total 28 percent of GDP that is public
consumption, 20 percent is the responsibility of local government. There are 289
municipalities and 18 county councils. The division of responsibility between county
councils and municipalities has it that tasks requiring a larger population base,
secondary health care and local transport for example, are handled by the county
councils, while the municipalities take care of the rest.108
The elements of performance management at the central level that will be
discussed below have all been part of the local government landscape for a long time,
longer than at central level. Most services are executed by institutions with a great deal
of managerial independence. Many local governments have implemented purchaser-
seller arrangements for much of its service provision even between public
institutions. Health clinics and hospitals, for example, are often reimbursed on a piece-
basis for services provided. Many local governments have introduced a system of
school vouchers allowing an individual student to attend any school with the local
government area, including privately run schools. Most community services like street
cleaning and local transport are outsourced. When they are not, their operations are
normally organized in community-owned companies. Standard private sector
management tools, like accrual accounting, have been used in local government for
decades.
The independence of local government is guaranteed by the constitution and
fiercely defended. There are, as a consequence, also considerable variations between
municipalities on how they are run. The independence of the local authorities has been
somewhat constrained in recent years with central government temporarily imposing
tax freezes and balanced-budget requirements while at the same time saddling them
with an increasing number of unfunded mandates. The Swedish Association of Local
Authorities, among other things, represents the local authorities vis--vis the Central
Government, for example on financing issues.
3.2.2. Decentralization to independent central government agencies

Sweden has a long history of strong independent agencies that are responsible for
the execution of the bulk of the budget. A sign of the importance of agencies in the
Swedish system is that 98 percent of all central government staff work in agencies and
108 There is a long-standing debate whether or not the county councils should be eliminated and its
responsibilities assumed by the municipalities singly or, where needed, in collaboration. Such a reform may
come in the next few years.
only 2 percent (excluding the Foreign Office) in the Government Offices, i.e. the
equivalent of traditional ministries. In the UK the proportion is 75 to 25 percent.
The responsibilities of an agency are specified in an Agency Ordinance. A
minister is not allowed to intervene directly in the operations of an agency or in any
decision within its area of authority made by it. To a degree, but in a subtle way
and behind the scenes, the independence of agencies has, however, become
compromised by an increasing politicization of the appointment of the director
generals of the agencies.
Block appropriations
In order to achieve the objectives set forth in the letter of appropriation an agency
is given one or several block appropriations, i.e. a total expenditure ceiling for the
area(s) that are its responsibility. If the agency is responsible for only one global task,
like for example the Prosecution Services, it receives only one single appropriation.
This allocation it is free to allocate to whatever inputs it deems necessary to carry out
its work. Agencies with a range of responsibilities may receive a number of
appropriations. The County Administrations, an extreme case, may finance their
operations form over twenty different appropriations. Agencies, like the National
Insurance Board which handle a number of entitlement programs, receive one
appropriation for their administrative expenditure and are responsible for execution of
the different entitlement appropriations.
Carry-overs and credit facilities

There are two types of block appropriations: framework appropriations
(ramanslag) and reservation appropriations (reservationsanslag). A framework
appropriation gives the agency the right to carry over unused funds to the following
year up to a limit. The agency has also the right to borrow against next years
appropriation up to a limit. A reservation appropriation only allows carry-overs.
Authority over payments

All payments are executed by the agencies. There is no central treasury or
payments office intervening in the process. In order to minimize the total cost of
liquidity to the government, all payments are, however, made to and from accounts
that are part of the government group held in one of the commercial banks.
Personnel
The agencies are free to hire whomever and however they desire and to set the
salary at the necessary competitive level. The latter right is, however, to some extent
tempered by the formal process of wage negotiations between the unions concerned
and the representative of the Government, the Swedish Agency for Government
Employers. There is of course also the restriction set by total available resources. An
agency will not, in principle, be compensated for cost increases resulting from wages
increases beyond index. Should it, however, choose to reallocate resources from other
types of expenses it is free to do so.
Facilities
Agencies rent their premises and pay market price even if and when the property
is owned by the Government. Government commercial type property is managed by a
public company which is expected to behave like a normal commercial operator.
Right of disposal of own revenue

Some 33 percent of central government consumption, i.e. services production, is
financed through fees, external grants and other types of direct revenue. The right to
levy fees is regulated in a government Fee Ordinance and regulations attached
thereto issued by the Swedish National Financial Management Authority. A general
principle is that fees should cover the full costs associated with the production of
services concerned. Once authorized to levy fees, the agency has full disposal of that
revenue.
3.3. Responsibility and accountability
To balance the large degree of freedom that the agencies have to shape their
operations and decide on the input mix, a number of instruments and processes have
evolved to foster responsible behavior and to uphold accountability. It is worth
pointing out that responsible behavior and accountability in Sweden does not depend
on one single mechanism but is rather the (hoped for) outcome of a number of factors
that together are expected to create responsible behavior.109 In the end, every public
institution is accountable to the public that it is to serve. That accountability can be
ensured by direct or indirect mechanisms. Here they are grouped into mechanisms that
involve the Government, mechanisms that involve the public directly or indirectly, and
mechanisms that involve the elected bodies directly.
3.3.1. vis--vis the Government
Agency ordinance
The Agency Ordinance (Verksfrordningen) establishes the general rights and
responsibilities of agencies.
Agency specific instructions

For every agency there is a specific instruction which establishes the concrete
responsibilities of the agency, its general reporting requirements, its organizational
structure in broad terms, its authority to levy fees and other matters of specific
relevance to the agency.
Letters of appropriation
Yearly letters of appropriation are prepared for single appropriation and/or for a
set of appropriations under the jurisdiction of a specific agency. The letters of
appropriation vary a bit in character depending on the nature of the appropriation, but
when it concerns the operations of an agency they contain sections on:
z objectives;
z results indicators;
z special reporting requirements;
z assignments (tasks) for the coming year; and
z financing of the activities of the agency.
The latter includes:
109 In the discussion of the reforms in New Zealand an important distinction has been made between
responsibility and accountability, the former being the broader concept relating to the spirit and not merely the
letter of the objectives set for an agency.
1. the value of the appropriation(s) for the year;

2. an indication of any planned use of carried-over appropriations;
3. a ceiling on the appropriation credit, i.e. the right to borrow against next
years appropriation;
4. an estimate of fee revenue that will be under the control of the agency;
5. a ceiling on borrowing for investment in fixed assets, a ceiling on credit to
meet short-term cash-flow needs; and
6. any explicitly authorized, special, longer-term or multi-year commitments.
The letters of appropriation establish or confirm the objectives that are set out for
the activities that the agencies carry out and spell out the results indicators that the
agencies are supposed to report on. The objectives defined for the activities have
tended to be rather vague in nature and the results indicators more or less ad hoc. What
has been missing in terms of preciseness has often been made up for by increasing the
number. The problems with objectives and results indicators are discussed further on.
In addition to the more or less formalized results indicators, the agencies may be
requested in the letters of appropriation to report on specific issues that are of interest
to the Government.
Statement of accounts and yearly performance report

Every year an agency is required to produce a yearly performance report including
a statement of accounts. The statement of accounts follows normal private sector
standards. Thus it contains, at a minimum, a profit and loss statement, a balance sheet
and a cash flow statement. The reports otherwise vary a great deal in size and ambition
but generally contain a rather extensive descriptive part. As can be expected after all
institutional self-preservation is a very pervasive force there is often tendency to
depict the situation in rather rosy colors.
Follow-up meeting with the responsible minister

Every year, each head of agency is to have at least one formal meeting with the
responsible minister in order to discuss the development of his or her agency. As there
exists no formal contract between the Minister and the Head of the agency cf. the
situation in some other countries like New Zealand and Denmark it is difficult to
assess the importance of these meetings.
External audits
Agency financial accounts are audited every year by external auditors.
Performance audits are carried at the initiative of the National Audit Office when
deemed important by the latter.110
Evaluations of financial management

Of a slightly different nature but nonetheless important when it comes to
upholding high standards in the financial management of the agencies, have been the
yearly evaluations of the agencies carried out by the Swedish National Financial
110 The National Audit Office was moved in 2003 from being under Ministry of Finance to being under
Parliament. Its modus operandi have, however, remained largely the same with the one exception that they are
now empowered to audit the Government including the consolidated accounts.
Management Authority. The agencies have been rated in a number of areas and then
ranked. The top-ranking agency receives a prestigious award.
3.3.2. vis--vis the public
Consumer choice
As pointed out above, much of the services produced by the public sector have an
identifiable consumer-citizen at the receiving end. Perhaps the most efficient way of
enforcing accountability in these circumstances is to give the consumer a choice as to
from whom he chooses to receive the service.
This latter philosophy has shaped much of the development at the local
government level. At the central level this tendency has not been quite as strong but
there are areas where it applies, for example higher education. At local educational
levels freedom of choice is ensured through the voucher system. This system is
partially introduced at university level when it comes to students wishing to study
abroad. Swedish universities are instead financed on a piece basis, i.e. per student year,
differentiated by type of education.
Application processes
Some services, in particular research, are financed by grants for which the
institutions have to apply and on the use and the results of which they have to report.
The same procedures apply to EU-grants made available to individual public
institutions.
The Parliamentary Ombudsman

Residents in Sweden have recourse to the office of the Parliamentary Ombudsman
in case of malfeasance by a public authority or an individual official. Other
ombudsmen exist but they do not focus primarily on the public sector.
Media
Media serves as a very powerful, albeit often a bit erratic, watch-dog keeping
public institutions and officials on their toes. A very potent instrument in the hand of
the media is the very far-reaching Swedish Freedom of Information Act.
3.3.3. vis--vis elected bodies
Agency annual accounts and performance reports

The agencies annual accounts and performance reports are public documents
which are relatively widely used by the members of the Riksdag. Parliamentarians tend
to specialize on the issues of concern to the committees of which they are members.
Consolidated Government annual accounts

Consolidated Government Annual Accounts have been produced since 1995. The
document is gradually evolving from a mere financial document to a more analytically
oriented one. It was originally produced by the National Financial Management
Authority (ESV) and, before the latter was created, by the National Audit Office. Now
ESV only produces the background material, in particular the financial part, and the
overall responsibility for the document has been taken over by the Government Offices,
concretely the Budget Department with in Ministry of Finance. In order to increase the
political relevance of the document, its tabling in the Riksdag has been moved up to the
spring. Earlier it was tabled too late to have any impact on the deliberations on the
Budget Bill.
3.4. Reassessment and prioritization
One of the core arguments for moving away from traditional input focused
management to performance management was the felt need to reassess and prioritize
government spending. There was a hope that, with the new approach and the new
tools, continuous reassessment and prioritization would become an integral part of the
normal budget process. That has not really happened. The faith put in the three-year
in-depth reviews cum budget proposals to be prepared by the agencies quickly proved
unfounded. It has happened that agencies have proposed their own elimination, but
those are rare exceptions. Normally institutional self-preservation or expansionist
ambitions tended to take over. The three-year in-depth reviews/budget proposals have
now been abandoned.
The traditional way of initiating structural changes has been and continues to be
the appointment of a committee giving it, sometimes, very clear terms-of-references
to the point of more or less prescribing its conclusions. When the Government is less
sure what it wants to do, the terms-of-reference are more open-ended. Committees may
or may not have parliamentary representation and draw more or less on external
academic and other expertise. Over 100 Swedish Government Official Reports
produced by the committees are published every year. Committees may deal with just
about any issue of public concern, and it does not necessarily have to be budget-
related.
Changes budget-related or not may also be initiated by appointing internal
working groups within the Government Offices. Their deliberations and proposals are,
when they are expected to translate into legislative proposals, normally presented in
the different ministries publication series.
Committee reports and the reports published in the publication series of the
different ministries are all formally referred to a large number of selected bodies for
consideration and comment. These bodies may be public authorities, local government
authorities, NGOs, advocacy groups or voluntary associations.
Impetus to change may also come from the semi-autonomous research institutes
that have been set up for a number of areas; for example the Swedish Institute for
Transport and Communications Analysis and the Institute for Labour Market Policy
Evaluation. An Expert Group for the Study of Public Economics, attached to the
Ministry of Finance but fiercely defending its independence, financed a large number
of critical studies on a wide range of topics.
3.5. Tools
Accrual accounting
The reforms carried out during the last thirty years have moved Swedish public
management in the direction of that of the private sector. The legal and regulatory
framework is still important, but agencies and their managers have been given greater
authority to manage their operations and their finances. In regard to the latter it is no
longer merely a question of executing the budget. Agencies are financed from a
number of sources and manage a wide range of programs and projects.
As complexity increased it became obvious that traditional budgetary accounting
was inadequate and the whole central government, without much ado, moved to
standard accruals in 1992/93.
Agencies are free to choose their own accounting system and their own chart of
accounts provided that they can report in the format of the central chart of accounts
used for the consolidated government accounts.
Central FIMS
A web-based central Financial Information Management System is now being
developed. The system, when its complete parts are up and running, will contain
budget, accounting, and results information.
Balanced scorecard, benchmarking, et. al.

