Public and Private Sector Defined Benefit : A Comparison

The public and private sectors differ in the benefits provided to covered employees. But these differences are less pronounced when factors such as employee contributions and Social coverage are considered.

BY ANN C. FOSTER lthough employer costs for Statistics’ Employee Benefits Survey employee compensation (EBS)4 to make a detailed compari- Adiffer between the public son of public and private pensions. sector ($25.73 per hour worked) and Comparisons are made between full- the private sector ($17.49 per hour time workers in State and local worked), the proportions allocated to governments and those in medium employee benefits are roughly and large private establishments comparable: 30 percent for public (those employing 100 workers or sector employees and 28 percent for more). These establishments are private sector employees. Pensions comparable in size to most State and made up 7.4 percent of the total local governments.5 benefits package in the public sector and 3.1 percent in the private Retirement plan coverage sector1. The dollar cost differential In 1993-94, 96 percent of full- between sectors reflects differences time State and local government in the work activities and occupa- workers were covered by a retire- tions in each sector2. ment plan compared to 78 percent of The public and private sectors workers in medium and large private also differ in the incidence of establishments. Ninety-one percent employer-provided benefits. In 1993- of public employees participated in 94, for example, the incidence of defined benefit plans paid sick leave, medical and dental compared to 56 percent of private care, and life was much employees. Among private sector higher among public sector employ- employees, however, defined ees than among private sector contribution plan coverage was employees. In the private sector, greater than among public employ- however, the incidence of benefits ees (49 percent compared to 9 for paid holidays and vacations was percent). Among workers with higher.3 defined benefit coverage, 3 percent Differences in the provision of of public employees and 45 percent retirement benefits between the of private employees were also public and private sectors has covered by a defined contribution received much attention. The focus plan.6 has often been the differential Social Security benefits are an between benefit amounts, without important component of most Ann C. Foster is an economist in the Division of considering the factors underlying workers’ retirement benefits. While Compensation Data Analysis and Planning, Bureau of Labor Statistics. Telephone (202) 606- these differences. This article uses Social Security coverage is universal 6222. data from the Bureau of Labor in the private sector, not all public

37 Compensation and Working Conditions Summer 1997 employees are covered. EBS data multiplied by years of service. For The impact of plan integration on show that in 1994, 76 percent of the majority of public sector partici- the benefits received depends on a full-time participants in public pants (61 percent), terminal earnings worker’s earnings level. Research pensions were covered by Social are the average of an employee’s using data from the 1991 EBS found Security.7 highest 3 years’ earnings. For most that for an employee retiring at age Therefore, a typical full-time private sector participants (83 65 with 30 years’ service and a final State and local government em- percent), terminal earnings are salary of $25,000 or less, integrated ployee is covered by a defined defined as the average of an plans replaced a lower proportion of benefit pension plan and probably employee’s highest 5 years’ earn- final earnings than nonintegrated Social Security. A typical full-time ings. plans. For a comparable employee employee in a medium or large The majority of participants in with a final salary of $35,000 or private establishment is covered by a plans with a terminal earnings more, integrated plans replaced a defined benefit and/or defined formula have a pension benefit greater proportion of final salary contribution plan and Social formula based on a flat percent or than nonintegrated plans.11 Security. factor that also varies by sector. The In the public sector, an The following sections compare average factor is 1.85 percent in employee’s required defined benefit provisions of defined benefit public plans and 1.48 percent in plan contribution and the formula pensions because most full-time private plans. If terminal earnings used to determine benefits vary by State and local government employ- of $50,000 are assumed, a public Social Security coverage status. ees are covered by these plans. employee with 30 years’ service Research using data from the 1990 would receive an annual retirement EBS found that the average benefit Employee contributions benefit of $27,750 ($50,000 x .0185 for public employees covered by A direct comparison of plan x 30); an employee in the private Social Security was 1.83 percent of benefits needs to account for sector with 30 years’ service would terminal earnings times years of employee contributions. EBS data receive a benefit of $22,200 service. For those not covered by show that in 1994, 72 percent of ($50,000 x .0148 x 30), or 80 Social Security, the average benefit public employees with defined percent of that received by the public was 2.18 percent times years of benefit coverage had to contribute to sector employee. service. Public employees covered its cost. The average participant by Social Security made lower contribution was 5.9 percent of Social Security average contributions to their earnings and ranged from 5.8 One way in which Social Security defined benefit pension plans (5.11 percent among white-collar workers affects pension benefits is through percent of earnings); their counter- to 6.2 percent among teachers; plan integration. For State and local parts without Social Security among blue-collar and service government workers, the employee coverage contributed 7.55 percent of workers the average was 5.9 percent. contribution and the formula used to earnings towards defined benefit In comparison, in 1993, 97 percent compute benefits vary by Social pension plans.12 of participants in medium and large Security coverage status. private establishments had benefits Plan integration occurs when Normal retirement entirely paid for by their employers. benefits are adjusted to account for Requirements for normal retire- Assuming the same employer employer Social Security costs. In ment differ considerably between the contribution, employees who 1993-94, 48 percent of full-time public and private sectors. For contribute to the cost of their private workers were in defined example, in 1993-94, two-thirds of coverage should expect to receive a benefit plans that integrated pension full-time public sector participants in larger benefit than comparable benefits with Social Security 8 defined benefit plans could retire at workers who do not contribute. compared to 4 percent of their public age 55 or earlier upon meeting counterparts.10 Benefit formulas service requirements and still receive Integration is usually accom- Virtually all full-time public normal (unreduced) pensions plished by varying the percent of employees (99 percent) covered by compared to one-tenth in the private earnings used to compute benefits. defined benefit pension plans were sector. (See table 1.) An example of a typical integrated subject to terminal earnings-based Forty-three percent of public formula is: formulas. In the private sector, only sector participants could retire at any 61 percent of defined benefit 1 percent of earnings up to age after satisfying a service require- participants were in terminal $25,000 x years of service ment, usually 30 years, much higher earnings-based plans.9 + than the 5 percent for participants in These plans compute benefits as a 1.5 percent of earnings over the private sector. Another 21 fixed percent of terminal earnings $25,000 x years of service percent of public sector participants

