RETIREMENT

RESEARCH

State and Local Pension Plans Number 20, September 2011

COMPARING COMPENSATION: STATE-

LOCAL VERSUS PRIVATE SECTOR WORKERS
By Alicia H. Munnell, Jean-Pierre Aubry, Josh Hurwitz, and Laura Quinby*

Introduction

The comparability of state-local versus private sector The discussion proceeds as follows. The first sec-
pay has become a major issue in the wake of the tion presents some basic data on wages and benefits.
financial crisis. Funded levels of public pension plans The second section, following the methodology of
declined sharply, and governments ability to make earlier researchers, estimates the relative wages in
required contributions has been severely constrained the state-local versus private sector, controlling for
by the collapse of state-local budgets. Politicians education, demographics, and other factors. The
everywhere are looking for ways to reduce pension results suggest that state and local workers in the
costs and increase revenues. Often such efforts are aggregate have a wage penalty of 9.5 percent. The
couched in terms of excessively generous existing third section explores the extent to which benefits for
compensation especially, current pensions. Dueling state and local workers offset the wage penalty. With
studies have appeared arguing that state-local workers appropriate modifications for pension contributions
are paid less or more than their private sector coun- and the addition of retiree health insurance, annual
terparts. Virtually all agree that wages of state-local public sector compensation including both wages
employees are lower than for private sector workers and benefits is about 4 percent less than that in the
with similar education and experience, but research- private sector. The final section concludes that, given
ers differ on the extent to which pensions and other the modest size of any differential between public
benefits compensate for the shortfall. This brief and private compensation, policymakers should look
builds on the recent wave of studies by refining the carefully at the specifics of their own state or locality
estimates of the value of benefits. before making significant changes.

* Alicia H. Munnell is the Peter F. Drucker Professor of

Management Sciences in Boston Colleges Carroll School of
Management and director of the Center for Retirement Research
at Boston College (CRR). Jean-Pierre Aubry is the assistant
LEARN MORE
director of state and local research at the CRR. Josh Hurwitz
and Laura Quinby are research associates at the CRR. The au-
thors would like to thank Andrew Biggs, David Blitzstein, Keith Search for other publications on this topic at:
Brainard, Peter Diamond, Elizabeth Kellar, Steven Kreisberg, crr.bc.edu
Jason Richwine, and Nathan Scovronick for helpful comments.
 2 Center for Retirement Research

The Basic Facts Figure 2. Average Employer and Employee Pension

Costs as Percent of Payroll, by Sector, 2009
While a full answer to the question of parity of
compensation requires careful comparisons between 8%
people with similar skills doing similar jobs, some 7.0% Employer - solid 7.0%
basic statistics are a good place to start. Average Employee - striped
wages for state-local sector workers between 25 and 6%
5.2%
64 even without controlling for education and other 4.9%
factors are lower than those in the private sector,
and the ratio of public to private sector wages has 4%
3.0%
been declining over time (see Figure 1).
2%

Figure 1. Ratio of Average Public to Private

Sector Wages, Ages 25-64, 1990-2010 0.0%
0%
State and local Private sector Private sector
104% defined benefit defined benefit defined
contribution

100% Note: The costs for defined benefit plans represent the
normal cost. State-local costs are for Social Security eligible
employees. The costs for those without Social Security aver-
96% aged 7.1 percent (employer) and 7.6 percent (employee).
Sources: Public Plans Database (PPD) (2009); Towers Watson
92% (2009); and Vanguard (2010).

88%
Finally, retiree health insurance is much more
prevalent in the public sector than the private sec-
84% tor (see Figure 3, on the next page). Unfortunately,
1990 1994 1998 2002 2006 2010 no data are readily available to confirm this pattern,
Source: Authors calculations from U.S. Department of so estimates are required. In the private sector, the
Labor, Current Population Survey (CPS) (1990-2010). Medical Expenditure Panel Survey provides information
on retiree health insurance offerings by firm size, and
the Census shows the distribution of workers by firm
On the other hand, pensions are more generous in size. Combining the two pieces of information yields
the public sector. First, a greater percentage of workers an estimate of private sector coverage of 18 percent.
has an employer-sponsored plan in the public sector In the public sector, our assumption is that the per-
than in the private sector 76 percent vs. 43 percent. cent of the state-local workforce potentially eligible for
Second, among those employers who do sponsor plans, retiree health is the same as that enrolled in employee
costs to the employer are higher in the state-local sec- health insurance roughly 65 percent.1
tor, despite significant employee contributions, than in
the private sector (see Figure 2).
 Issue in Brief 3

