Shaping Pension Reform In Poland: Security Through Diversity

UH 1X OM 3 H I UNVISISSV IVDOS 'SNOISN3d 'SJ3)IZIVW SIO8V1 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized )M N V 6661 lsnfnV l NsAxoN:jn- U0o14 Public Disclosure Authorized A41SISAIQ q6nojqI lpq3ilY PNZSOIUbV AmDaS :puuLo0d LUJOJOU £Z66-ON 'd]dVd UOISUad NOISSflDSIU - Uli bUld-PI4s dS !l Security through DivE Shaping Pensilon Reform in Poland: Security through Diversity AgnieszkaChlon, Marek Gora and Michal Rutkowski Abstract All over the world, pension systems have financing difficulties that need to be addressed.There are three ways of dealing with pension systems problems - finance it to a greater extent from general revenues, rationalise the system, which produces savings in the short run, or a full-fledgedreform, changingthe logicand foundations of the system. After several years of political and professional discussions,Poland decided to follow the latter path and introduced a new defined contribution mulitipillar system, consisting of a public Notional Defined Contribution, pay-as-you-gofirst pillar, a funded private second pillar, and voluntary funded third pillar. The new framework covers only retirement savings, while other benefits still remain under the old defined-benefit pay-as-you-goregime. The reform was launched on January 1, 1999. As of this date, the old definedbenefit pay-as-you-go system was terminated for workersyounger than 50. The new old-agesystem attempts to offer actuarially fair benefits, potentially creating incentives to increase compliance and postpone retirement. Minimum benefit provision for those who fall below the guaranteed level is cofinanced from general revenue. Diversificationof retirement savingsprovides greater security to the members, as labour market developments that determine the notional rate of return in the first pillar, and financial market developments that determine the second pillar rate of return are not perfectly correlated. This is why the reform package has been named Security throughDiversity. This paper presents the current situation of the pension system, the struggle for pension reform in the 1990s,structure, the long-term outlook of the new pension system and the main aspects of the system design as well as first experiences from the implementation process. Long-term projections show that the new system allows for greater financial stability of the public pension scheme and increases the savings rate with a positive impact on economic growth. 2 Table of contents INTRODUCTION .............................................. 5 1. RECENT PERFORMANCE OF THE PENSION SYSTEM................................................. 7 2. THE STRUGGLE FOR PENSION REFORM ................................................... 11 3. DESCRIPTIONOF THE NEW PENSIONSYSTEM.................................................. 17 3.1 Financing and operations .................................................. 17 3.2 The contribution rate ................................................... l8 20 3.3. The first pillar ................................................... 20 The accumulation phase ........................................... The benefit-distribution phase ........ ......... ......................... 23 Reserve policy in first pillar ........................................... 23 A new role for the social insurance administration (ZUS) ....................................... 25 3.4 The second pillar ........................................... Regulation and supervision .......................................... Pension fund operations ........................................... Funded pension benefits........................................... .. Second pillar pension funds in 1999 ................. ....................... 3.5 Minimum pension guarantee........................................... 3.6 The third pillar ........................................... The new occupational pension plans .......................................... Regulation, supervision and taxation .......................................... 3.7 3.8 4.1. 4.2. 4.3. 34 35 36 37 Estimated benefits from the mandatory pillars .......................................... 37 Early retirement and sector privileges ................................................ 42 PAYG SYSTEM- FINANCINGTHE TRANSITION................................................ 4. 27 27 28 32 33 45 45 Joining the second pillar ................................................ Early retirement schemes and impact ofphasing-out rules....................... 50 Overall savings rate in the pension system .................................... 53 5. LONG TERM IMPACTOF PENSIONREFORM.................................... 6. PRELIMINARY CONCLUSIONS AND LESSONS FOR OTHER COUNTRIES.59 7. REFERENCES 54 .61 APPENDIX .......................................... 63 Other benefits financed from the social security fund ........................................ Uniformed services pensions ......................................... 3 72 73 Table of figures Figure 1. Polish pension system in absenceof reforms....................................................................... 9 Figure 2. Recent developmentsin the Polish pension system............................................................... 10 Figure 3. Financingand functioningof the new pension system..........................................................18 Figure 4. Shareof initial capital in pension value, cohorts 1949-1974 ...................................................22 Figure 5. Demographicreserveinflow and assets, 1999-2020(%GDP)................................................24 Figure 6. Projectedreplacementratesin the new pension system, cohorts 1949-74 .............................38 Figure 7. Shareof funded part in total benefit, cohorts 1949-74...........................................................39 Figure 8. Projectedreplacementratesin the new system...................................................................... 40 Figure 9. Projectedreplacementratesin the new system - sensitivityanalysis.....................................41 Figure 10 Calendarof early retirement scheme withdrawal.................................................................. 44 Figure 11.Differencebetween PolishNDC and funded schemesaccumulation...................................46 Figure 12.Deficit/surplusin the PAYG old-agepension fund in 1999-2020........................................49 Figure 13.Accumulationof assetsin Polish open pension funds, 1999-2020 ........................................50 Figure 14Medium-termimpact of final pension rules vs. Securitythroughdiversityproposal...............52 Figure 15.Savingsin the Polish multipillarpension system, 1999-2020(% GDP)................................53 Figure 16.PAYG old-agefund expendituresand revenues,1999-2050(% GDP)..................................57 Figure 17.IT systemdesign................................. 67 Figure 18.Registrationof employeesin ZUS................................. 68 Figure 19.Registrationof pension fund members................................ 69 Figure 20. Contribution collection mechanism ................................. 70 Figure 21. Individualaccountsidentification................................. 71 Figure 22. Socialinsurancesystem expendituresand revenues,1999-2020(% GDP)............................74 Table 1. Riskswith different types of pension financing.......................................................................6 Table 2. Social-securitycontribution rates as share of gross wage.........................................................19 Table 3. Weightsof old and new pension in the mixedformula........................................................... 21 Table 4. Impactof fees on the realrate of return ..................................................................... 30 Table 5. Age structure of switchersby the end of May 1999:............................................................... 34 Table 6. Projectedpension revenuesand expenditures,transfer to pension fund and reservesin 19992006(% GDP)..................................................................... 48 Table 7. Old and new pension systemcharacteristics..................................................................... 63 Table 8. Initialreform proposal as in SecuritythroughDiversityand its changesduring the reform process..................................................................... 65 Table 9. Medium-termprojections assumptions..................................................................... 66 4 Shaping Pension Reform in Poland: SecuritythroughDiversity Agnieszka Chlon, Marek Gora and Michal Rutkowski Introduction All over the world, pension systemshave financingdifficultiesthat must be addressed. There are three ways of dealingwith pension systems problems.The most affluent countries, especially in the European Union, have tended to opt for continued subsidiesto pensions from general revenues. Countries that cannot afford large subsidieshave exerciseda second option: trying to rationalise their pension systems by seeking more revenues and reducing spending. The third option is fundamental reform, which is the only way of achieving a sustainable solution. But it requires a coherent vision for the design of the new pension system. Following the Chilean and other Latin-Americanpension reforms, it is often asserted that fundamental reform implies replacing a monopoly of a pay-as-you-go,defined-benefit system with a fully funded mandatory defined-contributionsystem. From the perspective of central and eastern European countries,this reform option, despitenumerous advantages,has two essential flaws. First, it does not really diversify risks. A funded monopoly merely replaces a pay-as-you-gomonopoly. Secondly, because of transition costs, this option is difficult to implementin countrieswith a sizeablepay-as-you-gosystem. This is why Poland determined that the pay-as-you-gomonopoly should be replaced with a multipillar system. Future retirement savingswill be diversified.Productivity growth 'Securitythrough Diversityteam.AgnieszkaChlonjoinedPolishpensionreformteam in January1997and was Deputy Directorof the Officefor PensionReformsinceSeptember1998.Marek G6ra was Directorof the Officefor PensionReformof the Polishgovernment sinceOctober1997until Officeterminationin April 1999 and waspreviouslyDeputyDirector(October1996to September1997).MichalRutkowskiis SectorLeaderfor SocialProtectionin Europeand CentralAsiaRegionof the WorldBankand formerDirectorof the Officefor PensionReformof the PolishGovernment(October1996to September1997).Wewouldliketo thankRobert Palaciosfor insightfulcommentsand the entireSecuritythroughDiversityteam for their brilliantwork. The viewsand opinionspresentedin the paperarethoseof the authorsand shouldnot be attributedto the World Bank,or anyotherinstitutionor government. 5 and capital market returns will play equally important roles through the pay-as-you-goand funded parts of the system for providing old age pensions. The long run objective of the reform is to have about 50 per cent (initially 62.5 per cent) of the mandatory old-age contribution going to a pay-as-you-gopillar, while the other 50 per cent (initially 37.5 per cent) being shifted to a funded pillar. Since each of the pillars has different types of risks (Table 1), especially after retirement, the system's overall risk will be better diversified. Of course, pay-as-you-goand funding both offer individuals a measure of certainty about their future, but neither method can insure against common aggregateshocks. One should not 'put all the eggs in one basket'.2 Finally, both pillars will operate in a defined-contribution-type framework. The aim is a transparent system, with pension based on lifetime income, fully adaptableto changing circumstances. Table 1. Riskswith differenttypes of pensionfinancing Risk Ageing population Unemployment Political bargaining Financial market crisis Inflation Pay-as-you-go Exposed Exposed Exposed Not exposed Less exposed Funded Not exposed Not exposed Less exposed Less exposed Exposed The Polish pension reform was launched in 1999. The implementation followed two years of preparation, during which two sets of laws describing the new pension systems were passed. The first set of reform laws, including the law on organisation and operation of pension funds (second pillar) and the law on Employee Pension Programs (part of a third pillar), passed in August 1997. The second set of laws, including the law on Social Security System and the law on the new pay-as-you-gopensions from the Social Security Fund, passed in October and December 1998,respectively. The new multipillar pension system consistsof a notional defined-contribution, pay-asyou-gofirst pillar, a mandatory defined contribution, privately-managed,funded second pillar, and voluntary employee pension plans in the third pillar. The pension reform was launched on January 1, 1999 with changes to the PAYG pillar. Open pension funds started their operationsin April 1999,although licensingprocesshad begun in August 1998. This paper looks at parts of the reform specific to Poland, but draws more general conclusions.We look at the motivation for reform, the struggle of reformers and politicians to advancethe reform agenda, changes made to the initial reform proposal, the architecture of the new system, and issues arising during the transition. The final section offers tentative conclusionsand lessonsfor other countries.3 2 See,for example,Barr (1998),chapter9 andJames(1998). For an overviewof reformsin othertransitioneconomies, seeRutkowski(1998). 6 1. Recent performance of the pension system4 The crisis of the Polish pension system had much in common with the problems in other countries' pay-as-you-go defined-benefit schemes, such as a worsening demographicdependency ratio. But it was exacerbated by a particularly inefficient set of special rules, some typical of many former centrally planned economies (such as low retirement age and widespread sector privileges) and other specific to Poland (such as generous disability provisions). Moreover, in the early 1980sand later in the early 1990s, additional privileges and special rules were added with little or no thought to the fiscal consequences. As a result of the latter policies, the average effective retirement age dropped to 57 years (55 for women and 59 for men), compared to legal standard retirement age at the level of 60 for women and 65 for men. Also in the 1990s, the gap between the system dependency ratio (measured as ratio of pensioners to contributors) and the demographic dependency ratio (measured as ratio of people 65 + to people 15-64) widened, and projections show than this tendency would continue in the future (Figure la). It became difficult to ensure the financial sustainability of the Social Insurance Fund (FUS). Consequently, the contribution rate was raised rapidly from 25 per cent in 1981 to 38 per cent during 1987-1989,at finally to 45 per cent in 1990. This was made necessary by: * * * a decrease in the number of contributors as a result of a decline in employment (Figure 2b) an increase in the number of new pensioners, especially in the early 90s (Figure 2c) growth in the real value of pensions and relative increase compared to the average wage, caused by wage-indexation until 1994 (Figure 2d) The old-age benefit formula introduced after 1991 offered replacement rates of 76 per cent of the last salary. There was no reduction in benefit for early retirement. The formula for the basic pension was: P = 0.242W + (0.013T + 0.007N)B Where: P - monthly pension X- set at 0.91in 1992and graduallyincreasedsince 1993,to 1.00from January 1999 W - average,gross,economy-widemonthly wagein relevant quarter T - total years of contributions N - other eligibleyears B - individualassessmentbase The old-age pension was split into three parts: * A flat component, equal to 24 per cent of the reference wage, adjusted by a coefficient (k) * An earnings-related component, equal to 1.3 per cent of the applicant's assessment base for each year of contributions paid 4 Thispaperdoesnot discussnor coverthe farmerpensionsystem(KRUS). 7 * A supplement of 0.7 per cent of the applicant's assessment base for each year of noncontribution during the career. Other eligible years (e.g. bringing up children, university education) may not exceed one third of contribution years. The individual assessment base equalled average monthly earnings over a period as indexed for inflation. In 1993, the employee chose the best three consecutive years from the last 12 years; each year since, one year was added to the averaging period, until in 2000 it reaches 10 years from the last 20. The pay in a chosen year is compared to the average, economy-wide wage for that year. The resulting ratio, capped at 250 per cent, is multiplied by the indexed figure for economy-wide earnings, reduced with X coefficient to derive the salary base for the averaging process. The minimum pension was established in 1991 at the level of 35 per cent of average salary. Indexed as other benefits since then, in 1999 it stood at PLN 450.71, which represented 33.4 per cent of average salary, net of social security contributions. Additionally, a supplement of 10 per cent of the average wage is payable from age 75 onwards. Expenditure on retirement and disability benefits' grew, from 12.6 per cent of GDP to 15.4 per cent by 1994 (Figure 2a). Pension expenditures in Poland were higher than those in the European Union, where in 1996 social security expenditures (old-age and survivors) amounted to 12.3 per cent of GDP (44.8 per cent of total social protection spending)6 . This share was different across member countries reflecting differences in the social protection systems, demographic change and other social, institutional and economic factors. The highest expenditure could be observed in Italy - 16.2 per cent of GDP, and the lowest in Ireland - less than 5 per cent, reflecting the demographic situation in the countries. The former has the oldest, while the latter has the youngest population in EU. Expenditures on old-age benefits are expected to increase, due to the population ageing process, experienced in most of the EU countries. The same situation can be observed in Poland. Demographic forecasts show a significant increase in the ratio of older people to working age population after 2010, as a result of post-war baby-boom cohorts, with a further increase in the dependency ratio after 2040, caused by falling fertility rates and the ageing of the baby-boom generation from 1980s. In the 1990s, more than 50 per cent of social security expenditures were on old-age pensions. In the absence of reform, projections show a small decline in spending on old-age pensions until 2003 and then an increase to the level of 8 per cent of GDP by 2020 and stabilisation at this level thereafter. In the same period, expected contribution revenues (at current rates) would drop to 4 per cent of GDP, creating annual deficit in old-age pension system of 4 per cent of GDP (Figure lb). The same pattern could be observed within the disability fund, where projected expenditures would rise even earlier, due to earlier impact of ageing population. The contribution rate, needed to finance only the old-age system would Includingboth employees'and farmers'socialsecuritysystems.The employeepension systemexpenditures peakedat 13.12%of GDP in 1994. 6 Eurostat(1999) 8 have to rise from the current 24 per cent to 42 per cent in 2050.By this time, the number of people aboveretirement age would be double the current level, while the number of people in working age would decreaseby one quarter. Not surprisingly, the crisis of the Polish pension system became apparent not only to social securityprofessionals,but also to the public and politicians. Figure1. Polishpensionsystemin absenceof reforms la. Demographicestimatesand projections,1995-2050 0,9 System dependency rate, 0,8 current retirement age 0,7 System dependenc rate, - 0,6 0,4 0,3 0,2 o,itDemographic ..HHHHHHHHHHHHH_..._ U( C 5 0 ( 0 N 05 0 50 (N (N v 0 CN dependency rate e 0 N 0 CN Cn (N N 00 C( (N e ( 0 N C 0 (N 0 (X q `N 0 N 0> C( (N ( lb. PAYG old-agesystemexpendituresand revenues,1999-2050 8% ': - 7%; 4"_ 6% L _ 5% ; 4% 3% 2% 1% 0% _ 2003 2008 2013 2018 2023 2028 2033 2038 2043 2048 revenues + expenditures ' Source: Office of the Government Plenipotentiary for Social Security Reform 9 Figure2. Recentdevelopmentsin the Polish pensionsystem 2a.ZUS pension expenditures, 1992-98(% GDP) 14%~ 120/ A 4% 2c. Number of new entrants to the pension system, 1990-98(thousand) 600 .