Compensation and Working Conditions Summer 1997 38 Table 1. Requirements for normal retirement, full-time employees, by establishment type, 1993-94 Medium and State and Medium and State and Requirement large private local Requirement large private local establishments governments establishments governments

Percent Percent

Total with normal Service requirement retirement provision ...... 100 100 15 years ...... (1) -

Younger than age 55 ...... ( 1) 1 Age 62 ...... 21 6 No service requirement ...... - (1) No service requirement ...... 3 (1) Service requirement Service requirement 5 years ...... - (1) 5 years ...... 5 (1) 20 years ...... - (1) 10 years ...... 7 5 25 years ...... - (1) 15 years ...... 1 (1) 30 years ...... (1) (1) 20 years ...... 2 - More than 30 years 21-24 years ...... (1) - 25 years ...... 1 (1) Age 55 ...... 4 21 30 years ...... 1 (1) No service requirement ...... - 1 More than 30 years ...... 1 - Service requirement 5 years ...... 1 2 Age 63-64 ...... 2 3 10 years ...... - (1) No service requirement ...... 1 - 20 years ...... (1) (1) Service requirement 21-24 years ...... - (1) 5 years ...... - 3 25 years ...... (1) 5 10 years ...... (1) - 30 years ...... 3 11 25 years ...... (1) - More than 30 years ...... (1) 1 Age 65 ...... 48 8 No service requirement ...... 26 2 Age 56-59 ...... (1 ) (1) Service requirement Service requirement 1-4 years ...... 4 (1) 20 years ...... (1) - 5 years ...... 15 3 25 years ...... - (1) 6-9 years ...... (1) (1) 30 years ...... (1 ) - 10 years ...... 2 3 20 years ...... (1) - Age 60 ...... 13 5 25 years ...... 1 - No service requirement ...... 3 2 Service requirement Service requirement only ...... 5 43 5 years ...... 3 2 Less than 30 years ...... (1) 7 6-9 years ...... - (1) 30 years ...... 4 29 10 years ...... 1 1 More than 30 years ...... (1) 8 15 years ...... (1) - 20 years ...... (1) - Sum of age plus service ...... 8 12 25 years ...... 1 (1) Less than 80 ...... 2 (1) 30 years ...... 4 (1) 80 ...... 1 4 More than 30 years ...... (1) - 81-89 ...... 3 1 90 ...... 1 4 Age 61 ...... (1) - More than 90 ...... 1 2

1 Less than 0.5 percent. NOTE: Because of rounding, sums of individual items may not equal totals. Dash indicates no employees in this category.

could retire at age 55, most com- retirement (with varying service Among private sector participants monly after satisfying a service requirements) compared to 8 percent with early retirement provisions, the requirement of 30 years. In the of public sector participants. greatest proportion were in plans private sector, 4 percent of defined allowing early retirement at age 55 benefit plan participants could retire Early retirement with 10 years of service (34 percent), at age 55 after satisfying a service Among defined benefit partici- but substantial proportions were in requirement, also commonly 30 pants in 1993-94, 95 percent of plans allowing early retirement at years. In comparison, 48 percent of private sector and 87 percent of age 55 with 5 years of service (21 private sector participants had a public sector employees participated percent) and age 55 with 15 years of minimum age of 65 for normal in plans allowing early retirement. service (13 percent).

