March 14, 1995
In recent years, the Minnesota Legislature has offered a variety of early retirement incentives for public employees, including state, city, county, and school district employees. Early retirement incentives are designed to reduce salary expenses, avoid layoffs, or increase productivity. Supporters of early retirement incentives contend that the public saves money by inducing older, higher paid employees to retire and replacing them with lower paid workers or leaving the positions vacant. Critics argue that the benefits do not last long and question whether the benefits justify the costs.
Our study examines statewide early retirement incentives that have been used since 1980. It focuses on the costs and benefits of the early retirement incentive established in 1993, the most recent statewide incentive. We asked:
What are the public costs of the 1993 incentive program?
What are the public benefits of the 1993 incentive, including salary savings, layoffs avoided, and organizational benefits? How do salary savings compare with costs?
Are early retirement incentives appropriately targeted and financed? What are the implications of demographic changes for early retirement incentives?
To answer these questions, we analyzed data on retirement trends and costs from the three state retirement associations and the actuary for the Legislative Commission on Pensions and Retirement. We also surveyed and interviewed officials from state agencies, schools, cities, and counties.
In recent years, the Legislature has established two types of early retirement incentives: (1) employer-paid health insurance until age 65 (when citizens become eligible for Medicare) and (2) higher pension payments. In 1990, 1991, and 1992, the state offered employer-paid health insurance until age 65 to eligible employees who retired between the ages of 55 and 65.
In 1993, the Legislature established a more generous incentive that provided higher pensions and/or employer-paid health insurance until age 65. Participation was mandatory for state agencies, but was optional for counties and cities. School districts were required to offer the incentive to teachers, but could choose whether to offer the incentive to other staff.
The early retirement incentives offered in 1993 varied among employers and retirement plans. Eligible state employees covered by the General State Employees Retirement Plan could choose to: (1) continue their health insurance until age 65 or (2) receive a higher pension. This second option would increase a retiree's pension by about 15 to 19 percent. For example, a state employee who retired at age 65 under the 1993 incentive program (with 30 years of service and an average salary of $36,000) would receive a pension of $18,900, instead of the normal $16,200.
Eligible county, city, and school district employees covered by the Public Employees Retirement Plan could choose between these two incentives, provided they were both offered. Participating local governments could offer either incentive or both. Public school teachers covered by state or local teacher retirement funds received both higher pensions and employer-paid health insurance, though the pension increase was about 60 percent smaller than that received by other early retirees with the same salary history and years of service.
The costs and benefits of early retirement incentives depend on the extent to which incentives cause employees to retire earlier than they otherwise would retire. For each of the three major public retirement funds, we compared the actual number of retirements which occurred during the year spanning the 1993 incentive with the "expected" number of retirements based on the experience of the preceding three years. Our estimates take into account changes in the number of employees and their age distribution. However, it is not possible to take into account other factors that might affect retirement rates. Thus, our estimates should be viewed as approximate. We estimate that:
We also estimated how long retirees would have kept working had there not been an incentive. Given the high degree of uncertainty, we made high and low estimates of the time retirees would have kept working. We estimate that:
The public costs of the 1993 incentive include the liability incurred by the retirement funds to finance higher pensions and the cost to employers to finance health insurance. Our cost estimates are based on the difference in the present value of retirees' pensions with and without the incentive. We estimate that:
These figures include about $82 million to $113 million in retirement fund costs attributable to the pens ion incentive. The range in cost reflects the uncertainty in estimates of how the incentive program affects the timing of retirements. In addition, the health insurance incentive will cost employers about $19 million. The proportion of cost directly paid by employers varies widely. Cities, counties, and schools will directly pay only 2 percent of the program's total costs for members of the Public Employees Retirement Association (PERA), whereas schools and colleges will pay about 32 percent of the incentive's cost for members of the Teachers Retirement Association. The reason that employers will pay such a small percentage of the cost for PERA members is that most employers offered the pension incentive but not the health incentive to PERA members. We estimate that:
The cost is highest for teachers because they received both the pension and the health incentive. The main reason that the cost is lowest for PERA members is that their average salary is lower than the average salary of other public employees.
One of the major objectives of early retirement incentives is to help employers save salary expenses. Early retirement incentives can provide salary savings if retirees are replaced with lower paid employees or retirees' positions are held vacant. The most that an early retirement incentive could save would be an employee's earnings between the time of early retirement and the time at which the employee would have otherwise retired.
We found that early retirees were replaced by lower-paid employees, but the average salary savings are considerably less than the average cost per retiree under the 1993 incentive program. Furthermore, overall salary savings could exceed total costs only if a high percentage of retirees' positions were kept vacant. We estimate that salary savings would exceed costs only if employers kept open more than 69 percent of the positions vacated by retiring PERA members. Similarly, 59 percent of state employee positions and 36 percent of TRA member positions would have to be left vacant in order for salary savings to equal the cost of the 1993 incentive program. These percentages are based on mid-range estimates of how much the 1993 incentive program affected the timing of retirements.
We reviewed employment trends for public employees and surveyed employers about whether they refilled retirees positions. We conclude that:
It is necessary to keep a high percentage of retirees' positions vacant in order for salary savings to exceed costs. However, employment trends suggest that this is unlikely in schools and counties.
Eighty-three percent of public school teachers are in school districts in which the number of teachers increased between 1993 and 1994. About 92 percent of counties experienced growing employment between 1988 and 1992. Together, schools and counties had about two-thirds of the 1993 incentive's participants.