The move to performance management has as much been a change of how
agencies manage their own operations as a question of the relationship between the
Government and the agencies. The responsibility for the execution of agency budget is
decentralized to the different departments, and agencies often develop results
indicators that are only used for internal management purposes. Some agencies have
chosen the balanced scorecard approach to packaging management information.
Benchmarking has spread, particularly in local governments where there are
plenty of nearby experiences to benchmark against. A central database to facilitate and
stimulate benchmarking, as well as independent research, is being developed. In
central government there are fewer opportunities for this.
4. Implementation
As already pointed out, the move towards performance management has been a
very drawn out process involving a number of successive initiatives. It is therefore
difficult to provide a simple and single answer to what has been driving it and where
the resistance has come from.
However, one can say what it has not been. It has not been a reform process driven
by politicians. There has been no Tony Blair seeing the political benefits to be drawn
from public sector reform. It does not mean that the politicians have always been on
the sideline. They have been supportive but the original initiatives have not come from
the political sphere but rather from within the administration. And politicians have
generally not, perhaps with the exception of the move to top-down budgeting, invested
much political capital in the reforms. Nor have the reforms, again with the exception of
the move to top-down budgeting, been the result of a crisis. The Swedish reform
process is thus very different from that of New Zealands.
When attempts at writing the history of the reforms are being made the
definitive one is still to be written it is customary to refer to the different committee
reports and to other initiatives like VESTA. International influences like program
budgeting in the US in the sixties and new public management in general and the
reforms in New Zealand in particular are also highlighted. Surprisingly little
influence is deemed to have come from developments in local Swedish local
governments despite their geographic and cultural proximity.111
The conceptual champions of change and external influence are important. But the
question is if it is not the thirst for authority by the institutions that actually do the
job and their acutely felt management needs that have been the most important force
111 For some reason, central governments seem to turn to each other for inspiration rather than to their
brethren in lower levels of government. Perhaps it is precisely because of this institutional pecking order.
motrice. It is worth noting that local governments have been many years ahead of
central government in their reforms and it probably has to do with the fact that they
are primarily service providers and as such in close contact with their clients.
At the central government level, the reforms have also started from below. The
move to accruals, for example, was driven by agencies most acutely suffering from the
inadequacies of traditional public sector accounting. Agencies within the central
government have a long history of constitutionally protected independence and they
enjoy having the authority. Perhaps as a reaction to this, the more recent reforms in this
field have actually been presented as a way of bringing the politician back in the
drivers seat rather than as a means to letting the managers manage. Agencies do not
need to be empowered; politicians do.
The one vociferous opposition to the notion of performance management has also
come from a group of more or less leftist-leaning academics who see performance
management as a technocratic exercise, depriving the politicians of their right to
govern. The fact that some of the management techniques and tools come from the
private sector and from abroad is also seen as suspicious and has made some
politicians wonder whether it is all a Trojan horse. The enthusiasm expressed by some
of the opposition parties has not helped allay the fears of Social Democrat
parliamentarians either. Given their rather lukewarm interest it is actually surprising
how far the reforms have got.
On the other hand no one, particularly not the academics who staunchly defend
that it is not their role, has been able to present an alternative to performance
management. It is like with post-soviet Russia, people gripe about it but very few
would want to return to the old ways.
Thus, while there has not been any strong political leadership there has not been
any active resistance either. The biggest brake on reform has rather been inertia
political and bureaucratic. Reforms have thus had to be introduced one step at a time.
The boon: easier acceptance. The drawback: a lack of consistency. When accrual
accounting was introduced, for example, its architects initially also wanted to introduce
accrual budgeting but decided against it because they feared that it would be too big a
bite for the Government Offices.
Another reason for why the reforms have tended to be piecemeal and gradual is
that the agents and managers of change have been outside the Government Offices at
least during most of the 1990s. Of particular importance during that period were the
National Audit Office and eventually the National Financial Management Authority
(ESV). The Swedish Agency for Public Management also played a role. Being at arms
length from the Government Offices and with a direct mandate to foster the
development of financial management at agency level, they were very well placed to
initiate and implement reforms that were largely limited to the agency sphere. Their
authority to drive reform processes that had a direct impact on the work of the
Government Offices was, however, much more limited; vide accrual budgeting. The
weakest link in the performance management chain has thus become the Government
Office which in turn has left the Riksdag in limbo. Despite the latters oft-expressed
desire to receive and factor in performance information in its decisions, it has been
given little chance to do so because of the Government Offices failure to deliver it in an
intelligible format.
In the Government Office the institutional responsibility for public management
reforms has shifted around. During the 1980s the Ministry for Public Administration
(now defunct) was the driving force. A remnant part of that ministry, responsible for
public reform in general, has had an itinerant life. For a while it was located in the
Ministry of Finance; then it was moved to Ministry of Justice. At the end of the 1990s,
the VESTA project based in the Budget Department took the lead and, in close
collaboration with ESV, drove the process for a while but not without opposition.
The ambition of VESTAs was to extend the performance management chain all the
way to the Riksdag and to tie up a number of loose ends left by the piecemeal reform
process. Some of the proposed reforms are moving ahead true to the piecemeal
tradition while others are deferred, mostly for fear112 of loosing control of cash.
Changes in the management framework have always been backed up by training.
ESV offers courses directed at the agencies and the Government Offices have internal
training units responsible for offering the needed training. When accrual accounting
was introduced, agencies sometimes had to recruit new blood from the private sector
but that demand was easily met.
5. Incentives
The attractiveness of performance management to agencies is the greater authority
accorded them. No special, direct financial incentives have been linked to the reforms
as such, but the greater freedom to set wage levels has had the effect of
narrowing/eliminating the wage gap between the public and the private sector.
The incentives to the staff of the Government Offices are less clear-cut. One
possible benefit could be work satisfaction. Rather than simply applying norms and
rules defined by others, all officials need to get involved in the substance which is more
intellectually stimulating. In the short term, a common, very human and difficult-to-
avoid reaction to any change one certainly seen in during the long course of reforms
in Sweden is that it will increase the workload. The consequences of that type of
resistance very much has depended on the leadership exercised by the ministers and by
the other high officials of the Government Offices. When taking a clear stance, the
political leadership of the organization has been able to count on the solidarity of the
organization; when not, things can quickly get bogged down.
6. Results
Most observers would probably agree that the Swedish version of public
performance management has led to greater efficiency and to greater responsiveness to
its citizen-clients.113 To statistically prove it is, however, next to impossible for want of
any counterfactual information. The fact that the reforms have been carried out over
such a long period also makes it very difficult to attribute efficiency improvements to
particular components of the reform. Attempts at calculating cost effectiveness at a
disaggregate level have, unfortunately, been abandoned. The global estimate in the
national accounts is that productivity in the public sector is growing by, on an average,
1 percent a year.
It is less clear whether the reforms have led to greater effectiveness. It has proven
difficult to develop mechanisms for systematically reassessing what the government is
112Unfounded in this authors opinion.

113This does not mean that all agencies are efficient. Greater authority does not automatically lead to good
leadership and other factors, such as shortsighted unions more concerned with the immediate well-being of its
members than with the services they are to provide to the public, may constrain the choices open to
management.
doing and for reprioritizing when society changes. The latter failure is, however, not, in
this authors opinion, the result of a failure of the basic conceptual framework. It is not
a design flaw.
It is rather a political failure; a failure to recognize that a more systematic use of
information on outputs and outcome on costs and effects, an unbiased search for
alternatives, and readiness to accept failures can be turned into good and vote-winning
politics.
7. Outstanding issues
The Swedish public sector has come a long way on its way from public
administration to public management. Nobody would, however, claim that the present
model is the final, definite one. There are a number obvious outstanding issues and
history will not stop even if and when the latter are resolved.
Performance budgeting
Most of the reforms, so far, have concerned the internal operation of agencies and the
relationship between the Government and the agencies. The relationship between parliament
and the Government in many ways is a vestige of pre-performance management days.
The Budget
The Central Government budget is still based on different principles and is
presented in a format and with a coverage which is different from that of the accounts.
A process is, however, on its way to change that. A shadow budget compatible with the
accounts will be presented in next years budget bill and eventually the old type
budget will be phased out.
The budget constraints

The budget constraints are also presently based on other principles than those of
the accounts and can thus not be followed directly in the normal accounts. A partial
reform is underway but will not fully solve the problem.
Programmatic structure
A three-tier programmatic structure has been elaborated and will be used for
performance follow-up. The budget is, however, still organized in expenditure areas
and appropriations. This duality forecloses one of the fundamental objectives of
performance management; the possibility of contrasting costs and results. Besides it is
untenable for practical reasons and single structure will have to be agreed upon.
Improved objectives and indicators

Arriving at clear objectives and good results indicators has proven one of the
Achilles heels of the Swedish performance management model; and the difficulties
experienced have been used as arguments against the approach in general.
Of course it should be acknowledged that in many areas, particularly for broad
policy areas, it is genuinely difficult to define good and clear objectives. Furthermore
there may very well be more than one objective for an area. It should also be
recognized that indicators can be dangerous if not handled with care. There is always a
risk of agencies focusing on what is measurable rather than on the broader result.
Many of the difficulties and much of the confusion surrounding objectives and
indicators, however, stem from a lack of coherent and commonly agreed conceptual
framework. The lack of a design yardstick has led to a proliferation of indicators

without much logical consistency and even less permanence. A deepening of
performance management and budgeting in Sweden will require much more time and
effort to be spent in this area without losing sight of the fact that objectives and
indicators are only the starting point for the important part the analysis.
Analytical capacity in the Government Offices

Like tango, performance management requires two partners equal, but one leads
and one follows. In the performance management dance, the Government Offices have
proven to be the weak partner, to the point that the agencies complain about it.114 They
enjoy the authority they have been given, but at the same time want guidance and a
strong partner to dialogue with. This they often feel they have not. They produce a lot
of information on their activities and on their results but often feel that they do not get
any response.
The reason is often a lack of analytical capacity at the Government Offices, in
quantitative as well as in qualitative terms. One study suggested that if the
Government Offices are to be able to serve as a strong owner vis--vis the agencies, it
probably would need to grow by one third, much of it in the form a staff with a
research profile.
Coordination
Performance management in Sweden has been based on a successively more
definite programmatic structure with programs sometimes spanning ministries and
agencies. At the same time agencies are the executive agents and need to be given
guidance as such. The internal processes within the Government Offices have not really
fully adapted to and developed the preparation and decision-making processes that are
necessary to handle this matrix.
Agency coordination also leaves a lot to be desired.
Global priorities
The lack of analytical capacity within the Government Offices and the still limited
capacity in semi-attached think-tanks has hampered the much-emphasized need for
reassessment and reprioritization. It could, possibly, be argued that a part of the
problem is that politicians are not always too terribly keen on independent evaluations.
In specific situations that may be true, but generally speaking it is not a tenable
argument.
8. Lessons
z Delegation of authority is easy. It is more difficult to put in place the routines

and mechanism that will ensure accountability and responsibility.
z Performance management is more challenging than rule definition and rule
adherence. Officials must be willing and capable of taking on those challenges.
z Willingness to take on responsibility is as important as the willingness to
relinquish immediate control.
114 ESV at the end of the nineties carried out an extensive survey to find out how the agencies looked upon
performance management and this was one of the conclusions.

z Other types of skills and other types of managers are needed; entrepreneurs
vs. administrators.
z A programmatic approach increases the need for coordination between
ministries and agencies.
z Performance management presumes professionalism. There is no room for
cronyism and patronage. Promotion should be made on the basis of ability
rather than seniority.
z Performance management implies a change of management culture from
hierarchical to participatory.
z Mobility within the public sector and between the private and the public
sector can facilitate change.
z Strong owners are needed to balance empowered agencies.
z Good analytical capacity in the ministries is essential.
z Tools important are important: modern accounting, customer relations
measures, etc..
z Clear objectives and results indicators are important but surprisingly difficult
to arrive at and easily become a stumbling block.
z Mechanisms for reassessing and reprioritizing are essential but do not
necessarily have to be part of the normal budget process.
z Other instruments have a complementary role to play: performance audit,
independent evaluations, academic research etc..
z It is desirable that reforms are carried out within a consistent framework but it
is not always possible.
z Strong political leadership makes the implementation of the reforms much
easier but it is not impossible without it; it just takes longer.
CHAPTER 16
Performance Budgeting in the United States
by
1. Outline
z Context
z GPRA Strategic Planning Requirements
z GPRA Performance Planning Requirements
z Key Implementing Guidance
z Lessons and Observations
z Future Directions
z Performance Budgeting in a Federal system: Collaboration
2. Context
z U.S. has a long history of trying to improve performance and assess what public
monies buy
- Planning-Programming-Budgeting System (1960s)

- Zero-Based Budgeting (1970s)
- Management by Objectives (1980s)
- National Performance Review (1990s)
- Government Performance and Results Act (1990s)
- Program Assessment and Rating Tool (PART) (2000s)
z Attempts have varied between using the budget process and directly through
management of programs
3. GPRA
z GPRA represents another step in the historical search to grasp whether agencies
are fulfilling their mission, and whether programs are effective
- a natural extension of legislative oversight
z Over-all objective: Linking resources to results.

z The GPRA statute defines two complementary performance activities, defined by
the short and long term
- Strategic plans, covering five year horizons

- Performance plans, covering annual goals to reach the strategic targets
4. GPRA Strategic Plan Requirements

GPRA [5 USC 306] requires each agency to produce a strategic plan by 9/30/97 which
includes:
- a comprehensive mission statement;

- general goals and objectives, including outcome-related goals and objectives,
for major agency functions/operations;
- a description of how goals and objectives are to be achieved (operational
processes, skills, technology, human, capital, information and other resources
required);
- description of how performance goals of performance plan relate to strategic
plan;
- identification of key factors external to the agency and beyond its control that
could significantly affect goal achievement;
- a description of program evaluations used in establishing/revising general
goals and objectives, and a schedule for future program evaluations; and,
- covers five years forward, to be revised at least every 3 years.
Additionally, agencies must consult with Congress and entities potentially affected by
or interested in the plan, and the plan must be drafted by Federal employees.
5. GPRA Performance Plans Requirements

GPRA [31 USC 1115(a)-(e)] requires the Director of OMB to require agencies to prepare
annual performance plans covering every budget activity for each agency. Plans must:
- establish performance goals and performance levels to be achieved for each

activity;
- express goals in objective, quantifiable, measurable form (unless OMB
authorizes otherwise);
- briefly describes operational processes, skills, technology, human, capital,
information, and other resources necessary to attain goals;
- establish performance indicators to be used to measure and assess outputs,
service levels, and outcomes for each program activity;
- describe the means to be used to verify and validate measured values.
Agencies are allowed to aggregate, disaggregate or consolidate program activities in

the plan, so long as it does not minimize major agency programs.
Again, Federal employees must draft these plans.
6. Key Implementing Guidance

z Goals should be quantifiable
z Outcomes over outputs
z Outcomes over process
z Input goals not acceptable
z Measures should be ones program managers use regularly, find useful, and are
part of normal management information systems
z GPRA should incorporate previous performance-based reforms
Chapter 16. Performance Budgeting in the United States 259
7. Lessons and Observations

z Clear model of program important
z Choosing the right goal and outcome measures
z Outcomes dont supplant other measures
z GPRA creates demand-pull for better data; will change data systems
z GPRA as culture change, rather than merely a paper exercise: effects will vary.
z If central budget institutions lead performance budgeting efforts, which is natural,
then a prerequisite of success is redefining the role of the central budget institution
from command-and control to collaboration.
z Shorter the time horizon, more process oriented the measure
z Performance budgeting is iterative
8. Future Directions
z Program evaluation must be integrated into GPRA performance planning: deep
look at why programs are or are not working.
z Auditing of data systems
z Auditing of performance reports
z Inclusion of performance data in annual financial statements
z Integration of performance and financial systems
9. Program Assessment Rating Tool

z Key assumption of GPRA was that agencies would themselves use performance
information to improve program management, decisions
- Currently little evidence this has happened
- PART developed in 2002 with advice from Presidents Management Council
z Intended to integrate performance and management
- Bridge gap between performance data and what to do with data

- Undertaken by the budget office, though with close collaboration with
agencies
10. Coverage of the PART