~~~~~~~~~~~~~~~~I . i0 I4 500 li - _ 900i~\ 200* 20 " U 0% 1992 1993 1994 1995 1996 1997 1998 1990 1991 gold-age []disability MSiRAtVOrs - 1992 1993 1994 --w old-age 2b. Number of contributors and beneficiaries, 1990-98 (millions) 1995 disablt 1996 1997 -- uriors 1998 2d. Average pension to average wage ratio, 1990-98 15 64% 12 62% 9 60% 56% 3 1990 1991 1992 1993 1994 U-Contributors (in) 1995 -- 1996 1997 1998 Beneficiaries (in) 52% - - 50% j1990 - 1991 0~~~~~~~~~~~~~~~~~1 Source ZUS statistics, analysis and forecasting department, Central Statistical Office 10 1992 1993 1994 1995 1996 1997 1998 2. The struggle for pension reform The most contentious point in the pension-reform debate among experts since 1989 was whether the systemshould remain a pay-as-you-go monopoly.Three groups emerged, roughlyreflectingthe world-widedivision of opinion.One group, the 'rationalisers',argued that the defined-benefitpay-as-you-gosystem should be cut back, but remain in the same form. A second group, the 'reformers', argued that the system should be fundamentally changed:reformed,eithertowardsa fully fundedsystemor a multipillarsystem.But both of thesegroups,until the early 1990s,were overshadowedby those who believedthat short-term preventivemeasures- such asunder-indexationof benefits- would be sufficient(the 'nonreformers').However,by 1992-93, this last groupwason the vergeof extinction. The rationalisers'argumentwas that a packageof reformsto the pay-as-you-go system would stop the increasein pension costs, particularlyunder conditions of high economic growth. This packagewould include consistentprice (or approximateprice) indexationof benefits, increasing the effectiveretirement age from its low level and extending the calculationbaseperiod. However,even with these reforms,projectionsshow that the system wouldbe on the edgeof financialsustainability.A significant,adversemacroeconomicchange would causeserious disruptionor even a breakdownof the system. Moreover,the system would not be immune to the demographicchangeexpectedafter 2006. There was a broad agreement- involvingboth the rationalisersand the reformers- that thesereformsshould be introducedas quicklyaspossible. However,the reformersargued rationalisationalone would be neithersufficientnor desirable.Social-insurance reformshould,they argued,not simplybe associatedwith cuts and stringency,but with new opportunitiesfor a generationwith many workingyears beforeit. New horizons should show not only clouds of change,but also rays of new opportunity, which would make the reform more politicallyacceptable.They felt that rationalisation wouldnot be desirablebecausethe crisis causedby populationageingwouldaffectany pension system purely pay-as-you-gofinanced, as it is decidedlymore subject to labour-market, economic,politicaland demographicpressuresthan fundedsystems. Reformers argued for a social-insurancesystem with new opportunities for its participantsand with a stabilisingmechanism,to resist demographicand macroeconomic pressures.In their minds,this requireda move from definedbenefitto a defined-contribution systemand the introductionof a funded componentin the pensionsystem.The first, pay-asyou-gopillar of the new system would be downsizedand made more transparent, by introducinga closer link betweenindividualcontributionsand individualbenefits.In the secondpillar, contributionswouldbe investedin an individualaccountto generatea return. Finally,the existingthird pillarfor additional,voluntarysavingswouldbe developed. The debate between rationalisersand reformersbecame quite heated after 1991.7 Initially, this had little impact on politicians. However, an increasinglynoisy debate, combinedwith growing difficultiesin paying pensionsput pressureon decision-makers. 7 Thedescriptionof the debateis basedon Hausner(1998). 11 Backward-lookingwage indexation of benefits meant that financial pressure sharply increased periodically. This was reflected, above all, in the inability to prepare a budget without a dramatic choice between a huge rise in the budget deficit - undermining recently recovered macroeconomic equilibrium and reversing disinflationary trends - and major spending cuts on important programmes. The successfuldefence of macroeconomic discipline by successive ministers of finance meant there was no choicein practice. It becamenecessary to under-index benefits periodically. With the help of so-called supplementary budget legislation, it was technically possible to limit the expected rise in retirement and disability pensions. In 1991, the revaluation act on pensions introduced several changes to the pension system. The most important ones includedwithdrawal of special benefitsdue to the work in hazardous conditions, restoration of some link between earning history and level of benefit, restrictions of possibilities of combining working with pensions (by imposing wage limitations), caps on the individual assessmentbase at the level of 250% of average wage and a new minimum guarantee of 35% of average wage. Most changes aimed at reducing expenditures, but those remained high and in 1992 exceeded 13% of GDP. Additional measures were taken in 1992by cutting the wage base used for calculatingbenefits from 100% to 91% of average wage. Pensioners and their representativesopposed this policy. It became a major political burden and partly accounts for the sharp fall in support for the post-Solidarity government, and its eventual collapsein 1993. The post-communistopposition that won the elections promised, among other things, a return to 'fair' benefits.However, the new government facednot only the same difficultiesas before, but also some new ones. Public protests had been accompanied by formal appealsto the constitutional tribunal which repeatedly ruled in favour of those who questioned the amended regulations. The new parliamentary majority could have formally overruled the verdicts of the tribunal. In most cases it did not, feeling bound by its election promises. The verdicts of the tribunal came into force and the state's unpaid debts to pensioners grew rapidly, forming a significantpart of public-sectordebt. The constitutional tribunal consistently ruled the practice of repeated temporary suspension of the state's commitments to pensioners as unconstitutional. At the sametime, it clearly stressed that this did not preclude the possibility of a permanent change in the regulations, provided there was appropriate legislation. Thus, it was only when legal and political factors prevented ad hoc manipulation of the pension system that the warnings of experts and the idea of major reform was taken seriously. The coalition pact of the SLD-PSLgovernment, which assumedoffice in the autumn of 1993, includeda very generalcommitment to reform the social-insurancesystem. The ideawas to improve the existing system rather than changeit radically.Therefore, the plan was geared more towards protecting accrued rights than limiting them. Only in June 1994 was the idea of radical reform - introducing a new mandatory funded pension - suggested, in the 'Strategy for Poland' economic programme. The 12 government and parliament acceptedthe strategy presentedby the deputy prime minister and minister of finance (G. W. Kolodko8 ). As a result, in 1994 the new law introduced price indexation of benefits, simultaneously increasing the X coefficient in pension formula from 91% to 93%,and then by 1% per year thereafter. Finally, the base reached the levelof 100%in 1999. The heateddebate between the minister of finance (the 'reformer') and the minister of labour (L. Miller - the 'non-reformer' turned 'rationaliser') lasted a year and a half (from mid 1994 to the beginning of 1996). It was mostly about their competing visions of reform. The ministry of labour set out a plan for limited rationalisation and reorganisation of the pay-asyou-go system,with a marginal role for the funded pillar. The plan consisted of X reducing the growth rate of pensions paid to the uniformed sector (police, army, etc.), by moving to the sameindexation rules applied to other employees * graduallyincreasingthe contribution rate to the agriculturalsystem * introducing new rules for disabilityqualification * gradually increasingthe retirement age These proposals, adopted by the government in May 1995, were then submitted for public consultation. However, the minister of finance regarded them as much too timid. Instead an alternative model was prepared, based on the Chilean reform. It envisaged the replacement of the pay-as-you-gosystem with a fully funded plan and the introduction of a minimum state pension. However, this programme was never submitted to government. It was presented only with the intention of stimulating a debatethat might produce alternatives to a modified pay-as-you-gosystem. It largely achieved this goal: public-opinion research showed the ministry of labour's plan was perceived as conservative and that people expected reform that would be more decisive.From the initial plan, only the change to disability qualificationrules was implemented9 . After the consultation process, the government recommended a revised programme giving a greater role to the funded component in the autumn of 1995. This was submitted to Parliament for debate at the end of 1995,but the minister of finance remained opposed. The debate took place in April 1996,by which time a new prime minister (W. Cimoszewicz)and a new minister of labour and social policy (A. Baczkowski)were in office. The latter was also appointed the first plenipotentiary for pension reform. He strongly supported fundamental pension reform, and so announced in parliament that the government's programme still needed final 'touches'. In particular, he announced his intention to develop the idea of funded pensions,which had been included in the opposition's plans (Solidarity's proposed solution). In 1996,as a result of protests from pensioners, the indexation rules were made more generous. Accordingto those rules, the real growth of pensions would be decided annually in SeeKolodko (1996). Accordingto newscheme,disabilityis grantedon the basisof incapacityto work, not healthloss as it was before. 13 the state budget. The successful passage of the bill through parliament was helped by key personnel changes at the ministry. Baczkowskiput a great deal of effort into getting this bill passedclearingthe way for fundamental reform by easing some of the tensionsconcerningthe pension system. The change in the cabinet meant there were now reformers in both essentialposts of labour and finance minister. They began to co-operate closely. Baczkowski held a unique position in the government because he had originally been a member of Solidarity. In 1992, while still an activist,he had been deputy minister of labour. He demonstratedhis negotiating skill during the post-communist coalition (SLD-PSL)government, and became chairman of the tripartite commissionon socio-economicaffairs,establishedat the beginningof 1994. Quarrels in the governing coalition had caused some members of the SLD leadership, including the prime minister to establish a dialogue with the opposition. Awarding Baczkowskia ministerial post was a good way of winning their trust. With the support of the prime minister, the finance minister and the encouragement of the opposition, he began work on a completelynew reform program. However, for political reasons, it was presented merely as an update and an expansion of the previous proposal. The office for pension reform - a team of experts assembled by Baczkowski - prepared the programme. The new pension reform programme, SecuritythroughDiversity was published in February 1997,three months after the shocking, sudden death of Mr. Baczkowski. Security throughDiversity was wholeheartedly embraced by Baczkowski'ssuccessors, Jerzy Hausner (February-September1997) and Ewa Lewicka, who took over in November 1997 after the return to power of the Solidarity-based coalition. Their sincere conviction regarding pension reform and their professional and political efforts made it possible for reform to proceed. The legislativeprocessthat followed the SecuritythroughDiversity program was also divided between the two governments. The first set of laws included: - Law of 28 August 1997on organisation and operation of pension funds - Law of 22 August 1997in employee pension programs * Law of 25 June 1997 on using privatisation proceeds to support pension reform Parliamentary discussion on the above laws was fairly short, mostly due to the fact, that the laws created new elements in the pension system, and did not change any of the existing rules. The laws gainedthe support of the tripartite committee, which also allowed for quick legislativeaction in Parliament. Though the rest of the bills were not legislated,the two that were passed announced the date of the reform introduction as of January 1, 1999. The reform calendar started in August 1998,when the licensingprocess for the new pension funds started. Everybody also agreed that privatisation revenues should be used to finance introduction of the funded component of the system". The second set of laws included: * Law of 13 October 1998on social security system 0 Sucha postulatewasalsoformulatedin an earlierSolidarityproposal. 14 * Law of 18 December 1998 on old-age and disability pensions from Social Security Fund It took half a year to formulate the draft laws. During this period the reform team focused on preparing detailed proposals, including re-drafting the old-age and disability pensions law in order to unify existing regulations by including all the arrangements in the old system". This was also a time of political consultations both within governmental departments and with the tripartite committee. The latter was especially important. As a result of those meetings several changes were introduced to the initial proposal. The most significant change concerned the retirement age. The initial reform proposal of equal minimum retirement age of 62 was controversial. Some conservative politicians and trade unions were attached to a more traditional view of women's role in society and proposed differential pension ages of 60 for women and 65 for men They argued, that already the effective retirement age would increase, if current rules remained (60 for women and 65 for men), due to the withdrawal of the early retirement privileges. This especially concerned women, who could retire at the age of 55, after contributing for at least 30 years. As a result, the final drafts sent to the Parliament included different retirement ages for men and women. Despite several attempts, this was not changed during parliamentary debates12 Another change grandfathered the old system for those retiring through 2006. This regulation was a result of demands fromntrade unions, not to withdraw early privileges from those who planned to retire within the next few years"3 . Also, this change was in line with Constitutional tribunal verdict to recognise accrued rights for early retirement. The government agreed that a person has a right to expect his or her retirement rules not be changed 8 years prior to retirement' 4 . This regulation was questioned by trade unions, who demanded a longer transition. Thefirst draft of the old-agepensionslaw includedonly paragraphsfor the populationcoveredby reformed PAYG.However,this opportunitywas usedto unify the existingregulations,as wellas to introduceseveral rationalizingchanges,regulatingexistingpractices,whichwerenot specifiedin the law.Unificationalsohelped to shapethe samesolutionsfor all coveredgroups(for instance,the eligibilitycriteriafor survivorbenefits wereunifiedfor minersand other occupations,someregulationsfrom decreeswere introducedto the laws, adjustinglegislationto Constitutionalrequirements). 12 However,upon the introductionof the reform,women realisedthat lower retirementage with actuarially reducedpensionsresultsin significantlylower pensionsat the ageof 60 comparedto age65. In mid 1999the Plenipotentiarystartedworkingon alternativeproposal,allowingwomen to retire at later agewith Labour Codeprotectionof the workplace,forbiddingemployersto firefromwork on basisof age. 13 Most occupationalgroups enjoying early retirementschemescould retire up to 5 years prior to legal retirementage.However,severalgroupshad loosercriteria.For example,minerscouldretireafter 25yearsof workingundergroundregardlessthe age,teachersafter 30 years of working,also without age limit, ballet dancerscouldretire at 38.Introductionof a year2006ruleallowsall those groupsand three cohortsof those retiringat 55to drawtheir pensionunder old regulations. 4 Approximately1/3 of requiredtenure. It wasdeterminedas a minimumvestingperiodfor acquiredrights, whichwas in 1999confirmedby the Constitutionaltribunal, that overruleda complaintfrom the railway workersthat the pensionreformchangedtheir acquiredrightsto earlyretirement. 15 This part of the legislative package was very difficult, both from the technical and social point of view. Intensive consultationsled to numerous modifications of proposals. On the grounds of lack of time to formulate final opinion due to constantly changing drafts, OPZZ refused to present an opinion within tripartite committee. The final opinions of trade unions were send directly to the Parliament. Both OPZZ and Solidarity expressednegative views on the changes in the PAYG old-agesystem, especiallyfocusing on early retirement issue. The government decided to continue the legislative process, despite the negative opinion of the trade unions. One of the goals was to have the general law enacted before the scheme replacing early retirement was negotiated. The proposals were presented to the Parliament in April 1998. Discussionsin Parliament were much longer than in the case of the first package.This was expectedfor two main reasons: * The laws changed the functioning of current system and as such, were more difficult to pass. Parliamentarians expressed their concern about the reduction of replacement rates and intra-generationalredistribution in the PAYG system * The opposition objected to the changes, by extending discussionsin the Parliamentary Committee. Because it could not formally object to the reform, as the SLD-PSL governmentinitiated the reform process,this could potentially slow-down and, eventually, postpone the reform * Trade union representatives in the Parliament continued to push for drafting solutions replacingearly retirement before the laws were passed Parliamentary discussion led to additional changes in the law. The major ones included: changes in the notional accounts indexation from discretionary rule as annually defined in state budget law to fixed 75%growth of wage bill, change in the benefit indexation rule from at least prices to at least prices plus 20 per cent of real averagegrowth, permission to combine full disability benefit and earningsin the labour market'5 , change in the coverage of military forcesfrom everybody under age30 to only new entrants. The Parliament accepted the laws by the end of 1998 allowing the reform to be launched in January 1999. 5Thegovernmentproposalalloweddisabledpeopleto work in labor-protected companies. 16 3. Description of the new pension system 3.1 FINANCING AND OPERATIONS Under the new system, old-age pension benefits can be derived from three pillars"6 . The first and second pillars will be universal and mandatory, the third voluntary. The first will be pay-as-you-go financed, the second and third funded. Contrary to the old system, the two mandatory pillars will be based on the defined contribution principle, where benefits are linked to accumulated lifetime contributions and earning returns based on either financial returns or wage bill growth. Benefits also will be affected by changes in life expectancy at the retirement age. Figure 3 summarises the differences in the financing and functioning of the two mandatory pillars of the new system. The fundamental concept underlying the reform is that security comes from diversification of the sources of pension income, hence 'security through diversity'. The first and second pillars are linked to the labour and capital markets, respectively. The rate of return in the first pillar is linked to the rate of growth of the covered wage bill, in the second and third, the rate of return on investments. There is some evidence that these are not perfectly correlated"7 . In this case the system is more stable. The long-term target is that half of the system will be funded and half pay-as-you-go. The target contribution rate for the old-age system was 18 per cent of wage, net of contributions. Half of this is shifted to the second pillar. The rest is paid to the pay-as-you-goNDC pillar. The first pillar contribution would be steadily reduced until reaching 9 per cent. This level of funding was both desirable from a diversification point of view and affordable from the fiscal perspective. Greater funding creates a short-term cash-flow deficit in the first pillar, financed, according to the law, from privatisation revenues. On the other hand, participation in the funded pillar reduces the accumulation of implicit debt for the baby boom cohorts, which helps to maintain balance in the pay-as-you-go pillar in the future. The current pay-as-you-gosystem is closed for those born after 1948 and converted to a 'Notional Defined Contribution' system, forming the new first pillar. All people born after 1968 are covered by both pillars and those born between 1949 and 1968 have a right to either participate in both pay-as-you-go and funded pillars or to choose to be in the new first pillar only. Both pillars are mandatory for new entrants to the workforce. Public opinion showed support for reform. Some 73 per cent of people agreed that pensions should be closely related to contributions and the length of time they were paid, and 68 per cent that pensions should be derived from employee contributions, capitalised over their working life"8 . 