So as to not introduce new issues, the dataset and

Figure 3. Percent of Workers Eligible for
variables for the nationwide analysis are the same as
Retiree Health Insurance, by Sector, 2009
used in the California studies. The data come from
80% the Annual March Supplement of the Current Popula-
65.4% tion Survey (CPS) for the years 2006-2010. The analy-
sis is limited to adult civilians working full time for a
60%
wage or salary during the whole previous year.3 The
variables include whether the employer was federal,
40% state, or local government and controls to standard-
ize for hours worked per week, years of education,
17.8% experience, experience squared, firm size, occupation,
20% immigration status, race, gender, marital status, years
to account for inflation, region, and some interaction
0% terms.
Private sector State and local sector
Before reporting the results, it should be noted
that two variables in these types of regressions are
Sources: Authors calculations from U.S. Census Bureau controversial. The first is firm size. The argument
(2008); and Agency for Healthcare Research and Quality for including firm size is that most state and local
(2009a and 2009b). workers are employed by large entities.4 Including
this variable means that public employees are be-
At this point in time, virtually all analysts agree ing compared mainly to employees of large firms,
that wages in the state-local sector particularly, as which for reasons not fully understood tend to pay
discussed below, when adjusted for the higher educa- higher wages and benefits.5 Omitting the variable
tional attainment of public sector workers are lower would make the wage penalty for working for a state
than those in the private sector. The big debate, as or locality somewhat smaller.6 Both California studies
will become evident, is the extent to which fringe ben- include firm size.
efits pensions, retiree health insurance, and other The other controversial variable is union status.
amenities offset the lower wages. The following One could argue that union status reflects the em-
sections explore each of these issues. ployees preference, implying that should the employ-
ee leave public employment he would seek a union
job. Therefore, union public sector workers should
Wages: State-Local versus Private be compared only to union private sector workers.
Sector The problem is that only a small percent of the private
workforce is unionized, so the exiting employee
A rash of recent studies has examined whether state- would be unlikely to find a union job. Therefore,
local workers are overpaid relative to their private controlling for union status does not seem relevant,
sector counterparts.2 All start with an examination of and indeed the variable has virtually no effect on the
wages, finding lower wages in the public sector, and coefficient for state-local workers.7 Both researchers
then make adjustments for fringe benefits and, in one leave this variable out of their California studies.
case, other amenities of public employment. The results of the wage regression for the nation
The following repeats for the nation an analysis of are shown in Figure 4 (on the next page). (The full
wages in California undertaken by two groups one results are presented in Appendix A.) The coeffi-
on each side of the debate of whether public sector cients for the continuous variables are the percentage
workers are overcompensated. Like all other recent increase in wages for a one-unit increase in the vari-
studies, both find wages lower in Californias state able. For example, an additional year of education is
and local sector. associated with an 8.5-percent increase in wages. For
the 0/1 variables, the coefficient shows the percent-
age increase associated with having the characteristic.
For example, women earn 15.7 percent less than men.
 4 Center for Retirement Research

Figure 4. Impact of Selected Factors on Wages of Full-time Workers, 2006-2010

S-L worker -9.5%

Federal worker 14.6%
Hours 1.4%
Education 8.5%
Experience 1.6%
Female -15.7%
Married 18.4%
Foreign born -6.0%
Black -14.0%
Hispanic -9.5%

-20% -10% 0% 10% 20%

Notes: For 0/1 variables, the bars represent the relationship between the characteristic and wages; for continuous vari-
ables, the bars represent the impact of a one-unit change on the wage. All coefficients are significant at the 1-percent level.
Source: Authors calculations from CPS (2006-2010).

The most important coefficient for our purposes is

that associated with being a state-local worker. After Figure 5. State-Local Relative to Private Sector
controlling for firm size, education, experience, and Wages, by Wage Tercile, 2006-2010
numerous personal and job characteristics, the results
show that state-local workers earned 9.5 percent less
Lowest tercile 2.0%
in wages than otherwise similar private sector workers.
To provide some sense of the heterogeneity of the
wage penalty by type of worker, the wage equation
was re-estimated by wage tercile. That is, a separate Middle tercile -0.5%
wage equation was estimated for the lowest paid
one-third in each state, the middle one-third, and the
highest one-third. The results confirm what other
Highest tercile
studies have found (see Figure 5).8 Public employ- -21.1%
ees in the lowest one-third of the wage distribution
are paid more than their private sector counterparts.
Those in the middle third are paid about the same. -30% -20% -10% 0% 10%
And those state-local workers in the top one-third are Source: Authors calculations from CPS (2006-2010).
paid about 20 percent less than private sector workers.
Despite the variation by wage level, the message
from the wage analysis is clear: state-local workers as not cancel out the wage penalty base their case on the
a group are paid less than their private sector counter- Bureau of Labor Statistics Employer Costs for Employee
parts. So far, researchers have no real disagreements. Compensation (ECEC) survey.9 This survey shows that
while benefits are much higher relative to wages for
state-local workers than for those in the private sector,
Benefits: State-Local versus they are not high enough to offset the wage penalty.10
Private Sector Therefore, they conclude that public sector workers
receive less total compensation wages and benefits
The controversy starts on the benefits side. The ques- combined than their private sector counterparts.
tion is the extent to which the value of the benefits The response by one set of critics is that the ECEC
provided to state-local workers offsets the wage penal- survey understates state and local employee compen-
ty. Several researchers who conclude that benefits do sation in three ways:11
 Issue in Brief 5