Each PART has four sections with questions designed to elicit specific information
z Program Purpose & Design to assess whether the programs purpose and design
are clear and sound.
z Strategic Planning to assess validity of long-term and annual measures and targets
for the program
z Program Management to rate agency management of the program, including
financial oversight and program improvement efforts
z Program Results/Accountability to rate program performance on measures and
targets reviewed in the strategic planning section and through other evaluations
Part III
Fiscal Transparency
CHAPTER 17
Coverage of Public Finance in Korea
by
Dong suk Oak, Incheon University
1. Introduction
The objectives of public expenditure management are generally viewed as
maintaining aggregate fiscal discipline, allocating resources in accordance with
government priorities, and promoting the efficient delivery of services. The fiscal
targets and indicators used most commonly in maintaining aggregate fiscal discipline
are the fiscal deficit and fiscal debt, regarding which many countries have laws and
rules to restrict fiscal policy.115 The establishment of explicit fiscal targets allows the
government to state clearly its fiscal policy and enable the legislature and the public to
monitor the implementation of government policy, and ultimately make the
government politically as well as financially accountable.
Fiscal deficit is defined as the difference between fiscal revenue and fiscal
expenditure. In order to measure the fiscal deficit, the government needs, among other
things, to determine the coverage of public finance in a sufficient and appropriate
manner. Determining the coverage of public finance implies that the government
should decide which public institutions, funds, and accounts are to be included in
measuring the fiscal deficit. The coverage of public finance is also instrumental in
measuring fiscal debt. Which public institutions, funds and accounts are to be included
in measuring the public debt?
Since every country has its own historical tradition and institutional system for
public finance, countries differ greatly in their governmental and fiscal coverage. While
the statistics on the coverage must reflect the institutions and practices of each country,
a common system of definitions and classifications is helpful to users because it allows
comparisons with other countries. It is desirable for the fiscal statistics to follow the
conventions and classifications in constructing the national accounts for the entire
economy, as laid out in the United Nations A System of National Accounts 1993 (SNA).
Following the fiscal coverage of the SNA, the European Union publishes more
specified criteria, in the Manual on Government Deficit and Debt in order to obtain
reliable and comparable statistics for evaluating fiscal convergence concerning both the
fiscal deficit-to-GDP and the debt-to-GDP ratios.
In the SNA five mutually exclusive institutional sectors are identified. They are the
nonfinancial corporations sector, the financial corporations sector, the general
government sector, the nonprofit institutions serving households sector, and the
household sector. The IMFs Manual on Government Finance Statistics (GFS) harmonized
with sectorization of the SNA, defines the coverage of the statistics to be a country's
115 Fiscal targets and indicators should cover three areas: current fiscal position (e.g., fiscal deficit), fiscal
sustainability (e.g., debt-, tax-, or expenditure-to-GDP ratios), and vulnerability (e.g., analysis of the
composition of the foreign debt).
264 Part III. Fiscal Transparency
general government sector, which is defined on the basis of institutional units. The
coverage can also be extended to the public sector by including the fiscal activities of
public corporations. The public sector is defined as consisting of the general
government sector plus government-controlled market entities, known as public
corporations.
In spite of the SNA's definition of the general government sector, the coverage of
government expenditure, being the baseline of government deficit and government
debt, varies in a number of countries. Especially many developing countries, having a
relatively short history in operating the modern public management system and rare
opportunities to share the experiences of other countries, have problems in
transparency, responsibility, and efficiency because their coverage of government and
public finance is not defined comprehensively and systematically.116
In the developed countries the coverage of government expenditure is not all
consistently the same through the stages of budget preparation, parliamentary votes,
funds release, accounts and statistics. A significant share of public expenditure is
usually managed through special procedures. The objectives of the special
arrangements include: avoiding excessive negotiation during the budget preparation
process, resolving difficulties in the legislative debates and scrutiny, and providing
greater flexibility in the use of funds by the executive.
This study draws lessons for the Korean government from the comparative study
of Korea, Japan, and the UK. The paper is structured as follows. Section 2 discusses
some key factors in determining the general government sector in the SNA and EU
Manual, and summarizes several concepts on the coverage of the public finance
according to the management process of public expenditures. Section 3 explains the
Korean government's coverage of the public finance, while Sections 4 and 5 provide a
broad overview on the UK and the Japanese governments coverage of the public
finance. Section 6 concludes with policy implications for reforming public expenditure
management in Korea.
2. Several concepts on the coverage of public finance

2.1. Determining general government sector
The SNA, a cornerstone for the other international manuals on statistics, divides
the economy into five sectors in addition to the rest of the world: (1) non-financial
corporations sector; (2) financial corporations sector; (3) general government sector; (4)
non-profit institutions serving households sector; and (5) households sector. The public
sector includes the general government sector plus all public non-financial and
financial corporations. The total resident economy except the public sector is the
private sector, so there should not be no gray area between the public and private
sectors.
The general government sector and the public sector are the basis for the
government finance and public finance statistics. In deciding if an entity is to be
included in the general government sector, the steps to be followed could be
summarized as three-staged. These are provided by the EU Manual, describing more
116 The two terms, public finance and government finance are generally used in the literature without
distinction. Fiscal deficit of the general government sector and the public sector, however, are defined as
government deficit and public deficit, respectively.
Chapter 17. Coverage of Public Finance in Korea 265
specifically the concepts and criteria of the SNA for the Excessive Deficit Procedure
required in the Article 104 of the Maastricht Treaty.
The first step to be taken is to ask whether the entity is an institutional unit.
Sectorization decisions in the SNA must be taken at the level of institutional units,
defined in the SNA as units having the autonomy of decision and a complete set of
accounts. Autonomy of decision means that it must be able to own goods or assets in
its own right; be entitled to make economic decisions and carry out economic activities
for which it is held directly responsible and accountable by law; and be able to incur
liabilities on its own behalf, to take on other obligations or further commitments and to
enter into contracts. In order to keep a complete set of accounts, a unit must keep
accounting records covering all its economic and financial transactions carried out
during the accounting period, as well as a balance sheet of assets and liabilities.
The second step is to ask whether the institutional unit is private or public. A
public unit is one controlled by the general government, whereas a private unit is one
not controlled by the general government. Control, defined as the ability to determine
general policy, is an essential criteria for sectorization between the public and private
sectors. The general government of a country consists of the public authorities, their
agencies and instrumentalities, which are established through political processes and
exercise a monopoly of compulsory power. The form that administration or accounting
for a entity takes whether budgetary, extra-budgetary, special account, etc.
should not affect its inclusion for the general government sector. When a public unit,
by virtue of special legislation, is recognized as an independent legal entity, it can be
considered as an institutional unit, to be classified in the general government sector,
and if it is not a public corporation.
The third step is to ask whether the public institutional unit is market or non-
market. When the principal function of the public institutional unit is to supply certain
public goods and services and fulfill certain public purposes not for commercial or
financial reasons, or, if of a commercial or financial nature, not on a major basis or not
primarily for profit, this unit is to be classified in the general government sector. In the
SNA, the distinction between market (i.e., for commercial reasons or for profit) and
non-market producers depends on whether or not prices charged are economically
significant. A price is said to be not economically significant when it has little or no
influence on how much the producer is prepared to supply and is expected to have
only a marginal influence on the quantities demanded. In practice, the public
institution whose production costs are not covered by more than 50 percent through
sales is classified as a non-market producer and general government.
The general government sector, which is determined by the three steps, normally
includes the ministries, departmental agencies, other decentralized public bodies, and
local authorities. The principal function of the general government sector is to carry out
public policy through the production of non-market services primarily for collective
consumption and the transfer of income, financed mainly by compulsory levies on
units in other sectors. The public institutions in the general government sector are
being called in different terminology in different countries, making international
comparative research difficult. Institutions other than the ministerial agencies are
sometimes named as authorities, autonomous bodies or organisms, quangos, non-
departmental public bodies, extra-governmental organizations, subsidiary
organizations, etc.
Public policy may be also carried out by public corporations outside the general
government sector, and these are sometimes called as quasi-fiscal activities. To the
extent that the quasi-fiscal activities are funded by grants from the general government
sector, those can be included in the coverage of the public finance. The total receipts
and expenditures of public corporations should not be treated equally to those of the
governmental institutions, because they collect sales proceeds only by incurring
corresponding production costs.
For the purposes of control and accountability, the general government sector is an
appropriate coverage for the public resources and finance. The general government
sector includes all the ministerial departments, departmental agencies, the public and
private law non-market bodies which are directly and indirectly controlled by the
ministers. The finances of those are ultimately the responsibility of the government, so
those are to be answerable and accountable to the legislature and managed and
controlled by the ministers. Public corporations, not included in the general
government but in the public sector, are usually private law bodies whose owners have
limited liabilities and can control them only with voting shares. However, some
considerations should be given to public corporation established by public law, since
the law provides implicit guarantee and financing for it seems to be the responsibility
of the government.
2.2. Coverage of public finance at management stages
The expenditure management process of various institutions included in the

general government sector has four major stages: planning, approval, execution, and
reporting. The coverage of the institutions in each stage of the process may be greatly
different because it is necessary to allow special management arrangements to some
special institutions, funds, or accounts on historical, legal, political, and administrative
grounds. Owing to the special arrangements, the coverage of government expenditure
is not all consistently the same through the stages of budget preparation, parliamentary
votes, funds release, accounts, and statistics.
Figure 17-1. Stages of expenditure management
Planning
(Budget Preparation)
Reporting Approval
(Accounts and Statistics) (Parliamentary Votes)
Execution
(Funds Release)
In order to use the resources of the general government sector in a responsible,

efficient and effective way, the government needs to define the budget as a
comprehensive plan for the operations of the sector. Different countries have different
budgeting coverage, each with its own specific features consistent with the overall
administrative culture. Not many countries seem to require budgets to cover all the
public institutions of the general government sector defined in the SNA. 117 For the
coverage of the budget to be comprehensive, the budget should include all revenues
and expenditures of the general government regardless of the arrangements for
separately managing some particular programs, the legal provisions for authorizing
expenditures, and the financing source. If the budget excludes major expenditures of
the general government sector, there can be no assurance that scarce resources are
allocated to priority programs and that legal control and public accountability are
properly enforced.
Different actors are involved in budget preparation: the Central Budgeting Bureau
(CBB) of the Administration, budget departments of line ministries, and, within line
ministries, budget departments of subordinate spending agencies and decentralized
public bodies. The CBB reviews budget requests from line ministries, departmental
agencies, and decentralized public bodies, and is responsible and accountable for
drafting budget bill submitted to the parliament. Depending on the country, the
coverage of budget bill, which the CBB is involved in preparing, may not fit the general
government sector. Special arrangements on the budget process may allow line
ministries to have all the power and responsibilities on some budgetary resources
without reviews by the CBB, which weakens accountability. Some decentralized public
bodies may not be under the budgetary control of the CBB and be just under that of the
line ministries. In this study, the budget that is controlled and prepared by the CBB is
denoted as Budgetctr, and the Budget, controlled by the government in general, includes
the Budgetctr plus the other budgets controlled only by the line ministries.
The Budgetctr is submitted to the legislature for approval some months before the
start of the fiscal year, to allow budgetary debates to be completed before the start of
the fiscal year. Many countries have adopted procedural rules to regulate and limit
legislative debates on the Budgetctr. These usually pertain to (i) the coverage of voting
on the Budgetctr; and (ii) the legislature's powers to amend the budget.
The Budgetctr, submitted to the legislature by the CBB, has two parts; Votes and
Non-Votes. Votes are the budgetary resources which should be approved by legislature
for its execution, and Non-Votes are the ones just for scrutiny by the legislature. The
de-facto differences between Votes and Non-Votes lie in the legal powers of the
legislature to propose amendments to budgetary items.
In addition to the annual voting process, the legislature may also have access to
administrative information through regular departmental briefings and expenditure
reports from the Comptroller and the Auditor General. The resource included in the
Budget other than the Budgetctr will be scrutinized by the legislature through briefings
and reports. For example, some decentralized public bodies do not require approval by
the legislature for their budget, instead report budget executions to the legislature on a
regular or ad-hoc basis.
Budget execution needs the release of budgetary resources from funds and
accounts. Spending institutions must be provided with the funds necessary to
implement the budget in a timely manner, and the cost of government borrowing must
117 Since public corporations in general do not use tax or other compulsory levies, they are not included in
the budget. Only grants to them will be included in the budget.

be minimized. In most countries, the government has a number of funds and accounts
to administer public money irrespective of the institutional unit. Every country has a
consolidated fund or account, the main fund into which all receipts are deposited and
from which all payments are made, and some earmarked special funds. The central
bank is, in most countries, the main cashier of the government, so the consolidated
fund or account is held at the central bank. Special or earmarked funds and accounts
may be held in commercial banks, whether the funds and accounts are automatically
swept at the end of each day or not.
To minimize borrowing costs or maximize interest-bearing deposits, operating
cash balances should be kept to a minimum. Idle balances of special earmarked funds
or accounts increase the borrowing needs of the government, which must borrow to
finance the payments of some funds, even if other funds have excess cash. Whatever
the institutional arrangements, the centralization of funds and accounts is required and
should cover as many of the government funds and accounts used for payment
transactions. In this paper, funds and accounts held at the central bank will be denoted
as FA1, and funds and accounts held at commercial banks, as FA2. The coverage of
FA1 and FA2 will vary greatly according to the country.
The last stage of the expenditure management will be government accounting and
financial reporting. Government finance statistics can be distinguished from two other
data systems. The CBB or other ministerial departments will publish the government
finance statistics from the government budgets and accounts, maintained by operating
units for purposes of control and accountability. In the national accounts, data for the
government are recast in a form common to all sectors so as to measure and aggregate
their production, income, consumption, capital accumulation, and finance, and
delineate inter-sectoral transactions. Even if the office responsible for preparing the
national accounts will vary according to the country, the statistical guidelines and
concepts of the SNA should be observed. In this paper the de-facto coverage of
government finance statistics published by the CBB or other ministerial department
will be denoted as FS (Financial Statistics), and the coverage of national accounts NAS
(National Accounts Statistics).
3. The coverage of public finance in Korea