16 Disability and survivor pensions remain in the public pay-as-you-goscheme. 17 jagannathan and Kocherlakota (1996) cite evidence from the US. Palacios (1998) shows the correlation between annual wage growth and equity returns to be close to zero for four OECD countries. 18 Results of opinion polls, conducted on the request of Government Plenipotentiary for Social Security Reform, April 1997. 17 Figure3. Financingand functioningof the new pensionsystem PAYG DB FUNDED old system DC 1 St pila 2 nd pillar NEW SYSTEM 3.2 THE CONTRIBUTION RATE Under the old system, contributions of 45 per cent of earnings plus subsidiesfrom the general budget (approx. 1.5 per cent of GDP) financed retirement, disability, and other benefits, such as work injury, occupational diseases,sicknessand family allowances.Over half of expenditures (about 24 percentage points of wage bill) were for retirement pensions. In the new system, nine percentage points of the contribution are diverted to mandatory funded pensions. The individual's notional defined-contribution account will be credited with 15 percentage points of the contribution. The remaining 21 percentage points will finance other pay-as-you-gobenefits'9 . Until the end of 1998,employers paid the entire 45 per cent contribution, based on the total wage bill in the companies. Under the new system, the contribution payment system became more complex. First, the contribution was divided into separate parts, reflecting differentlong-term and short-term risks and second, it was divided between the employeeand employer. Starting from January 1999each pays equally for old age and disability insurance, work injury will be the responsibility of the employer and sickness of the employee. Employees' earnings were grossed up to reflect their new contribution, so the change is neutral with regard to net wages and total labour costs The intention was to make financing more transparent to employees. After the reform, the total contribution rate therefore falls to 36.59 per cent (45/(100+employee'scontribution)). Table 2 shows the structure of contributions to the new system. Each part of the contribution will be allocated to separate funds within the socialsecurity fund (FUS). Each sub-fund will prepare its own actuarial forecast, improving the transparency of the system and allowing for better management of all components of social security. These include disability, survivors, work injury and sickness benefits. The paper focuses only on the old-age part of the social security system. A brief description of the rest of pay-as-you-go scheme can be found in the Appendix. 9 18 Table 2. Social-securitycontributionratesas share of grosswage Contribution Total old age disabilit &survivor sickness&maternity work injury | Emplo ee Em lo er 19,52% 13,00% 9,76% 6,50% 9,76% 6,50% 2,45% 2,45%0/ 0,4% to 8,12% - 0,4% to 8,12% The old age contribution is divided between the pay-as-you-go pillar (12,22%) and the funded pillar (7,3%). The contribution is collected centrally and transferred by the Social Security Institution. Contributions are registered on individual accounts in both pillars, forming the base for future benefits. All contributions are tax deductible and pensions paid are taxed in both pillars. One aim of the reform is to cut 'the contribution rate from its current, very high level. This effect is achieved by increasing the effective retirement age from its current level of 57 and lowering the average replacement rate by introduction of notional accounts, which reduces benefits by shifting the longevity risk to beneficiaries. An actuarially fair benefit formula increases incentives to postpone retirement decision. The reform introduces an upper limit for contributions of 250 per cent of average earnings2 0 . Furthermore, the contributions to third pillar employee pension plans are social contribution deductible, up to the level of 7% of individuals' earnings21 . The above changes are expected to have little impact on the contribution revenues. Only small fraction of the population earns more than 250% of the average wage and, with existing cap in the benefit formula, most of them did not report higher earnings to social security system. The 7 per cent reduction also existed in the previous arrangement, where employers could deduct contribution to group life insurance schemes up to 7 per cent of company average salary from the wage bill reported to ZUS. The total effect of the above revenue losses should not exceed 0.3% GDP annually. The social security system covers employed and self-employed. Additional contributions are transferred from the state budget for periods of national military service, nursing disabled child and parental leave. The periods of unemployment are covered from the Labour Fund. Those intergovernmental transfers are estimated to run about 0.2 - 0.3 per cent of GDP annually. 20 This correspondsto the maximumearningscap taken into accountin the old pension formula. Becausethere is no favorabletax treatment in third pillar arrangement,deducting contribution to employee pension schemesfrom socialsecuritycontribution base allowsemployersto establishthird pillar plans without additionallabour cost. 21 19 33. THE FIRSTPILLAR The accumulation phase 12,22% of employee's gross earnings, paid by both employer and employee will be registered in the individual's notional account. These contributions will then be indexed in line with 75% of the quarterly growth of the covered wage bill. Indexation in line with the wage bill, rather than average wage growth allows for better stabilisation of the pay-as-you-go system, as the liabilities grow in line with revenues, affected both by the average wage growth and growth of labour forcer. The sum of uprated contributions, 'virtual' or 'notional capital' will then form the basis for the individual's pension. Indexation in line with the wage bill is designed to give contributions paid in early years similar weights in determining the overall pension value as those paid just before retirement. However, because the notional interest rate is below the growth of the covered wage bill, the contribution rate can be reduced in the future, without sacrificing the financial stability of the system. Once the system is mature, the notional rate of return could be increased to 100 per cent wage bill growth. The benefits in many pay-as-you-go systems are related to earnings only over a short In the old Polish system, earnings prior to 1980 would not influence the individual wage used for pension calculation. Earlier periods would only be reflected in the number of contribution years accrued in the pension formula. The result is redistribution from people with longer working lives to people working for a shorter period, and from people with flat age-earnings profiles (generally manual workers) to people with steeper earnings paths (generally professional and managerial workers). The new system ensures that contributions count throughout the working life, and so removes an undesirable and unintended redistribution. period2 3 . Each participant annually, by the end of March, will receive information about his or her virtual capital account balance. ZUS will provide standardised estimates of the pension value under different assumptions of retirement age24 . Initial capital in the first pillar. People who started their working life before 1999 have initial capital' added to their accounts in recognition of pension rights accrued under the old system. Initial capital will be calculated to deliver the same pension benefit as the old formula (adjusted for age and contributable years), as if everyone retired on the last day of the old system. There will be no differentiation of initial capital between those who participate in both pillars or in one pillar only25 . 22 See:Valdes-Prieto (1999) 23 See Disney and Whitehouse (1999 - forthcoming) 24 25 Old-age benefit estimates will be provided after 2003, when all contributors will have calculated their initial capital. This was not the case, for example in Hungary, where acquired rights were reduced for those who decided to switch to funded pillar. 20 This approach to the pension rights recognition was mainly a result of the lack of appropriate individual data. In the old pension system, the Social Security Institute received individual information only upon retirement. Becausemost of the individual records prior to 1980 were destroyed,this method provided a way to deal with initial notional account status. Also, this allowsfor the gradual reduction of replacement rates in the pay-as-you-gosystem, as the initial capital portion of the notional account decreasesover time. Due to difficultiesin finding appropriate records, the law sets a period of 5 years to calculate the initial capital for all contributors in the new pension system. The formula for the initial capital calculation is: Initial Capital (C0 ) = Po* G6,2,where: G62 unisexlifeexpectancyat the ageof 62 in 1998(209months). Po old-agebenefit calculatedas of December31, 1998,with constantelement 26. adjustedfor work experienceand ageof a worker The G-valueused for the calculation uses one retirement age. If G-valuesfor 60 and 65 were used to calculate the initial capital, women with identical work history would receive 30% higher initial capital. To avoid this, the G-value was set at the average level of age 62. That, combined with lower retirement age for women would create significant drop in pension value between women retiring in the old and new systems in 2008 and 2009 respectively,i.e. women lose from conversion. In order to compensate for this, the first five cohorts in the new system, will receive their pensions according to another transition rule. Namely, pensions granted in the years 2009-2013will be calculated according to mixed old-new pension formula. This formula appliesto those women, who will not participate in the funded pillar. Table 3. Weightsof old and new pensionin the mixedformula Year Old pension New pension 2009 2010 2011 2012 2013 80% 70% 55% 35% / 20% 20% 30% 45% 65% 80% Constant element in the formulais multipliedby adjustmentfactorp equal to 26 min{ A- ,18 where Ai=individual'sageat the end of 1998,A, = retirementage (60for women and 65 for men),C, years of 4 / L)), Cr = required years of contributing (20 for women and contributing at the end of 1998 (=Min((L+S);( 3 25for men). 21 The initial capital calculation is rather generous, because the old system offered significantly higher replacement rates than the new system. However, using current level of life expectancy to calculateinitial capital (as opposed to projected life expectancies)and lower indexation of the accounts allow for reduction of the implicit debt for the cohorts covered by the reform, which results in additional reduction of replacement rates in the new system. Initial capital will count for more than 60 per cent of the benefit for the oldest cohorts covered by the new system, falling gradually for subsequent generations (Figure 4). The impact will be smaller for younger cohorts, until it disappears for new entrants to the labor market. Figure4. Share of initial capital in pensionvalue, cohorts1949-1974 100% .2 ' 80% CL 0. 60% 0, 3240%. *2 40% :,'20% 0% _ i 1 _ l 1949 - 1954 1959 1964 1969 1974 year of birth *i initialcapital[ contributionsfrom 1999 Note: Benefit share from contributions from 1999 includes first and second pillars Assumptions: Average wage earners, starting career at the age of 25 40 years of working career Average wage growth: 4 % Rate of return from funded pillar: 6% Administrative fee from contribution: 5% Administrative fee from assets: 0.6% Annuity company fee: 6% Rate of return on annuity: 2% 22 The benefit-distributionphase The system will have a minimum pension ageof 60 years for women and 65 years for men. The old-agepension is calculatedaccordingto the formula: n p n E, Ci fI i=k j-i (I + rj ) ( . * where: Pn old-agepensionat agen c; contributionin year i rj rate of returnin yearj k ageof enteringto socialsecurity Gn averagelifeexpectancyat retirementagein the calendaryear of retirement This formula adjuststhe level of benefit both to the value of contributions paid during entire working career and life expectancyat the retirement age.The formula still redistributes between men (livingshorter) and women (livinglonger), by using the unisex life expectancy tables. However, it automaticallyadjustsfor increasinglife expectancythat is observed in the Polish population, increasingthe stability of the system. Due to the actuarialcalculation of the benefit,the new system has stronger incentives to continue work afterthe minimum retirement agethan the doesold system. Each additional year of work and contributions will be rewardedwith a clear increasein the net present value of pension benefits,as the accumulatednotional capitalincreasesand life expectancydecreases. Benefits in payment in the first pillar will be indexed to at least consumer prices increased by 20% of real wage growth. Regulations, however, allow for more generous uprating of pensions,which stays at a discretionof annual state budget law. Notional capitalis simply an account of rights and it cannot be liquidated at any point in time. The accountsof deceasedpersons are terminated and form an inheritance gain,which 2 7. is used to increasethe revenuesof the pay-as-you-go system Reserve policy in first pillar The defined-contributionsystem is notional in the sensethat funds are not built up: it is still a pay-as-you-gosystem. Thus, at any time it is dependent on the cohort of workers paying for the benefitsof the cohort of pensioners,and so is vulnerableboth to economic and demographic shocks. While the former cannot be anticipated, the latter, to a certain extent can, and some measures can be taken to make the system less vulnerable to demographic Thisis not the onlysolutionpossible.For example, in Sweden, the notionalcapitalof deceased personis dividedbetweensurvivors andregistered on theirindividual accounts 27 23 changes. To stabilise the contribution rate in the system in the face of demographic fluctuations, reserves will be set aside. Reserves will be built up when a large, 'baby-boom' cohort is working, and drawn down when it retires. The reservesystem is equivalentto partial funding of the pensions' systemsfirst pillar. The 'buffer fund' or so-calleddemographicreserve fund, will consist of any surplus in the first pillar and one percentagepoint of wage bill (approx. 0.35 per cent of GDP) transferred to the demographic reserve fund in years 2002-2008.The demographic reserve can be additionally supplemented from privatisation revenues, if stated in separate legislation. Interest, and any extra revenues, will also be added to the fund. According to estimates, assets of the reserve fund will reach 14per cent of GDP by 2020 (Figure 5). Figure5. Demographicreserve inflow and assets, 1999-2020(% GDP) 2.0% 1.8% -1.6% 1.4% - a S - 14% 12% - 10% 1.2%- L8% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 1999 -6% -4% -2% g 0% 2004 2009 2014 2019 accumulatedassets (right-handscale) surplus in pensionfund (left-handscale) -contributions 2002-2008(left-handscale) Source:Social Budget Model, Assumptions: rate of return: 3.4 - 3.2%, no decline in contribution rate The aim of such reserve policy is to ensure that the pension system is entirely selffinancing, will not need subsidies from the general budget (aside from those, that are purposely designed)and that contribution rates could be steadily decreased in the future. The demographic reservefund will be managed by the social insurance administration until 2002. After this date, fund management will be contracted out to private asset managers,following a tender procedure.The law allows one assetmanager to manageup to 15%of total assets of the demographicreservefund. The investmentlimitations are similar to those of the open pension funds in the second pillar (see section 3.3). Additionally, the investment policy should by guided by a 50 year forecast of revenues and expenditures of the pay-as-you-gopension fund. This requirementwas introduced to ensure an appropriate level of liquidity in the system. The periods of surpluses and deficits in the pay-as-you-go system continue for several years, following demographic developments. The reserve accumulation and investment period should followthe projected needs of the pay-as-you-gopension fund. 24 A new rolefor the socialinsurance administration (ZUS) The role of the social insurance administration also changed following the reform to the pension system. Under the new system, the social insurance administration is responsible for: * Managingindividual'snotional accounts * Calculatingand paying out first pillar pensions Managingold pension system for people born before 1948, calculating and paying out a pensions * Managingother parts of social security system - disability, survivor, sickness, maternity, work injury and other benefits 2" * Collectingall social security contributions * Keepinga databaseof all contributors29 * Keepinga databaseof all employers and other contribution payers * Transferringcontributions to open pension funds * Supervisingthe (contractedout) management of the demographic reservefund. ZUS needs substantialrestructuring to meet the challengeof the new pension system. By 1998 ZUS employed around 40,000 people in its headquarters, 51 branches and more than two hundred inspectorates. Each of the branches had significant independence, including separate Supervisory Boards. There was little or no communication between branches". Lack of qualified personnel also resulted in a weak position of ZUS in Labour Court, which almost always took side of beneficiaries31 . The social security system law changed the institutional structure of ZUS. As of January 1, 1999 the institution gained legal entity status and was no longer a part of public administration system. This change was a prerequisite to introducing better resource management and proper accounting principles3 2 . One of the most important changes was the creation of a more centralised management structure. The new social security law abolished the Supervisory Boards in all ZUS branches 2B ZUS also collects health care contributions and Labour Fund payments from employers, 29 Also for the purpose of health care system. 30 As a result, for example, there were cases when people would draw two or more pensions. 31 3 For instance, ZUS lost a case, when plaintiff suffered a work injury working in one place, simultaneously receiving sickness benefit from another employer. Though ZUS argued, that one should not work on sickness leave, the Labour Court ordered to pay both sickness and work injury benefit. For example, as a part of public administration, ZUS was not obliged to account for amortization of fixed assets,which lowered reported costs. 25 and introduced one General Supervisory Board that consists of 15 people (5 representatives of: government, employers and trade-unions and pensioners organisations).The new Board was established in April 1999. The responsibilities of the Board cover the following: * setting the rules of management operations * assessment of the management members * approving yearly financial reports of ZUS * evaluation of annual financial plans of FUS (social insurance fund) and demographic reserve * evaluation of draft laws in the field of social security * assessment of the salary structure of ZUS employees (including management) * choice of the auditor * evaluation of the candidate for the President position * evaluation of ZUS statutes. Introduction of the individual accounts and transfer of contributions to the second pillar required the design of a new IT system in ZUS. Earlier attempts to introduce the individual accounts were altered, in order to fit the requirements of the reform. The most important change led into the introduction of a centralised database to be able to instantly process the information received. According to the law, ZUS has to transfer contribution to the second pillar within two days from receipt of information and contributions from the employers The decision to leave collection of the funded pillar contributions to ZUS was made for several reasons. The most important was that ZUS reports better collection than the tax revenue service in Poland. Both systems required significant changes to allow for monthly information registration, and ZUS was already advanced in preparation of a computer system that could cope with such a task. The Polish government also decided to collect health care contribution through the social insurance administration, which lowers the overall cost of social security management and administration3 4 . In the future, contribution collection and benefit payment could be separated under two independent institutions, which would help to divide between the clearinghouse role of ZUS and the role of pay-as-you-go social security system manager. One of the most important implementation issues was the identification of contributors and employers in the system, as there is no unique social security number for Polish citizens. At the moment there are two databases for individuals and two for companies operating in Poland. It was decided, that ZUS will use those databases to identify participants in the social security system, rather than introducing a new identification number. These are: It is 5 days in 1999, 4 days in 2000, 3 days in 2001 and on two days from 2002. ; See Demarco and Rofman (1998) for discussion of contribution collection and transfer. 