It omits retiree health since employers generally Retiree health

do not prefund these plans and therefore do not
make payments for active employees. Plus, cov- Clearly, retiree health should be added. Nationwide,
ered employees can buy retiree health insurance the normal cost for retiree health in 2009 was 7.6
at group rather than individual rates, which raises percent (see Appendix B). However, a 2011 survey
the value of these benefits above the employers of state and local governments reports that many re-
normal cost. This omission is relevant because spondents were cutting back on their commitments,
retiree health benefits are much more prevalent in shifting more costs to employees in the form of high-
the public sector than the private sector. er premiums, co-payments, and deductibles.13 Given
the uncertainty of eventual payment, we re-estimated
Contributions to defined benefit pensions and to the 2009 normal cost using a certainty equivalency
401(k) plans are not comparable. Public sector factor of 50 percent.14
plans in essence guarantee participants a return of This adjustment reduced the applicable normal
8 percent, whereas 401(k) plans provide no such cost to 3.9 percent.15 This figure was then increased
guarantees. Taking this higher guarantee into ac- by 25 percent to reflect the fact that retirees could
count increases the value of public sector pension purchase in a group rather than the individual mar-
contributions. ket.16 Finally, the scant available data suggest that the
normal cost in the private sector is roughly equal to
Public sector workers have much greater job that in the public sector.17 Adding retiree health in-
security than their private sector counterparts, and surance increases public sector compensation much
this advantage has a baseline value of 6 percent. more than private due to the higher cost and more
For California, the value increases to 15 percent extensive coverage in the public sector.
because California public employees are assumed
to be highly risk averse and enjoy a substantial Pensions
compensation premium compared to their private
sector counterparts.12 Comparing ECEC pension data across the public
and private sectors involves two problems. First, the
Their study concludes that a proper accounting for ECEC contributions to defined benefit pension plans
retiree health and defined benefit pensions gener- do not separate the normal cost and the amortiza-
ates a state-local pay premium of about 15 percent for tion payment to reduce unfunded liabilities. As the
California. Adding in an additional 15 percent for job employee only earns the normal cost, including the
security raises the premium to 30 percent. amortization payment overstates public sector com-
Our assessment of the debate falls between the pensation. Second, contributions to private sector
two sides. We accept the importance of adding retiree 401(k) plans and public sector defined benefit plans
health insurance and agree that adjustments are are not comparable. The public sector contribution
required for pensions, but reject the notion that job guarantees a return of about 8 percent, whereas no
security is higher in the state-local sector once educa- such guarantee exists for 401(k)s. Thus, the public
tional attainment is taken into account. sector contribution understates public sector compen-
Before addressing each of these issues, it is impor- sation.
tant to be clear about the role firm size plays in the Given the limitations of the ECEC data, we began
benefit calculation. Since nearly 90 percent of state- with total normal cost of 13.4 percent (liabilities dis-
local workers are employed by entities with 100 or counted at 8 percent) from the Public Plans Database
more employees, we adopted that category from the (PPD). This number was then multiplied by the
ECEC survey for the private sector benefit rate. Be- state-local payroll coverage rate of 85 percent to reflect
cause the results are sensitive to controlling for firm the fact that the PPD relates only to payroll for those
size, an alternative estimate is presented in the BOX covered by a pension while the ECEC number refers
on page 7 that excludes firm size from the wage equa- to total public sector payroll.18 This number was then
tion and uses the average benefit rate for all firms. adjusted to reflect the implicit guarantee. Our initial
thought was that employees could be guaranteed only
the riskless rate on their 401(k) investments. But the
recent academic literature suggests that a defined
contribution account can earn a certainty equiva-
 6 Center for Retirement Research

lency return of 1.23 percentage points more than the Putting aside job security, the calculations show
risk-free interest rate by allowing for investment in that state/local benefits nearly offset the private sec-
equities.19 Therefore, we re-calculate the public plan tor wage premium, but compensation in the public
total normal cost using an interest rate of 6.23 percent sector is 4 percent less than that in the private sector
(5 percent riskless rate + 1.23 percent).20 We then (see Figure 6). Given all the assumptions required,
subtract the employee contribution. The amount by the best way to describe the respective compensation
which the re-estimated employer normal cost exceeds levels is roughly equal.
the ECEC contribution number was added to em-
ployee benefits.
Figure 6. Total Compensation, as a Percent of
Job security Private Sector Wages, by Sector, 2010

The remaining issue is job security in the public sec- 160% 148.1%
147.0% 142.3%
tor. The argument is that job security, like wages and 138.2%
benefits, is a major goal of collective bargaining. To
120%
the extent that workers have security, they should be
willing to accept less in wages or benefits. During 100.0%
this recession, employment in the state-local sector 90.5%
80%
is down 3.1 percent since its peak, compared to 5.6
percent in the private sector. However, state-local
workers should be expected to fare better given that
40%
52 percent have a college degree a category where
employment has continued to grow compared to
only 35 percent in the private sector. In fact, the
0%
peak-to-present drops in employment for state-local
Private sector State and local sector
and private sector workers can be projected almost
perfectly based on the educational attainment of the Wages ECEC benefits
respective sectors (see Table 1). Moreover, public Pension adjustment Retiree health
sector employment continues to decline while private
Sources: Authors calculations from U.S. Bureau of Labor
sector employment appears to have stabilized. Thus, Statistics (2010); CPS (2006-2010); and PPD (2009).
it is not clear that public sector workers have any
greater job security than their private sector counter-
parts after accounting for their education level.21

Table 1. Projected Drop in Employment Based on

Educational Attainment, by Sector, Peak-to-July
2011
Educational % distribution of workers Change in
attainment State-local Private employment
Less than high school 2.3% 8.1 % -18.6 %
High school 19.9 29.3 -8.4
Associates degree 25.4 27.8 -4.7
College degree 52.4 34.8 0.4
Addendum:
Projected drop in -3.1 % -5.1 %
employment
Actual drop in -3.1 -5.6
employment

Sources: Authors calculations from CPS (2010) and U.S.