The coverage of the consolidated government finance in Korea is traditionally
recognized according to the criteria of accounting and fund units such as the General
Account, Special Accounts, and Funds. The General Account is the main consolidated
account with the Bank of Korea. Most of the central government's expenditure is
financed from this account, and most taxes and other receipts deposited there. Special
Accounts, whose total number is 22, and Funds, whose total number is 55, are all
created for earmarking to some special programs and bypassing traditional
appropriation management rules, such as rules for transfers between line items
(virements) and the cancellation of appropriations at the end of the fiscal year.
The coverage of the government accounting and fund units, however, do not
correspond with that of the general government sector in the SNA, being defined on
the basis of institutional units not of the administrative, legal or accounting types.
According to the SNA, the central government of the general government sector in the
Korean resident economy should include all the decentralized public bodies as well as
ministerial departments and departmental agencies. Decentralized public bodies, most
of which are public law administrations, are controlled by the ministerial departments
and departmental agencies. They have their own legal personality, while ministerial
departments and departmental agencies have no separate legal identities from the state
and their staff are comprised of civil servants. They are equivalent to NDPBs of the UK.
The resources of which the decentralized public bodies have control are not
reflected in the government accounts except government's grants to them. The
coverage of the General Account, Special Accounts and Funds do not reach all the
expenditures of the decentralized public bodies. The coverage of the General Account
and Special Accounts is approximately equal to the coverage of ministerial
departments and departmental agencies. However, the 55 Funds do not cover all the
decentralized public bodies.
Some NDPBs are managing government Funds, and some are not. In the case of
NDPBs managing Funds, their expenditures funded from the Funds can be reflected in
the government accounts. Since the expenditures of NDPBs have never been published
in Korea, it is impossible to identify the share of the resources of the decentralized
public bodies (NDPBs in the Figure 17-2) covered in the resources of Funds. All the
resources of the NDPBs' own accounts, whose amount is usually alleged as equivalent
to that of the General Account, are excluded in the coverage of government accounts.
Figure 17-2. Public expenditures in Korea
NDPBs Government departments and NDPBs

(Managing Funds) departmental agencies (Not Managing Funds)
General Account
Fund Own
Account
Own Special Special Special

Account Account Account Account
While the budgetary resources in the coverage of the General Account, Special
Accounts and some Funds are under the control of the CBB (the Ministry of Planning
and Budget), the resources of the NDPBs' own accounts are outside the control of the
CBB and approved just by line ministries. In other words, the budgetary resources of
the general government sector are all controlled by the government, but some are not
controlled by CBB. That means the Budget does not equal to the Budgetctr. Additionally,
some funds are not controlled by the CBB, so are not included in the Budgetctr. Those 10
Funds are related to financial intermediation activities, and are managed by the
Ministry of Finance and Economy. They are not legally included in the Votes. Their
operational plans are submitted and scrutinized by the National Assembly, but their
approval is not legally required for execution.
Most of the Funds are held at the Bank of Korea, like the General Account and
Special Accounts. While it is legally binding that the Bank of Korea holds General and
Special Accounts, Funds' account can be held either at the Bank of Korea or at
commercial banks. Some Funds and NDPBs' own accounts are all held in the
commercial bank owing to banking services.
Since the coverage of public finance has long been recognized as the General
Account, Special Accounts and Funds, both the Government Finance Statistics
published by the CBB and the National Account Statistics published by the Bank of
Korea reports only the Budgetctr. The coverage of the public finance according to the
management stages can be shown as follows.
Table 17-1. The coverage of public finance in Korea
General Special Local

Funds NDPBs
Account Accounts Authorities
Budget
Budgetctr
Votes
non-Votes - -
FA1 - -
FA2 - -
NAS1)
PFS
Note: 1) Only a few NDPBs are included in the National Accouns, but those can be neglected.
4. The coverage of public finance in the UK

In the UK, decisions on which sector an actual or proposed body should be
classified in are made by the Public Sector Classification Committee (PSCC), which is a
small committee consisting of about 10 staff members and external experts within the
Office for National Statistics (ONS). The PSCC's purpose is to help formulate decisions
about the sector classification of bodies/institutions. Straightforward classification

decisions may be taken by the HM Treasury on the basis of agreed classification rules
between the Treasury and the ONS. Decisions must be consistent with the principles of
the international statistical manuals, i.e, ESA (European System of Accounts) and SNA,
and supporting manuals and case laws. Sector Classification for the National Accounts was
published in 1998, listing all the institutional units allocated to each sector including
the general government sector and public non-financial corporations.
The UK government sets targets for themselves based on a concept of national
accounts. Fiscal policy targets are based on two strict rules, namely the Golden Rule
and the Sustainable Investment Rule. 118 The government has chosen to present the
government accounts using principles from the national accounts and to control public
expenditure according to the rules based on national accounts. The political debate on
fiscal policy focuses on the statistics based on the national accounts principle.
Table 17-2. Key fiscal indicators in the UK

(unit: million)
1997 1998 1999 2000 2001
A. Surplus on Current Budget (B-E) -11,777 8,260 15,108 20,252 17,365
B. Gross Saving plus Capital Taxes (C-D) 615 20,846 27,714 33,183 30,754
C. Current Receipts 305,959 333,968 349,872 375,797 391,558
D. Current Expenditure 305,344 313,122 322,158 342,614 360,804
E. Depreciation 12,392 12,586 12,606 12,931 13,389
F. Net Investment 4,664 5,876 5,136 4,836 8,481
G. Total Managed Expenditure (D+E+F) 322,400 331,584 339,900 360,381 382,674
H. Public Sector Net Borrowing (F-A) 16,441 -2,384 -9,972 -15,416 -8,884
I. Net Debt ( billion) 357.6 353.8 352.7 317.0 318.2

Source: ONS, United Kingdom National Accounts: The Blue Book 2002, p.291,
The Blue Book, United Kingdom National Accounts, shows the surplus on current
budget (net saving plus capital taxes), net debt, net borrowing, and the public
expenditure measures. The surplus on current budget measures achievement against
the Golden Rule, and the public sector net debt is related to the Sustainable Investment
Rule. General government net borrowing is closely related to government deficit in the
context of the Maastricht Treaty, which requires governments to keep their deficits
below 3 per cent of GDP. The public expenditure measures in the UK is denoted as
total managed expenditure (TME). TME is the present government's main measure of
118 The Golden rule is that over the economic cycle, the government will borrow only to invest and not to
fund current spending. In other words current spending must be matched by current receipts such as taxation
and social security contributions. The Sustainable Investment Rule is that net public debt as a proportion of
GDP will be held over the economic cycle at a stable and prudent level. The current government's intention is to
hold this ratio below 40 percent.
public expenditure. It represents the consolidated current and capital expenditure of

the general government sector and excludes capital grants to public corporations and
adds in certain expenditures (mostly capital) by public corporations.
On Budget day, the Treasury publishes the Economic and Fiscal Strategy Report
(EFSR) and the Financial Statement and Budget Report (FSBR) in order to describe the
government's fiscal policy. Those two now appear together within a single budget
document, sometimes called the Red Book. The Treasury also publishes Main Supply
Estimates, Supplementary Budgetary Information, and Public Expenditure Statistical Analyses
(PESA), which are command papers to be submitted to the Parliament. PESA is a
compendium that brings together recent outturn data, estimated outturns for the latest
year and budgetary plans covering a whole range of UK public expenditure. Further
information on each governmental department's spending is made available in the
individual Departmental Reports. The Red Book and PESA present TME as a
government measure of public expenditure, similar to the Blue Book.
Table 17-3. TME by spending sectors

(unit: million)
2002-03
2001-02 2003-04
estimated
outturn plans
outturn
Central Government Own Expenditure 285,811 309,620 330,600
Local Authority Expenditure 98,424 105,736 117,900
Central Government Support 74,416 81,081 88,955
Local Authority Self-financed Expenditure 18,125 18,923 21,168
Others1) 5,882 5,732 7,746
Public Corporation Expenditure 5,344 5,669 7,200
Central Government Support 2,883 3,182 4,544
Public Corporation's Own-financed Capital Expenditure 1,540 2,161 2,579
Accounting and Other Adjustments 921 327 100
Total Managed Expenditure 389,579 421,025 455,700

Note: 1) Includes Locally Financed Support in Scotland, accounting and other adjustments.
Source: HM Treasury, Public Expenditure Statistical Analyses 2003, May 2003.
Some discrepancies in the numbers can exist between the Treasury statistics and
National Accounts. Needless to say it reflects the differences of the fiscal year and the
calendar year on the statistical period and the fact that the Treasury's accounting and
budgeting concepts are based on the UK's generally accepted accounting practice while
the National Accounts follows the international guidelines of the SNA. Those
discrepancies are very minor.
The UK's budget covers all the public institutions of the general government sector
and the capital expenditures of public corporations. Vote, approved by the Parliament,
covers the resources drawn from the Consolidated Fund, which is the main Treasury
fund. Therefore, total expenditures of NDPBs, public corporations and local authorities
are not included in the Vote, but grants to them are included in the Vote.
The main funds in the UK are the Consolidated Fund and the National Loans
Fund, but there are other important funds such as the National Insurance Fund and
Exchange Equalization Account and the Paymaster General's Accounts, National Debt
Commissioners' Accounts (the National Lottery Distribution Fund Investment Account,
the Court Funds Investment Account, the National Savings Bank Fund, Crown Estate
Account, Donations and Bequests Account, Insolvency Services Investment Account,
National Endowment for Science, Technology and the Arts, National Heritage
Memorial Fund, Northern Ireland Court Service Investment Account, Northern Ireland
National Insurance Fund Investment Account), the Northern Ireland Funds and
Accounts, and a host of smaller accounts (Customs and Excise Account, Inland
Revenue Account, etc.).
Table 17-4. Voted and non-voted expenditure in TME

(unit: resources, million)
2002-03
2001-02 2003-04
estimated
outturn plans
outturn
Total Voted Expenditure in TME(A) 261,757 283,452 275,672
Central Government Spending 189,909 199,178 191,976
Central Government Support for Local Authorities 62,081 65,770 72,151
Central Government Support for Public Corporations 1,014 1,257 2,239
Accounting Adjustments 8,753 17,247 9,306
Total Non-Voted Expenditure in TME(B) 127,822 137,573 180,000
Central Government Spending(including spending by NDPBs and Devolved Executives) 108,525 114,882 132,237
Central Government Support for Local Authorities 12,335 15,312 16,804
Public Corporations 1,869 1,924 2,305
Others1) 5,094 5,455 28,650
Total Managed Expenditure(A+B) 389,579 421,025 455,700
Voted Expenditure not in TME(included within "Non-Budget" in Estimate)(C) 52,245 62,627 96,955
Grants to NDPBs to finance their expenditure 19,688 24,115 28,205
Grants to Devolved Executives to finance their expenditure 31,134 35,954 36,643
Other non-public expenditure 1,423 2,558 32,107
Total Voted Expenditure(A+C) 314,002 346,079 372,627
Note: 1) Include DEL not allocated by spending sector, Central Government Debt Interest, AME Margin, Locally Financed
Expenditure, Non-Voted Net Payments to EU, and accounting and other adjustments.
Source: HM Treasury, Public Expenditure Statistical Analyses 2003, May 2003.
The Consolidated Fund was inaugurated in 1787, which collects every stream of
public revenue and issues the supply of funds for every service. The legislation
governing both the Consolidated Fund and the NLF ensures that government finances
are centrally accountable to Parliament. Generally speaking, all money collected by the
government must, unless Parliament specifies otherwise, be paid into one of these two
main funds. And in general no money can be paid out of the Consolidated Fund or the
National Loans Fund without the specific legal authority of Parliament, and the prior
approval of the Comptroller and Auditor General. Expenditures voted by Parliament

are mostly financed by the Consolidated Fund.
The Consolidated Fund, NLF, EEA and Paymaster General's accounts are all held
at the Bank of England, while NDPBs' accounts are held at commercial banks. This
concentration of funds at the Bank enables the government to minimize both its
holdings of non-interest bearing bank deposits and its gross level of borrowing. It also
facilitates the measurement of the CGBR (Central Government Borrowing
Requirements), which is based on the transactions undertaken by central government
funds and accounts.
The coverage of the public finance according to the stages can be shown as follows.
Table 17-5. The Coverage of Public Finance in the UK
Department Agency NDPBs Local Authorities
Budget
Budgetctr
Votes
Non- Votes - -
FA1
FA2 - -
NAS
PFS
5. The coverage of public finance in Japan

In the National Accounts published by the Japanese Government, the central
government of the general government sector is composed of the General Account,
Special Accounts and Special Entities (SEs, tokusuhouzin). The coverage of the General
and Special Accounts are assumed to coincide with the ministerial departments and
departmental agencies, whose employees are all civil servants. SEs are defined as
'institutions established directly by the government departments or established
according to special public laws. Those are public law administrations equivalent to
the NDPBs in the UK. The Japanese central government has approximately 40 Special
Accounts and 92 SEs, which are classified between the general government and the
public corporations sector according to their principal function. The local governments
of the general government sector are also composed of their general account, and
special accounts including some entities established by themselves.
It seems likely that there exist a number of public institutions in Japan which are
not classified into the general government sector. Japanese government is controlling
many non-profit institutions, and those can be classified into the general government
sector according to the SNA. Sectorization decisions in the Japanese national accounts
are not taken systematically at the level of institutional units. That could be the reason
to underestimate the size of the Japanese public sector.
There are three types of public institutions which may be classified into the public
sector; Related Enterprises of SEs (RESs, kankeikaisha), Joint Enterprises between SEs
and private corporations (JESs, shusshikaisha), Appointed Entities (AEs, siteihouzin).
RESs are the enterprises on which SEs exercise a power in determining their general
policies. The number of RESs of which SEs' ownership is more than 50 percent is 923 in
1997, while that of RESs below 50 percent is 91. All RESs should be classified into the
general government sector or public corporations, since they are controlled by the
central government. JESs are normally used as a policy tool for channelling funds into
public-private partnership programs. The number of JESs of which the proportion of
SEs' ownership is above 20 percent is 240 in 1997. Some of them can be classified into
the public sector according to the nature of SEs' ownership and control.
AEs are nonprofit institutions which perform activities related to administrative
procedure for public policies. They may charge administrative fees or be funded by the
government. Nonprofit institutions which are established by private law are called
Public Interest Entities (PIEs, kouekihouzin). PIEs which perform executive, advisory,
tribunal, inspection and authorization role by the law, order or other administrative
actions will be regarded as AEs. Those, numbering 720 in 1997, can also be classified
into the general government sector.
The total number of RESs, JESs, AEs, which are currently classified into private
sector but controlled and mainly financed by government are approximately 2,000. If a
systematic classification would be performed observing the SNA manual, most of
RESs, JESs, AEs would be classified into the public sector. Therefore, there is no
government publication in Japan which can show the consolidated statistics of the
general governmment and the public sector.
The coverage of public finance statistics published by the Japanese government is
only the General Account and Special Accounts. The budget, which is submitted to
Parliament for its approval, includes 11 Government Related Institutions (GRIs,
seibukankeikikan) in addition to the General Account and Special Accounts. Eleven GRIs
are all SEs which principally engage in financial intermediation for the purpose of
public policy. Budget and public finance statistics do not include any other SEs than
GRIs to say nothing of RESs, JESs, AEs. Government deficit and government debt,
reported annually by the Japanese government, are not observing the international
standard of the SNA and GFS.
6. Conclusions: lessons from the two countries

To be an effective instrument of policy, the coverage of the public finance should
be comprehensive. In many countries, a significant share of government expenditure is
excluded from the coverage of the public finance. These issues concerning the
comprehensiveness of the public finance are discussed in this paper.
European countries are observing the SNA and GFS manual in determining the
coverage of public finance, which require it to be the general government sector. The
European Union has published Manual on Government Deficit and Debt in order to obtain
reliable and comparable statistics for evaluating fiscal deficit and government debt. In
the UK, especially, there exists an institutionalized independent decision-making
process for the sectorization of the general government sector and the public sector.
In the far eastern economies, Korea and Japan, there has been a long tradition that
the coverage of the public finance is recognized not in institutional context but in
accounts and fund units. The government has long exercised its power to establish a
number of the decentralized authorities, as a measure to accomplish policy goals. This
tradition has made the coverage of public finance be underestimated to a considerable

extent.
The most important task by which the government is confronted in strengthening
the management of public expenditure is to rearrange the coverage of the public
finance. In making arrangements for the scope and system of the consolidated public
finance, what is important is that international standards of the SNA and GFS should
be applied on every government and national accounts statistics. As budgetary,
parliamentary, fund, and accounting practices serve many different legal, political, and
administrative purposes, they cannot be taken as a guide to the inclusion of particular
institutions in the general government sector or some other sector.
All the coverage at each expenditure stage, i.e, budget preparation, parliamentary
votes, funds and accounts release, reporting, and statistics need not be equal to the
coverage of the general government. In both budgeting and reporting, the coverage of
the public finance should be the general government sector. The CBB should manage
and control all the budgetary resources included in the general government sector.
Every decisions taken on public finance should be based on the reports and statistics of
the general government sector.
The scope and contents of the budget submitted by the government to the National
Assembly, undoubtedly, should correspond with those of the general government and
the public sector. However, the coverage of the parliamentary approval may be
determined by the allocation of budgetary power between the Administration and the
parliament. The coverage of funds and accounts may be also determined by the
accounting practices in each country.
References
UK Cabinet Office, Public Bodies 1998, 1999.