26 * Personalidentificationnumber (PESEL)for individuals * Tax identificationnumber (NIP) for individualsand employers * REGON - enterpriseidentificationfor employers. Becauseall the systems include mistakes, ZUS IT system (KSI)is designedto use two identificationkeys for both databases.Yet, the first months of implementation proved that even using two separate keys does not eliminate all mistakes. Problems with proper identificationof individualsstill occur. Centralised collection and running individual accounts required solving such issuesas registration of individuals in ZUS, registration of pension fund members, contribution collection mechanismand individualaccounts identification.Final solutions adopted for those issuesand the elements of the IT system are presentedin the appendix. 3.4 THE SECONDPILLAR Nine percentagepoints of salary net of contributions (7.3%of gross salary) is diverted to a pensionfund chosen by the participant. Each person can select only one fund. There is a free choice between the funds: they are not permitted to refuse entry or restrict the right to transfer to another funds, either directly or indirectly,through the imposition of charges. A retiree will be mandated to buy an annuity. Annuities will be provided by 35 . The chosen option is similar to the one in Argentina, where specialisedannuity companies the contribution collection is centralisedand there is a link between the pension fund and the managingcompany.Differencesoccur in the disbursementperiod, as in Polish system there is a mandatory annuitisation, whereas in Argentina participants can either buy an annuity or go for the scheduledwithdrawal option36 . Regulation and supervision Licencesare issued to both managing companies and pension funds by the pensionfund supervisionoffice (UNFE).Pension fund managersmust meet a number of requirements: * a minimum of Euro 4m paid-upcapital * requirementsfor the probity of shareholdersand board members of managing companies (for example,they may not be convicted criminals or in arrears with tax or social security payments) * shareholdersmay not directly or indirectly hold stakes in more than one pension-fund company The annuitylawwasnot legislatedby mid-1999,andit ispossiblethat the finalsolutionwillbe different. See L. Thompson (1999) and Rofman (1999) 27 * individuals holding influential positions in capital markets cannot serve as a director of a fund or work for the supervisory agency Any changes to the fund manager's shareholders, board members, articles of associationor custodian must be reported immediately to the supervisionoffice. Initially, each company will be able to manage only one pension fund, except in the case of liquidation or merger, when more than one fund may be operated for a transitional period of a year. The fund's articles of association must be submitted to the supervision office for approval. Any proposed amendments must be published five months before introduction, and again must be approved, except when a shorter period would be in the members' interest. Pensionfund operations Pension funds operate much as other open investment (mutual)funds. Contributions are converted into 'settlement units' (or a share of the fund) on a date of conversion, at least four times a month. This generates a relatively smooth flow of assets into the fund and prevent monthly cycles in securities markets because of periodic demand from funds. (There will be a substantial surplus of contributions over benefit withdrawals for at least 20 years.). The settlement unit is valued daily and published in major newspapers. The fund's value is assessedprimarily on the basis of market prices, according to rules set by regulating decree and the supervision office. The balance on retirement will be calculated as the number of accumulatedshares (or settlement units) multiplied by the unit value five days before the funds are withdrawn. The legal form of pension funds. Pension funds are legal entities. This clarifies ownership, and the rights and obligations of participants and managers. The alternative that the fund is commonly owned by the participants - would demandmajor modificationof the joint property concept of the civil code. Currently, Polish law regulates trusts through contract law, but this does not adequately cover the relationship between beneficiariesand trustees. Having the fund as a single legal entity ensures that the property of the fund, the participants and the pension-fund company are all kept separate.It strengthens participants' rights in the case of insolvency of the managing company. Regulationsaffectingthe managing company should not affectthe fund's property. Each fund is responsible for running individual accounts for its member. This can be performed either in-house or it can be contracted out. The decision depends on the pension fund managers.Most of the establishedfunds, decided to contract-outthis activity, to separate transfer agents that will handle databasesof individual accountsof pension fund members. Portfolio decisions. The investments of the pension funds are determined by each managing company, within investment limitations specified by decree by the Minister of Finance. Assets must be bank deposits or publicly traded securities,including securitiesissued or guaranteed by the treasury or the central bank (the National Bank of Poland) and investment funds. Funds may also invest in non-traded bonds, but not derivatives,except as a means of limiting exchange-raterisk in foreign investments. 28 Pension funds are not allowedto hold more than 5 per cent of assetsin the securitiesof one issuer, exceptfor mutual funds, short-term bank depositsand public-sectorsecurities.This is designedto ensure a prudent level of diversification.To avoid possible conflictsof interest, the fund may not be invested in securities issued by a pension-fund company or its shareholders, as well as their controlled, controlling or associated entities. Funds are not allowedto invest in real estate or commodities. Limits set out as to where they can locate their investments include: 40% in quoted stock, 5% in foreign securities, 10% in the secondary stock market, 10% in National Investment Funds (NFIs), 10% in National Bank of Poland papers and 15% in municipality bonds, 10%in close-endedinvestmentfunds, 15%in open-endedinvestment funds. Starting on 1 January 2005, every pension companywill be able to operate two types of fund. Type 'A' will invest as above, while type 'B' will be restricted to fixed-income securities.People approaching retirement age will then be able to select a lower risk fund, albeit at the cost of lower expected returns. An individualcannot split his capital between 'A' and 'B' funds. Fee structure. A fund manageris allowedto chargetwo types of fees: 3 3 a managementfee from the fund' assets, that must not exceed0.05 per cent of asset value per month (0.6 per cent annually) and the fee must be definedin the articlesof association. Most of the funds in their articles of associationdecidedto chargethe maximum amount definedpercentagecommissiondeductedfrom contributionsby the company,which must be the same level for all participants,with reductionspermitted for contributors who stay longer. In early 1999 those fees amount to 7 to 9 per cent of the contribution and are usually reduced to around 5 per cent after two years of participation in the fund. The law doesnot specify a maximum value of this fee Additionally, fund assets may only be used to finance some of the fund's operations, such as capitalmarket activities and safekeepingof assets,includingthe custody fees.The goal is a transparent fee structure, to allow members and potential membersto compare costs. The discount for long tenures in a fund is to discouragetransfers. The real rate of return depends on the length of participationin the fund (Table4). For those who accumulatein the pension fund for 10 years, feesreducetheir actual rate of return by almost 2 percentagepoints, while actual returns of those who save for more than 30 years is lower by approximately 1 percentage point. Taking this into account, people with shorter accumulationperiods will not gain as much by switchingto the funded pillar. 29 Table 4. Impact of fees on the real rate of return Fees: Administration fee on contributions 6.0% Administration fee on assets 0.6% Rate of interest 5.0% Real rate of return vs. years of savin s: 10years 3.22% 15 years 3.63% 20 years 3.83% 25 years 3.95% 30 years 4.03% 35 years 4.09% Assumption:flatearningsprofile Custody. The fund's assets must be held by a custodian or depository, selectedby the fund and confirmed by the supervisory agency. The depository must be a bank with at least Euro 100m of assets and no capital affiliation with the pension-fund company. The national securities depository is also allowed to play this role3 7. The custodian is liable for damages resulting from the pension fund's failure to comply with legal requirements, and must inform the supervisory agency of any irregularity. This should guard against misappropriation of pension-fund assets, with additional security provided by the assetsof the custodian. Rate-of-return guarantee. Pension funds are subject to a relative rate of return guarantee, based on the average return of all pension funds. At the end of each quarter, the supervisory agency will calculatethe averagerate of return, weighted by size of fund, achieved for the last 24 months by all pension funds in operation. Any fund management company which fails to achieve50per cent or four percentagepoints (whicheveris the lower) below the averagenominal return for all funds will immediatelymake additional payments to the fund. These payments will be made in the first instance from a specialreserve of between one and three per cent of total fund assets, depending on the size of the fund. These assets are managed as an integral part of the fund. If the reserves are not sufficient, the fund manageris obliged to pay from its own assets. If the reserve and the assets of the fund-management company do not meet the shortfall in the return, then the fund manager will be declared bankrupt. In the case of insolvency,the guaranteefund will make up the shortfall. The custodian will take over managementof the assetsand participants will then be free to choose another fund38 . In practice,however,noneof the pensionfundmanagerschosethe nationaldepositoryas custodian. 38 From2005,the returnsfortype 'A' and type 'B'fundswill be calculated separately. 3 30 The relative rate of return mechanism in the Polish system has been somewhat relaxed compared to Latin Americanpension funds, as the rate of return is calculated quarterly based on rolling 24-month average(compared to initial Chilean monthly calculation based on 1239). However, it is argued by pension funds that this requirement month rolling average discouragesfrom investingin stock market, which is volatile in short-run and may affect the rate of return guarantee. Guarantee fund. A separate guarantee fund will be established. t will be managed by the national securities depository. Guarantee fund assets come from pension funds payments and returns from accumulatedassets. The total value of the fund cannot exceed 0.1 per cent of all pension funds net assets. A guarantee fund finances shortfalls in the minimum rate of return and other lossesof a pension fund assetsthat cannot be attributed to the pension fund manager. In the case of the deficit in the guaranteefund, state budget covers all its liabilities. Disclosure. Pension fund companies are obliged to inform both participants and UNFE of their activities. Participants have the right to a prospectus containing the fund's articles of associationand the rights of fund participants (as defined by UNFE). They must also be told of changesin the prospectus and the financial results of the fund. Every 12 months and on demand, the fund must give participants a statement of account showing the number of units held and their total value. Reporting to UNFE includes information about the state of assets and the results of investment policy. UJNFEcan also demand other information related to the fund's activity. Pension-fundassetsof married couples.Accumulated pension assets constitute a part of a married couple's common assets. In the case of divorce, the family and guardianship code will specify the division of assets. Assets will not be paid out, but transferred to the spouse's account with the fund. In case of participant death, half of the assetswill be transferred to the spouse's account with the fund. The other half will be paid as a lump-sum to beneficiaries specified by participant, or to the familymembers (spouse,children, grand-children,parents and siblings). During the first 12 or so years after reform, a spouse who does not participate in any fund will be entitled to participate in a divorced or deceasedparticipant's pension fund, unless he or she will have no right to participate in funds. In that case,the inherited funds will be paid as a lump-sum. Transfers between funds. Transfers of all assets accumulatedin one fund to another fund will take place only on the last day of each quarter. This allows a clearing-house mechanism for settlement of transfers out of funds net of transfers into a fund, restricting the need to sell assets to finance transfers. The clearing house will be the national securities depository, which alreadyacts in this capacityfor brokers. If a fund member decides to change the fund earlier than 24 months after the enrolment, he will be obligedto pay a transfer fee to the fund he is leaving. The transfer fee 3 Currently Chile is discussing a change to 36-month rolling average. 31 depends on the number of months of the participation in fund and amounts from 40 per cent to 5 per cent of the minimum wage. Taxation. Contributions are tax deductible, as in the first pillar. The fund's investment earnings are tax exempt. The profits of the pension-fund company are taxed Assetstransferred to claim pension benefits on retirement are not taxed. Only when participants beginto receive their pension will they begin to pay personal income tax on the benefits on a current basis. Thus, the second pillar has an expenditure-tax or EET (exempt-exempt-taxed)treatment. Following the death of a participant, assetstransferred to the spouse'saccount in a fund will be exempt from tax. These assetswill be taxed when the spouse begins to collectbenefits. In all other cases,such as payments to other beneficiaries,including closefamily members, the assets will be subjectto an inheritance tax at 20 per cent rate. Fundedpension benefits Accordingto the draft proposal, being sent to the Parliament, membersmust purchase an annuity from an insurance company when they retire (defined as the time they draw their first-pillar pension). Only licensed insurance companies will be allowed to participate. This market, too, will be regulated strictly, becausepension assetswill accumulatetax-freeand the public sector will guarantee the benefits through an insurance guarantee fund (alongthe lines of the pension benefit guarantee corporation in the United States), that will ensure full payment of benefits in the event of an insurance company's bankruptcy, underwritten by treasury guarantee. The plan at time of writing is to require a licensed annuity company to meet the following conditions: * fully paid-upshare capital of at least Euro 25m required to obtain a licence (which can be increased as the discretion of the council of ministers depending on the insurance company's commitments * insurancecompanies offering annuities would not be ableto sell other types of insurance * annuity-company licences will not be issued until one year before the first participants reach the minimum retirement age (around 11 years after the implementation of the reform) * in addition to prudential norms defined in the Insurance Act, the council of ministers should be able to introduce investment limits for insurers * the state insurance supervision office (PLNI) would be able to monitor insurance companies'reserves and order capital increasesand restrict investments as it seesfit To protect the pension's purchasing power, benefits must be indexed at least to consumer prices, although indexation increases up to average wages growth would be permitted. The annuity rate offered can vary only with the age of the purchaser. Annuity companies will be obliged to use the same life expectancytables, not varied by gender,health or region. Companies cannot refuseto provide an annuity. All companies would have to offer a set of standard pension benefits as follows: 32 * single life annuities, paid until the death of the annuitant * guaranteed (or survivorship) annuity, where benefits will be paid out for at least ten years, to the annuitant's survivors in the case of death during that period * joint annuity, paid until the death of the second spouse, with survivors' pensions at least 75 per cent of the original annuity * joint, guaranteed (or survivorship) annuity, where benefits are paid out for at least ten years, even in the case of the death of both spouses during that period Longer periods of guaranteed benefits and different spouses' benefits will also be possible. However, all contracts must be lodged with the insurance supervisory agency (PUNU), which will have the right to prohibit certain contracts. Individual annuities may only be sold with the written consent of the (uninsured) spouse. At the request of an annuitant, the insurance company is obliged to convert an individual to a joint life annuity. For example, if one spouse retires while the other continues working, the couple may choose to take a single life annuity until the second spouse retires. Second pillar pension funds in 1999 By June 1999 UNFE granted licenses to 21 pension funds and refused to give licenses to 3 companies. There are no other applications waiting, as newly established pension funds would not have a chance to win a significant share of the market. The first mergers are already expected, due to consolidation movements in the Polish banking industry40 . Also smaller funds, that could breakeven are considering mergers. The total size of the market is estimated at 11.5 million members, including 3.8 million people born after 1968, who are obliged to join a pension fund and 7.7 million people born between 1949 and 1968, who may participate in the fund on voluntary basis. The authors expect, that in total around 9 to 10 million contributors will- join 2 nd pillar funds (including both mandatory and voluntary participants). An inflow of 400 thousand new members annually is projected for subsequent years. By end of August 1999 around 6 million people joined pension funds. The structure of the market is concentrated. The top 3 funds have around 70% of all pension fund members and the top 7 have 95% of the market. This situation will lead to a number of mergers and acquisitions in the pension fund sector in the future. At the end, there should be from 8 to 10 pension funds in Poland. However, the final shape and the size of the market will be known by the end of 1999, when everybody makes their decisions about participation. Most of those who joined pension funds at the beginning of the year, were younger people whose participation in the pension funds is mandatory. However, they expressed their Thepensionfundslawdoesnot allowone companyto be a shareholderof morethan onepensionfund. Thus, any mergerof two banksbeingshareholdersof separatepensionfundsresultsin mergerof pensionfunds.In mid-1999 two majorwholesalebanks- Bank Handlowyand BREBank- both beingshareholdersof separate pensionfundsannouncedmerger.This must be followedby mergerof the pensionfunds. 