Comment S T U A R T DORSEY
U.S. Department of Labor and Western Illinois University Professors Barth and Cordes (BC) have greatly simplified my task. I have no major problems with their technical analysis other than some minor points which do not affect their principal conclusions. Therefore, I shall ignore these fine points and concentrate on the more enjoyable and productive job of dealing with some of the issues raised on the fringes of their analysis, but not fully addressed in the paper. BC come to three main conclusions: (1) Pursuit of nontraditional investment objectives will lower the economic performance of pension funds, narrowly defined over risk and return, (2) market forces will impose this loss on workers, and (3) the aggregate impact on this phenomenon is likely to be modest. I have no quarrel with the thrust of these findings. However, these are quite straightforward and could have been derived in a few short pages. Much of the effort at modelling adds little to the analysis. On the other hand, this space could have been devoted to some important questions about this kind of behavior, the fundamental being: Why do we observe pension investing that fails to maximize return, given employees risk preferences? A derivative of this is: what characteristics of pension funds and/or unions would imply this kind of behavior? The latter is particularly relevant, because if we conclude that nontraditional investing is socially inefficient, pension policy should be aimed at reducing the incentives to these actions. That there is pressure for diverting pension funds is not surprising. Union members and their leadership have many goals which they would pursue if there were no cost to their actions or if the cost could be passed on to a third party. (I will ignore the issue of whether the leadership accurately reflects the sentiments of the membership. This is an important issue regarding diversion of pension funds and should be considered elsewhere, but the analysis is simplified by assuming a common purpose.) An example is increased investment in the union sector. However, this describes only the demand for social investing. What determines the degree to which these goals actually will be pursued? A good first start would be to assume that unions behave rationally, i.e. weighing costs and benefits when determining the emphasis to be placed on social goals. Deviating from optimal return-risk investing imposes a cost upon workers. BC make this conclusion but do not explicitly explain why this will occur. The competitive employer must manage pension investments to maximize return consistent with the reveaIed risk preferences of his workforce. In so doing, total labor costs are minimized because any inferior investment policy would require compensating wage increases or higher pension contributions in order to attract the desired workforce. If, however, the fund manager is forced to accept a lower J O U R N A L OF L A B O R RESEARCH Volume n, Number 2 Fall, 1981
STUART DORSEY
249
return one of two results is possible: (1) The value of the retirees' annuities must decline, or (2) Given that most pension benefits are based upon a fixed formula that is unrelated to the market earnings o f pension assets, i.e. defined-benefit plans, wages will decrease. This is because the employer must increase contributions to offset reduced fund earnings. (Of course, if only one competitive firm is so affected, it will be unable to pass backward this cost and will have to absorb it or be forced out of operation.) Clearly, the worker has an incentive to limit this type of behavior to the extent that it is costly. As Donald Martin remarked in his comments on this paper " . . . the demand curves for ethical and social investments are downward sloping." In nontechnical jargon, there is only so much the rank-and-file will be willing to sacrifice to censure producers o f infant formula. An exception to this result should be noted. If the pressure to change investment policies comes from consumers, firms will not be able to pass backward the cost to workers. For example, consumers may boycott a c o m p a n y that invests in South African enterprises. In the unlikely event that diverting this portion o f the pension's assets lowers the performance o f the portfolio (because o f the availability of a wide range of substitutes), consumers would bear the cost in the f o r m of higher prices. Or, if competitive forces prohibit, the firm may have to absorb the loss as a cost of public relations. While union pursuit o f social investment goals may be privately rational, whether it is socially efficient is another matter. It will be as long as the costs are borne by union members. They will be trading-off some o f their economic rents for ethical goals or other kinds of economic benefits. This is socially efficient if the union leadership reflects the rank-and-file's best interests. However, there are two exceptions. The first arises if there are external costs to the union's " c o n sumption" of nontraditional goals. BC provide one example: Increased union power which may result from channelling funds into the union sector. Another is moral hazard. Given that a portion o f a firm's pension liabilities are insured by the Pension Benefit Guarantee Corporation, the cost o f excessive riskiness of pension portfolios may be borne by taxpayers. In other words, there can be a moral hazard problem which subsidizes the choice o f high-return, high-risk assets. Note, however, that this problem is not merely the consequence o f social investing. Any rational portfolio manager concerned with maximizing the value of the pension fund also will choose relatively risky assets if pension liabilities are insured. Further, this problem is best handled by a careful structuring o f the insurance mechanism to minimize the incentives for this kind o f behavior, whether it comes from a desire to generate superior earnings or to pursue some ethical or social goal. In this regard, the P B G C does not insure the full value o f a firm's pension liabilities. In addition, one-third o f the firm's net worth is liable to meet the obligations of a terminated pension plan. Thus there is some incentive for workers and their firms to keep the pension fund solvent. The idea of union use of employee pension funds suggests an interesting hypothesis about the relationship between unions and pensions. We know that pension coverage is much more extensive among union members than in the nonunion sector. Yet the explanations for this result are not fully satisfying. Examples of these are that union members are less mobile or that the union leadership justifies its existence by providing different, if not necessarily better,
