Pension Plans and Employee Performance: Evidence, Analysis and Policy.

By Richard A. Ippolito.

Chicago: University of Chicago Press, 1997. Pp. xiv, 259. $34.95.

Richard Ippolito has written 13/15 of a very good book. I am not so sure about the remaining 2/15.

Let us begin with the first 13/15. The author, who is chief economist for the Pension Benefit Guaranty Corporation and a former pensions expert for the U.S. Department of Labor, develops a thorough analysis of how pension plans might affect worker behavior and how firms might be expected to use these effects to shape and motivate their labor forces. Theory and empirical analysis are judiciously combined in a way that makes the book an excellent recommendation to someone beginning an applied economics Ph.D. thesis and searching for a style to emulate.

So how do pension plans affect worker behavior? Building on work by Burkhauser and others, Ippolito starts from the view that defined benefit pension plans are contracts designed to influence worker tenure in a way wage profiles cannot. Initially, the pension acts as a bond both to encourage worker performance and to reduce quitting, but at some point, ideally from the firm's perspective when the worker's value to it begins to fall, the plan also provides an incentive to retire because the value to the worker of the increase in the eventual annual pension benefit will be offset by the shortening of the potential collection period. Ippolito also provides convincing empirical evidence, both on the incentive effects of pensions on retirement and of the importance of more generous private pension early retirement benefits, in explaining the trend since 1970 toward early retirement by males.

After first establishing that there is considerable variation in pension provisions across firms, Ippolito then proceeds to explore how firms can use pension structure to affect their work force. He suggests that the recent shift to defined contribution pension plans (which now cover about one half of all workers covered by a pension plan, compared to roughly one sixth in 1979) is too large to be explained by changes in industry composition or by changes in regulatory cost. Instead, a good portion of the change is due to the introduction of 401(k) plans, which are like previous defined contribution plans except that they allow additional voluntary employee contributions and allow firms to match those contributions.

Ippolito shows that defined contribution plans will be the choice of some firms that do not need or expect long-tenured employees but still want the tax advantages of a pension plan as part of their wage offers. But he also draws a distinction between workers with low-discount rates and high-discount rates, arguing that the former will likely be (unobservably) better employees. Hence, some firms may want to use defined contribution plans to winnow out high-discount-rate workers who will tend to grab the cash and leave because defined contribution pension plans are relatively attractive to early leavers. The 401(k) approach also has the advantage of allowing low discounters to take advantage of the matching provision, while high discounters do not take advantage of the provisions and hence essentially volunteer to receive less pay. The author estimates using the 1988 Current Population Survey (CPS) that 40% of workers eligible for matching 401(k) contributions nonetheless do not contribute.

A variety of empirical analyses using CPS and federal government employee administrative data is used to demonstrate that low-discount workers (identified, e.g., by nonsmoking behavior or 401(k) contributions) are, on average, better employees. (More reason to avoid smoking during job interviews!) The model does not explain why all individuals do not have a marginal discount rate equal to the market interest rate, but that gap could easily be filled. In any case, all that is required for Ippolito's sorting model is that employees who exercise 401(k) matching provisions be better quality - whether that is because they are more patient, have more foresight, or are smarter doesn't matter.

While I find the Ippolito approach insightful, I nonetheless wonder how much of the growth of 401(k) plans comes from firms, especially new firms, discovering that they can get away with them as a cheap substitute for old-style pensions. In the U.S., pensions for university academics are standardized, but in Canada they are not; yet in my experience, even economists seem to compare only salaries across institutions, apparently not even bothering to understand the large differences in pension provisions. If academic economists pay little heed to such matters, what can we expect of typical workers?

The final part of the book is devoted to policy analysis, and Chapter 13 is a largely first-rate use of the earlier analysis to make the case for easing the restrictions on the funding of defined benefit plans and eliminating the nondiscrimination rules that try to ensure relatively uniform 401(k) participation across wage levels. The essential argument is that these rules are unnecessary interference with voluntary contracts.

The policy analysis continues in the last 2/15 of the book, Chapters 14 and 15, and here I am less enthusiastic. While the analysis in the book has concentrated almost entirely on employee pension plans, the vast majority of the policy recommendations are in these two chapters and are concerned with social security. The basic theme is that social security should be more like private pension plans. Broader government goals and income distribution get barely a mention. Given the basic theme, it is not surprising that the recommendations include eliminating the 50% spouse benefit, drastically reducing the survivor benefits provided before retirement, eliminating the supplemental social security income program (which provides income to the otherwise destitute elderly who would be transferred to state welfare systems), fixing the retirement age at 67 with no early retirement option, lowering replacement rates paid by the public Disability Insurance (DI) program, cutting the number of DI beneficiaries to a fixed number equal to about 50% of the current level and, last but not least, switching Medicare and Medicaid to a voucher system in which only minimum care would be guaranteed and federal law would need to be amended "to insulate hospitals and doctors from liability for terminating treatment not covered by patients' insurance policies (p. 203)."

It is odd to see such a sharp reduction in the analysis/recommendation ratio at the end of this book and to find such a sweeping set of policy proposals with so many potential problems swept under the carpet. There are many interesting ideas, yet just as I do not have space here to record all my queries and objections, the author did not have the space to articulate his programs and their costs and benefits more fully. It would require another book, which I would hope would deal with my two main objections. First, just as the tax code, social security cannot be thought of solely in efficiency terms: It has redistributive goals. Second, individuals plan for their retirement relying on government programs: Proposed reforms without transition arrangements are of little use.

The book is free of typographical errors, is heavily but usefully footnoted, and has an index of below-average quality. It will be a very valuable addition to the library of any researcher working on U.S. private and public pension issues and their interaction with public policy, both because it brings together a number of disparate theoretical ideas into a single narrative and because of the many new and interesting empirical results it contains.

Michael R. Veall McMaster University