, Public Bodies 2002, 2003.
Eurostat, European System of Accounts: ESA 1995, European Commission, 1996.
, ESA Manual on Government Deficit and Debt, 2002 edition, European Commission, 2002.
Gill, Derek, Signposting the Zoo From Agencification to a More Principled Choice of
Government Organisational Forms, OECD Journal on Budgeting, 2002, pp. 27-79.
HM Treasury, Public Expenditure Statistical Analyses 2003, 2003.
, Central Government Supply Estimates 2003-04: Main Supply Estimates, HC 648, 2003.
, Central Government Supply Estimates 2003-04: Supplementary Budgetary Information, Cm
5797, 2003.
, Budget 2003, HC 500, 2003.
IMF, A Manual on Government Finance Statistics, 1986.
, Monetary and Financial Statistics Manual, 2000.
, A Manual on Government Finance Statistics, revised version, 2001.
Oak, Dong-Suk, The Public Sector in Korea: Coverage, Measurement and Classification, Korean
Journal of Public Finance, Vol. 12, No. 2, Korean Association of Public Finance and
Economics, 1997 (in Korean).
, Institutional Reform for Improving Public Expenditure Management in Korea,
Korean Institute of Public Finance, 2003 (in Korean).
UK Office for National Statistics, Sector Classification for the National Accounts, 1998 edition, 1998.
, Introducing the European System of Accounts 1995 in the UK, 1998.
, United Kingdom National Accounts: The Blue Book, 2002.
Schiavo-Campo, Salvatore and Daniel Tommasi, Managing Government Expenditure, Asian
Development Bank, 1999.
United Nations, Handbook of National Accounting: Public Sector Accounts, Series F No. 50, 1988.
, System of National Accounts 1993, 1993.
CHAPTER 18
Information Flow for Successful MTEF Operation and its Implication for
the Korea Government
by
Junghun Cho, The World Bank119
1. Introduction
The Korean government is now preparing to introduce a medium-term
expenditure framework (MTEF) and other public expenditure reforms. These reforms
are intended to increase efficacy and transparency of the way public resource are
allocated and spent in Korea. Among the many challenges for the Korean government
to succeed with these reforms and generate anticipated benefits is to have its financial
management information system (FMIS) updated to meet the new, additional
information needed to operate MTEF.
Usually, MTEF requires significant, wide-ranging changes in the way a county
allocates public resources. As a fundamental component to a countrys decision-
making process, the FMIS must provide the necessary financial and performance
information if MTEF reform is to succeed. A successful MTEF will clearly define multi
years total spending and sectoral ceilings, delegate more budgetary discretion to line
ministries and, in return, hold them accountable for their performance. For this to
occur, the governments financial information system must supply necessary and
timely information throughout the budgetary process on macro-forecasting, sectoral
expenditure analysis, and costing and performance management of programs/projects.
The existing Korean FMIS is inadequate to meet the new challenges of MTEF due
to the fragmented networks between ministries and low level of utilization for high-
level decision-making. Also, the underlying Korean regulatory framework, including
budget classification, coverage, and chart of accounts is not suitable for MTEF
operation, and needs substantial changes as well. Finally, to minimize reform fatigue,
the Korea government should pay careful attention to sequencing and pacing of its
efforts to introduce MTEF and the updated FMIS, and carefully consider the
institutional, organizational, and cultural elements impacted by the reform.
This paper consists of five sections. Following the Introduction in Section One, Section
Two explains how successful reform of MTEF and FMIS also depends closely on other
public sector reforms, and introduces the Korean MTEF framework that is the basis for
subsequent FMIS analysis. Section Three introduces a stylized chart that illustrates the
preferred information flow between the central budget office and line ministries for MTEF
operation. Section Four evaluates the adequacy of the existing Korean FMIS to meet the
information needs identified above in Section Three. Finally, Section Five proposes an
119 This paper incorporates substantial contributions from Dong Yeon Kim, Bill Dorotinsky, Mike Stevens,
and Il-Whan An of the World Bank, and Sang Dae Choi, Ho-Joong Shin, Seok-Kyun Lee, Su-Min Park of Korean
Ministry of Planning and Budget, Kwang-Ho Kwon of the Korean Ministry of Finance and Economy, and Allan
Gusfasson, a former Director of Budget Department of Swedish Ministry of Finance.
implementation strategy to improve FMIS and MTEF that will minimize internal resistance
and avoid reform fatigue.
2. MTEF, FMIS, and other public expenditure reforms

2.1. Medium-term expenditure framework
During the 1990s, many countries, both developed and developing, have
introduced medium-term expenditure framework (MTEF). Most OECD countries have
adopted the full plethora of features of MTEF: 73 percent have multi-year expenditure
estimates incorporated into their budgets and 67 percent have a medium-term fiscal
framework with ceilings for expenditures, deficits, and revenues over multiple years as
part of their budget. 120 MTEF is also widely spread out to many middle- and-low-
income countries including 19 African countries.121
However, specific framework of MTEF significantly differs among countries and it
reflects different purposes of MTEF and countrys unique environment. While MTEF is
meant to improve all three goals of public expenditure management122 a) maintaining
fiscal sustainability by controlling aggregate, b) linking budget allocation with national
policy priorities, c) improving efficiency of public service delivery specific
framework of each country differs according to its focus among three objectives. For
example, Sweden adopted MTEF to contain and reduce fiscal deficit; since the
introduction of MTEF, fiscal balance has improved from the deficit of 10.8 percent of
GDP in 1994 to the surplus of 4.8 percent in 2001. 123 On the other hand, the U.K.
intended to increase public investment that continued to decrease; through MTEF,
public investment has increased from 20 UK billion pound in 1997 to 31 UK billion
pound in 2003.124
Strengthening the linkage between budgetary allocation with national policy
priorities is the main purpose for the Korean government to introduce MTEF.
According to the National Resource Allocation Improvement Plan, the Korean
governments main purpose of MTEF is to change a rigid expenditure allocation
structure, meeting social demands that have changed significantly since the 1970s. It is
self-diagnosed by the MPB that current budgeting process is in effect unable to change
swiftly sectoral allocation according to changing environment.
2.2. Financial management information system.
Financial management information system (FMIS) is a basic foundation for

countrys public expenditure management. Information system provides public sector
managers a modern set of tools to assist them in performing a variety of tasks including
a) designing fiscal and monetary policies in responding to changing macroeconomic
120 World Bank/OECD Survey on Budget Practices and Procedures, 2003. Available at: http://ecde.dyndns.org.
As of December 2003, 27 (of 30 total) OECD and 17 developing countries are included in the database
121 Benin, Burkina Faso, Chad, Ethiopia, Gabon, Gambia, Ghana, Guinea, Kenya, Lesotho, Malawi, Mali,
Mozambique, Namibia, Rwanda, South Africa, Tanzania, Uganda and Zambia.

122 Schick (1997).
123 Swedish Ministry of Finance (2003), p. 7.
124 The number is expressed in nominal term (Passman, 2003, p.4).

Chapter 18. Information Flow for Successful MTEF Operation and its Implication for the Korea Government 281
conditions, b) conducting policy analysis to effectively allocate resource among sectors,

and c) delivering service with efficiency and accountability.
Integrating central budget offices FMIS with those of line ministries can more
effectively provide the types of information that MTEF requires. Integration of public
financial management means that the core functions of a financial management system
like budgeting, accounting, cash/debt management, and auditing/reporting need to
work as one entity even though it consists of multiple organizations with geographical
distribution. 125 Under an integrated system, the information outflow from one
component is inflow to all other related components so all components share a
common financial evaluation and measurement basis. This consistent information
flows and linkage among various components in financial management increase public
administrations overall efficiency.
2.3. MTEF and other public sector reforms; performance management and
fiscal decentralization
Line ministries performance on projects/programs must be regularly fed into the

central budget office; it allows the central budget office to monitor the progress and
serves as an accountability mechanism as MTEF delegates more budgetary discretion
to line ministries. In order for performance management to work successfully,
ministries and agencies need to monitor their performance indicators very closely and
the central budget office must have access to the performance information. The types of
information and frequency of reporting, however, vary by country, depending on a
countrys regulatory framework.
As a country becomes more fiscally decentralized, FMISs coverage under MTEF
must be expanded to include expenditure spent by sub-national governments. Fiscal
decentralization indicates that more expenditure is channeled through subnational
governments. Managing only central governments expenditure shows only fraction of
nations overall expenditure and MTEFs function of maintaining nations fiscal
sustainability is seriously impeded. Thus, a countrys FMIS should encompass sub-
national level expenditure by linking local governments and municipalities
information system with the central budget offices FMIS.
2.4. MTEF in Korea
To draw the list of information requirements and implementation strategy, this

paper is based on the MTEF framework designed by the MPB. While the final plan of
the Korean MTEF framework is yet to be finalized by the MPB, this paper bases its
analysis upon the document, National Resource Allocation Improvement Plan
prepared by National Resource Allocation Improvement Team of the MPB. The
stylized chart below shows overall process that the MPB is preparing. For more detail,
please see the Annex 18-1.
125 Hashim and Allan (2002).

Figure 18-1. Stylized chart for Korean MTEF process

Total
Sectoral Annual
Macro- Expenditure
Fiscal Ceiling Budget
economic Setting
Target Set for Formulation
fore casting For multi -
multi-years
years
Updated cost New sectoral Sectoral

estimate of
demand for t+2 Budget
existing
(priority/cost) Preparation
policy/program
3. New information needs for MTEF operation

3.1. MTEF, a new paradigm for budgeting process
As the medium-term expenditure framework expands horizontal budgetary

coverage and strengthens top-down approach, decision-makers information needs
change at all stages of budgetary process. Providing necessary information to decision-
makers is the essential prerequisite for successful MTEF operation in a timely manner.
Integration of financial information systems operated by each ministry is essential for
MTEF to link nation-wide policy with budgetary allocation and to ensure timely
feedback of performance for accountability. Also, for information provided by FMIS to
be useful and relevant for MTEF, underlying framework of budgetary classification,
coverage, and chart of accounts must be aligned with MTEF framework. The following
provides types and flow of information needed for each stage of budgetary process.
3.2. Information needs for budget formulation
3.2.1. Macroeconomic forecasting

The development of macroeconomic framework needs to be seen as a continuous
process with the framework updated frequently. According to the MPBs plan, the
MPB will prepare a Long-Term National Fiscal Management Plan to draw a long-term
sectoral allocation strategy and this Plan will be updated every year, which will
necessitate updating macroeconomic framework every year. However, more frequent
update will be necessary for preparing supplementary budget and cash management.
Macroeconomic framework needs to be linked with other systems, including debt
management system, treasury system and national accounts data for frequent update.
As a countrys macro economy is intricately affected by numerous variables national
debt level, cash balance, foreign reserve, interest, and exchange rate among others - and
these variables are managed by different agencies and ministries, predicting
macroeconomic condition involves linking relevant information systems so that data
from each system flows in and out of the rest of systems on a timely basis.
The central budget office needs to formulate a separate, conservative

macroeconomic forecast by a) combining multiple sources of forecasts and b) creating
pre-determined mechanism to lower economic growth rate to ensure fiscal
sustainability. Countries often conservatively modify general-purpose macroeconomic
forecasting for budgeting-purpose forecasting.126 It seems to be a prudent exercise for
the central budget office to base its revenue estimate and other budgetary decision on a
conservative set of assumptions. For this purpose, the central budget office should have
access to necessary macroeconomic information as well as technical capacity to analyze
the results of macro forecasts presented by other institutions, such as the central bank
and private and public research institutes, and then come up with its own version.
Figure 18-2. Information flow under MTEF in Korea
126 For example, the Canadian Department of Finance systematically revises the private sector forecasts
downwards as a further measure of prudence. This takes the form of the government adding 50-100 basis
points (0.5-1.0 percentage points) to the average private sector economic forecasts for interest rates and then
feeding this through its entire econometric model, thus producing lower forecast economic activity. This
provides a buffer in order to maintain the governments fiscal objectives (OECD, 1999, p.9).
3.2.2. Fiscal target and total expenditure setting for multi-years

Setting fiscal target and total expenditure ceiling involves both economic and
political consideration. Interpreting the status of current economy within economic
cycle is very important, because the fact that the economy is viewed as in its downturn
of economic cycle more easily justifies central budget office to create fiscal deficit to
boost up consumption and investment. However, precise and accurate interpretation of
economic cycle is extremely difficult and often vulnerable to non-economic factors.
To maintain fiscal sustainability, therefore, the central budget office needs to
design a decision-making mechanism based on objective and quantifiable indicators
and multiple scenario analysis. Reducing room for the central budget offices
capricious intervention on macroeconomy can actually increase the automatic
stabilizing effect of public expenditure and increase the credibility and integrity of
central budget office. Also, the central budget office needs to develop several scenarios
optimistic, pessimistic, and medium regarding economic growth and other macro
indicators and then decide which scenario to take through a committee that consists of
well-known scholars and policy makers.
3.2.3. Sectoral ceiling setting for multi-years