39 Compensation and Working Conditions Summer 1997 Table 2. Requirements for early retirement, full-time employees, by establishment type, 1993-94 Medium and State and Medium and State and Requirement large private local Requirement large private local establishments governments establishments governments Percent Percent

Total with early 15 years ...... (1) - retirement provision ...... 100 100 20 years ...... (1) - 25 years ...... (1) - Younger than age 55 ...... 9 20 30 years ...... 1 - No service requirement ...... (1) - Service requirement Age 60 ...... 7 3 5 years ...... 1 8 No service requirement ...... (1) - 6-9 years ...... (1) (1) Service requirement 10 years ...... 2 1 5 years ...... 2 (1) 15 years ...... 1 1 10 years ...... 3 1 20 years ...... 1 5 15 years ...... 1 - 25 years ...... 2 1 20 years ...... (1) 2 30 years ...... (1) 3 Age 62 ...... 3 (1) Age 55 ...... 72 46 No service requirement ...... (1) - No service requirement ...... 2 - Service requirement Service requirement 5 years ...... (1) - Less than 5 years ...... 1 3 10 years ...... 2 (1) 5 years ...... 21 11 More than 30 years ...... (1) - 6-9 years ...... (1) 1 10 years ...... 34 9 Service requirement only ...... 5 26 11-14 years ...... (1) - Less than 30 years ...... (1) 16 15 years ...... 13 5 30 years ...... 4 9 20 years ...... 2 3 More than 30 years ...... (1) (1) 25 years ...... (1) 10 30 years ...... - 2 Sum of age plus service ...... 3 3 Less than 80 ...... 2 3 Age 56-59 ...... 1 - 80 ...... (1) (1) No service requirement ...... - - 81-89 ...... (1) - Service requirement 90 and above ...... (1) - 10 years ...... (1) -

1 Less than 0.5 percent. NOTE: Because of rounding, sums of individual items may not equal totals. Dash indicates no employees in this category.

The most common age and ser- better indication of the degree of ment rates and private employees vice requirements for public sector income protection provided by a the lowest. participants were age 55 and 5 years retirement plan than absolute benefit Because virtually all public of service (11 percent), age 55 and 25 amounts. A replacement rate is the sector participants are in plans years of service (10 percent), and age proportion of a retiree’s pre-retire- with a terminal earnings formula, 55 and 10 years of service (9 percent). ment earnings during his or her final pension only replacement rates Twenty percent of public sector year of work that is “replaced” by the vary little as final earnings levels participants were in plans allowing pension benefits received. increase. For private sector early retirement before age 55 with When defined benefit pension participants, replacement rates varying service requirements, plan benefits are considered in decrease with earnings levels. compared to 9 percent of private isolation, public employees usually One reason is that 22 percent are sector participants. Twenty-six have higher replacement rates than in plans with a dollar amount percent of public sector participants private employees. When Social formula that provides the same were in plans with a service require- Security benefits are considered, benefit to all workers at equal ment only and could potentially however, a different pattern emerges. levels of service regardless of retire before age 55. Five percent of As may be seen in table 3, earnings histories. This results in private sector participants were in average replacement rates increase a decrease in the replacement rate such plans.13 with years of service. 14 When as earnings increase.15 considering pensions only, however, Social Security benefits replace Benefit comparisons public employees without Social more pre-retirement earnings for Wage replacement rates are a Security have the highest replace- those at lower earnings levels.