33 preferencesby participatingin selectedfundsearlierthan requiredby law. Somepeopleborn between1949-68seemedto delaytheir decisionuntil the end of the year and in the meantime, their contributionscontinueto flow into the socialinsurancefund and are registeredon the individualaccounts. Table5. Age structureofswitchersbythe endof May1999: Men Women Total Born after 1968 Born between 1959-68 787 514 473 919 942 163 1 729 677 508 825 982 744 Born between 1949-58 140409 209 269 Born between 1949-68 Total 614 328 1 401 842 % of agegroup 45,2%* 17,2% 249 678 (total cohorts 718 094 1 332 422 1 660 257 3 062 099 born between 1948-68) Source:ZUS Mustchoose2ndpillar fund by endof September1999 Note:By August 1999the numberof pensionfundmembersdoubled By the end of July 1999,the structureof pensionfund portfoliosincluded80 per cent in treasurybondsand other government-backed bonds and an averageof 12per cent in stocks and shares.However,this share variedacrossfunds from 1.72per cent to 24.67per cent of total assets.For the first years of pensionfund operations,the portfoliostructureshould be similar,asfundstake into accountthe relativerate of return guarantee,which is encouraging 4". In the future, the shareof portfolio investedin stocks them to investin less volatileassets should increaseto provide significantdiversificationof retirementsavingsand prospectsfor better returns. Investment only in governmentbonds may not provide competitivenet returnscomparedto the notionalrate ofreturn. '* 3.5 MINIMUM PENSION GUARANTEE The new systemhasa guaranteedminimumpension,set at the samelevelas in the old system4 2 . It will be paid at the retirement age to people who have contributed for a minimum of 20 years (women) or 25 years (men). This benefit will top up pensions (the sum of both first and second pillar) to the minimum level. It will be financed from general revenues, not from contributions to the pension system. This policy is designed to separate the redistributive role of the system from the lifecycle reallocation of income. It means that the financing of minimum pensions will be on broader tax base - including capital and transfer as well as labour income - than the rest of the pension system. This is contrary to the old pension 41 42 The first relative rate of return will be announcedin two years time, when the fund members will be allowed to changea fund without a transfer fee. Since 1 July 1999equal to PLN 451,11 (approx.33% of averagewage,net of contributions) monthly, indexed accordinglyto generalbenefit indexationrules. 34 scheme, where minimum benefit was a part of the pay-as-you-goscheme, not a separate guarantee. Due to the reduction of intragenerational redistribution in the new system, the number of pensioners covered by a minimum pension guarantee is expectedto increase.If the indexation rules for the minimum benefit are the same as for other benefits, the share of pensioners covered by this guarantee is estimatedto peak at 17 per cent in 2035and gradually decreaseto 7 per cent thereafter. However, close to price indexation of a minimum benefit would lead to reduction of its poverty protection role, as gradually it would become ineffectivein reduction of relative poverty rates. If the minimum benefit was linked to a fixed percentage of average salary, the share of pensioners covered by this guaranteewould likely remian at the level of around 15per cent of all beneficiaries. Most retirees covered by minimum guarantee would be low-incomepeople with short 4 3. General revenue transfers to finance the minimum are estimated between working careers 0.05 per cent of GDP if the indexation of minimum benefit follows general benefit indexation to approximately 0.1 per cent of GDP, if it remainsequal to 30 per cent of averagewage. 3.6 THETHIRDPILLAR A voluntary third pillar will supplement the universal, mandatory part of the pension system (the first and secondpillars). It will consist of a number of long-term savingsplans and occupational-pension programmes. This makes the system more flexible, allowing each individual to reallocate income across the lifecycleaccording to their own preferences and needs. The third pillar is more flexible than the first two pillars, with choice over the timing and amount of saving and the ability to bequeaththe capital without restriction. Many employers alreadytook out group insurancewith a life insurancecompany, with the plans negotiated individuallyfor each workplace. This insurance is attractiveto employees becausethey avoid the information cost selecting from available products and assessingthe risks of different insurers. Becauseof the pooling of risks among the employeesof a particular company, the adverse selection risk for the insurer is reduced, so there is no need for individual medical interviews or health examinations. The cost is deductible for employers, but employees are taxed on the employer contribution as a benefit in kind. This system will continue as part of the employeepension programs. Growing awareness of the uncertainty over the real solvency of the present pension system is the reason why employees are increasingly willing to agree to exchange current wages for future pension benefits. Group life insurance policies with a set time for benefit withdrawal with optional life insurance are becoming popular. Nevertheless, because it requires employers voluntarily to establish schemes,it is limited in scope. 43 In the Polishcasemostlywomenwith lower education,asthey haveshorterthanaverageworkingcareersand theirsalaryincomeis lowerthanthoseof highereducatedpeopleand menin general. 35 The new occupationalpension plans After the reform, employers will have the right to direct employee contributions to group insurancewith a joint-stock or mutual life-insurancecompany, an occupationalpension fund or to open investment funds. An occupational plan must meet the following criteria * all eligibleemployeeshave an opportunity to participate * eligibilityconditions can only cover the employee'stenure in the company and at leasthalf of employeesmust be eligible e payments on behalf on an individual employee cannot exceed seven per cent of earnings assessablefor social security contributions * benefitsbe paid out from age 60, exceptin casesof death or permanent disability The detailed rules for the functioning of an occupational plan must be defined in a company pension contract negotiatedwith employees'representatives. Contributions paid on behalf of the employee will be a deductible expense for the employer. They will be included in personal income when the employee is taxed, but up to the seven per cent ceiling,they will not be subjectto social security contributions. Additional contribution can be paid by employees, but it is neither income tax nor social security contribution deductible. Although the terms of group insurance or payments to investment fund will be negotiated betweenthe employer and the plan managementcompany,there will be some legal requirements to qualify as an employeepension programme. Occupational pension funds will operate along similar lines as second-pillarschemes. However, there will be fewer portfolio restrictions, no minimum rate of return, more influence for participants on the fund's investment policies, including the possibility of investing all the assetsin an open investmentfund. In the case of setting up an employeepension fund, the employer is also able to make employee participation depending on the contribution of a (uniform) proportion of the company's shares receivedduring privatisationof the enterprise within five years. This limits the risk that a large part of the company's shares return to the market following an initial public offering after the two-year waiting period imposed on employees. This overhang of shares is a significantfear among potential investors. The legislation also aims to make occupational pensions portable when employees changejobs. The employee will have the right to transfer the assetsto the pension plan of the new employer or to an insurance company, open investment fund or a non-employersponsoredplan. 36 Regulation, supervision and taxation Since the third pillar involves a range of different institutions, supervision will be spread across different authorities: bank supervisors, the state insurance supervision office (PUNU) and the securities and exchange commission (SEC). The pension-fund supervision office (UNFE) will cover employees'rights in occupationalschemes. The third pillar will be taxed using the pre-paid expenditure tax approach, where contributions are made out of taxed income, but investment returns and benefit withdrawals will be tax free (i.e., taxed-exemptexempt or TEE). Although this means a similar or even the same present value of tax will be paid, it brings forward the revenues to government to the time contributions are made rather than the time benefitsare received (seeWhitehouse, 1998). 3.7 ESTIMA TEDBENEFITSFROM THEMANDATORY PILLARS The actual value of the benefit from first and second pillars depends on the developmentof labour and financialmarkets. In this section, we provide some estimates of the value of pensions for different cohorts covered by the reform, as well as for different assumptionsof wage growth and financialmarket returns. Replacement rates are constant for people with different wage levels, up to the maximum earnings cap, assumingthat the relative ratios of individuals'wagesto average wage are constant over the contribution period. Projections are based on averagewage earners, but the sameestimate appliesto other levelsof earnings (up to a maximum earningscap). The baseline assumption set for calculationsincludes: * * * * * Average wage earners, starting career at age25 Average wage growth: 4%, thus notional rate of return 3% Rate of return from funded pillar: 6% Administrative fee from contributions: 5 % Administrative fee from assets:0.6% K Annuity company fee: 6% of balance at retirement * Rate of return on annuity: 2% * Current life expectancy. In the transition period, the replacement rates (measured as a ratio between the first benefit received from the old-age system and the last salary) depend not only on the development of the new system after reform implementation, but also on the value of the initial capital.Thus, replacement rates change for each cohort, as presented in Figure 6. 37 Figure6. Projectedreplacementrates in the new pensionsystem, cohorts1949-74 Mixedformula 90% 2008-203 X 50% 0 40% --- 30%20% | 30% ~~~~~~~women| i 1949 1954 1959 1964 1969 1974 year of birth ------ pillars1&2 pillar1 only Due to differentiation in retirement age, replacement rates differ for men and women. The difference between the genders is stable at the level of approximately 20 percentage points, which resultsfrom 5 year differencein retirement age. Additionally, for women born in 1949-1953,adjustment in mixed formula causesincrease in replacement rates, comparedto the initial proposal. For men, replacement rates drops from 76 per cent in the old system, through almost the same level for the cohort 1949, to approximately 60 per cent for new entrants to the system.For women, replacement rate drops sharply from almost 70 per cent in the old system to slightly above 50 per cent for cohort 1954(first that is not covered by a mixed pension formula), then decreaseis more stable and projected replacement rate for new entrants is below50 per cent level. Under reasonable assumptions, replacement rates are higher for those who decideto participate in the funded pillar. The gap between first pillar only and first and second pillar participationincreasesfor younger people, as returns from financialmarket are assumedto be higher than the notional return in the pay-as-you-goscheme. For the youngest cohort with non mandatory participation in the funded pillar, difference in replacement rates is approximately10 per cent The other elementthat depends on the age of participants in the system is the share of the benefit generated by the funded component. Obviously, it will be increasing for younger participants. The increase however, does not reflect the share of contribution diverted to the secondpillar. For the cohort born in 1949,the estimated share of the second pillar annuity in the total benefit reaches 14 per cent and increasesyear by year to exceed 50 per cent for new entrants (compared with 32.5 per cent of old-age contribution shifted to the second pillar). The estimates do not take into account changesin the real benefit value in first and second pillars, that occur due to the indexation of benefits. Draft annuity law enforces annuity 38 indexation at least to prices, whereas in the fist pillar, indexation should not be lower than 20 per cent of real wage growth. Figure7. Share of funded part in total benefit,cohorts 1949-74 100% 80% 'c3 60% 16 40% 20% 0% 1949 1954 1959 1964 1969 1974 year of birth * secondpillarD3first pillar Note: First pillar share refers to sum of initial capital and accumulated notional contributions. Contribution rate assumed to remain constant As the benefit formula is actuarially adjusted, replacement rates increase with the retirement age. A person starting to work at age 25, and continuing until retirement without any breaks can expecta replacement rate of around 44 per cent at the age of 60, 62 per cent at the age of 65 and finally, almost 90 per cent at the age of 70. The marginal increase in replacement rates also depends on the retirement age and varies from 2.5 per cent for 60 year olds to more than 5 per cent for 70 year olds. Thus, both first and second pillars benefit formulae encouragepostponing retirement decision, offering higher replacement rates year by year. This does not necessarilymean that the net present value of the benefit increasesfor each additionalyear, as individual'spreferencesmay differ. 39 Figure 8. Projected replacement rates in the new system j 600%_ ~~80% 6 40%~ 20% 0% 60 61 62 63 64 lEsecondpillar 22.9% 24.5% 26.2% 28.0% 30.0% *first 218% 23.2% i 24.7% 26.3% 28.0% 1 29.9% pillar 65 66 67 66 69 70 37.0% 39.7% 42.6% 45.9% 318% 1 33.9% 36.2% 38.7% 413% 32.2% i 34.5% retirement age Assumptions: see text Estimates of the replacement rates are sensitive to assumptions. In order to investigate the influence of work duration, longevity rates and rates of return, we projected benefits with alternative scenarios (Figure 8). If a working career is 5 years shorter than in the initial projection, the replacement rates are reduced by 6 to 10 per cent, depending on the retirement age and they fall to the level of 40 per cent for retirees at age 60 to 77 per cent for those who retire at 70. In the case of higher gross returns (8 per cent, compared to baseline 6 per cent), replacement rates increase by 10 to 28 per cent for 70 year old retiree and range from almost 55 per cent to 115 per cent. Also, the share of funded component in the initial value of benefit increases to 62 per cent for 65 year olds. Changes in longevity also have an impact on replacement rates. If life expectancy increases by approximately 2 years, replacement rates are lower by almost 5 to 9 per cent and fall to 40 per cent for those retiring at 60 and 78 per cent for 70 years old. 40 Figure9. Projectedreplacementrates in the new system- sensitivityanalysis longevity 9c.Increased 9a.Shorterworkingcareer 100~~~~~~~~~~~~~V% -- 130% _0% 80% ii 0 60%. 4G0 l_. ci 60% 0 40% | f~~~~~~~~~~~~~~~~~~~~~~~ | | 111 1 l140 ! 60 |second pillar |3.9%o Ffirst pillar *f 20% 20% . 61 62 63 64 65 66 2seco 292% 20.3% 219% 23.5% 25.3% 27.1Y 26.7% 286% I 23.4% f0 205% 219° retirementage r 67 60 70 9 68 nd pillar 20.7o 393% 314% 33.8% 36.o Iirst 30.6Yo 32.7% 35.f% 37.6% pilar 193Y 61 63 62 64 65 67 66 68 - 69 70 29. 312% 33.5% 359% 38.6% i16% 23.7% 25.4% 2. 37.0% 302% 32.3% 20.5% 219% 23.3% 24.9Yo_265%_28.3% retirementage 222% 1S4.5Y to baseline ratechangescompared 9d. Replacement 9b.Highermarketreturns ~~~ ~ ~~ ~ ~ ~ ~ ~ Age ~ ~ ~~~_______ ~ ~ Z_________ ~ I 60 2nd increase pillar |_ longevity -6.6% 10.0% -4.7% -6.9% 11.1% -5.0% 62 -7.2% 12.4% -5.3% 63 -7.5% -7.8% 13.7% 64 15.2% -5.7% -6.1% 401 II! 65 -8.2% 16.9% -6.4% 201' 66 -8.5% 18.8% -6.8% 61 * - } 1^ 80% . * * } .. 60% X in Shorter career |Return 0 61 613 64 60 62 6 165 67 68% 69 - ____ ~~~~~~~~ 68 70 68.4% 744 453%49i-9 53.3% 578% 62.9% 86% 418% 260% 329% 35.6% pillar |218% 1111second 29.9% 318% 33.9% 36 2% 38.7%. 413% 263% 232Y 24.7% Mfirstpi(iar retirementage _ Z _ 69 70 41 _ -8.9% -9.4% -9.8% _ -10.3% I ~ 20.9% ~~~~~ 23.2% ~ ~~~~~~~~~~~~67 -7.3% -7.7% 25*7% -8.2% 28.6%| -8.7% 3.8 EARLY RETIREMENTAND SECTOR PRIVILEGES The reform aims to eliminate all privileges in the old universal pension system, with equal treatment of all participants with regard to retirement age, means of paying contributions and calculating benefits. Sector privileges were largely an inheritance from the planned economy era - many were introduced in the 1980sduring the period of martial law - but some had a much earlier provenance (railway workers were given privileges by the Austro-Hungarian emperor in the 19th century). It is estimated, that approximately 24 per cent of the workforce enjoyed some kind of special privilege. Additionally, women could retire at age 55, if they had at least 30 years of work experience. Successful elimination of all privileges would improve the pension system's finances enormously. The 'cost' of early retirement alone is equivalent to 12 percentage points of the total 45 per cent contribution. The problem of privilege-based early retirement is compounded by the possibility to continue working after drawing early pensions, almost without restriction. There is a strong incentive to take the early retirement pension, but then continue to work. This was for example the case of teachers. Most of them after retirement continues working at schools, receiving both salary and old-age pension. In the new system, first and second pillars do not include any special privileges as pensions will be paid for those reaching the minimum retirement age. For those, who work in special conditions, that cannot be performed until retirement age, the period of permissible time of work will be sanctioned in law. After this period a worker will have to change his/her work position or profession and could be provided by help in finding a new work or in changing the qualifications. The medical criteria to acknowledge certain working conditions as 'special' were presented in the report of an independent Medical Committee. On the basis of the statistical and medical data, the Medical Committee determined for selected jobs a maximum period of work, which will not cause a detriment of health in particular professions and positions'.. For the transition period, for those who started their working career before 1999 and could expect early retirement another solution is envisaged. All those, who stay in the old system (i.e. born before 1949) will have a right to retire at lower retirement age and their pension will be calculated according to old rules. Additionally, also those working in professions eligible to lower retirement age (provided they work there until they reach stated earlier retirement age), born between 1949 and 1968, who: 4 Plenipotentiary, Ewa Lewicka asked leading home institutes and international organizations to delegate experts, who could participate in the work of independent committee. The committee presented its report in April 1999. 42 * fulfill all three conditions (age, total working period and working period in special conditions)until 31 December2006will be eligibleto earlierretirementaccordingto the old rules (thisappliesalsoto women) e on January 1, 1999had fulfilledtwo conditionsrelatedto work experience,but they will not haverequiredageuntil 31December2006willbe entitledto earlierretirementagebut their pension will be calculatedon the basisof the new formula.Initial Capital of such people will be increasedby adding to contributableperiods the differencebetweenthe generalretirementageandtheir retirementage.E.g.Pensionofthe personthat will retire 5 yearsbeforelegalretirementagewill be calculatedasif that personworked5 years longer than actuallydid. The old-agepensionof suchpersonwill be lowerthan that of a person retiringat minimalretirementagewith the samenotionalcapitalvalue. If somebodychoosesearlierretirementon the basisdescribedabove,he or she willnot be permittedto join the secondpillar. Secondpillar arrangementsare not suitablefor solving rights from the old system.People that decideto participatein one of the capitalpension funds(2nd pillar) automaticallyforfeitthe right to earlierretirement. All of these groups have the right of choice.The final shape of the solutions is discussedbetweenrepresentativesof government,employersand trade unions,which aim to prepareadditionalsolutionscopingwith specialworkingconditions. For givingup the right to earlierretirement,employeescan join a pensionfund pillar, continueworkinguntil legalretirementage and additionallyhavefinancialcompensation- in differentforms dependingon the list of professionsrequiringearliercessationof the activity, similarto those, who do not havea right to retireearlier. Those born between1949and 1968,who did not fulfil requiredconditionsat the momentof the beginningstartingreformwill not haveright to retire earlierin the new social insurancesystem.Insteadthey will be offeredone ofthe two compensations below: * If the MedicalCommitteedecidedthat the job performedby thesepeoplerequiresearlier cessation of professionalactivity (e.g. miners, pilots, train drivers) special bridging arrangementswill be created,which allow them finishtheir professionalactivityearlier. Sourcesof financingsucha solutionwill be subjectof negotiations.It will not be allowed to combinethis benefitwith employment * If the professionwill not be on the list workedout by the MedicalCommittee,employees willreceivecompensationbasedon the fact,that this job wasrecognizedin old systemasa professionrequiringearlierretirementage.The initialcapitalwill be increasedin the first pillar of a new system,and this will leadto a higherpensionin the future at the normal retirementage(e.g.teachers,railwaycompanyemployees) Bridgingpensionswillbe financedfrom separatecontributionsaccumulatedin a special fund,calledthe 'bridgingpensionsfund'.The fundwill pay a 'bridge'pensionfrom the earlier age at which working career is finished to the minimal retirement age. The additional contributionsfrom the day when the bridgingpensionslaw is enactedwill be financedby employers,whilethe contributionrepresentingthe rightsaccruedin the past willbe paidfrom 43 the state budget. This will make the trade-offs in granting special privileges clearer. The bridging fund managementwill be contracted out to private asset managers. This system will be accessiblefor those, who work in conditions, which by medical criteria cannot be performed until retirement agebecauseof the threat of health loss (e.g. working underground, underwater or in changing microclimate) or because they require high psycho-physical condition (e.g.pilots, engine drivers). Bridgingpensions would be set up as a defined benefit scheme. The actual value of the benefit will be negotiated between employers, trade unions and government. Figure10 Calendarof early retirementschemewithdrawal 01/01/1999 2001(') 2026 2018 01(O1!2007 2045 yearl to Earlyretirementaccording old system rules (2) Early retirementwith new formula (3 pensions Brdging (1)Assumedenactmentof bridgingpensionslaw; (2) It is enough if a person fulfills conditions prior to 2007, and then can retire even later; (3) Maximum bridge: 20 years, this includes also people that undergo retraining; Assumption: work career starts at age 20 Negotiations of bridging pensions started in June 1999. At the time of writing, they were still continuing, thus it is not possible to give any reasonable financial estimates of required benefit level and state budget financing, as the level of bridging pensions is not known. This decision will determine the actual rate of employer and state budget contributions necessary to finance bridging pensions. Additionally, until the individual accounts in ZUS are established,there is no information on number of people covered by specialarrangements(both bridging pensionsand initial capitalcompensation). The envisagedsolution will increase the overall costs of the pension expenditures, as state budget will be required to make additional transfers to bridging pension fund and pensions paid for those who will receive compensation will be higher than assumed in projections. Becauseof the lack of appropriate information, these effectsare not included in the projectionspresentedin further sectionsof this paper. 