Bureau of Labor Statistics (2011).
Issue in Brief 7

Conclusion
Total Compensation without
Controlling for Firm Size The decline in the funded status of pensions in the
wake of the financial crisis has put state and lo-
The baseline analysis controls for firm size to estab- cal governments under great pressure just as their
lish the most direct comparison of similarly situated budgets were decimated as a result of the ensuing
state-local and private sector workers. The controls recession. The response all over the country has been
occur in both the wage equation and in calculating to increase employee contributions, cut benefits for
the ECEC benefit rate. We believe that controlling future employees, and in some cases cut cost-of-living
for firm size is correct. But given the sensitivity of adjustments for current employees and retirees. To
the outcome to firm size, the following figure shows justify these changes, the story is that public employ-
the results without the control (see Figure 7). ees are overpaid and their pensions are a particularly
egregious example of that overpayment.
Figure 7. Total Compensation, as a Percent of At this point, observers generally agree that wages
Private Sector Wages without Controlling of similarly situated workers are lower in the state-
for Firm Size, by Sector, 2010 local sector than in the private sector. The disagree-
ment hinges on the extent to which benefits offset the
160% wage penalty. Our re-estimation of the much-used
149.1%
142.4% 144.9% wage equation plus adjustments for proper valuation
141.6%
of pensions and retiree health insurance indicates
120% that the two roughly balance out. The estimated
difference nationwide is about 4 percent in favor of
100.0% private sector workers.
94.9%
In short, for the nation as a whole the difference
80%
between public and private sector compensation
appears modest. The relatively modest differential
should make policymakers cautious about massive
40% changes without carefully studying the specifics of
their particular situation.

0%
Private sector State and local sector
Wages ECEC benefits
Pension adjustment Retiree health

Sources: Authors calculations from U.S. Bureau of Labor

Statistics (2010); CPS (2006-2010); and PPD (2009).

Omitting firm size from the wage equation

reduces the state-local wage penalty from 9.5 to 5.1
percent, lowers the private sector benefit-to-wage
ratio from 47.0 percent to 41.6 percent, and lowers
the private sector retiree health coverage rate from
26 percent to 18 percent. These changes raise state-
local total compensation and decrease private sector
compensation to 149.1 percent and 142.4 percent of
private sector wages, respectively. Without the firm
size controls, compensation in the public sector is
4.7 percent higher than the private sector. This dif-
ferential, which represents a maximum on the high
side, reinforces the contention that compensation
of state-local and private sector workers is roughly
equal.
APPENDICES
 9 Center for Retirement Research

Appendix A. Wage Regression in the Current Population Survey

The baseline wage regression uses 2006-2010 data from the Annual March Supplement of the Current Popula-
tion Survey. The results are shown at the end of this Appendix. The dependent variable is the log of annual
earnings. Those with imputed earnings are dropped from the sample. The sample is at the individual level and
imposes the following restrictions:
Age 16 to 64;
Works for a wage or salary and receives at least $9,000;
With at least one year of work experience;
Full-time, working 52 weeks per year;
Not a member of the armed forces or the postal service; and
Currently living in the United States.

Table A1. Summary Statistics for Regression on

Annual Wages, 2006-2010

Standard Minimum
Mean Maximum
deviation
Annual wage 51,132 49,037 9,000 706,117
S-L worker 0.1454 0.3525 0 1
Federal worker 0.0301 0.1708 0 1
Hours 43.0910 7.2222 35 99
Education 13.7780 2.7770 0 21
Experience 21.6670 11.1250 1 58
Female 0.4468 0.4972 0 1
Married 0.6430 0.4791 0 1
Foreign born 0.1775 0.3821 0 1
Black 0.1083 0.3108 0 1
Hispanic 0.1600 0.3666 0 1

Note: Additional controls not depicted include occupation,

firm size, year, and region dummies. Some respondents
report less than 10 years of education. They comprise only
a small portion of the total sample, and removing them
does not alter the results.
Source: Authors calculations from CPS (2006-2010).
 10 Center for Retirement Research

Table A2. Regression Results on Log of Annual

Wage, 2006-2010
Variable Coefficient Standard error
S-L worker -0.0949 *** 0.003
Federal worker 0.1459 *** 0.006
Hours 0.0140 *** 1.8 x 10-4
Education 0.0853 *** 0.001
Experience 0.0164 *** 0.002
Female -0.1575 *** 0.004
Married 0.1838 *** 0.003

Foreign born -0.0604 *** 0.003

Black -0.1404 *** 0.005
Hispanic -0.0949 *** 0.004
Experience squared 0.0001 *** 3.3 x 10-5
Experience x ed 0.0011 *** 1.2 x 10-4
Exp squared x ed -0.0001 *** 3.0 x 10-6
Married x female -0.1622 *** 0.004
Black x female 0.0774 *** 0.006
Hispanic x female 0.0314 *** 0.005
Constant 8.6132 *** 0.023
R-squared 0.4647
Number of observations 290,125