250
JOURNAL OF LABOR RESEARCH
compensation packages. The BC paper suggests another: The pension is a solution to a public good problem. Suppose there is some investment strategy which would benefit most workers - - for example, targeting investment into depressed heavily-unionized sectors, such as construction in the Northeast. This strategy will not be followed if left to individual, voluntary decisions. Each worker's savings represents such a small portion of the total funds required, that it will not be privately rational to make the investment at a reduced, or more risky return. However, a mandatory pension solves the "free-rider" problem. Everyone must contribute to the achievement of the collective goal. Again, this process requires that the union leadership directs the funds in accordance with the preferences of the membership. An implication of this hypothesis is that small pension plans will be unlikely to attempt nontraditional investment strategies. The marginal impact of their policies will be small; yet the costs will be no less than for large portfolios. Another implication of rational nontraditional investing is that the decision to pursue divergent goals will be a function of the characteristics of employees. Not all union members will have the same goals. A classic difference will arise between older workers nearing retirement and younger employees. The former will be much less willing to "play games" with pension assets. Solvency of the pension fund will be their primary interest. Younger workers, however, would receive the bulk of benefits from investment policies designed to preserve union jobs (and rents). This is because the number of years they will be reaping these benefits will be greater. Clearly, the potential for conflict exists. However, a reasonable prediction is that the older the median worker, the less likely is pension investing to diverge from traditional economic criteria. As mentioned, the BC paper concludes that the aggregate impact of nontraditional investment of pension funds is likely to be small. They base this on the fact that pension saving is only a portion of the total flow of investable funds and that, given efficient capital markets, any action by one investor is likely to be offset by another. I have offered another reason: nontraditional investing may be costly and thus there is an employee incentive to limit its use. It is surprising that BC barely mention the one other factor that may be the most salient of all. That is, the ERISA regulations as interpreted by the U.S. Department of Labor could deter nontraditional investing in some cases. The "prudent man rule" requires that only economic motives, i.e. the traditional risk and return calculus, are allowable in pension investing. (See the definition of prudence in Current Federal Regulations, Sec. 2550.404a-1.) Therefore, it is possible that the prudent man rule itself has partly constrained union use of pension funds. Does this imply that the issue of divergent investing for social goals need not concern us? I think not. The analysis presented here predicts that the larger the pension fund is, the more likely it will be the focus of attempts to use the funds for social goals. Recently, the President's Commission on Pension Policy released its study of retirement policy. Its most ambitious recommendation is for a mandatory minimum universal pension system (MUPS). It is proposed that all employees 25 years or older receive a pension contribution equal to 3 °7o of earnings. Yet for the present discussion, the most interesting aspect of MUPS is that it would establish a central portability fund which would receive and invest these contributions. At present, estimates are that just short of 50°7o of all employees
STUART DORSEY
251
participate in some type of retirement program. Thus, unlike a pay-as-you-go system, a tremendous pool investable funds would be created. I believe we can expect the pressure to divert these contributions into various definitions of socially acceptable investments would match the enormity of the fund. Those whose pet projects have failed the market test for the allocation of credit may succeed in the political market. The President's Commission has recommended that the fund be managed by an independent Board of Trustees, but its independence likely will be severely tested by political pressures. Finally, I would have preferred that BC omit the final section on inflationary implications of nontraditional investing. The analysis was technically correct: less efficient investing could reduce productivity growth and, given a fixed rate of increase of the money stock, worsen inflation. However, they conclude that such an impact is likely to be minor and I certainly agree. Currently, there seems to be a feeling that any economic matter must be related to our inflation problem if it is to be legitimized. This is one example; deregulation of business and "supply-side" tax cuts are others. Both of these latter policies may convey significant economic benefits such as increased productivity. However, I don't believe these should be viewed as anti-inflationary policies. Treating inflation by attempting to augment productivity begs the basic question: Why is it that the money stock is being allowed to rise faster than output? If the rate of growth in real output is 3-4% per year, why is the money supply growth rate in the range of 10% or better? While I do not propose to tackle this issue here, it is clearly the fundamental inflationary problem. Thus it may be that if productivity is increased, the money supply will simply grow more quickly with no resulting effect on inflation. Simply put, any supply-side policy to reduce inflation is doomed to failure unless monetary policy is complementary. This does not mean that the supply-side approach to raising output should be abandoned. Greater productivity means greater economic welfare. But such policies should be implemented for that purpose -- to increase production. Inflation is a monetary disorder and should be treated as such.