Compensation and Working Conditions Summer 1997 40 Table 3. Average replacement rates for specified final earnings and years of service for retirement at age 65,1 full-time employees, 1993-94 (in percent) Years of plan participation2 Final annual earnings 10 20 30

Private employees - pension only

$15,000 ...... 12.2 24.5 36.8 35,000 ...... 10.1 20.1 29.2 65,000 ...... 9.6 19.2 28.9

Private employees - pension plus Social Security3

$15,000 ...... 39.3 62.8 84.2 35,000 ...... 28.1 47.3 64.6 65,000 ...... 22.2 37.7 49.7

Public employees with Social Security - pension only

$15,000 ...... 17.3 34.1 51.0 35,000 ...... 17.2 34.1 51.0 65,000 ...... 17.3 34.2 51.0

Public employees with Social Security -pension plus Social Security3 $15,000 ...... 44.8 73.7 100.0 35,000 ...... 35.8 62.1 86.8 65,000 ...... 30.4 53.7 73.0

Public employees without Social Security - pension only $15,000 ...... 20.8 41.8 62.6 35,000 ...... 20.8 41.7 62.6 65,000 ...... 20.8 41.7 62.5

1A replacement rate is computed by dividing a pen- and for Social Security benefit computation, a worker is as- sion benefit by earnings in the final year of work. Private sumed to have retired after having paid into Social Security for employees are assumed to have retired on January 1, the same number of years as years of plan participation. Com- 1993 and public employees on January 1, 1994. Final putations exclude participants in cash-account pension plans earnings for private employees are for 1992 and public or plans with benefits based on career contributions. employees for 1993. Earnings histories, necessary for 2Time spent satisfying plan participation requirements was applying pension formulas, were constructed for each excluded from replacement rate computation unless the plan final earnings level using data provided by the Social specified that such time was to be included in benefit computa- Security Administration. tions. For pension formulas integrated with Social Security 3Excludes benefits for spouses and other dependents.

This is why replacement rates for Post-retirement pension Some employers provide a pension plus Social Security for both increases discretionary or ad hoc increase to private employees and covered Inflation can seriously erode the adjust retiree benefits for inflation. public employees decrease as final purchasing power of a retiree’s In 1993-94, 13 percent of public earnings levels increase. defined benefit pension. To prevent participants were in plans that had When Social Security benefits are such erosion, some plans specify an granted an ad hoc increase during considered, replacement rates for automatic cost-of-living increase, the previous 5 years; among private private employees tend to be higher usually based on changes in the participants, the proportion was 6 than those for public employees Consumer Price Index. In 1993-94, percent.16 without Social Security coverage. 54 percent of public pension partici- Social Security retirement Public employees with Social pants were in plans that provide benefits also have an automatic Security, however, have the highest these automatic increases; in the annual cost-of-living adjustment combined replacement rates of all private sector, 4 percent of partici- (COLA), based on changes in the groups. pants had such provisions. Consumer Price Index.

41 Compensation and Working Conditions Summer 1997 Therefore, most private sector of public and private pensions will plans with cost-of-living adjustments workers can expect their defined likely continue. Public sector than are private employees. benefit pension benefits to remain workers are more likely to be Although defined benefit replace- unchanged, thus decreasing in real covered under a defined benefit ment rates are higher for public terms during retirement. Their pension plan and less apt to have employees, they are more apt to have Social Security benefits, however, Social Security coverage than are to contribute to their pensions than would increase with inflation private sector workers. Public sector private employees, most of whom because of the COLA. Many public participants in defined benefit plans have wholly employer financed employees will have both pension are much more likely to be able to coverage. and Social Security benefits indexed retire at or before age 55 and still A pension is only one factor in an for inflation.17 receive normal (unreduced) retire- employee’s compensation. The value ment benefits than are private sector of other employer-provided benefits Conclusions participants. Public employees are and an employee’s wage rate are also The argument over the generosity also more likely to participate in important.

