Comment JOHN RAISIAN*
Bureau of Labor Statistics I.
Introduction
The primary purpose of this comment is to extend some of the provacative ideas put forth by Northrup and Northrup (hereafter referred to as N-N) and to offer additional perspectives on the proposed practice of divergent investing of pensions by unions in behalf of their constituents. N-N are to be complimented for their thorough treatment of issues that arise as a result of potential changes in pension fund management. My intent is to couch many of the same issues in a slightly different setting so as to explore a different perspective. As defined in their paper, divergent investments refer to financial outlays which are deemed "socially responsible" - - in this case by unions proposing to utilize the assets o f members' pensions. N-N pursue the effects that these practices are expected to have on employer/employee relations. The basic framework offered in this comment distinguishes the effects of union divergent investment policy into two separate considerations: efficiency and wealth redistribution. Indeed, effects such as these could very well affect relations between employer and employee as postulated by N-N. II.
Efficiency Consideration
A pension system is termed efficient if it receives the highest expected return on investment (net of management costs) for a given level of risk. Any change in an existing pension system which alters the expected return of pension fund investments net of management costs and holding risk constant (irrespective of who's returns are affected) is termed an efficiency effect. There are four potential parties to consider in a pension arrangement: a firm, a union, a pension fund manager, and an employee. Suppose that a firm currently retains a financial institution to manage its pension contributions. If the union obtains the right to control these funds and chooses the same financial institution to handle this management, no efficiency loss would result from the changed pension structure (assuming that transaction costs between parties are relatively small). Alternatively, if it is clear that the union can either obtain a higher return on invested funds or economize on management costs, an efficiency gain is attainable. This could occur if the union has a special talent (relative to the firm) for picking financial institutions that yield greater expected returns (holding risk constant), or if the union itself is able to obtain higher yields on investments relative to its own management costs. *Senior Economist, Office of Research and Evaluation, Bureau of L a b o r Statistics. T h a n k s are due to colleagues in this office for comments and suggestions. Points o f view or opinions stated in this document do not necessarily reflect the official position or policy o f the U.S. Department o f Labor. JOURNAL OF LABOR RESEARCH Volume II, Number 2 Fall, 1981
JOHN RAISIAN III.
215
Wealth Redistribution Consideration
Suppose that pension fund management is currently efficient in the sense just described. Any change in the allocation of the returns net of management costs is termed a wealth redistribution. If the union obtains the right to manage pension funds, the result could lead to wealth redistribution even in the absence of efficiency effects. The union probably will pursue one or a combination of two actions: obtaining a greater share of returns for itself or for the employee it represents. Implicit in these goals is a redistribution of returns from the firm to either union or employee. It is not at all apparent that the firm currently shares in the return accruing from pension investments. For example, in a cost-based pension plan, the firm contributes a fixed amount per worker toward the pension fund. At the time of retirement, the eligible worker then obtains a lifetime annuity, based on the total return that has resulted from prior contributions. The only situation imaginable, whereby a firm obtains a portion of the pension return, requires an assumption of extremely imperfect information. That is, suppose the only signal regarding pension benefits available to the employee (when deciding whether to work for a firm) is the per capita contribution toward the pension fund. This would then allow a firm to announce relatively large contributions, and then skim a portion of those returns at a later date. Yet, so long as prospective workers can observe actual benefit levels of current retirees, it is not clear that this maneuver is viable to the firm. Nonetheless, if the union perceives that the firm currently reserves a portion of the return to pension investments for itself, acquisition of control may then lead to attempted redistribution of the firm's portion to either employee or the union itself. Should the firm have nothing for redistribution, the only available transfer results from employee to union. This might occur in a rather subtle fashion whereby the union skims part of the return much like what is proposed for the firm. Once again, it would seem to require the assumption of imperfect information.
IV.