Note: Additional controls not depicted include occupation,

firm size, year, and region dummies. Coefficients are sig-
nificant at the 1-percent level (***).
Source: Authors calculations from CPS (2006-2010).
Issue in Brief 11

Appendix B. Normal Costs for Public Sector Retiree Health Plans

Pay-go Fully-funded
State Plan name
Fiscal year Normal cost Discount rate Normal cost Discount rate
U.S. Average weighted by payroll 2009 7.6 4.5 0.2 7.9
AK Alaska PERS Postemployment 2009 5.76 4.7
Healthcare Plan
Alaska Teachers Postemployment 2009 4.15 4.5
Healthcare Plan
AL Alabama State Employees Health 2009 8.28 5
Insurance Plan (SEHIP)
Alabama Public Education Employees' 2009 7.43 5
Health Insurance Plan (PEEHIP)
AR Arkansas State Employees Postretirement 2010 4.95 4.5 2.41 8
Health Plan
AZ Arizona State Retirement System 2010 0.43 8
Health Benefit Supplement (HBS) Plan
CA California Teachers Medicare Premium 2010 0.01 4
Payment (MPP) Program
Contra Costa County Other Post 2009
Employment Benefit Plan (OPEB)
LACERA OPEB Program 2010 15.17 5
San Francisco City and County Retiree 2010 8.38 4.25
Health Plan
SDCERA Health Insurance Allowance 2008 0.01 8.25
CO Colorado PERA Retiree Health Care 2009 0.21 4.5
Trust Fund
CT Connecticut State Other Post-Employment 2008 21.76 4.5 8.72 8.25
Benefits Program
Connecticut TRS Retiree Health 2010 2.78 4.5
Insurance Plan
DC DC Other Post Employment Benefit 2009 4.91 7.25
Plan (OPEB)
DE Delaware Other Post Employment 2009 14.26 5 6.54 8
Benefit Plan (OPEB)
FL Florida State Employees' Health 2009 0.61 4 0.22 7.75
Insurance Program
GA Georgia Schools Personnel Post- 2009 5.3 4.5
Employment Health Benefit Fund
Georgia State Employees Post- 2009 5.68 4.5
Employment Health Benefit Fund
HI Hawaii State Employees OPEB 2007 11.8 5
(Employee-Union Trust Fund)
Hawaii State Teachers Association 2007 10.37 5
(HSTA) Voluntary Employees
Beneficiary Association (VEBA)
IA Iowa Postretirement Medical Plan 2008 0.71 4.5
12 Center for Retirement Research

Pay-go Fully-funded
State Plan name Fiscal year Normal cost Discount rate Normal cost Discount rate
IL Chicago Teachers Pension Fund Retiree 2010 5.35 4.5
Health Insurance Program
Illinois Comm. College Insurance 2009 15.33 4.5
Program
Illinois State Employees Group Insurance 2009 14.15 4.5
Program (SEGIP)
Illinois Teachers Retirement Insurance 2009 9.18 4.5
Program (TRIP)
IN Indiana State Personnel Healthcare Plan 2009 23,080 * 4.5
KS Kansas Health Policy Authority Post- 2009 0.76 3.85
retirement Medical Plan
KY Kentucky TRS Employee Health Plan and 2010 5.58 4.5 2.32 8
Medicare Eligible Health Plan
LA Louisiana State Post-Retirement Benefit 2009 407,139 * 4
Plan
MA Massachusetts Postemployment Benefit 2009 16.41 4.5
Plan
MD Maryland State Retiree Health Plan 2009 12.13 4.3 4.98 7.75
ME Maine State Employee Retiree Healthcare 2008 25.12 4.5
Plan
Maine Teachers Retiree Healthcare Plan 2008 4.87 4.5
MI Michigan Public Schools Retirement 2009 12.7 4 5.51 8
System's Retiree Health Plan
Michigan SERS Retiree Health Plan 2009 12.21 4 4.54 8
MN Minneapolis Retiree Health Insurance 2008 5,423 * 4
Plan
Minnesota State Postretirement Medical 2008 1.99 4.75
Plan
MO Missouri Consolidated Healthcare Plan 2008 7.5 1.97
Missouri DOT and Highway Patrol 2009 43,169 * 4.5
Medical Plan
MS Mississippi State and School Employees 2010 0.6 4.5
Health Insurance Plan
MT Montana State Employee Group Benefits 2009 3.93 4.25
Plan
NC North Carolina State Health Plan 2009 11.95 4.25
ND North Dakota PERS Retiree Health 2010 0.4 8
Insurance Credit Fund
NH New Hampshire State Postemployment 2008 18.26 4.5 7.22 8.5
Welfare Benefit Plan
NJ New Jersey Local Employees Health 2009 16.44 4.5
Benefits Program
New Jersey State & School Employees 2009 12.26 4.5
Health Benefits Program (SHBP & SEHBP)
Issue in Brief 13