44 4. PAYG system - financing the transition4 5 The pension reform had a significant impact on the functioning of institutions of the social security system. Starting from the first year, it also affects the revenues of the pay-asyou-go scheme, as part of the old-age contribution is transferred to the funded pillar. This effect depends on the size of switching to the second pillar. The more people decide to switch, the deeper cash-flow deficit in the short term. Savings caused by the pension reform appear only in medium-term perspective, as old system rules are preserved by the end of 2006. Savings generated in the short run result from close to price indexation of benefits, introduced in 1994 (see section 2). Savings resulting from increase in retirement age and reduction of benefits appear after 2006. The level of savings is lower than initially expected, due to several transition regulations which were introduced in the legislative process. This section analyses both changes resulting from the introduction of the funded component in retirement savings and financial implications of transition rules in the enacted regulations. 4.1. JOINING THE SECOND PILLAR Voluntary participation in the second pillar for people born between 1949 and 1968 introduces an element of choice, which is socially, economically and politically desirable (Palacios and Whitehouse, 1998). However, this raises the problem of predicting how many people will choose different options and makes it more difficult to get a clear picture of longterm obligations (Holzmann, 1997). Predicting the decisions of people aged between 30 and 50 on whether to switch from the first pillar to the mixed first and second pillar option is a difficult task. One possible source of information is opinion polls. These suggest that most people will choose to join the second pillar. A second source is experience in other countries, of which the most relevant is Hungary4 6 . Opinion polls there suggested people were very optimistic about funded pensions Projections and forecasts presented in this paper are based on enacted legislation and long-term proposals included in pension reform program (such as lowering contribution rates). For the short and medium-term analysis we used Social 'Budget Model, prepared in Gdansk Institute for Market Economics in co-operation with ILO and Polish Ministry of Labour and Social affairs. It produces 20 year forecast of all social funds, including pension system, based on final legislation outcome. For long-term projections a model prepared in the Office of the Government Plenipotentiary for Social Security Reform is used. This allows for comparison of initial proposal and final version of the pension reform legislation package. 46 However, there are significant differences between the new pay-as-you-go pillars in Poland and Hungary. In Hungary, it remained within defined benefit regime, downsized for those who switch to the pension fund. In Poland, regardless the decision, contributors are covered by the NDC first pillar. 45 and pessimistic about the state scheme (see Palacios and Rocha, 1998 and Palacios and Whitehouse, 1998). In order to analyseincentivestowards switching, one has to analysethe return on 7.3 per cent of contribution in the first and second pillar, taking into account such elements as wage growth and notional account indexation, projected market returns, pension fund and annuity company fees (as discussed in section 3.3, fee structure in Poland affectsto a higher extent pension fund members with shorter savingsperiods). Decision whether to switch or not depends on a differencebetween the projected accumulated capital from 7.3% contribution in the first and second pillar accounts, given anticipated returns on contributions. Projected differences between first and second pillar value of accumulatedpension rights or capitalare presented on Figure 11. Figure 11. DifferencebetweenPolishNDC and funded schemesaccumulation 250,000 - 200,000 ___ 150,000 100,000 / 50,000 -50,000 - 1 - - - - - -pension 6 11 fund (4%) 16 - 21 26 31 36 pension fund (5%) -- 41 pension 46 fund (6%) Assumptions: Wage growth: 4% Notional interest rate: 3% Adrministrativefee on contribution: 6% Administrativefee on assets:0.6% Annuity company fee:6% of accumulatedassets According to projections, if gross returns in the funded pillar are equal to wage growth, it is not worth switching, because the notional rate of return (after costs) is higher. Generally, international evidence shows, that in the medium term, returns of private pension funds are higher than wage growth4 7 . If it is the case in Poland - the breakeven point depends on the difference between the wage growth and rate of return. In the case of one per cent difference, it is 18 years, if the difference is 2 per cent, than this period shrinks to 11 years. 47 See: Palacios (1999) and OECD (1998) for 1984-1996rates of return and wage growth. 46 However, in either case the difference between first and second pillar accumulated assets is not that significant for those with shorter accumulation period from the annuity point of view (projected difference in the annuity value is less than PLN 20 (approx. USD 5) per month. If the differences between the value of the annuity in the first and second pillars do not differ significantly, people for diversification reasons should participate in the second pillar. If the expected market gross rate of return is higher than the notional rate of return, all those who do not have any early retirement privileges should switch. The incentive to switch is higher than assumed in earlier analyses4 ' due to the reduction in the notional rate of return below wage bill growth. Such regulation affects to a larger extent those who will not switch, because their full old-age contribution will be indexed at a lower rate. Those who switch will earn full financial market returns on their contributions in the second pillar. As a result, more switching will take place and this will increase short run social security deficit. The financial effect of switching in 1999 is lower than in later years however, because of the time schedule of switching decisions. Until one switches, his full contribution remains in the pay-as-you-go pillar, increasinig the notional capital and therefore pay-as-you-go liabilities4 9 . As a result, the pace at which people switch to the second pillar has a large impact on the finances of the system. During the first months of pension funds enrolment, the number of pension fund members increased by approximately one million per month (in the period of March - August 1999 total of 6 million participants). In order to estimate the financial impact of switching decisions to PAYG pension system deficit, two projections of participation were prepared (Table 6). One assumes that everybody who is entitled switches, and the other assumes that participation will decline with age. The latter bases on the fact that compound-interest effect means that younger workers will get a higher return from the funded pension, and so have a larger incentive to switch. This age-related pattern also occurred in all other countries introducing a reform involving some element of individual choice: Argentina, Chile, Colombia, Hungary, Peru, the United Kingdom and Uruguay (Palacios and Whitehouse, 1998). However, in the discussed examples switching population covered all age cohorts, while in the case of Poland, switching is limited to those below 50 years of age and people who have a choice still have at least 10 years of contributing ahead. 4S 4 SecuritythroughDiversity(1997c) From the year 2000,the new entrants to the labor market are forced to switch within 3 months. During the decision period, the contribution is kept on the separate account, and after they choose a fund, the nominal contributions are transferredto the selectedpension fund. 47 Table 6. Projectedpensionrevenuesand expenditures,transfer to pensionfundand reservesin 1999-2006(% GDP) YEAR 50% of switchers Pension revenues Pension expenditures Deficit /surplus 2nd pillar transfer 100%of switchers |1199912000 12001 12002 12003 12004 12005 12006 Pension revenues 5.19% 4.37% 4.38% 4.05% 4.04% 4.05% 4.06% 4.06% 5.64% 15.14% 15.14% 4.78% 4.75% 4.73% 4.70% -4.67% 5.92% 1575% 5.59% 5.42% 5.25% 5.09% 4.95% 4.82% -0.28% -0.62% -0.46% -0.64% -0.50% -0.36% -0.24% -0.15% 0.53% 1.22% 1.36% 1.49% 1.61% 1.73% 1.86% 1.97% Pension expenditures 5.92% 5.75% 5.59% 5.42% 5.25% 5.09% 4.95% 4.82% Deficit /surplus -0.74%-1.38% -1.21% -1.37% -1.20% -1.04% -0.89% -0.76% 2nd pillar transfer 0.85% 1.89% 2.05% 2.20% 2.33% 2.46% 2.58% 2.68% Source: Social Budget Model If the number of switchers turns out to be higher than assumed by the Polish government (roughly 50%), the deficit in the Social Insurance Fund (FUS) will be higher than claimed. According to the social insurance law, it must be covered from the state budget. Given the constrained budget deficit, other budgetary expenditures would have to be reduced50 . The final outcome, however, will not be known until the end of the 1999, when the switching process has ended. If participation exceeds expectations, the effect will not be disastrous. As argued in the original Security through Diversity proposal, much of the revenue diverted to funded pensions will return to the state treasury through increased demand for bonds, allowing noninflationary financing of the increased deficit. The other part will be invested in the capital market. In other words, higher household saving offsets the reduction in public sector net saving. These arguments were supported by the OECD's (1998) analysis of the reform in its economic survey of Poland. In 1999 international rating agencies increased country rating for Poland partly as a result of pension reform introduction, Privatisation revenues will be used to help finance the deficit in the social-security system, as specified by the first part of the pension reform package (1997). According to the treasury's plans, PLN 53bn will be available to support the reform (about 14 per cent of GDP), with PLN 4bn allocated in the 1999 budget. This amount will be sufficient to fill the gap in the Social Security Fund, even if more people than initially estimated by Polish government switch to the pension funds. s Accordingto both the Polishconstitutionand the Maastrichtcriteria,the state budgetdeficitmay not exceed the limit of 3 % of GDP. Accordingto 1999state budget, the deficit should not exceed2.4per cent of GDP. The Polish government agreed in July 1999, that next year's figure should be less than 2 per cent of GDP. 48 The deficit in the pay-as-you-goretirement scheme is expected to decrease, due to reduction of expenditurescausedby earlier changesin indexation of benefits and by reduction of early retirement policiesin the new system. As a result, after 2010the public schemeshould have a surplus. The projected surplus is smaller in the case of higher switching and appearsin 2012. Figure 12. Deficitisurplusin the PAYGold-age pensionfund in 1999-2020 2.50% 2.00% 1.50% _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1.00Y cL 0.50% l 0.00% ___________________ -0.50% -___- -1-00% -1.50% , ._ _ -2.00% 1999 2004 .. 2009 2014 2019 everybodyswiches .. e,~age-dependent switching Source:SocialBudget Model As the deficit in the pay-as-you-goschemeis financed from privatization proceeds, and in the case of deficit, mainly from taxes (givenstate budget deficit constraints that cannot be exceeded),the projected deficitfigure ignores accumulatedinterest in the assumption that it is not debt financed. The switching decision has a significantimpact on pension funds' asset accumulation. In 2020, depending on the number of switchers, total value of accumulated assets may vary from 23 per cent of GDP to 35 per cent of GDP (Figure 13). 49 Figure 13. Accumulation of assets in Polish open pension funds, 1999-2020 40% -_- 35% T_ , _ _ 20% ___ _ _ _ _ _ _ _ _ _ _ ___l_l _______511 _lli ____ 30°/ S 10% __l!!_l_ii________i_ 2004 1999 ! 2009 everybody sw itches E age-dependent swiching 2014 2019 * only up to 30 Source: Social budget model As only a small fraction of assets can be invested abroad, accumulation of pension funds will strengthen the Polish financial market. Pension funds are expected to create the strongest group of institutional investors, as other forms of savings are not that developed, due to voluntary participation. As up to 40 per cent of assets can be invested in stocks, the Polish securities market is expected to increase its volume. This process will be accompanied by increased privatization, as privatization proceeds will be needed to finance deficit in the payas-you-go pillar. Pension fund managers plan to invest a significant share of assets into government bonds, which is merely a change from implicit debt to explicit one. Even this effect can positively influence Polish economy, as budget deficit cannot exceed 3 per cent of GDP threshold. In a long run, Polish market will not be able to absorb capital flows generated by the pension funds. Thus, changes in investment limitations in the area of international diversification will be required. Additionally, as a part of EU accession process Poland would need to adjust the financial market regulations on capital movements. 4.2. EARLY RETIREMENT SCHEMES AND IMPACT OFPHASING-OUTRULES Another important element that had an impact on medium-term expenditures are the changes to the initial reform proposal, mostly related to early retirement privileges. In the final regulations, there are three major changes compared to the initial reform proposal, all aimed to smooth the transition between the old and the new system. Those changes create additional expenditures for the pension system. They include: a right to retire according to old-rules for those who accrue pension rights to the end of 2006. This regulation covers: women born in 1949-51,retiring at age 55, miners with more than 17 years of work experience under ground (can retire after 25 years of working) and teachers with 25 years of experience (can retire after 30 years of working) - introduced during the consultations in Tripartite commission; 50 * a right to retire at early retirement age after 2006 for those who fulfilledwork experience criteria (20(w) or 25(m)years of working, including 10 to 15 in specialconditions) prior to reform introduction. This regulation covers part of women older than approx. 38 and men older than 43 - introduced during Parliamentarydiscussions; . mixed pensions for retirees in 2009-2013. This increases replacement rates for the mentioned groups - introduced during Parliamentary discussions. Relative to the initial reform proposal, these changes have an impact on both expenditures and revenues of the old-age system. This impact can be observed, according to projections between 2003 and 2020 (Figure 14). Prior to 2003, there are no changes in expenditures, because all people above 50 years of age are covered by the old system regulations. The revenues of the pension system depend on the number of switchers. Projections in this section assume that the switching process depends on age and approximately half of the contributors born between 1968and 1949decidedto switch. The most significant differences in the pension system balance between the initial proposal and legislatedsolutions can be observed in the period between 2003 and 2009, after that the differencesbecome smaller, as:cohorts covered by reform start to retire according to legal retirement age (60for women and 65 for men). Transition rules implemented in the Polish legislation have only a medium term impact on the stability of the PAYG pension system. This increasedexpenditure will create additional seven years of deficit in the old-agepension fund. The annual deficit in the period, however, should not exceed 1 per cent of GDP (unless more people switch to the funded pllar). This is mainly a result of favorable situation in the labor market and growth of the covered wage bill. This happens for two reasons. First is the expected growth of wage bill above GDP growth, a result of increasing wage bill/GDP ratio5". In 1998it was equal to 33 per cent and by 2020 it is expectedto rise to 40 per cent. The second reason is an increase in labor force supply, as a generationof 1980sbaby-boomis entering to workforce52 . The drop in the old-agepension system revenues observed in 2002 is related to the transfer of the part of the old-age contribution to the demographic reserve fund, as a precaution against worsening balance of the system after 2020, caused by rapid increase of dependency rates. As discussedearlier, the pension reform introduces savings in expenditures after 2006. According to projections, the old system would run deficits after 2015, while the reformed system in the its final shape starts to create surplusesin 2011. Stemming from increase in labour productivity and broadening the contribution base. For example, in Poland, company cars are not covered by social insurance contribution, which is expected to be changed in the future. 5 This demographic wave was quite significant in Poland, because of the marshal law in the early 1980s. During these period Poland experienced very high fertility rates, partially explained by introduction of the curfew after 8 p.m. 51 Figure 14 Medium-term impact of final pension rules vs. Security diversity proposal through 14 a. Expenditures of the PAYG old-pension fund (% GDP) 7% / 6%Y 5% / 0 4%; ; r g 3% 2% 1% _ 0% 1999 2004 2009 frhitial proposal 0 2014 2019 Rnal regulations no reform 14 b. Revenues of the PAYG old-pension fund (% GDP) 3% 7% 6% _ 5% CD 4% S3% 2% . 1% _ 0% _ 1999 2004 2009 Finalregulations Initialproposal 2019 2014 , no reform 14c. DeficiVsurplus of the PAYG old-pension fund (% GDP) -_- 2.5% 2.0% 1.5% 105%A .0% _ _ -1.0% -1.5% _ 1999 -4Initial 2004 proposal 2009 2019 2014 -- Final regulations n no reform Source: Social Budget Model (final regulations), Office of the Government Plenipotentiary for Social Security Reform (initial proposal, no reform) 52 4.3. OVERALL SAVINGSRATEINTHEPENSIONSYSEM" 3 The impact of pension reform on national savingshas several sources.These are private savings accumulated in pension funds and public savings in PAYG pension system, that according to projections appear after 2011. Figure 15 shows the combined impact of the two elements on the total savingsrate. The additional transfer to demographicreserve fund is also presented. This transfer does not generate any savingsin the economy, as it increasesPAYG system deficit in 2002-2008,however becausethose funds are invested in the capital market, it may have some impact on public savings,given state budget deficit constraint. In the projection period, the total savings rate is positive already in 1999. This is a result of expected surplus in the pay-as-you-gosystem without any transfer to the funded pillar. Thus, combined public and private savings reflect this surplus. Savingsgenerated by private funded pillar increase annually, as more contributors participate in pension funds. By 2020,the expected net inflow to the second pillar reaches 3 per cent of GDP. Savingsincrease significantlyafter 2011, when the pay-as-you-gopension fund reachessurplus. Overall annual savingsrate grows from almost zero in 1999to 5 per cent of GDP in 2020. Savingsin the old age system are presented as a gross flow. Pension reform introduces additional general revenue transfers to pay-as-you-gosystem, in the form of contribution for military serviceand maternity leave. These are estimatedat the level of approximately0.2 per cent of GDP annually. Also, after 2010 state budget will finance minimum pension guarantee for pensioners, which may decreasethe overall savingsrate by between 0.05 and 0.1 per cent of GDP"4 . Figure15. Savingsin the Polish multipillarpensionsystem, 1999-2020(% GDP) 6% 5% 4% fi3% ,w7~ 1R 1%-. 