Compensation and Working Conditions Summer 1997 42 —ENDNOTES—

1 For more information on pay and benefit cost Disparity between Private and Public Pensions,” pants are included, 17 percent of private employ- differences, see the news release, “Employer Costs Monthly Labor Review, April 1994, pp. 3-9. This ees and 49 percent of public employees could po- for Employee Compensation—March 1996,” article is an expansion and elaboration of infor- tentially retire before age 55. For a more detailed USDL: 96-424, U.S. Department of Labor, Octo- mation in Ann C. Foster, “Comparing Public and comparison of public and private sector differences ber 10, 1996. Private Pensions,” in Retirement Benefits: Pre- in early retirement provisions, see Ann C. Foster, 2For example, required government activities serving Benefits in Changing Times, Selected “Early Retirement Provisions in Defined Benefit like public education and safety necessitate a large Proceedings of the NEA Retirement and Benefits Pension Plans,” Compensation and Working Con- proportion of white-collar professionals and highly Forum, November 16-19, 1995, Washington, DC: ditions, December 1996, pp. 12-17. skilled service occupations, respectively. Private National Education Association, 1996. 14 The replacement rates discussed are for industry sectors such as retail and wholesale trade 6For more information, see Employee Benefits a straight-life annuity or periodic payment for have a large proportion of occupations, such as the life of the retiree with no additional pay- sales, with comparatively lower compensation 1993 Bulletin 2456, Bureau of Labor Statistics, ments to survivors. For married employees, costs. Employee Benefits in State and Local the standard benefit prescribed by law is the For a detailed examination of how these dif- , Bulletin 2477, Bureau of joint-and-survivor annuity, which provides ferences in industry activity and occupational mix payments to a surviving spouse after a retiree influence pay and benefits, see Bradley R. Braden 7 Employee Benefits in State and Local dies. For more information, see Looking Out and Stephanie L. Hyland, “Cost of Compensation . for #2: A Married Couple’s Guide to Under- in Public and Private Sectors,” Monthly Labor Research examining differences between the standing Your Benefit Choices at Retirement Review, May 1993, pp. 14-21. More information from a Defined Benefit Plan, Publication on occupational pay differentials between the pub- systems accounted for employee contributions. For 1566, U.S. Department of the Treasury, Inter- lic and private sectors may be found in John E. Federal Civil Service Re- nal Revenue Service, 1991. Buckley, “Pay in Private Industry and State and 15 It should be noted that dollar amount for- Local Governments,” Compensation and Work- and Private-Sector Retirement Systems mulas are much more common in plans covering ing Conditions, September 1996, pp. 22-26, and EPW, Congressional Research Service, June 5, blue-collar workers. For example, in 1993, 22 Michael A. Miller, Jr., “The Public-Private Pay percent of all private workers covered by a de- Debate: What do the Data Show?,” Monthly La- 9 fined benefit pension plan were in plans with dol- bor Review, May 1996, pp. 18-29. minal earnings plans included 22 percent subject lar amount formulas. Seven percent of profes- 3 The lower incidence of vacation coverage in sional, technical, and related employees and 11 the public sector is due to the large number of to a percent of career earnings formula. Three percent of clerical and sales employees were in teachers who rarely receive paid vacations. For such plans compared to 36 percent of blue-collar more information, see Ann C. Foster, “Employee count pension formula and 2 percent to a percent and service employees. While 61 percent of pri- Benefits in the United States, 1993-94,” Compen- vate sector participants were in terminal earnings- sation and Working Conditions, Spring 1997, pp. see based plans, 74 percent of professional, technical, 46-50. Private Establishments and related employees and 73 percent of clerical 4 The Employee Benefits Survey (EBS) is a 10 and sales employees were in such plans compared study of the incidence and characteristics of em- cent of private sector workers with defined ben- to 48 percent of blue-collar and service employ- ployer-provided benefits. The EBS is conducted ees. For more information, see Employee Ben- in three stages during a 2-year cycle. Data for pension benefits not reduced to account for Social efits in Medium and Large Private Establish- small private establishments (fewer than 100 ments, 1993. workers) and for State and local governments are 73 percent could receive Social Security and pen- 16 The incidence of ad hoc increases was higher collected in even numbered years, while data for in previous years when inflation was higher. For medium and large private establishments (100 Security. For more information, see instance, in 1989-90, 16 percent of public pen- workers or more) are collected in odd numbered Benefits in Medium and Large Private Establish- sion plan participants and 22 percent of private years. The EBS (1993-94) covers about 15 mil- and participants were in plans that had granted at least lion public sector and 83 million private sector Local Governments, 1994. one ad hoc increase in the previous 5 years. For workers. For more information, see Avy D. Graham, more information, see Employee Benefits in State 5 An earlier article used 1990-91 EBS data to and Local Governments, 1990, Bulletin 2398, compare public and private pension benefits. In cial Security,” , March Bureau of Labor Statistics, 1992, and Employee addition to using more recent EBS data, this ar- Benefits in Medium and Large Firms, 1989, Bul- ticle examines actual replacement rates computed 12 letin 2363, Bureau of Labor Statistics, 1990. from all pension plans; the earlier article used av- liam J. Wiatrowski, “On the Disparity Between 17 For examples of this effect on the purchas- erage and typical pension benefits to illustrate dif- ing power of public and private sector retirees, ferences. This article examines aspects, such as Participants in plans with a sum of age plus see Ann C. Foster, “Comparing Public and Pri- differences in early and normal retirement require- vate Pensions,” or William J. Wiatrowski, “On ments, not covered in the earlier article. For more also be eligible to retire before age 55 if they joined the Disparity Between Private and Public Pen- information, see William J. Wiatrowski, “On the sions.”

43 Compensation and Working Conditions Summer 1997