While countries allocate resource to sectors/ministries using various processes,
this paper bases its analysis on the Korean governments proposal outlined in the
National National Resource Allocation Improvement Plan. According to the MPBs
plan, sectoral allocation is done in three stages; Stage I is setting national strategy and
sectoral priority; Stage II is reviewing line ministries expenditure update,127 and Stage
III is allocating resource to sectors. The following analyzes each stage in terms of the
role of FMIS and the degree of computerization of FMIS.
Stage I may not need to be computerized as it involves political and strategic
process. The central budget office consults with the Presidents Office to incorporate
the Presidents policy priorities. Also, the central budget office comes up with its own
policy analysis to prioritize public spending for the coming three years. As the Korean
MTEF covers three to five years span, this analysis needs to have a medium-term
perspective and be agreed by sector ministers and the ruling political party.
Stage II, on the other hand, requires significant level of computerization as it
involves massive amount of information coming to the central budget office from all
line ministries. An extensive information network between the central budget office
and all ministries and agencies need to be set up so that line ministries expenditure
update is transferred through the network. The central budget office needs to provide
standardized template to line ministries so that line ministries can easily aggregate and
analyze all the updates submitted. The template also needs to fix all the macro
indicators to be used for updating line ministries cost estimate so that all line
ministries apply the same indicators to their calculation.
In addition to data transfer capacity, the central budget offices FMIS should
provide in-depth analytical tools to verify the reasonableness of line ministries
expenditure forecast. Given the massive amount of information submitted and the
limited resource and time within the central budget office, a computerized analysis
system is a critical prerequisite to this task. Given that, under MTEF, more autonomy
127 Expenditure update usually consist of a) cost estimate of existing programs/projects for year t and t+1,
using updated indicators such as price index, unemployment rates, estimated number of students by grade,
and b) new policy proposal for year t+2 and its cost estimate.
will be given to line ministries once sectoral ceiling is agreed, central budget offices
critical review at this stage becomes more important.
Stage III can be done using a simple set of linked spreadsheets until sophisticated
software is developed. Once national policy priority is set and the line ministries
budget update is analyzed, the central budget office begins to categorize all
expenditure in terms of priority. It will calculate all statutory spending and highest
priority spending and then move down to the next priority level spending until the
sum of expenditure reaches pre-determined total aggregate ceiling for ensuring
accountability of public spending.
For this task, codifying each expenditure account by priority, functional
/programmatic classification, 128 and netting out all internal transfers will be a
prerequisite. Especially, it will be important to clarify and simplify the flow of
expenditure among the general account, special accounts and funds within a sector. In
many countries, program expenditure usually consists of direct spending as well as
transfer to other accounts or institutions and these transfers are often counted twice by
both giving and receiving institutions. In order to make the sum of all sectoral
spending equal to total aggregate ceiling, netting out all the internal transfers within a
sector should be readily available from the FMIS.
3.2.4. Annual budget formulating

This section assumes that the total expenditure aggregate and the sectoral
/ministerial allocation is already finalized. Detailed information on decision-making
process can be found in the MPBs National Resource Allocation Improvement Plan.
The central budget office prepares an electronic Budget Circular template that
prefixes sectoral allocation and major macroeconomic indicators and line ministries
submit their budget request on line. Preferably, an electronic Budget Circular has an
easy-to-use template with macro variables fixed so that line ministries can easily fill in.
Line ministries prepare their own budget within the provided ceiling for upcoming
three years. As mentioned above, the FMIS should have a function to categorize
requested expenditure by function or programs and netting out internal transfers.
The FMIS should allow the central budget office to detect any request that exceeds
prefixed ceiling or any uses of different assumptions by line ministries to calculate
budget request. If sectoral allocation is hard constraint, the FMIS can be set in a way
that prohibits line ministries from putting more than the ceiling. Alternatively, if some
kinds of adjustment and negotiation on sectoral ceiling are allowed, the FMIS can be set
to detect and signal all the requests that exceed the ceiling.
In addition to financial information, the FMIS under MTEF should have capacity to
record and monitor cost and performance information of major, if not all, programs
and projects. MTEF, in return for allowing line ministries greater budgetary decision
power, needs to establish an accountability mechanism. While manual, separate follow-
up on performance may be one option as MTEF starts to operate, the FMIS should
eventually capture performance and cost information. In particular, cost information
for major programs and projects is an essential element for performance evaluation and
is in effect feasible to trace only with well-developed computerized cost accounting
system.129
128 According to OECD Guidelines, functional or programmatic classification is most suitable for sectoral
allocation under MTEF.

129 However, costing in public sector is recognized to be more difficult due to complex overhead cost
sharing and unclear valuation of government assets.

The central budget offices FMIS should be linked to treasury system and other
execution system. Once the National Assembly approves the draft executive budget,
the approved budget needs to be automatically registered in the treasury system. In
this integrated network, if situations such as low revenue collection, cash shortage, and
natural disasters arise so initial disbursement plan needs to change in terms of both
amount and timing of allocation the central budget office can immediately effect this
change by restricting line ministries spending authority through FMIS, which is a
necessary control mechanism in a fiscal emergency.
3.3. Information needs for budget execution and reporting
While effective budget execution and fiscal reporting requires many types of
information, what is needed under MTEF is not much different from the existing
system. Thus, this paper deals with only information that needs to change or add
under MTEF. For more detailed information regarding the role and working
mechanism for execution process, please refer to the Hashim and Moon (2004).
Even though MTEF allows more discretion to line ministries in operating their
budget within ceiling, the central budget office should have access to detailed accounts
of line ministries throughout a fiscal year, not just at the end of fiscal year. The fact that
the central budget office has access to what and where line ministries are spending
serves as a check-and-balance system for line ministries increased budgetary power.
However, line ministries are very likely to resist sharing information with the central
budget office. Therefore, legal mandate and political support is a critical element for
central budget offices accessibility to these information.
As maintaining fiscal sustainability is one of the objectives of MTEF, an FMIS
should cover expenditure spent not just by central ministries but also by agencies and
sub-national governments. This is particularly important, for many countries now
delegate spending decision into local government and central ministries spending is
often made through agencies and local governments. While integrating local
government information system into a central FMIS is the most desirable setting,
manual data transfer may be the initial step to feed local government spending
information into a central FMIS.
More important action than the network integration is the integration of budget
classification and chart of accounts between central and local governments. For central
budget office to effectively monitor and control sub-national level expenditure, it is
very important to ensure local governments use same classification and chart of
accounts as those of central government. Otherwise, central budget office will have
great difficulty aggregating and analyzing them.
4. Inadequacy of current FMIS to support MTEF

The current FMIS in the MPB and other ministries is not appropriate to provide
necessary information in a timely basis due to three major reasons; a) fragmented
system among ministries and agencies, b) inappropriate regulatory environments, and
c) low utilization for policy-level decision-making.
4.1. Fragmented systems among ministries, agencies, and sub-national governments
The MPB, the central budget office, and Ministry of Finance and Economy (MoFE),
the treasury, have two separate information systems and they are not sufficiently
integrated. Currently, the MoFE is in charge of macroeconomic and treasury
management, and the MPB is in charge of budget formulation and execution. For an
effective public expenditure control, two systems need to be integrated so that the
central budget office can regularly monitor budget execution progress, cash balances,
and other important indicators and enact any changes on spending amount and
schedule, if necessary.
An information system that ministries use to submit their budget request is
different from the one that ministries use for budget execution such as allotment,
commitment, and cash rationing. Within the MPB, two separate software are used for
line ministries budget submission and budget execution. There is hardly any
justification to operate two systems for budget formulation and execution other than
that two softwares were developed separately in different period. With two separate
systems, line ministries tend to care more on formulation and less on execution and
this can potentially lead to brakeage and inconsistency when transferring budget
information from one to another.
The e-procurement system in the Public Procurement Service (PPS) needs to be
linked with the MoFE and MPB systems to reduce opportunity cost from idle cash. The
PPS has its own e-procurement system,130 a very advanced feature for effective and
transparent public procurement, but it is not sufficiently linked with neither the
MoFEs nor the MPBs network. For example, e-payment function of the PPSs e-
procurement system is not linked with the treasury system so that PPS needs to keep
large amount of cash in its account to ensure sufficient fund to pay goods and services
it purchases, which incurs significant opportunity costs from idle cash balance. If the
PPSs network is linked with the treasury, PPS doesnt need to keep cash in its account.
The Internal Tax Office and the Customs use separate and stand-alone information
systems. These systems need to be linked with the central budget offices system to
give updated cash inflow information, which is critical for effective cash management
and budget execution.
A financial information system for sub-national governments does not exist and is
now prepared by the Ministry of Home Affairs, with little consideration of integrating with
the MPBs information system. While developing an information system that overarches all
local government spending is worth pursuing in the long run, it is very important to ensure
the local FMIS is compatible, if not integrated, with the central governments FMIS to ease
the burden of aggregation and consolidation. MTEFs function of maintaining fiscal
sustainability doesnt work well if significant amount of public expenditure is spent on
sub-national level and the central budget office has a very poor handle on it.
4.2. Low utilization for policy-level budgetary decision making
It is usually acknowledged that senior-level officials make little use of FMIS to

make policy-level decisions. 131 While the MPBs FMIS has strength in transferring
130 It is called Narajangtu which means national market (http://www.g2b.go.kr/index.html).

131 The author visited the Ministry of Planning and Budget in October 2003. During the visit, both senior
officials and information specialists mentioned that the MPBs FMIS doesnt have enough functions for policy
analysis.
information between ministries on line so that most of paper works get eliminated,
its analytical features are still underdeveloped and inadequate for sectoral and other
aggregate-level spending analysis even though FMIS can potentially equip the central
budget office with strong analytical tools.
Both senior officials lack of interest and understanding of FMIS and FMIS
inability to provide aggregate analysis information attribute to the low utilization of
FMIS. Like a vicious circle, lack of analytical tools in FMIS disinterests senior officials
from using FMIS and, in turn, lack of interest of senior officials makes it hard for
information specialists to gather inputs and feedback from senior officials to
strengthen FMIS analytical features.
National debt management system is not integrated with the MPBs information
system. The Korean government expects the national debt level to increase
significantly in the coming decades due to the rapidly aging population and potential
unification costs. Effective debt management requires accurate forecasting and
prudent management on macroeconomics, cash balance, and capital market
monitoring. The current MPBs debt management system is, however, far from
comprehensive and integrated, thus needs much work in this area.
4.3. Lack of capacity to record and monitor costing and performance information
The FMIS currently does not produce any performance information, output or
outcome of each program/project, on which decision-makers need to base their
judgment to make in-year and out-year budgetary adjustment. Past efforts of the
Korean government to implement performance budgeting has been mixed.132 A few
selected ministries and agencies prepared and submitted performance reports
manually only at the end of fiscal year. During the fiscal period, however, central
budget office doesnt have any control over line ministries performance. Also,
central budget office hardly makes any budgetary decision based on the performance
report. As a system is as strong as the weakest link, the lack of performance
information feedback to budgetary process can considerably weaken the MTEFs
overall trustworthiness.
Also, the FMIS doesnt have built-in functions to calculate costs of major, if not
all, projects or programs so the MPB staff has to manually compute the costs.
Accurate and timely cost information for major projects/program is an indispensable
source for performance evaluation. The Korean governments expenditure system has
an extremely complicated accounting structure and flows of fund between accounts
and special funds. Manually sorting out all the internal transfers and double
accounting is in effect prohibitive, and therefore, computerized FMIS is the
prerequisite for cost accounting in a modern government expenditure system.
4.4. Inappropriate regulatory framework
4.4.1. Complex budget classification and chart of accounts

Current Korean accounting system doesnt provide a functional (or programmatic)
classification, which is the appropriate classification for MTEFs sectoral allocation.
Classifying expenditures is important for policy and budget formulation and
measuring the allocation of resources among sectors. According to OECD (2001),133
132 Young-Sun Koh (2003).

133 OECD (2001), p.121.
functional or programmatic classification is appropriate for efficient allocation of

resource, which is the primary purpose of the Korean MTEF. In this regard, the UN
Classification of the Functions of Government (COFOG) can be one benchmark for
Korea to introduce functional classification.134
Government accounts are too detailed and rigid and there exist complicated
internal transfers between general accounts, special accounts, and extra budgetary
funds. Currently, more than 6,000 accounts exists in public expenditure system, and
this excessively detailed accounts make expenditure very rigid. It leaves very little
room for line ministries to manage funds flexibly to meet unexpected situations. Also,
extremely complicated internal transfers between accounts and funds make it difficult
to categorize expenditure by sector. 135 For an effective MTEF operation, the central
budget offices span of control needs to be move upward to a more aggregate level of
chart of accounts.
4.4.2. Insufficient budget coverage

Current Budget Law does not include in the budget, which requires parliamentary
approval, some extra-budgetary funds (EBF),136 and often exclude them from statistical
analysis. For purpose of fiscal control and efficient allocation of resources, which are
core objectives of MTEF, the coverage of the budget should be comprehensive.
Also, expenditure by certain institutions, which has defining characteristics of
public sector, is not included in the budget at this moment. For example, the Korea
Development Institute (KDI), whose funding comes almost 100 percent from the
government, fits the GFS description of general government.137 However, KDI doesnt
count in the general government in the Korean budget. While acknowledging some
shadowy areas exist in defining the public sector, the Korean public sector is defined
too narrowly.138 In principle, all expenditure spent by all institutions that fit a general
definition of public sector needs to be brought into consolidated financial account for
both comprehensiveness and transparency purposes.
4.4.3. Inappropriate government statistics

Currently summing up all expenditure allocated to each sector in the budget
document far exceeds the government total spending due to the practice of double
accounting of program/project expenditure affecting multi-sectors. One of the
reasons for this double accounting is the pressure to inflate spending on certain
sectors so spending on certain program/project counts twice depending on
programs sectoral coverage. While the complicated accounting structure and
classification have in effect prohibited anyone outside the MPB from detecting this
flaw, introducing MTEF will require the sum of all the sectoral spending equals to
total aggregate, especially to draw the table below as identified in the MPBs Plan.
Table 18-1. Analysis of budget composition
134 For more information, please refer to Annex 18-2.