Current Union Proposals
N-N refer to an AFL-CIO study which asserts that current pension fund managers have a record of poor investments. However, the same study purports to replace current investment strategies with divergent investment policies while maintaining financial returns to the employee. This suggests a perceived efficiency gain of transferring control of pension funds which will be distributed to benefit the union and/or employee. However, maintaining employee benefits at current levels suggests that divergent investment policies are (1) efficient relative to the firm's investment policy with efficiency gains distributed solely to the union, or (2) inefficient but no more inefficient that current investment policies. In the latter case, no redistributions occur between parties relative to the initial situation; however, investment strategies change to the perceived benefit of society. Furthermore, it should be emphasized that these societal benefits are in
216
JOURNAL OF LABOR RESEARCH
essence at the expense of the union and employee in terms of foregone returns if the union instead were able to choose a more efficient policy without regard to society's welfare. As N-N point out, there is no hard evidence to support the claim that current investment policies are either inappropriate or inefficient. If current investment policies are efficient and divergent investment policies indeed maintain returns to the employee, there can be no wealth transfer (assuming the firm was not obtaining a portion of the return prior to union control). This presumes that divergent policies are also efficient. If divergent investments are inefficient relative to current policies, lower returns to the employee result leaving the union open to question about its motivation for divergent investment policies. I can envision two qualifications the union might offer in an attempt to justify control of pension funds even in the absence of enhanced efficiency. First, it might claim that future income maximization is too narrow a criterion and that the employee is better satisfied by foregoing some future income for other things which include divergent investments. In this situation, the union acts merely as an agent of the employee attempting to obtain a more desirable set of attributes for given pension outlays. Validity of this scenario, however, requires a large leap of faith. Second, the union might claim that what is good for the union is good for its constituents. The employee can be thought of as taking a portion of pension contributions and investing them in union activities which promise future rewards, whether monetary (higher future wages or pension bonuses) or nonmonetary (working conditions). In theory, this has merit, yet clearly must stand up to empirical verification prior to employee certification. What activities might a union pursue under the heading of divergent investments? Are they activities which would enhance the well-being of the union and its current members? N-N mention several possibilities. These include denying capital to nonunion employers and employers with large overseas workforces while making capital available for such "social" purposes as subsidized housing for the aged, health care and the like, particularly if provided using union labor. Two fundamental questions must be addressed to evaluate the likely effects of such practices. Does anyone benefit from them? Does the union benefit from them? The first question is obviously broad yet pertinent. If the set of divergent investments under consideration by the union are efficient as defined here, then all such investments will be funded in the marketplace, presupposing that investors are strict income maximizers. In this situation, shifts in investment behavior by unions are offset by compensating shifts elsewhere, and no economic effects should result. Alternatively, divergent investment practices may be inefficient (in the narrow sense defined above) relative to current practices. Certain projects that would ordinarily be funded may now receive attention in this new setting. These practices must offer lower expected rates of return, otherwise they would be funded in the marketplace. Why is it rational for the union to consider such projects? It appears that there is a perception by the union that while it may sacrifice foregone returns to investment, it gains other things by creating more union business. Again, invoking the principle that a stronger union means greater benefits to con-
JOHN RAISIAN
217
stituents suggests a conclusion that divergent investment could lead to results comparable or better than current practices. While lower pension levels result, the hope is that better non-pension benefits will compensate for this sacrifice.
V.
Other Considerations
Implicitly, much of this discussion has focused on a cost-based pension plan. This entails fixed contributions with uncertain returns to the employee. The major alternative is a benefit-based pension plan which (to a degree) fixes the return to the employee. That is, the size of the pension annuity, while depending on salary and tenure levels, is known (generally speaking) well before time of retirement. While this plan is structurally different, it is nonetheless within the scope of the previous discussion. Any pension plan, whether based on costs or benefits, involves outlays of funds. The issues we ponder relate to whether these funds are more efficiently invested or more satisfactorily distributed at time of retirement. Therefore the structural aspects of the plans do not appear on the surface to be o f essential relevance to this discussion. Another issue that mertis attention is employee eligibility and job turnover. Before one can address this issue, the concept that pensions are merely deferred wages to the employee must be qualified. This concept has validity, but only in a present-value sense for the employee that has provided extensive years o f service to the firm. The rationale for this stance hinges on the presence of firm-specific skills which require training outlays in the early stages of employment. If job termination occurs at a relatively early stage, net benefits resulting from specific training expenses are often negative. However, they are expected to become less negative and eventually positive as job tenure increases, thus meriting a return to the worker for acceptable performance and job stability. This can take the form of a vested pension. This issue can complicate the contractual relation between firm and union if the union were to gain control of pension funds. In particular, it is my impression that many unions claim a narrower view of the deferred wage concept. That is, worker value marginal product is seen as exceeding wage levels without regard to tenure level. This would indicate a pension contribution which need not be turned back to the firm in the event of early job separation. Even if the union were to accept the broad deferred wage concept, establishment of contribution levels per worker would seem to be a problem. For example, both parties would have to agree on expected turnover rates so as to accumulate reasonable total amounts for those workers expected to reach eligibility. Alternatively, the firm could contribute to the union full support levels per worker and be refunded the contributions (plus interest) if job separation were to occur prior to vesting. However, an issue of job separation versus union separation arises. One can foresee a union arguing against the return of funds if the employee separates from a firm but obtains employment elsewhere without having to terminate union enrollment. Despite this, the firm may feel that the opportunity cost of such funds is higher than the amount to be refunded by the union. While it would seem that a bargain could be struck with regard to this issue, it would not be without its cost and uncertainty. In fact, on the surface, it appears to be inefficient for unions to control pension funds if the issue of eligibility and job turnover was the only consideration.
JOURNAL OF LABOR RESEARCH
218
VI.
Conclusions and Impressions
I think the spirit of these comments conform to the arguments and evidence outlined by N-N. To a degree, my comments are a restatement of propositions by N-N, yet I have tried to repeat them using a different analytical structure. At the same time, I have attempted to offer new propositions and raise new questions. Since the stated purpose of N-N's research is to evaluate the effect of union control of pension funds on employer/employee relations, I conclude With my impressions to the following question: Under what circumstances would the firm not object to the union controlling employee pension funds? I think the answer is simple. As long as the divergent investment scheme is efficient, not redistributive, and not subject to differences in viewpoint with respect to the eligibility and tenure issue, the firm is indifferent to union control. The likelihood of these conditions being met is nil suggesting that added friction in employer/employee relations can be expected.