Pay-go Fully-funded
State Plan name
Fiscal year Normal cost Discount rate Normal cost Discount rate
NM New Mexico Retiree Health Care Authority 2010 4.59 5
OPEB Program
NV Nevada Public Employees Benefits 2009 7.31 4
Programs Retiree Health Insurance Plan
NY New York City Health Benefits Program 2008 14.26 4
New York State Health Insurance 2009 1,013,836 * 4.24
Program (NYSHIP)
OH Ohio PERS Retiree Health Care Plan 2009
Ohio Police & Fire RS Retiree Health 2009
Care Plan
Ohio SERS Retiree Health Care Plan 2010 3.09 5.25
Ohio TRS Retiree Health Care Plan 2008 3.57 4.9
OK Oklahoma State and Education Employees 2007 1.95 3.5 1.02 7.5
Group Insurance Board
OR Oregon Retiree Health Insurance 2009 0.08 8
Account (RHIA) and Health Insurance
Premium Account (RHIPA)
PA Pennsylvania Retired Employee Health 2009 6.56 5
Program (REHP)
Pennsylvania Retired State Police 2009 17.96 5
Program (RPSPP)
RI Rhode Island State Employees and Electing 2009 4.07 3.57 1.84 7
Teachers OPEB
SC South Carolina State Employee Insurance 2009 4.93 5.5
Program (EIP)
SD South Dakota Postemployment Benefit 2008 0.67 3
Plan
TN TN Local Education Employee Group 2009 1.81 4.5
Plan
TN Local Gov. Group Plan 2009 0.38 4.5
TN State Employee Group Plan 2009 4.1 4.5
TX City of Austin OPEB plan 2008 12.96 4.21
Houston Postretirement Medical & Life 2009 10.16 4.5
Plan
Texas Employees Group Benefits 2010 10.02 5.5
Program (GBP)
Texas TRS-Care 2010 3.59 5.25 2 8
UT Utah Postretirement Medical & Life Plan 2008
VA Fairfax County Government Post- 2009 1.55 7.5
Employment Benefit Plans
Virginia State Health Insurance Credit 2009 0.31 7.5
Program
Virginia Teachers Health Insurance 2009 0.23 7.5
Credit Program
 14 Center for Retirement Research

Pay-go Fully-funded
State Plan name
Fiscal year Normal cost Discount rate Normal cost Discount rate
VT Vermont SERS Postretirement Benefit 2010 10.11 4 4.17 8.25
Plan
Vermont TRS Postretirement Benefit 2010 3.77 4 1.44 8.25
Plan
WA Washington K-12 School Districts OPEB 2008 3.71 4.5
Program
Washington LEOFF Plan 1 OPEB Plan 2007 10.01 4.5 6.59 7.5
Washington Pol. Sub. OPEB Program 2008 4.05 4.5
Washington State OPEB Program 2008 3.68 4.5 1.81 7.5
WI Wisconsin State Postretirement Medical 2008 2.71 4
Plan
WV West Virginia Public Employees 2009 13.73 3.56 4.57 7
Insurance Agency OPEB Benefits
WY Wyoming State Employee Group 2009 1.28 5
Insurance Retiree Benefit Plan

* Payroll information was not readily available. Normal costs are reported in thousands of dollars.
Note: Unlike pensions, normal costs and annual required contributions (ARC) for retiree health benefits are generally
reported in dollar amounts and not as a percent of payroll. However, payroll numbers are often provided separately from
the normal costs and ARC. Using these payroll numbers we are able to estimate the normal costs as percent of payroll. For
example, in the January 2010 actuarial valuation for the Massachusetts Postemployment Benefit Plan, the normal costs for
FY 2010 are reported on page 9 as 604.4 million dollars. Separately, on page 11, the total covered payroll as of January 2010
is reported as 3,684.1 million dollars in the plans schedule of funding. Using the two, we are able to estimate normal costs
as a percent of payroll equal to 16.41 percent.
Source: PPD (2008-2010).
 Issue in Brief 15

Endnotes
1 Agency for Healthcare Research and Quality, Medi- 12 See Biggs (2011 a and b) for a more detailed dis-
cal Expenditure Panel Survey (2009b). cussion of the value of public sector job security.

2 See Allegretto and Keefe (2010); Belman and Hey- 13 Center for State and Local Government Excellence
wood (2004); Bender and Heywood (2010); Richwine (2011).
and Biggs (2011); Borjas (2002); Braconi (2011); Keefe
(2010); Keefe (2011); Schmitt (2010); and Thompson 14 The normal cost accounts for the probability that
and Schmitt (2010). employees meet the vesting requirement for the
retiree health plan. The 50-percent certainty equiva-
3 See Appendix for a full description of the sample. lency factor represents the probability that retiree
The 52-week restriction could potentially cut state and health benefits and the share of premiums paid by the
local teachers who have summers off. Removing this employer will remain unreduced through retirement.
restriction and controlling for weeks worked does not See Mas-Colell, Whinston, and Green (1995) for the
significantly alter the results. theory behind the certainty equivalency, and Center
for State and Local Government Excellence (2011) for
4 Nearly 90 percent of state and local workers are em- statistics on recent changes made to retiree health
ployed by entities with 100 or more employees, based plans.
on calculations from the CPS.
15 We typically assume a risk-free rate of 5 percent
5 Hypotheses for the large-firm premium include 2 percent real return and 3 percent inflation which
greater unionization, economies of scale in non-labor is 50 basis points higher than the rate used in the
costs, firm age (larger firms tend to be older, and a valuations of most of the retiree health plans in our
correlation exists between employee compensation sample. However, in this case we did not attempt to
and firm age), and compensating differentials for rediscount the liabilities due to the complexity of the
bureaucratic work environments. See Brown and calculations for a minimal expected gain in precision.
Medoff (1989).
16 Buntin et al. (2003).
6 The results presented show a wage penalty of 9.5
percent. Omitting firm size from the equation re- 17 In 2006, monthly premiums for private sector
duces the penalty to 5.1 percent. retiree health coverage were $552 for retirees under
age 65, and $270 for those aged 65 and over (McArdle
7 Including union status in the equation increases et al., 2006). A survey of the 10 largest state-adminis-
the wage penalty from 9.5 percent to 9.7 percent. tered retiree health plans found that the public sector
premiums were, in 2009, $655 and $220, respectively.
8 See Borjas (2002); Fogel and Lewin (1974); Katz Based on the similarity in premium levels, we assume
and Krueger (1991); Poterba and Reuben (1994); and that retiree health costs for the two sectors are about
Schmitt (2010). equal.