135 For more information, please refer to Annex 18-3.
136 Korean Ministry of Finance and Economy ( 1999).
137 According to IMFs GFS (p.15), the public sector consists of the general government and public
corporations. The general government includes the central, state, and local governments, and public
corporations include financial public corporations, non-financial public corporation, monetary public
corporations including the central bank, and non-monetary financial public corporations.
138 International Monetary Fund (2001).
2005
By Function
Education
Health Total = 100
Defense
.
By Policy
Wage/benefit Total < 100
IT Investment (only major
R&D programs)

By Organization
Ministry of Education
Total = 100
Ministry of Construction
---------
5. Implementation strategy for improving FMIS and MTEF

Both introducing MTEF and integrating and improving FMIS require significant
amount of information flowing frequently between the central budget office and line
ministries. Thus, it is critical to engage all the relevant stakeholders during the
implementation process. Unless stakeholders have the sense of ownership through this
process, the new FMIS will have very little use and impact and will be unable to
provide necessary information to support MTEFs smooth operation.
Both introducing MTEF and integrating and improving FMIS require considerable
time and resource, and should be viewed as a long-term process. As evidenced from
the World Bank-supported projects, the average project duration to introduce FMIS is
6.7 years.139 Also, changing budgetary process must be carefully approached in order
not to impede annual budgeting process as budgeting is a continuous process and cant
afford any stop or break. At the same time, integrating FMIS can be successful only
after the underlying business process is reviewed and updated to fit into a newly
integrated environment. Many ICT-oriented government reforms have been
unsuccessful because they ignored or skipped business process reengineering and just
computerized the old, and often flawed, process.140
Therefore, a modular approach would be more appropriate since some elements of
information needs for MTEF are available in existing system and, more importantly,
the Korean government needs to update related laws and regulation to improve budget
classification and coverage, charts of accounts, and statistical methodology.
The following is a list of factors which deserve close attention in introducing MTEF
and improving FMIS in Korea. While this list is not comprehensive, it may raise
139 Dorotinsky and Cho ( 2004).

140
Department for International Development (1997).
awareness to reformers that more often than not projects with good intention and
design fail when the implementation strategy is not well thought out.
5.1. Leadership and management support
Continued commitment and support at the highest level to improve and make better
use of FMIS is a primary success factor for successful change. This is arguably the most
important success factor for any public sector reforms. Any trial without continued
commitment and support is in effect most likely to fail. While it is likely that MTEF reform
gets sufficient supports from the highest level in the MPB141 and the Office of the President,
the importance of improving FMIS to successfully implement MTEF seems to be less
appreciated within the government. The fact that current FMIS suffers from low utilization
for policy-level budgetary decision-making reinforces this concern.
One way to increase the support for FMIS reform is to piggyback with national e-
government initiative that the President puts high priority during his term. Good example of
the effectiveness of piggybacking strategy is e-procurement reform, which was one of the
eight e-government reforms that Korean government initiated since 2001.142 As one of the
Presidents high priority policy, the e-procurement reform received lots of media attention
and political support during the reform. Today, it is generally agreed that the e-procurement
system in Korea is one of the most successful in the world.143 Likewise, the FMIS reform,
framed as a national issue, not a MPBs, has a better chance to get much needed support from
line ministries.
5.2. Inter-agency coordination
Successfully integrating the network of information systems between the MPB

and line ministries to provide necessary information for MTEF is crucially dependent
on the cooperation of diverse users in ministries and agencies. While the MPB may
initiate to introduce MTEF and take lead in rolling out its plan, the MPB should engage
all levels of players in implementing MTEF and FMIS. However, the idea of setting
sectoral ceilings in advance can be seen as a mere return to the old control regime
under the Economic Planning Board during 1961 to 1994, with an empty promise of
autonomy within the ceiling. To overcome this suspicion and reluctance, various types
of awareness building efforts, including training, seminar, and on-the-job training,
needs to be provided to line ministries from the early stage of implementation. Also,
strong support from the President and political parties can increase interest and
participation from line ministries.
In particular, a careful coordination and collaboration between the MPB and MoFE
is necessary. The organizational evolution between the MPB and MoFE once one
institution and now divided into two separate institutions create the sense of both
cooperation and competition. 144 Thus, the MPB needs to be particularly careful in
designing macroeconomic forecasting function in MTEF, because the MoFE can
perceive the proposal of creating macroeconomic forecasting function within the MPB
141
For MTEF, the current vice-minister of MPB is a champion of reform and he makes a strong drive to
introduce MTEF into the Korean Government.
142
PPS started to computerized procurement in 1998 but it became nation-wide project in 2001.
143 It receives UNs Public Service Award (PSA) in 2003, http://www.unpan.org/dpepa_psaward.asp
144 For historical evolution of two ministries, please refer to Annex 18-4.
as an intrusion of its territory and hold back any information generated by its treasury
system and other information systems.
5.3. Institutional mindset
While the Korean public sector seems to have technical capacity and infrastructure
sufficient to manage complicated information system, cultural bias against technology
specialists needs deliberate efforts and time to overcome. As evidenced by the lack of
FMIS usage among senor officials, Korean public sector seems to have imbedded bias
against technical specialist over generalist. This issue has a long historical root and is
unlikely to change over night. However, efforts should be made to bring deeper
understanding and appreciation for technical specialist. Getting rid of separate job
grading between generalist and specialist positions and rotating all staff between the
budget department and the information system department may gradually diminish
the bias.
5.4. Avoiding reform fatigue
Careful sequencing and phasing of reforms, using a modular approach, will

reduce the risk of reform fatigue from the simultaneous implementation of MTEF and
FMIS. The MPB and line ministries often complain that they lack resources, mostly
staff, to exclusively commit to reform agenda while keeping up with heavy daily
routines. Also, so many reforms have been tried out since late 1990s without much
success or recognition, which creates somewhat cynical sentiment among public
servants and poses big challenge to MTEF and FMIS reforms. Therefore, careful
sequencing and phasing, using a modular approach seems the most feasible and
desirable way to mitigate this challenge.
The first step should be getting the basics right; improving budget classification,
coverage, charts of accounts, and statistical mythology. As Allen Schick (1998) rightly
pointed out, getting the basics right should be the beginning of any advanced public
expenditure reforms. This is more so to the types of reforms as comprehensive as
MTEF. Classifying expenditure by program, aggregating the chart of accounts, and
netting out all internal transfers within the same sector and within public sector
boundary should be the first action to be taken and require significant amount of
time and resource.
The second step is to review and reform the budgetary process itself (business
reengineering). This step includes identifying shortcomings of the current budgetary
process and draws a blueprint for a new budgetary process under MTEF, including
cabinet budget meeting, sectoral ceiling setting, and performance information
feedback. The MPB may initially implement only parts of complete blueprint of MTEF,
but having a master plan from the onset of reforming process will keep the MPB
focused and directed. This process will involve extensive consultation with other
stakeholders, especially the Office of the President, which shouldnt be bypassed for
the argument of time-saving.
Simultaneously, the MPB can initiate to integrate various information systems
located in multiple ministries and agencies to provide information necessary for
MTEF operation. As a new budgetary process is being drawn, the MPB can initiate
reviewing technical aspects of FMIS and incorporate new information needs
identified from the MPBs master plan. Once all the information needs and new
budgetary process is finalized, the MPB can update its system by adding new
functions, and integrate its network with other ministries network.
Finally, having these prerequisites met, a new budgetary process can be

introduced more smoothly by building confidence among the MPB staff and creating
a learning curve effect. As pointed out before, introducing MTEF usually takes
several years before fully operational. As Dong Yeon Kim (2004) points out, MPB and
line ministries staff can, through this modular approach, build understanding and
confidence of the new system as well as benefiting from a learning curve effect. As
budgeting cant afford any halt and the MPB is currently short of its staff, a gradual
approach seems to be more acceptable and desirable to the Korean government.
ANNEX 18-1. The Korean budgetary process comparison
Before MTEF Month After MTEF
12
MPB provides Guideline
for long-term fiscal plan
Estimate of new and on-going 1
projects spending request
Line ministries submit of long-term
no line ministrys input
spending plan and major project exp. plan
2 MPB draft total/sectoral ceiling for three
years
Cabinet Budget Meeting

a) total aggregate, b) sectoral ceiling for 3
MPB provides Budget Circular years, c) major program spending plan
3
for this years formulation
MPB provides Budget Circular for this years
formulation
MPB reports to N.A. and party 4 MPB reports to N.A. and major parties
Ministries submit budget 4-5 Ministries prepare budget request
Budget formulation
MPB reviews proposals
- review of individual
6-8 sectoral/ministerial overview
projects
update of Marchs macro conditions
Consultation with ruling party

Consultation with ruling party
8 2nd Cabinet Budget Meeting
- finalize next years budget
MPB submits draft budget and coming three

MPB submits draft budget and
10 years total/sectoral ceiling to N.A
coming three years
11 MPB Updates long-term fiscal management plan

ANNEX 18-2. COFOG: Breakdown by division and group

01. GENERAL PUBLIC SERVICES 06. HOUSING AND COMMUNITY
01.1 Executive and legislative organs, financial AMENITIES
and fiscal affairs, external affairs 06.1 Housing development
01.2 Foreign economic aid 06.2 Community development
01.3 General services 06.3 Water supply
01.4 Basic research 06.4 Street lighting
01.5 R&D general public services 06.5 R&D housing and community amenities
01.6 Other general public services 06.6 Other
01.7 Public debt transactions
01.8 Transfers of a general character between 07. HEALTH
different levels of government 07.1 Medical products, appliances and equipment
07.2 Out-patient services
02. DEFENCE 07.3 Hospital services
02.1 Military defense 07.4 Public health services
02.2 Civil defense 07.5 R&D health
02.3 Foreign military aid 07.6 Other
02.4 R&D defense
02.5 Other 08. RECREATION, CULTURE AND
RELIGION
03. PUBLIC ORDER AND SAFETY 08.1 Recreational and sporting services
03.1 Police services 08.2 Cultural services
03.2 Fire-protection services 08.3 Broadcasting and publishing services
03.3 Law courts 08.4 Religious and other community services
03.4 Prisons 08.5 R&D recreation, culture and religion
03.5 R&D public order and safety 08.6 Other
03.6 Other
09. EDUCATION
04. ECONOMIC AFFAIRS 09.1 Pre-primary and primary education
04.1 General economic, commercial and labor affairs 09.2 Secondary education
04.2 Agriculture, forestry, fishing and hunting 09.3 Post-secondary non-tertiary education
04.3 Fuel and energy 09.4 Tertiary education
04.4 Mining, manufacturing and construction 09.5 Education not definable by level
04.5 Transport 09.6 Subsidiary services to education
04.6 Communication 09.7 R&D education
04.7 Other industries 09.8 Other
04.8 R&D economic affairs
04.9 Other 10. SOCIAL PROTECTION
10.1 Sickness and disability
05. ENVIRONMENTAL PROTECTION 10.2 Old age
05.1 Waste management 10.3 Survivors
05.2 Waste water management 10.4 Family and children
05.3 Pollution abatement 10.5 Unemployment
05.4 Protection of biodiversity and landscape 10.6 Housing
05.5 R&D environmental protection 10.7 Social exclusion
05.6 Other 10.8 R&D social protection
10.9 Other
Source: United Nations. Classifications of Expenditure According to Purpose, 2000.
ST/ESA/STAT/SER.M/84. New York.
ANNEX 18-3. Example of internal transfer within accounts and funds

ANNEX 18-4. Organizational evolution between the ministry of planning

and budget and the ministry of finance and economy
Ministry of Finance
(1948)
Ministry of Finance Economic Planning

(1961) Board (1961)
Ministry of Finance and

Economy
(1994)
Ministry of Finance and Ministry of Planning

Economy (1998) and Budget (1998)
References
Allen, Richard and Daniel Tommasi, (eds.), Managing Public Expenditure, OECD, 2001.
Department for International Development (DFID), Good Practice in Developing Sustainable
Information Systems, London, UK, 1997.
Dorotinsky, Bill, and Junghun Cho, World Banks Experience With Financial Management
Information , System (FMIS) Project, the World Bank, Washington, D.C., 2003.
Hashim, Ali and Richard Allan, Information Systems for Government Fiscal Management, the
World Bank, Washington, D.C., 2002.
Hashim, Ali and Allister J. Moon, Treasury Diagnostic Toolkit, World Bank Working Paper
No.19, 2004, the World Bank.
IMF, Korea: Further Reforms in Fiscal Management, Washington, D.C., 2001.
Kim, Dong Yeon Key Issues for Introducing MTEF and Top-down Budgeting in Korea, the
World Bank, Washington DC, 2004.
Koh, Youngsun, Public Expenditure Management in Korea, Korea Development Institute, 2003.
Korean Ministry of Finance and Economy, National Financial Management Information System:
Overview, Seoul, Korea, 2003.
, Government Finance Statistics, 1999.
Korea Ministry of Planning and Budget, Current Status and Future Issues on National Budget
Information System, Seoul, Korea, 2003.
OECD, Budgeting in Canada, France, Paris, 1999.
Passman, Liz, Presentation on Public Expenditure Planning and Control, UK Treasury, November
2003.
Schick, Allan, The Changing Roles of Central Budget Office, OECD, 1997.
Swedish Ministry of Finance, Budget Statement, 2003.
UNs Public Service Award (PSA), http://www.unpan.org/dpepa_psaward.asp.
World Bank, Public Expenditure Management Handbook, Washington, DC, 1998.
World Bank/OECD, Survey on Budget Practices and Procedures, 2003. http://ecde.dyndns.org.
.
CHAPTER 19
Medium -Term Expenditure Frameworks, Performance, and
Transparency
by
1. Introduction
The connections between these three public sector reform areas may not be
immediately obvious. Medium-Term Expenditure Frameworks (MTEFs) are about
improving budget decision-making processes and incentives to lead to better
outcomes. Performance budgeting and management are about getting more output or
impact from any given level of public spending. And transparency, broadly, is about
greater provision of information on public sector operations, finance and performance.
How might these relate?
This paper briefly explores the relationship between the three areas. The paper
first provides a brief overview of each of the three topics and their salient features. The
remainder of the paper explores the relationships between a MTEF, performance, and
transparency.
2. Medium-Term Expenditure Frameworks (MTEFs)