0% -2% 1999 2004 . . Totalsavings rate -n-- Netflowto pensionfunds 2009 2014 2019 PAYG system deficit -A- demographicreservetransfer Source: Social budget model 5 This section does not cover additional expenditures generated by bridging pension systems, as discussed in section 3.8 54Minimum pension for beneficiaries in the old system is financed from Social Security Fund, which is captured by overall deficit/surplus of the pay-as-you-go system. 53 5. Long term impact of pension reform In order to evaluate overall impact of the pension reform on the stability of PAYG system, several projections were prepared. The baseline situation is based on the current legislation and long-term goals of the pension reform. Assumptions in the baseline scenario include: - GDP growth of 3.5% annually - decrease in contribution rate to pay-as-you-go pillar from 12.22 per cent to 7.3 per cent between 2013 and 2030 - retirement age at current level - 60 for women and 65 for men - increase in longevity to current Swedish rates by 2030 - 65 per cent of the population born 1949-68participates in funded scheme - indexation of benefits at CPI plus 20% of average real wage growth - notional accounts indexation at 75% of wage bill growth - Increase in wage bill / GDP ratio to 40 per cent - Covered wage bill: 80 per cent of total wage bill (current estimates) - increase in participation rates in labor force by 20 per cent According to baseline scenario, the initial period of deficit, caused by shifting part of the contribution to funded pillar and later, by the worsening demographic situation will last until 2001. After that, the old-age fund will have an annual surplus of approximately 0.7 per cent of GDP. The surplus will be smaller in the 2020s, due to retirement of the post-war baby boom generation and additional reduction of contribution revenues, because of the retirement of cohorts that do not participate fully in the funded pillar vs. cohorts entering labour market who split contribution between first and second pillars. Approximately by 2035 all contributors will be participating in the funded pillar and changes in revenues level after this date can be attributed only to changes in labour force and average wage. During this period, the revenues of the pay-as-you-go old-age fund will drop from current 6 per cent to 2 per cent of GDP, and expenditures will drop from 6.5 per cent to 1.5 per cent of GDP. The alternative scenarios assume changes in the assumptions to investigate impact of economic and demographic development on the pension system. a. Constant contribution scenario Compared to baseline assumption, the contribution rate was kept on the current level: 19.52 per cent until 2008 and 18.52 per cent thereafter (current legislated level). The most significant difference concerns the revenue side, as there is no additional reduction of revenues caused by reduction of contribution. At the end of projection period, revenues will be on the level of 4 per cent of GDP. Changes in contribution rates have also 54 impact on expendituresside. Highercontributionswill be registeredon the notionalaccounts and create higher implicit debt to working population. This will cause increase in expendituresby additional0.5 per cent of GDP comparedto the baselinescenario.Theoverall old-agefund, exceedingthe result of this projectionis a significantsurplusin the pay-as-you-go level of 1 per cent of GDP annually. b. Increasedretirement age scenario In the caseof increasedretirementageto 65 for women and 66 for men, revenuesof the PAYG old-agefund will increasecomparedto baselineand expenditureswill drop down. The most significantimpact of this changecouldbe observedbetween2010and 2030,when the systemstill pays obligationsto cohorts in old systemand transition period, when the replacementrates are still on a high level.Increasein retirement ages does not have that significantinfluenceby 2050,both as a resultof actuarialadjustmentsin pensionpaymentsand reductionof systemliabilitiesin the newfirst pillar. c. current survival rates Assumingno changesin the longevityrates for Polish population, one can observe additionalimprovementof PAYG old-agefund surplus.This changeincreasesthe surplusby approximately0.2 per cent of GDP, whichis not very significant.However,further increase in longevitycan havesubstantialinfluenceon the expendituresofthe system. d. higher switching Scenarioassumesthat 90 per cent of the populationborn between 1949and 1968 decidedto switchto the fundedpillar.This hasa negativeimpacton PAYGsystemduringthe first yearsof the new systemfunctioning,asold PAYG systemstillpaysits obligations,while revenuesare significantlyreduced.In the long run, howeverthe expendituresof the old-age system are lower in years 2015-2040,as part of the old-ageliabilitiesfor those cohorts was shifted to the privatesector. The total presentvalue of higherswitchingshould be positive, but from the point of view of the short term financialliquidity of the systemit is less desirable,asdiscussedin previoussections. e. higher indexation of benefits If the benefit indexation was increasedto 50 per cent of average wage growth, expendituresof the PAYG systemwouldincrease,especiallyduringfirst decadesafterreform introduction,as liabilitiesof the publicschemeare still very high. In the long run this effect has a smallerimpacton expenditures,increasingthem by approx.0.3 to 0.5 per cent of GDP annually.Changesin indexationprinciplesdo not haveany impacton. f. lowereconomicgrowth In this scenariowe assumeda growthrate at 1.5%annually.Lower economicgrowth affectsmostlythe expendituresof the pensionsystem, as the modelassumesa link between GDP growthand wagegrowth. Expendituresare increased,as the resultof lower indexation both of benefitsand notional accountsare not that significant,as it is in the caseof higher growth rates.In this scenario,the publicPAYGsystemis in deficitin the period of 2020-2034, 55 where there is the worst combination of economic and demographic conditions for the pension system. g. higher indexation of notional accounts The last scenario assumesfull indexation of notional accounts,to the wage bill growth. As a result, expenditures increase after 2014, when people covered by the new system are retiring. This change leads to lower surplus, compared to baseline scenario, but still, the system is not creatingany deficit after 2014. Generally, the long term prospects of the public PAYG scheme after the introduced pension reform are quite good. However, the system is not immune to developments of economy, such as lower growth, as well as political factors, such as changes in benefit or notional accounts indexation. If the assumed characteristics will be followed by real developments,the goal of reducingthe contribution to the target level of 14.6per cent for oldagepensionscan be achieved,without creating deficits. The pension reform by itself does not guarantee long-term stability of the pension system. It lays the ground for better management of the public pension scheme, however it needs to be followed by continuous monitoring of economic and demographic development in the country. That is why the government of Poland plans to create an office, responsible for long-termprojections of a social security system (including old-age,disability and health care systems).Existence of such an office and public discussion of the actuarialstability of the social securitywould check possiblead hoc manipulation of the pension system design in the future. It may be especiallytempting for exampleto raise benefits when reservesare supposed to be accumulated,when the reserve assetsneededto sustain the stability of the pay-as-you-go schemewill amount to a significantshare of GDP. 56 Figure 16. PAYG old-age fund expenditures and revenues, 1999-2050 (% GDP) 16a.Baseline 7% 16c.Increasedretirementage 7 7% Lo 6%*. --- 4% - - 3% P; ---- -6%~ - - ............ 2% - 1% -- 3% ; : - 4% - '*@~--. . 02% .- . 2004 2009 .........- i- 2014 2019 2024 2029 2034 2039 2044 2049 2004 1999 revenues- . - - expendouress WdeficitIsurps t1 , - -- 2009 2014 ,deficiVsurplus 16b.Constantcontribution 2019 - 2024 revenues -. ... 2029 2034 2039 2044 expenddu(es 2049 basene| 16d.Currentsurvivalrates - 7%- 6 6% 61'--- > -3.- 2% ;% .2% 1999 7% - -- ;3% -E-L 3%._- - 62% 2 0% `1% -2% %1955 2004 2009 - -p~Jdeic0/surpk.s - 2014 2019 -revenues 2024 2029 2034 2039 . - - - expendrtures- _ 2044 base(ue -2%---_. 1999 2049 1 | l 2004 I, 57 2009 2014 defici surplus - 2019 2024 revenues - 2029 2034 2039 . - expenditures 2044 - baseine| 2049 16e. Higher switching 16g. Lower economic growth 7% 7% 6/ 6%, :7 I5%it 5% 4% 3% S LI 4%,4% IL-*j,3% z 2% 2% ~~~~~~~~~~~~~~~~~~~~~~~1% 1% 0% 0% S*X S_ -2% -2% 1999 2004 | 2009 2014 deficit/surplus - 2019 2024 2029 2034 2039 revenues - - - - expenditures 2044 2049 1999 baselne 2004 E|f1 16f. Higher indexation of benefits 2009 2014 deficit/surpus - 2019 2024 revenues- _. 2029 2034 2039 - - - expenditures - 2044 2049 baseUneI 16h.Higher notional accounts indexation 7% 7% 6% **6% . 5% \ _ 5%\L 4% 4% 31% 2%~~~~~~~~~~~~~~~~~~~~~~~~~~2 1999 2004 Erz 2009 2014 deficit/surplus - 2019 2024 revenues - 2029 2034 - - expenditures 2039 2044 2049 1999 I baseline 58 2004 do 2009 2014 il/surplus - 2019 2024 revenues- 2029 2034 - - - expenditures 2039 2044 bas fine 2049 6. Preliminary conclusions and lessons for other countries Pension reform in Poland touched almost all aspects of retirement provision, starting from institutional changes in the social security administration, through diversification in financing of the future benefits to shifting the system from a defined benefit to defined contribution regime. The most important elements introduced by reform are: * reduction of demographicrisk of the finances of the system by introducing defined contribution regimein mandatory system; D introduction of notional defined contribution first pillar * introduction of funded second pillar, managedby private pension fund managers D allowing market mechanismsto increaseefficiencyof management * introducing strong link between contributions and benefitsand as a result, stronger incentivesto postpone retirement decision; * creating more room for individual choice, both in the accumulation and disbursement phase * separating retirement savingsfrom other parts of the social security system * setting a foundation for the further changesin the socialprotection area As shown in the paper, the old-age system reform sets foundations for long-run stability of the old-age pension system. It also makes it more immune to ad hoc political changesand debates, by reducing the necessary participation of the state budget. The final outcome includes some transitional regulationsthat reduce the fiscal savings as presented in initial proposal. Those changes,however, do not compromise the long-run financial balanceof the pay-as-you-gopillar. The reform process is not finished, and there are still some components to be put in place. Those include legislatingthe bridging pensions arrangements for people working in specialconditions, and the annuities law that establishesframework for benefit payments in the second pillar. The new social security system should also be a subject to long-term forecasting, in order to prevent any changes that could threaten the stability of the social security system. Thus, proper actuarial supervisionshould be establishedand institutionalized. Also, other elementsof social protection need to be adjustedto the new old-agesystem. Those include especiallydisability scheme, as in its current shape there is a threat of leakage from old-ageto disability,as disability schemeoffers more generousbenefits. The reformed system started operating on 1 January 1999and the first pensions will be paid about ten years from the implementation. It is too early to draw conclusions about the system performance. However, it is well worth underlining the factors that enabled such a 59 fundamental reform to be put together and legislatedafter a long period of fruitless discussions. First, the contents of the reform package. There was broad popular support for a pension reform that includeda closerlink betweencontributionsand benefitsand a greater role for the privatesectorat the expenseof the socialinsuranceinstitution(ZUS). Secondly,the leadership. The governments of prime ministers Cimoszewiczand Buzek recognisedthe need for a plenipotentiary for pension reform with an office independentfrom politicalinfluencesand free of the task of day-to-daymanagementof the pension system. The plenipotentiaries- Baczkowski,Hausner and Lewicka - have successfullyshieldedthe officefor pensionreform from politicalfights,enablingit to focuson its professionaltasks. Thirdly, co-operation with trade unions. SecuritythroughDiversitywasconsistently supportedby the Solidaritytrade union and OPZZ (at leastuntil the end of 1997)and both were intimatelyinvolved through severalconsultations.Solidarityhad publishedits own reformproposalin 1995,which includedthe creationof mandatoryfundedpillar. Fourthly,moving quickly to grasp opportunities. All three of the plenipotentiaries understoodthe need to movequicklyand decisivelyto takeadvantageof the publicconsensus behind pension reform, as the old system became discreditedand the implicationsof demographybecamewidely known. Mistakescan alwaysbe corrected later. It is vital to remainaheadof opponentsto avoidthe reform beingpostponedindefinitely. The mainconclusionfrom Polishpensionreformis that when an opportunitypresents itselfit shouldbe seized.Politiciansshould foster appropriateconditionsboth for expertsto designand implementthe proposalandfor a broad consultativeprocessto fostersupport. 60 7. References Barr, N. (1998), TheEconomicsof the WelfareState,third edition, Oxford University Press. Eurostat, Statisticsin Focus,Population and Social Conditions, no. 5/99, "Socialprotectionin theEuropean Union,Icelandand Norway" Hausner, J. (1998),Security throughDiversity: Conditionsfor SuccessfulReform of the Pension Systemin Poland,Collegium BudapestWorking Paper, March. Holzmann, R. (1997), 'On the economic benefits and fiscal requirements of moving from unfunded to funded pensions', American Institute for Contemporary German Studies, Research report no. 4, Washington D.C. Jagannathan, R. and N. Kocherlakota, "Why Should Older People Invest Less in Stocks than Younger People", in Federal Reserve Bank of Minneapolis Quarterly Review, Vol. 20, No. 3, Ummer 1996,pp. 11-23 James, E. (1998), 'New models for old-age security: experiments, evidence and unanswered questions', WorldBank ResearchObserver,vol. 13,no. 2, pp. 271-301,August. Kolodko, G.W. (1996),Poland 2000:The New EconomicStrategy,Poltext, Warsaw. Ministry of Finance (1995),pension reform proposal [in Polish]. Ministry of Labour and SocialPolicy (1995),pension reform proposal [in Polish]. OECD (1998),EconomicSurvey:Poland,OECD, Paris. Office of the Government Plenipotentiary for SocialSecurity Reform (1997a),PolishPension Reform Package:Part One, Warsaw. Office - - (1997c),Security through Diversity: Reform of the Pension System in Poland, Warsaw. Office - - (1998),PolishPensionReform Package:ExecutionLaw (Decrees),Warsaw, English version forthcoming. Office- - (1997b), Polish Pension Reform Package:Part Two, Warsaw, forthcoming after adoption by Parliament. Palacios,R (1998),A note on diversificationbetweenfundedand PA YGpensionschemes. Palacios, R. and R. Rocha (1998), 'The Hungarian pension system in transition', Social Protection DiscussionPaper no. 9805,World Bank,Washington, D.C. Palacios, R. and E. R. Whitehouse, (1998),'The role of choice in the transition to a funded pension system', SocialProtection Discussionpaper no. 9811, World Bank, Washington D.C. Palmer, E. (1998),The SwedishPensionReform Model-Frameworkand Issues,mimeo. Rofman, R. (1999),Argentinean Multipillar PensionSystem after 5 Years,Social Protection DiscussionPaper, forthcoming, World Bank, WashingtonD.C. 61 Rofman, R and G. Demarco (1999), 'Collecting and Transferring Pension Contributions', SocialProtection DiscussionPaper no. 9907, World Bank, Washington, D.C. Rutkowski, M. (1998),'A new generation of pension reform conquers the East in transition economies', Transition,vol. 9, no. 4, pp. 16-19,August. a taxonomy Social BudgetModel,Gdansk Institute of Market Economics, with co-operation from ILO and Polish Ministry of Labor, Warsaw 1999 Thompson, L. (1998),Older and Wiser:TheEconomicsof PublicPensions,The Urban Institute Press, Washington,DC. Thompson, L. (1999),Administering Individual Accounts in SocialSecurity: TheRole of Values and Objectivesin ShapingOptions,The Retirement Project, Occasional Paper No. 1, The Urban Institute, Washington, DC Whitehouse, E.R. (1998),'The tax treatment of funded pensions', Social Protection Discussion Paper, World Bank,Washington, D.C. 62 Appendix Table7. Old andnew pensionsystemcharacteristics Formula Old system New system P= 0.24W + W*I 0.013:-L + W I 0.007*S P= K/G W -nationalaveragewagefor previousquarter K - pensioncapitalof insured,composedof I -individualwageindex imputed,registeredand old-agecontributions L -total lengthof service G - lifeexpectancycoefficientat pension S-additionalyearsacceptedfor insurance allotment benefits Lengthof service20for women,25 for men Any contributingperiod,but minimum pensionguaranteeafter20yearsof ______________ _contributing for womenand 25for men Minimum 15forwomen,20for men,withouta rightto n.a. lengthof service a minimumpensionguarantee Qualifying service ir Employment,self-employment, military service(nonprofessional), timerepressed, unemploymentperiod,Additionalperiods: education,maternity,takingcareof disabled child Contributingperiods(employedand selfemployed),alsounemployment,maternity, takingcareof disabledchild.Eachnonworkingperiodhasprovisionsfor contributionsto SIFfrombudgetor elsewhere : Additional Specialmultipliersfor minersand railway Onlythroughthird pillarEmployeePension n credits workers Programarrangement Minimumretirementage:60for womenand g Normalpension 60for women,65for menwith lot of g age exclusions. Averageretirementagein 1998- 65for men 55for womenand 59for men EarlyretirementGrantedfor disabled,miners,teachers,railwayNo earlyretirementin the system.Bridging workers,peopleworkingin specialconditionspensionsfinancedfrom additional (listincluding250differentcategories)contributionsfor peopleworkingin re-defined approx.one quarterof populationcovered. specialconditions.Help in retrainingfor new Alsoearlyretirementat 55for womenwith at entrantsto the labourmarket ________________ least30yearsof contributing Creditfor No specialcredit,only increaseof 0.0013of Actuarialadjustment _ deferred pension individualwageper eachyearworked formula,with Indexationof Since1996- at leastprices.The realgrowth From 1999:mixedprice-wage benefits definedannuallyin the statebudgetlaw 20%shareof wages Taxationof benefits Taxed Taxed Afterreachingretirementage- allowed WorkingpensionersAllowedwith wagelimitations,pension withoutlimitations.Pensionrecalculatedby recalculated by addingextracontributory addingadditionalcontributionsdividedby life yearsto the formula expectancyat the recalculation moment.No fundedpillar participationafterretirement A choicebetweenbenefits Transferbetween A choicebetweenbenefits disabilityandold-age benefits 63 Old system Newsystem Depends on the number of people eligibleDepends on the number of people eligiblefrom 80%to 90%of the benefit that a late from 80% to 90%of the benefit that a late person would have receivedor was receiving person would havereceivedor was receiving (eitherdisabilityor old-agepension) (either disabilityor old-agepension). Pension split equally betweenpeople eligibleto receive benefit Replacementrates at 76% Approx. 59% 65 for average worker Minimum guarantee Minimum: In nominal terms, indexedas other Minimum: In nominalterms, indexed as other pensions for every pensioner that worked for a pensions for every pensionerthat worked for a qualifyingperiod, paid from the Social qualifyingperiod, topping up pension from InsuranceFund. In 1998- approx. 