9 Allegretto and Keefe (2010). 18 Authors calculations from CPS (2010).

10 Benefits include paid leave, such as vacation, 19 Gollier (2008) and conversation with Peter Dia-
holiday or sick pay; supplemental bonus pay, such mond. The Gollier model assumes a portfolio that
as bonuses and overtime; insurance, such as life is rebalanced annually over a 40-year investment
and health coverage; retirement and savings, which period to maintain a constant equity allocation of
include employer contributions to defined benefit 40.4 percent. The risk-free rate is assumed to remain
and defined contribution plans; and legally required constant over time and future stock returns are com-
benefits, such as Social Security and Medicare. pletely independent.

11 Richwine and Biggs (2011).

 16 Center for Retirement Research

20 See Munnell, Aubry, and Quinby (2011) for a

description of this calculation.

21 In fact, the difference in cost for supplemental

unemployment insurance suggests that the premium
for job security in the public sector amounts to only
2.4 percent of private sector wages. See Biggs (2011a).
However, other data do show that, controlling for age
and experience, the likelihood of unemployment in
the public sector is consistently lower than that of the
private sector. Regardless, any analysis including job
security would also need to incorporate all other non-
monetary aspects of the work environment.
 Issue in Brief 17

References
Agency for Healthcare Research and Quality. 2009a. Brown, Charles and James L. Medoff. 1989. The
Medical Expenditure Panel Survey. Summary Table Employer Size-Wage Effect. Journal of Political
1.A.2.e(2009). Washington, DC: U.S. Depart- Economy 97: 1027-1059.
ment of Health and Human Services. Available at:
http://www.meps.ahrq.gov/mepsweb/data_stats/ Buntin, Melinda Beeuwkes, Jos S. Escarce, Kanika
summ_tables/insr/national/series_1/2009/tia2e. Kapur, Jill M. Yegian, and M. Susan Marquis.
pdf. 2003. Trends and Variability in Individual Insur-
ance Products. Health Affairs Web Exclusive.
Agency for Healthcare Research and Quality. 2009b. Available at: http://content.healthaffairs.org/con-
Medical Expenditure Panel Survey. Table III.B.2.b tent/early/2003/09/24/hlthaff.w3.449.full.pdf
(2009). Washington, DC: U.S. Department of
Health and Human Services. Available at: http:// Center for State and Local Government Excellence.
www.meps.ahrq.gov/mepsweb/data_stats/summ_ State and Local Government Workforce: 2011 Reali-
tables/insr/national/series_3/2009/ic09_iiia_g. ties. Washington, DC.
pdf.
Fogel, Walter and Lewin, David. 1974. Wage De-
Allegretto, Sylvia A. and Jeffrey Keefe. 2010. The termination in the Public Sector. Industrial and
Truth About Public Employees in California: They Labor Relations Review 27(3): 410.
are Neither Overpaid nor Overcompensated.
Berkeley, CA: Center on Wage and Employment Fronstin, Paul. 2005. The Impact of the Erosion of
Dynamics at the University of California, Berkeley. Retiree Health Benefits on Workers and Retirees,
EBRI Issue Brief No. 279, Washington, DC: EBRI.
Belman, Dale and John S. Heywood. 2004. Public-
Sector Wage Comparability: The Role of Earnings Gollier, Christian. 2008. Intergenerational Risk-Shar-
Dispersion. Public Finance Review 32: 567-587. ing and Risk-Taking of a Pension Fund. Journal of
Public Economics (92): 1463-1485.
Bender, Keith A. and John S. Heywood. 2010. Com-
paring Private and Public Sector Compensation Katz, Lawrence F. and Alan B. Krueger. 1991. Chang-
Over 20 Years. Washington, DC: Center for State es in the Structure of Wages in the Public and
and Local Government Excellence and National Private Sectors. Working Paper No. 3667: NBER
Institute on Retirement Security. Cambridge, MA.

Biggs, Andrew. 2011a. The Value of Public Sector Keefe, Jeffrey. 2010. Debunking the Myth of the
Job Security. The American Enterprise Institute. Overcompensated Public Employee: The Evi-
Available at: http://blog.american.com/2011/07/ dence. Washington, DC: Economic Policy Insti-
the-value-of-public-sector-job-security/. tute.