An MTEF supports (1) macrofiscal discipline and stability (avoiding public finance
crises and supporting economic growth and stability), (2) strategic allocation of
resources, and better matching government policy with programs, objectives, and (3)
technical or operational efficiency (getting the most from each won spent).
An MTEF is about creating an orderly public financial management process and
appropriate incentives, promoting sound decisions and outcomes. Generally, an MTEF
tries to integrate policy with resource allocation decisions in a multi-year context to
better reflect the effects of decisions on public finances, and generally lead to improved
policy decisions. It also involves structuring the decision-making process to encourage
economy and efficiency, and focus spending and policy on attaining the governments
objectives. So a MTEF could be said to be about strategic allocation of resources in line
with government priorities and the opportunity costs of decisions.
3. Performance budget/management
Many different approaches to performance budgeting and management have been
tried over the years. Reforms from program budgeting, performance budgeting,
program-planning-budgeting systems (PPBS), management by objectives (MBO), and
zero-based budgeting (ZBB) fall within this genre of reforms. Program evaluation and
impact analysis are also initiatives that properly fall within this genre as well.
These initiatives fall into three types of reforms:
z Reforms that try to assess whether programs are working or can work better,
to provide incentives for resources to be used more efficiently through
linkages with resource allocation, and to reallocate resources from mess to
more effective uses (e.g. impact analysis);
z Reforms that directly trying to improve use of resources through management
oversight and action (e.g. MBO); and
z Reforms that attempt both.
Many early attempts at performance budgeting fall into the first category, focusing
on developing and publishing information on performance for government activities.
In this respect, organization of activities within ministries into programs can assist in
easing the challenge by grouping like activities with common objectives into one
program, and then developing measures of impact or at least output.
With lessons of experience, it has become clear that producing the information is in
itself not sufficient. The act of measuring performance can have a positive impact, but
more along the lines of the Hawthorne Effect, where the novelty of measurement
modifies performance of those being measured, but only for a limited period of time.
The collection and production of information does not result in a sustainable cultural
change or modification of incentives. Some reforms have attempted to address this by
developing a process of performance measure development that is done by the line
ministries and agencies themselves (e.g. U.S. prohibited use of consultants to develop
the GPRA performance measures or plans).
The more enduring issues that have emerged in performance management involve
the following questions: What is done with the performance information once
produced? Who uses the information, for what end, and how? How can decision-
makers productively use the information? How can the system be structured to
produce more enduring changes in incentives for better performance? How can the
information be made more user-friendly to policy officials?
Some selected lessons from international experience indicate both the challenge
and promise of performance management and budgeting. For example, implementing
effective performance monitoring145 requires:
z Setting organizational incentives to support performance monitoring;

z Getting performance monitoring consistent with organizational culture;
z Need for a central unit to play active and effective leadership role in defining
criteria and implementing practical performance monitoring;
z Linking ongoing performance measurement with more periodic program and
policy evaluations strengthens the culture of evaluation and program
improvement; and
z Attention to client/customer/citizen in developing performance monitoring
can strengthen value and use;
Further, integration of performance and financial management146 is easier where:
145 Mayne and Zapicogoni (1997), pp.22-24.

146 OECD(1999a), p.9.
Chapter 19. Medium-Term Expenditure Frameworks, Performance, and Transparency 301
z Strategic/target objective setting is linked to resource allocation;

z Global or output-based budgeting is in place;
z Full-cost activity accounting is in place;
z Programs consist of tangible and measurable products or services;
z Integration is attempted at the level of program management and operational
management;
z The impact of programs can be seen soon after delivery; and
z The results can be attributed to the program with high degree of confidence;
With respect to program evaluation, it is first critical to understand that it

generally refers to deeper, more extensive review of the impact of a specific program
using relatively more advanced techniques, such as impact analysis or randomized
sampling. It can be an invaluable adjunct to performance information, but is not a
replacement. From experience, issues for maximizing the use and value of evaluations 147
include:
z Gaining support from the top of Government;

z Generating effective demand for evaluations;
z Setting realistic expectations;
z Systematizing evaluation activities;
z Linking with the budget process;
z Meeting user needs;
z Ensuring relevance;
z Involving stakeholders;
z Communicating findings;
z Monitoring and follow-up; and
z Recognizing needs of staff for training and support.
From reviewing these lists, some common patterns emerge. First, groups or
stakeholders that are to be served or are expected to use the information are many and
varied. To enable full use of the information, it must be disseminated broadly. It must
be transparent.
Further, use of the information must be clear: who is using it, for what end, and
how? This applies to agencies, policy officials, finance staff, etc.
And in terms of linking with the budget process, full-costing of programs and
having a program structure and an MTEF provide a vehicle for pursuing these.
4. Transparency
The theme of transparency has many dimensions. At a broader conceptual level,
it is about having open decision-making process and easy access to information by the
public. It is an enabler of accountability and performance assessments, and a
prerequisite for sound decision-making (e.g. MTEFs).
More specifically, the IMF has developed an instrument for assessing many
dimensions of transparency, operationalizing the concept. For the IMF Fiscal ROSC,
improved transparency is a necessary basis for improving the efficiency and
effectiveness of fiscal management through:
147 OECD (1999b), p.22-24.

z Better and broader information, leading to better decisions;

z Broader access to information, and better decisions by actors within and
outside government;
z Stronger public and parliamentary oversight, and potential for improved
governance and accountability; and
z Improved business environment and attractiveness to FDI.
However, transparency is a public good and merit good gainers dont pay and
losers dont support. Taking the first steps of providing the information and improving
accountability must be an exercise in leadership. Once done, market analysts and civil
society organizations can demand the information, making it very difficult to turn
back. More importantly, private sector and civil society organizations, as noted
previously, can exert pressure on government for better decisions. The broad objective
is to create a virtuous cycle, whereby better information strengthens policy and leads to
demands for continuing improvement in information and policy making.
The focus of Fiscal ROSC is on data quality improvement and improved definition
of public sector roles and scope of public finances.
The Code of Fiscal Transparency includes many principles and questions, but
some examples will bring greater clarity. For example, IMF ROSC questions 4.1.1,
2.1.2, 3.1.3, 3.1.1, 3.1.4, 3.1.3, and 3.1.5 all address the extent of a countrys medium-
term budget framework. Question 2.1.3 address off-budget activity. And, questions
2.1.1, 3.2.1, 3.3.1, and 4.1.2, address issues of accounting and data quality. As is evident,
the ROSC actually tracks the key components of an MTEF, and through it one can
monitor country progress in implementing the MTEF.
It should also be evident that data and accounting quality are essential for effective
performance management or monitoring. Bad data do not allow accountability or
performance assessment.
5. MTEF and performance

From this brief over-view, the MTEF is directly about improving the performance
of the public financial management system along the three objectives. But equally
important, an MTEF involves several dimensions that seek to enable improved
First, in terms of removing barriers to program performance by providing more
predictability in resource flows to programs. One of the most common barriers to
public sector performance is the erratic flow of resources to programs, undermining
service delivery. Managers spend time handling weak staff morale, angry vendors, and
dissatisfied customers, and are not able to spend time improving service delivery.
Second, with more predictable resource flows, planning and budget take on more
relevance. If the budget includes resources for the program, the program will receive
them. Plans developed in advance can be executed. Better planning can improve
Third, the MTEF is also about changing incentives in the budget process, including
for spending ministries and under them, service delivery programs. With hard budget
constraints, spending ministries must re-examine all spending and ways of conducting
business to find savings for reallocation to new or higher priority programs. Within the
total ceiling, spending ministries have greater discretion in allocating funds to

programs under their jurisdiction, and during budget execution, have modest
discretion to reallocate across economic categories of spending to meet program needs.
Line agencies are more empowered, hopefully have greater ownership over the
budget and outcomes, and certainly have more authority over managing their funds
and resources. This is fundamentally about shifting bureaucratic cultures from a
traditional administrative approach to a managerial approach innovative managers
continually pursuing program improvements and efficiencies. Certainly, absent greater
authority over resources, there can be no accountability for results --- there is always an
excuse for poor performance (e.g. I did not get the resources in the budget, or I have
no control over my resources so cannot achieve better results).
6. MTEF and transparency

Transparency more open decision-making processes and accessibility of
information is integral to an MTEF. Indeed, greater transparency is a prerequisite for
successful MTEFs good budget classification, good accounting and reporting and
key to changing incentives in the budget process and also enabling performance and
accountability.
First, in terms of improving the budget process, spending ministries under an
MTEF have a greater role in developing sectoral policy within resource constraints, and
through cabinet in engaging in and understanding the macroeconomic constraints
facing the entire public sector. For this to work, spending ministries need better
information on spending past, current, and planned as well as better information
on spending during the year to enable active management of finances.
Second, greater availability of accurate public finance information is important for
the general public and for markets. If citizens have a clearer picture of the economic
and public finance situation, they are better positioned to modify expectations of
government services, and also more receptive to any needed retrenchment in
spending.148
Third, greater transparency is a key technique for changing incentives and
reinforcing good practice in public financial management. For example, as noted in
another conference paper in this volume,149 weak macroeconomic estimates are one of
the most common problems leading to weak public finance system performance.
Beyond better forecasting techniques or more skilled staff, approaches for improving
macroeconomic forecasts include: comparison with private or other forecasts during
forecasts development (Poland, Canada, U.S); vetting of forecasts through academic or
non-government fora or blue-ribbon panels (Poland); early release of forecasts to
Parliament and the public (U.K., Sweden); technical rules (publicly announced) for
making the forecasts more conservative (Canada); economic updates prior to budget
enactment that allow modifications to the budget proposals to fit emerging trends
(Finland, U.S.); and, inclusion in the budget of a table comparing the key economic
variables of the government forecast with several reputable private or public forecasts,
with explanations for significant variation for median forecasts (U.S.).150 The leitmotif
148 For more on the IMF Code of Fiscal Transparency and markets, see Petrie (2003).
149 See Chapter 1 in this volum.
150 IBID. Pages 2-3.
running through these is transparency: providing more information, earlier, to more

people, to allow accountability to work.
The IMF Code of Fiscal Transparency allows an assessment to be made of
transparency practices in countries, including among other issues whether components
of an MTEF are in place, and the general availability of information, openness of
processes, etc. While a more focused and applied definition of transparency, it can be a
valuable vehicle for monitoring progress in improving transparency.
7. Weaving it all together

There are clear synergies between the three public sector reform concepts of
MTEFs, performance, and transparency. Each reform reinforces the others, and allows
the strengths or improvements from each to be fully recognized. Where pieces of each
are missing, less than the full benefit of the reforms is likely.
The MTEF promotes fiscal sustainability, removes barriers to program
performance, and is designed to also empower line ministries and managers (within
limits).
Performance budgeting and management focuses on the performance of
government activities and spending, is easier to undertake where fiscal sustainability
exists, and can more fully achieve its benefits if there is a good decision-making process
(MTEF) for using the information and incentives for agencies to use the information
more effectively in setting priorities and making funding decisions.
Transparency, as a general concept, can broaden the constituency for reform, and
assure information is available for those seeking to use it. Transparency amplifies the
effects of performance information, and is essential for an effective MTEF. Narrowly, in
terms of the IMF Fiscal ROSC operational definition of transparency, it tracks country
implementation of an MTEF and the openness of decision-making and accuracy of
information.
References
Mayne, John and Eduardo Zapico-Goni, Monitoring Performance in the Public Sector: Future
Directions from International Experience, Transaction Publishers; London, 1997.
OECD, Integrating Financial Management and Performance Management, July 1999a. p. 9.
OECD/PUMA, Improving Evaluation Practices: Best Practices Guidelines for Evaluation and
Background Paper, January 1999.
Petrie, Murray, Promoting Fiscal Transparency: The Complementary Roles of the IMF,
Financial Markets, and Civil Society, IMF Working Paper WP/03/199, October 2003.

Reforming The Public Expenditure Management System (MTEF)

Uploaded by

Copyright:

Available Formats

Reforming The Public Expenditure Management System (MTEF)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Reforming The Public Expenditure Management System (MTEF)

Uploaded by

Copyright:

Available Formats

Public Disclosure Authorized

Reforming the Public Expenditure

The World Bank and Korea Development Institute

Welcoming Remarks by Choogsoo Kim (President, KDI)

Keynote Speech by Byung-il Kim (Minister of Planning and Budget)

PART I. MEDIUM-TERM EXPENDITURE FRAMEWORK

CHAPTER 1. Developing a Medium-Term Expenditure Framework ......... 1

CHAPTER 2. MTEF Implementation in Korea .............................................. 17

CHAPTER 3. Key Issues for Introducing MTEF and Top-down Budgeting

CHAPTER 4. Successful Installation of MTEF to the Korean Fiscal System .... 41

4. Concluding Remarks ............................................................................. 55

CHAPTER 5. MTEF in Sweden .......................................................................... 57

CHAPTER 6. The Swedish Budget Process: Expenditure Targets and Top-

CHAPTER 7. Medium-term Financial Planning in the Federal Republic of

CHAPTER 8. Medium-term Perspectives in Budgeting: The U.S. Case .... 93

PART II. PERFORMANCE MANAGEMENT

CHAPTER 9. An Overall Framework for Considering Performance

CHAPTER 10. Introducing Performance Management .............................. 121

CHAPTER 11. Performance Management in the Health Sector ................ 147

CHAPTER 12. Performance Management in Education: A Call for

CHAPTER 13. Establishing a Performance Management Strategy:

4. Designing a performance management system .............................. 209

CHAPTER 14. Results Based Budgeting in Australia ......................................... 227

CHAPTER 15. Performance Management and Budgeting: a Case Study of

CHAPTER 16. Performance Budgeting in the United States .............................. 257

PART III. FISCAL TRANSPARENCY

CHAPTER 17. Coverage of Public Finance in Korea .......................................... 263

CHAPTER 19. Medium-Term Expenditure Frameworks, Performance, and

Table 7-1. Preparation of the budget 2005........................................................... 90

Table 9-1. Difference between performance and traditional budgeting....... 113

Table 10-1. Key Differences between evaluation and monitoring................. 122

Table 12-1. Performance test ............................................................................... 170

Table 17-1. The coverage of public finance in Korea ....................................... 270

Table 18-1. Analysis of budget composition..................................................... 289

Figure 4-1. Business cycle and fiscal balance...................................................... 44

Figure 6-1. Annual GDP growth 1980-2002 ........................................................ 71

Figure 8-1. Functional Classification.................................................................. 100

Figure 10-1. BOD levels of four major rivers .................................................... 123

Figure 11-1. Chain of objectives in the health field.......................................... 150

Figure 13-1. Sample Performance Management System schedule for each

Figure 14-1. Improvement in net debt ............................................................... 228

Figure 17-1. Stages of expenditure management ............................................. 266

Figure 18-1. Stylized chart for Korean MTEF process..................................... 282

2. Korea's efforts to enhance the public expenditure management system

2.2. The participatory government's efforts to reform the public expenditure

2.3. Revising the coverage of public finance to meet international standards

March 15, 2004 (Monday)

Issue 1: Medium-Term Expenditure Framework

09:00 10:00 Registration and Networking

10:00 10:30 Opening Ceremony

10:30 11:00 Break

11:00 12:00 Session 1-1: Conceptual Framework

12:00 14:00 Lunch

14:00 15:30 Session 1-2: Introducing MTEF in Korea

15:30 16:00 Break

19:00 20:30 Welcoming Dinner (invitees only)

March 16, 2004 (Tuesday)

09:00 09:10 Registration and Networking