70%of first and second pillar and financedfrom the minimum wage State Budget. Survivorpension Maximum:replacement rate not higher than No maximum benefit 100%,individual'swage factor not higher than 250%of averagewage Financing Paid by employer, not dividedinto different Paid partially by employerand employee, risk categories divided into: old-age,disability,sickness and work injury contribution, contributions tax exempt First pillar Mandatory PAYG system - 45% of wage 12.22%contribution to PAYG old-agefund, o 17.07%- other benefits(disability,survivor and short-term benefits) Note: Wage increasedin 1999by 23% to ______________ ______________________________________ compensatefor the split of contribution Secondpillar n.a. 7.3% of wage Ceiling and floor Minimum base:minimum wagefor workers, Minimum base:minimumwage for workers, levels 60%of averagewage for self-employed; 60%of averagewagefor self-employed; no maximum Maximum: 250%of averagewage Maximumbenefit Contribution SocialSecurityInstitute (ZUS) collection ag PAYG pillar SocialSecurityInstitute 1. Secondpillar n.a. > (accumulation) i ____________ ___________________________________ > Secondpillar n.a. (benefits) _ Specialsystems Armed forces(army, police, border guards, firemen),farmers,judges and prosecutors Third pillar Mostly life-insurancecombined with investmentfund 64 ZUS collects contributions all social security purposes, including 2nd pillar SocialSecurity Institute Open pension funds and pension fund managers,supervisedby StateSupervision Agency Mandatory annuity in one of the private annuity companies Farmers, judges,prosecutors,armed forcesin force prior to January 1, 1999 Employee pension programsin four basic forms Oiveinsurance,investment fund, mutual insurance, employeepension fund). Contribution up to 7% not covered by social security tax, but coveredby income tax. Benefits not taxed. Benefitsavailablefrom the lageof 60. Table 8. Initial reform proposalas in SecuritythroughDiversityand its changesduring the reformprocess Retirement age Age groups covered Initialproposal 62 - 62. Equal for men and women Final solution 60 for women and 65 for men Everybody up to 50 Excludingthose who accrue pension rights before the end of 2006 Judges,prosecutors and those who were in force before January 1, 1999are excluded. Reform coversthose who join militaryforces after 01/01/99 _________________ Occupation Workers, self-employed, groups covered judgesandprosecutors, military forcesup to 30 (army,police, border guards,prison guards, Consequences Differencesin benefitsbetween men and women, higher expendituresand lower revenues of a pension system Medium-termincreasein expenditures Increasein expendituresof a state budget (financingpensions for those remaining outsidethe system) firemen). Early retirement No early retirement Disability pensions All disabledconverted into old-agepensioners at age of 60 Workingon disability pensions Forbidden for fully disabledpeople outside labour protected companies Allowed Notional accounts indexation Coveredwage bill growth 75% of coveredwage bill growth Benefits indexation Initial capital To prices Mixedprice-wagewith 20% weight of wages Accruedpension as of 62 years of age K/G for all groups, except mixedold-new system formula for retirees in 2009-2013 Pension formula Accruedpension as of 65 years of age K/G for all groups Early retirement for those who worked 20 or 25 years (depends on gender)prior to 1999,with adjustment in benefit formula (addingadditional rights) At the retirement age - free choiceof benefit Medium-termincreaseof expendituresand reduction of revenues Probably higher expendituresof disabilityfund and lower of pensionfund. Not solvedissue of 2ndpillar accumulation Increasedincentivesto draw a disabilitypension and work on the same time, what leadsto higherexpendituresof pension system Initialproposal imposesno changesin contribution rate, final solution allowsfor reduction of contribution rate Increasedexpendituresof pensionand disabilityfunds Higher implicit debt of the pensionsystem Higher expendituresof pension system,smoother transition betweenold and new system retirees 65 Table 9. Medium-termprojectionsassumptions Year GDPgrowth Real wage growth 1999 4.80% 5.12% 1.35% 4.50% 2000 5.10% 5.38% 1.15% 3.80% 2001 5.40% 6.19% 0.38% 3.70% 2002 5.70% 6.36% 0.28% 3.50% 2003 5.80% 6.46% 0.38% 3.40% 5.60% 5.90% 0.67% 3.40% 2005 5.40% 6.07% 0.19% 3.40% 2006 5.20% 5.61% 0.38% 3.30% 2007 4.90% 5.52% 0.19% 3.30% 2008 4.60% 5.24% 0.19% 3.20% 2009 4.40% 5.15% 0.10% 3.20% 2010 4.20% 4.87% 0.19% 3.20% 2011 4.10% 4.87% 0.10% 3.20% 2012 4.00% 4.78% 0.10% 3.20% 2013 4.00% 4.78% 0.10% 3.20% 2014 3.90% 4.78% 0.00% 3.20% 2015 3.90% 4.78% 0.00% 3.20% 2016 3.90% 4.78% 0.00% 2017 3.80% 4.78% Employment | Real interest rate growtg 2004 r 2018 1 3.80% 4.78% 2019 1 3.80% 4.78% 3.80% 4.78% 2020 e 1 66 1 3.20% -0.10% j 3.20% -0.10% I 3.20% -0.10% 3.20% -0.10% 3.20% Figure 17. IT system design Health funds EGNIP Existing ; PESEL Pension funds regstratio 0 systeEmlExteral data Regional Computing Center _ _ | Central database verification system Mail Cooutput Data Datat Inpaut | X Benefits system servicing ~~Contribution Payer 1 ~~~Software E; i { Banks _ ZUS * t 67 ! 7 branches t ~~~~~~~~~~~~~Contributors Figure 18. Registration of employees in ZUS EMPLOYEE Getshiredby an employer 5 Employer shall inform employeeof Iheoutcomeof the process 4 *ZUSshallinformthe employerof the outcomeof theprocess *(processnotdefined) EMPLOYER l ZUS zus * -l Registeremployees (Predefined ZUS format)withthefollowinginformation: +Registration number(PESELor other) +Employer tax number +Centralregisterof the employee +Centralregisterof the employer +Centralregisterof membersof pensionfunds +Otherrelevantinformation OTHER +Monthlydataonthe contribution +Pensionfund +Medical care society +Contributions paid +Other 68 3 Shallverifydata to avoid multiaffiliation. Registerstheinformationin thedatabase Figure19. Registrationof pensionfund members 2. * Verifies documentation and consistency of information. 7 * Creates individual account * Confirms registration to member 8 Confirms registration to member MEBE MEMBER _ _ _ _ _ __ _ _ _ _ _ _ ____ _ _ _ 1.~~~~~~~~~~~~~~~~~~~~~~~3 1. * Selects a Pension Fund and registers through a sales agent. * Presents proper documentation: Fills a contract with the required information _ * _ FundManager 3__ the § ensionFund ~~~~~~~~~~Captures information in an electronic form P _und Answersrequestto the fund if rejection (no confirmation for acceptance) 4. Transfers registration request in an electronic formwithin6 daysfr s Zus erlmnent _ _ ~~~~~~~~~95 * Reports the result of the process or provides access to the information * 69 Verifies the registration request: Figure 20. Contribution collection mechanism 7 3.1 The banks shall verify the correctness and completeness of tne information. Registers payments in databases (employers and employees) 8 _ Post or ZUS offices 2.1 Provides information on employer and for individual employees EMPLOYER : f Proot Throughpredefinedformats fillsthe information for the monthly contribution Could be in paper or * WORKER zus Centraldatabase of 3.2 peyment b0 - Commercial 7777a::7 Com eraBank . v 2.2 Through special deposit slips, pays the contribution in a commercial bank _ -7-::::L:V0zS0:t Transmitthe informatlonthroughKIR i Through KIR, banks transfer/ the money to ZUS accounts In the NBP 3.3 Informs the employee on the contributions made +r process for the correction of errors. f ,~~~~~~~~~~~~~" Shall match the money deposited against the information provided and initiate a Nation Bankof thedeposits received of Poland 9 UNFE shall be informed of the aggregated amounts collected and also on the exceptions encountered. PROCESSFOR OF IDENTIFICATION INDIVIDUAL ACCOUNTS 70 Figure 21. Individual accounts identification Banks 7 UNFEshallbe informedof all transactionsmade betweenZUS and the PTE(transferagentsand custodianbanks) 3.1 * vViththe Informationof the paymentsZUS shaltinform the Banks ofthe differencesencountered. * ZUSperformsanalysisof employerswhodidn't payto initiatethe processof collection. PROCESSFOR 4.1. Ordersthe transferof moneyfromthe NBP to the PTEs(NBPwill useKIR for these transters) zus IDENTIFICATION OF INDIVIDUALACCUNTSBP 1.\ From the information received from the banks and Ihe employersthat doesn't presenterrors,and together wih the informationin is central database, it proceedsto identify individualemployees. 2. It performsthe individualization process,identifying: + Employeeswiththeir selectedPTEsand the amountoftheir contributions. 4 Employeesthat havenot selecteda PTE andthe amountof their contributions, + Thepaymentswith problemswithinformationand the correspondentamounts, 4.2 * InformseachPTE through their TransferAgents,the amountof moneythat will be transferredand the supporting individualinfornation. 3. * Calculatesthe amountsto be transferredto PensionFunds and the amountsto be registerin its owndatabase. * Also calculatesthe amountsto be keptin the accountsfor employeesthat havenot selecteda PensionFund. 8, UNFEshallreceiveinformationfromthe PTEs(Transfer AgentsandCuslodlanBanks)on all transactionsmadewith ZUS.This informationshouldbe providedin a standardformat and as oftenas necessary(everytime the processlakes place). 71 ThroughKIR,sends the moneyto the PTEs PTE's (Transfer Q U N FE.,~ _ Agont) PTE's tod B k (CUSodan ank) ~~~~~~~~~~~~~~~~~~~~~~~ the informationand the moneytransferredand registerthe contributionsIn individualaccounts. * Fromthis point,the PTE should investthe moneyaccordingto the investmentregime. Other benefitsfinanced from the social securityfund The social security fund (FUS) will establish separate funds for other benefits paid from socialsecurity. It includes: * Disability fund e Fund for sickness (includingmaternity leave and rehabilitation), e Fund for work-related illness (including accidents) There are also two reserve funds - one for sicknessand disability and the other for work injury fund. Having separate reserve funds is related to planned changes to work injury insurance. According to those, employers will pay contribution related to the probability of occupational disease or work injury in the sector. Work injury fund will be a subject to actuarialadjustment each year. The rules for disability benefits were also rationalised.Disability benefits are granted for permanent or temporal incapacity to work (not with relation to health detriment as it was before).If there is any possibility that the individual'shealth might improve, then a temporary benefit will be paid. There are two levels of disability: full and partial incapacity to work, based on the judgement of a social security doctor, employed by ZUS. Partial disability benefit amounts to 75% of a full benefit. A pre-pension rehabilitation system was also established.Disabled people are allowed to work, regardlessthe level of disability, not only in labour protected workplaces, but also on a regular labour market. Attempts to change this rule caused protests from disabled, who argued that working is a way to rehabilitation and additionally, they need salary income to pay for the living expenses. But this solution is still discussedin the government. At the moment, disability expenditures in Poland exceed 3 per cent of GDP, while 55 . The disability pensioners amount to 38 per cent of beneficiariesof social security system social security system allowsfor a choice of the benefitfor those who are eligibleto more than one benefits (e.g. old-ageand disability pension). If this situation was continued, in the future there could be a significantleakagefrom old-ageto disability.Thus, the next step in reforming the welfaresystem in Poland is to resolve the issueof the disability system". Also the area of short-term benefits requires significant changes. Sickness benefit system was a source of increasing deficit. In 1998, the number of sickness days was by 9.7 per cent higher than in 1997 and 13 per cent higher than in 1996,of which, number of days financed by social security system5 7 were 14.5per cent and 20 per cent respectively. Absenteeism figures also vary within the groups of insured people. Number of sickness days in 1998 increased by 5 This figure does not take into account that Polish systemallowsto draw disabilitypension also after retirementage,which meansthat the real numberof disabilitypensionersin Polandshouldbe smaller. 5 Initialstepstowardsreformingthis elementweretakenby Polishgovernment in 1999.The reformproposalis plannedto be formulatedin early2000. Accordingto the law,employersin Polandfinancefirst 35daysof employee'ssicknessbenefit. 72 8.7 per cent among employed and by 15.5 per cent among self-employed. In mid 1999 Polish parliament legislated new sickness benefit law, which introduces several measures to lower sickness spending. Treating doctors, who provide sickness certificates will need a registration, which can be withdrawn in case of irregularities. At ZUS the number of social insurance doctors will be increase, to ensure better supervision and screening of sicklisted clients. During the initial period of sickness absence (obligatory wage payment period), employers may request medical re-examinations by ZUS in case of doubt about the legitimacy of work incapacity. Some additional measures are still discussed. For example, introducing a three day period, when worker does not need any certificate to proof sickness period (without a right to salary). Polish government plan for 1999 also includes preparing a new proposal for work injury system, leading to diversification of work injury contribution, based on the risk of work injury or occupational sickness in industries. Short and medium-term financial projections The highest share in expenditures is attributed to the disability fund (3 per cent of GDP in 1998). In the future, expenditures of disability fund should decrease, mostly as a result of benefit indexation principles. After 2010 disability expenditures may grow, as a result of ageing population an inflow to disability of post-war baby boom generation. With regards to short-term benefits, without changes, both sickness and work-injury fund are expected to have a deficit, each approximately at the level of 0.3 per cent of GDP annually (Figure 22). Because those benefits are based on the current wages, the only way to improve their balance is to either increase the contribution rates or to cut expenditures. The first option may be exercised under the new work-injury law, when contribution rates will be calculated according to actuarial principles. The new sickness benefit law is expected to create savings at a level of 0.3 per cent of GDP, which will reduce deficit in this part of the social security system.. Uniformed services pensions The uniformed services, mainly the army and police, have a special pension system with completely different rules. The scheme is non-contributory and benefits are paid directly from the general state budget. The reform covers those, who start their service after January 1, 1999 by including them in the general system, all those who were in service before keep their existing pension entitlements. Although the uniformed services do have different pensions needs, an entirely separate system reduces mobility between the uniformed sector and other jobs. Including the uniformed services in the universal system does not impose substantial additional costs. 73 Figure 22. Social insurance system expenditures and revenues, 1999-2020 (% GDP) 22a.Disabilityfund 6.0% 401/ (s 30% 2.0% _______________________ 1.0% 0 0% _ 1999 2004 2009 : revenues .m. 2014 2019 expendtures _2b.Sicknessand maternityfund 1.6% 1.4% 1.2% =S-= 0.8% 0.6% 0.4% 0.2%/. _ _ - 0.0O/_ 1999 2004 2009 !sness - revenues -- 2014 2019 sikness expendirutes 22c.Work injuryfund 09% 0. 8% - p=- 0.7% * 0.6% 0.5% p S 0.4% - - _ 0.3% 0.2% - 0.1% -- 0.0% 1999 2004 -4-work 2009 injury revenues -- work-injury Source: Social Budget Model 74 2014 expenditures 2019 Social Protection Discussion Paper Series No. Title 9924 Pension Plans and Retirement Incentives 9923 Shaping Pension Reform in Poland: Security Through Diversity 9922 Latvian Pension Reform 9921 OECD Public Pension Programmes in Crisis: An Evaluation of the Reform Options 9920 A Social Protection Strategy for Togo 9919 The Pension System in Singapore 9918 Labor Markets and Poverty in Bulgaria 9917 Taking Stock of Pension Reforms Around the World 9916 Child Labor and Schooling in Africa: A Comparative Study 9915 Evaluating the Impact of Active Labor Programs: Results of Cross Country Studies in Europe and Central Asia 9914 Safety Nets in Transition Economies: Toward a Reform Strategy 9913 Public Service Employment: A Review of Programs in Selected OECD Countries and Transition Economies 9912 The Role of NPOs in Policies to Combat Social Exclusion 9911 Unemployment and Unemployment Protection in Three Groups of Countries 9910 The Tax Treatment of Funded Pensions 9909 Russia's Social Protection Malaise: Key Reform Priorities as a Response to the Present Crisis 9908 Causalities Between Social Capital and Social Funds 9907 Collecting and Transferring Pension Contributions 9906 Optimal Unemployment Insurance: A Guide to the Literature Social Protection Discussion Paper Series continued No. Title 9905 The Effects of Legislative Change on Female Labour Supply: Marriage and Divorce, Child and Spousal Support, Property Division and Pension Splitting 9904 Social Protection as Social Risk Management: Conceptual Underpinnings for the Social Protection Sector Strategy Paper 9903 A Bundle of Joy or an Expensive Luxury: A Comparative Analysis of the Economic Environment for Family Formation in Western Europe 9902 World Bank Lending for Labor Markets: 1991 to 1998 9901 Active Labor Market Programs: A Review of the Evidence from Evaluations 9818 Child Labor and School Enrollment in Thailand in the 1990s 9817 Supervising Mandatory Funded Pension Systems: Issues and Challenges 9816 Getting an Earful: A Review of Beneficiary Assessments of Social Funds 9815 The Quest for Pension Reform: Poland's Security through Diversity 9814 Family Allowances 9813 Unemployment Benefits 9812 The Role of Choice in the Transition to a Funded Pension System 9811 An Alternative Technical Education System: A Case Study of Mexico 9810 Pension Reform in Britain 9809 Financing the Transition to Multipillar 9808 Women and Labor Market Changes in the Global Economy: Growth Helps, Inequalities Hurt and Public Policy Matters 9807 A World Bank Perspective on Pension Reform 9806 Government Guarantees on Pension Fund Returns 9805 The Hungarian Pension System in Transition 9804 Risks in Pensions and Annuities: Efficient Designs Social Protection Discussion Paper Series continued No. Title 9803 Building an Enviromnent for Pension Reform in Developing Countries 9802 Export Processing Zones: A Review in Need of Update 9801 World Bank Lending for Labor Markets: 1991 to 1996 SummaryFindings All overtheworld,pensionsystems havefinancingdifficultiesthat needto be problems - finance addressed. Therearethreewaysof dealingwithpensionsystems rationalise thesystem, whichproduces it to a greaterextentfromgeneralrevenues, savings in theshortrun,or a full-fledged reform,changingthelogicandfoundations of thesystem. Afterseveral yearsof politicaland professional discussions, Polanddecidedto a newdefinedcontributionmulitipillarsystem, followthelatterpathand introduced consisting of a publicNotionalDefinedContribution,pay-as-you-go firstpillar,a fundedprivatesecondpillar,andvoluntaryfundedthird pillar.Thenewframework coversonlyretirement savings, whileotherbenefits stillremainundertheold definedbenefitpay-as-you-go regime.Thereformwaslaunchedon January1, 1999.As of for workers thisdate,theold definedbenefitpay-as-you-go systemwasterminated youngerthan50.Thenewold-agesystemattempts to offeractuarially fairbenefits, potentiallycreatingincentives to increasecomplianceandpostponeretirement. levelis coMinimumbenefitprovisionfor thosewho fall belowthe guaranteed providesgreater financedfromgeneralrevenue.Diversification of retirement savings securityto themembers, aslabourmarketdevelopments thatdetermine thenotional thatdetermine rateof returnin thefirstpillar,andfinancialmarketdevelopments Thisis whythe reform thesecondpillarrateof returnarenot perfectlycorrelated. packagehasbeennamedSecuritythroughDiversity. thestrugglefor thecurrentsituationof thepensionsystem, Thispaperpresents the long-term outlookof thenewpension pensionreformin the1990s,structure, systemandthe mainaspects of thesystemdesignaswell asfirstexperiences from showthatthenewsystem allows theimplementation process. Long-term projections thesavings and increases for greaterfinancialstabilityofthepublicpensionscheme ratewith a positiveimpactoneconomicgrowth. HUMAN DEVELOPMENT NETWORK Th Woi B*n peso r*- v rne an tt~ 'vd .'.'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ C in plementin penso reon.:saroi4t he rl.i Poll m *a ed tlxew Th Wi l Ba k s t o8 i i o l ~t a ir me jii- i i n ~ n t 2>-S..-S<.n-Ss 08-;.<-S-<--< -2-gi-2.,--S.,-S.n. <,si<2--< <-. a,--<205. -22-<S,N:< tlons a- 2ai2>-< Cva ^ialuo nthela E2 i otherinternatlonal i-sfit of';theWod D-<< Baka '= nd..>--S- f:reed coi;;. i <ofihispapkaiX>fer pee corsta theio ....,... rla<, PM oi dv em w illi d usI Dl frn - tn o re or ^n A eti g h ki A i r- i~ Poi-.cSi es-i., ..oi . ....... . ~i e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ . It o^ss i--< < <- Prte ionAdior i. - -^-l iii i.-.:.9.u -ni j-D;i.i.:ii NW, goorn(JB-1 m: a5irt$f--nD 20433-IOl .i0 i:l5.4 1 TelephneD22 i idi iiiSheions @kowok.r n B erive, T-^eWrd Rat unk 1DD'H SDe'i&j 5-ns. I. (".O D7)&j4 s E-ni