Biggs, Andrew. 2011b. Lining Up for Government Keefe, Jeffrey. 2011. Are Wisconsin Public Employees
Jobs. The American Enterprise Institute. Avail- Over-Compensated? Washington, DC: Economic
able at: http://blog.american.com/2011/02/lining- Policy Institute.
up-for-government-jobs/.
Mas-Colell, Andreu, Michael D. Whinston, and Jerry
Borjas, George J. 2002. The Wage Structure and the Green. 1995. Microeconomic Theory. Oxford: Ox-
Sorting of Workers into the Public Sector. Work- ford University Press.
ing Paper No. 9313. Cambridge, MA: National
Bureau of Economic Research. McArdle, Frank, Amy Atchison, Dale Yamamoto, Mi-
chelle Kitchman Strollo, and Tricia Neuman. 2006.
Braconi, Frank. 2011. Municipal Employee Compen- Retiree Health Benefits Examined: Findings from
sation in New York City. New York, NY: Office of the Kaiser/Hewitt 2006 Survey on Retiree Health
New York City Comptroller John. C. Liu. Benefits. The Kaiser Family Foundation and
Hewitt. Available at: http://www.kff.org/medicare/
upload/7587.pdf.
 18 Center for Retirement Research

Munnell, Alicia H., Jean-Pierre Aubry, and Laura U.S. Census Bureau. Statistics of U.S. Businesses.
Quinby. 2011. Public Pension Funding in Prac- Table 2a. Employment Size of Employer and
tice. Journal of Pension Economics and Finance Nonemployer Firms, 2008. Washington DC:
10(2): 247-268. Available at: http://www.census.gov/econ/small-
bus.html#EmpSize.
National Conference of State Legislatures. 2011.
State Employee Health Benefits. Washington, U.S. Department of Labor. Current Population Survey,
DC. Available at: http://www.ncsl.org/portals/1/ 1990-2010. Washington, DC.
documents/health/IndivPrem09.pdf.
Vanguard. 2010. How America Saves 2010: A Report
Poterba, James M. and Kim S. Reuben. 1994. The on Vanguard 2009 Defined Contribution Plan
Distribution of Public Sector Wage Premia: New Data. Valley Forge, PA.
Evidence Using Quantile Regression Methods.
Working Paper No. 4734. Cambridge, MA: Na-
tional Bureau of Economic Research.

Public Plans Database. 2009. Center for Retirement

Research at Boston College and Center for State
and Local Government Excellence.

Richwine, Jason and Andrew Biggs. 2011. Are Cali-

fornia Public Employees Overpaid? Washington,
DC: Heritage Foundation Working Paper.

Schmitt, John. 2010. The Wage Penalty for State and

Local Government Employees. Washington, DC:
Center for Economic and Policy Research.

Thompson, Jeffrey and John Schmitt. 2010. The

Wage Penalty for State and Local Government
Employees in New England. Center for Economic
and Policy Research and the Political Economy
Research Institute at the University of Massachu-
setts, Amherst.

Towers Watson. 2009. Employer Commitment to

Retirement Plans in the United States. Avail-
able at: http://www.towerswatson.com/assets/
pdf/649/Employer%20Commitment%20to%20
Retirement%20Plans%20in%20the%20Unit-
ed%20States%281%29.pdf.

U.S. Bureau of Labor Statistics. 2011. Employer Costs

for Employee Compensation March 2011. Wash-
ington, DC: U.S. Government Printing Office.
Available at: http://www.bls.gov/news.release/
ecec.toc.htm.

U.S. Bureau of Labor Statistics. 2007-2011. Employ-

ment Statistics. Available at: http://www.bls.gov/
data/#employment.
 Issue in Brief 19

About the Center Affiliated Institutions

The Center for Retirement Research at Boston Col- American Enterprise Institute
lege was established in 1998 through a grant from the The Brookings Institution
Social Security Administration. The Centers mission Massachusetts Institute of Technology
is to produce first-class research and educational tools Syracuse University
and forge a strong link between the academic com- Urban Institute
munity and decision-makers in the public and private
sectors around an issue of critical importance to the
nations future. To achieve this mission, the Center
Contact Information
Center for Retirement Research
sponsors a wide variety of research projects, transmits
Boston College
new findings to a broad audience, trains new schol-
Hovey House
ars, and broadens access to valuable data sources.
140 Commonwealth Avenue
Since its inception, the Center has established a repu-
Chestnut Hill, MA 02467-3808
tation as an authoritative source of information on all
Phone: (617) 552-1762
major aspects of the retirement income debate.
Fax: (617) 552-0191
E-mail: [email protected]
Website: http://crr.bc.edu

Visit our:

pubplans.bc.edu

2011, by Trustees of Boston College, Center for The CRR gratefully acknowledges the Center for State and
Retirement Research. All rights reserved. Short sections of Local Government Excellence (SLGE) for its support of this
text, not to exceed two paragraphs, may be quoted without research. The SLGE (http://www.slge.org) is a proud partner
explicit permission provided that the authors are identified in seeking retirement security for public sector employees,
and full credit, including copyright notice, is given to part of its mission to attract and retain talented individuals
Trustees of Boston College, Center for Retirement Research. to public service. The opinions and conclusions expressed
in this brief are solely those of the authors and do not repre-
sent the opinions or policy of the CRR or the SLGE.