We found that the proportion of early retirees' positions that were kept vacant was considerably less than that required to cover the cost of the incentive program for schools, counties, and cities. The proportion of positions left vacant was 6 percent for school administrators, 11 percent for school district employees covered by PERA (those who do not hold a teaching license), 22 percent for county employees, and 31 percent for city employees.
The benefits of the 1993 early retirement program may have exceeded the costs for two large state agencies that together account for nearly half of the state's early retirees: the Minnesota Department of Transportation (MnDOT) and the Department of Human Services (DHS). Both agencies needed to achieve significant reductions in payroll in the 1994-95 biennium. MnDOT's budget required holding more positions vacant than could be accomplished without making many layoffs. DHS had very limited hiring flexibility because it could not lay off employees even though it was closing certain programs.
Another purpose of early retirement incentives is to enable employers to avoid layoffs. Layoffs are costly to organizations as well as to the people involved. We surveyed public agencies to find out how many layoffs were made by public employers between 1991 and 1994, and how many layoffs were avoided because of the 1993 early retirement incentives. We found that most public employers did not make any layoffs in the years from 1991 to 1994. The percentage of school districts, cities, and counties with no layoffs is over 90 percent in most years, and is below 80 percent in only one year. In addition:
Employment in state government has not grown, in contrast to school district and county government. More departments made layoffs during the 1991-1994 period, and a few departments in state government, including the two largest, the departments of Transportation and Human Services, needed to reduce staffing levels and create job vacancies in order to stay within their budgets. Still, most departments did not make any layoffs between 1991 and 1994, nor did they report that they avoided layoffs as a result of early retirements.
We surveyed state agencies, counties, cities, and school districts to find out how personnel directors felt about early retirement incentives. We found that:
Three-quarters of employers or more said that the overall impact of the incentives is positive. Most personnel administrators in state government and in cities, counties and school districts said that early retirement incentives have produced salary savings. Personnel administrators were more likely to feel that there was a productivity or quality gain than a loss, particularly for teachers. About two-thirds of school administrators felt that the early retirement program improved the quality of teachers, while only 6 percent felt that there was a loss in quality. Administrators for state agencies and cities were about ten times more likely to report a productivity gain than a loss, though about half of these administrators felt there was no change. County personnel administrators were the least likely to feel that there was a productivity gain. Only 15 percent felt there was a gain, while 11 percent felt there was a loss and 74 percent felt there was no change. Of course, as we have noted, most of the cost of the early retirement incentives is not directly borne by employers; they benefit from early retirement incentives but do not directly bear the full cost.
Early retirement incentives are popular with employers and produce tangible benefits. We conclude that early retirement incentives have a useful function in specific circumstances, but that these conditions are not typical or widespread. Offering an expensive option to employers who do not need it is not cost effective. The problem with offering incentives like those offered in 1993 is that it is doubtful that benefits outweigh the cost, and the cost is considerable.
It is helpful to keep in mind that an estimated half of all early retirees would have retired anyway in the same year they took early retirement. Many others would have retired in the next year or two. Early retirement does not create employee attrition, it borrows it from the near future. The beneficial use of early retirement incentives would appear to be restricted to situations where an employer is facing a one-time need to cut back or reorganize. This means that frequent repetition of early retirement incentives is not likely to be cost effective.
An argument against targeting early retirement incentives is that it is inequitable because it treats similar groups of public employees differently. However, broad early retirement incentives such as the 1993 incentive program are inequitable because they grant higher benefits to employees who retire during a particular time at the expense of employers and employees who must finance these incentives in the future. Future employees may have to help finance early retirement incentives for current retirees, but may not receive these incentives when they retire. As a result, we recommend that:
Benefits are likely to be high in relation to costs if an employer needs to reduce staffing levels beyond that permitted by normal attrition, which is around 9 percent per year in public employment. Certainly the closure of residential treatment centers places the Department of Human Services in this situation. The Department of Human Services has negotiated memoranda of understanding with its public employee unions that permit other severance arrangements that are not available to all state employees. Such a mechanism could be used to permit early retirement incentives in specific circumstances.
Furthermore, we conclude that the present method of financing early retirement incentives is flawed. For most retirees, the higher pensions in the 1993 incentive are financed by statewide retirement funds. To the extent that these costs increase the unfunded liability of retirement funds, they are financed by all employees and employers covered by these funds over the next 25 years. This arrangement is undesirable for several reasons. First, it is unfair to employers who choose not to participate because they still must pay their share of the cost of the higher pensions offered by other employers. Second, it gives employers an incentive to participate regardless of whether they believe the benefits outweigh the costs. Finally, participating employers get the benefit now and leave the financing burden to employers and employees in the future. Some of the benefits of early retirement are subjective and difficult to measure from a distance. Therefore, employers are in the best position to weigh the costs and benefits of early retirement incentives. We recommend that:
If an employer is undergoing a major one-time reorganization, it may be that special funding has to be provided specifically for this purpose. In any case, future generations should not be burdened by the cost of higher pensions for the present generation of retirees beyond the unfunded liabilities that already exist. The demographic structure of Minnesota and the United States is changing in a direction that makes it unwise to further require future workers to pay for current retirees.
On June 22, 1994, the Legislative Audit Commission directed the Program Evaluation Division to conduct an evaluation of the costs and benefits of Minnesota's recent early retirement incentive programs.
For a copy of the full report, entitled "Early Retirement Incentives" (95-04), published on March 14, 1995, please call 612-296-4708, e-mail Legislative.Auditor@state.mn.us, or write to the Office of the Legislative Auditor, 658 Cedar St., St. Paul, MN 55155.
Staff who worked on this project were Dan Jacobson and Elliot Long (project managers), and Carrie Meyerhoff.