Just as for social security incentives, the key issue is this: If the receipt of benefits is delayed for a year, are benefits increased enough to offset their receipt for one less year? Is the benefit formula actuarially fair? Under the typical employer-provided DB plan in the United States, the answer is no. Benefits are not increased sufficiently to offset receipt for one less year. Thus, the plans impose a large implicit tax on work, just like the implicit tax described above for social security defined benefits. Indeed, the analysis of retirement incentives inherent in employer-provided DB plans in the United States led to the international comparison project that produced the results discussed above. The following discussion provides some information on the incentive effects of employer-provided plans. It does not attempt to draw from all the many studies that address this issue.

Until the mid-1990s, one of the important economic trends in the United States was the withdrawal of older persons from employment. This trend was surely made possible by the advent of the Social Security program and by the concurrent spread of employer-provided pension plans. Most of the pension plans were DB plans in which the benefit at retirement depended on years of service and earnings (usually those during the last years of employment).

Stock and Wise (1990a and 1990b) used company data to develop and estimate a formal model of retirement based on the option value retirement model. The model’s central feature involves recognition of the significant effect on retirement of the future accrual pattern of pension benefits. A series of subsequent papers found similar behavioral responses to plan incentives among men and women and among different types of employees (see, e.g., Lumsdaine, Stock, and Wise, 1997).

Looking at social security systems in different countries was analogous to looking at the effects of different employer-sponsored pension plans (with varying provisions) in the United States. Indeed, the retirement responses to social security plan incentives to retire correspond very closely to the incentive inherent in the employer-provided DB plans.

401(k)-Like Plans Have None of These Incentive Effects

Since the early 1980s there has been a rapid conversion to 401(k) plans in the private sector and a movement away from DB plans. The data are shown in Figure 5-14. People can of course participate in more than one type of plan. The 401(k) plans and other personal accounts such as individual retirement accounts (IRAs) have none of the incentive effects of DB plans and in particular none of the early retirement incentives inherent in DB plans. Since DB plans had typically provided a strong inducement to retire early, it should be expected that the decline in DB plans reduced the incentive to retire early. Thus the advent of personal accounts probably was

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement