Investment Decisions for Public Pension Funds

This figure includes the Federal Civil Service Retirement System but excludes the Social Security System (OASDI). Securities and Exchange Commission, Statistical Series #1978, June 4, 1964.

Ibid., and U.S. News and World Report, June 24, 1963, p. 106.

For a thorough discussion of these and other methods, see Cottle C. C. Whitman W. T. , Investment Timing: The Formula Plan Approach (New York: 1953).

For purposes of the present paper, we will ignore yield differentials among different classes of bonds and use the yields on high-grade utility bonds as our measure of bond yield.

It is presumably obvious that we are concerned here not with the accountants' “stated earnings,” but with the economists' concept of income. To the extent that these two concepts do not coincide, published earnings figures would need to be modified. In practice, it would only be necessary to adjust stated earnings in instances where the two differed by substantial amounts.

Two of the companies (General Electric and Standard Oil of New Jersey) had increases in market value of the common stock outstanding in 1946 of over $4 billion each, whereas the average for the remaining 42 was $90 million, and the next highest was only $665 million. If one or both of these were a growth company as defined by Malkial Burton G. (“Equity Yields and Growth,” American Economic Review, Dec., 1963, pp. 1004–1031), and there were no “decay” company of comparable absolute size, their inclusion would obviously distort the results.

To avoid any possible misunderstanding, we must emphasize that we are not suggesting that, because company A reinvests earnings of $1 million in a given year, its common stock outstanding will be valued $1 million higher at the end of the period than at the beginning of the period. Rather, we are suggesting that, for a diversified group of issues, those issues in the group which have reinvested larger amounts of earnings will enjoy a greater increase (or a lesser decrease) in their values than those which reinvest smaller amounts of earnings and that if we make our comparison between two points in time in which earnings-price ratios are comparable, the differential in value will, on the average, approximate the differential in reinvested earnings if the dollar figures are corrected for inflation.

See, e.g., Baumol William J. , “Comment on Dividend Policy, Growth and the Valuation of Shares,” Journal of Business, Jan. 1963, pp. 112–115; Gordon Myron J. , The Investment, Financing and Valuation of the Corporation (Homewood, Ill.: 1962); Miller Merton H. Modigliani Franco , “Dividend Policy, Growth and the Valuation on Shares,” Journal of Business, Oct. 1961, pp. 411–433; and Friend Irwin Puckett Marshall , “Dividends and Stock Prices,” American Economic Review, Sept. 1964, pp. 656–682.

See Burns Arthur F. Mitchell Wesley C. , Measuring Business Cycles (New York: 1946), pp. 56–114.

In the remainder of this paper, we will arbitrarily use a three-year moving average in our calculations for sake of uniformity and simplicity.

Again we must recognize that neither the prices of individual issues nor average stock market values will promptly reflect the impact of inflation. Day-to-day or even year-to-year movements in market prices are so strongly influenced by other factors that the impact of inflation is almost certain to be completely masked, but over a period of years it can be expected that dollar earnings will rise in proportion to the degree of inflation so that “real” earnings will be maintained. See Bronfenbrenner Martin Holzmann Franklyn D. , “Survey of Inflation Theory,” American Economic Review, LIII (Sept. 1963), 646–652, and the references cited therein.

In symbolic form, where x is the fraction of the fund to be invested in stocks, our decision rule would be:

If A = B, x = S/(S+L)

If A > B, x is the greater of S/(S+L) and (A–B–L)/(A–B)

If A < B, x is the lesser of S/(S+L) and S/(B–A)

See Cottle Whitman , op. cit., chaps. 6–10.

The executive officer of one of the largest public pension funds has objected that our decision rule would not be practical for a very large fund since it would require, both initially and subsequently, very large purchases and sales and that these might have an influence on the movements of the market. Where markets are so thin that purchases or sales by a single fund might significantly affect the market price, it would be necessary to use the average price that could be obtained for the stock (or at which it could be bought) in the relevant quantities, rather than the current market price. The same problems would, of course, plague a fund using the constant- or variable-ratio approaches. Under strict dollar-averaging, the problem would not arise; but this advantage is bought at the price of a rigidity which we find repellent—a rigidity which requires us to say—regardless of how low expected yields on stocks may be, we will never sell them, rather we will continue mechanically to put x per cent of our new money into them.

In situations where legislative fiat or the existence of certain specific obligations required that the pension fund should always hold at least some minimum percentage of its portfolio in bonds (or stocks), e.g., if the probability of forced liquidation were substantial and the fund (or the responsible public jurisdiction) would then be required to meet certain fixed dollar obligations, this proviso could simply be added as a further constraint to the percentages produced by our decision rule.

See Greenough William G. , “Pensions Meeting Price Level Changes,” McGill Dan M. , ed., Pensions: Problems and Trends (Homewood, Ill.: 1955); or his A New Approach to Retirement Income (New York: 1951).

See, e.g., Howell Paul L. , “Management of California Pension Funds,” California Management Review, V:1 (Fall 1962), 38–40.

It is interesting that Benjamin Graham in his analysis of T.I.A.A.—C.R.E.F. comments, “the people behind College Retirement Equity Fund (C.R.E.F.) are eminently wise in insisting that its beneficiaries have at least an equal dollar stake in bond as in stock investment” (“Investment Aspects of Accumulation through Equities,” Journal of Finance, May 1962, p. 214) and that Eli Shapiro in his “Discussion” (ibid., p. 218), seems to suggest that an even higher percentage of the fund should be in stocks. In fact, a participant in T.I.A.A.—C.R.E.F. on a “50:50” basis may find himself with a very high percentage of his funds in stocks (74 per cent of our example after only 17 years' accumulation) and at a time when the stock market is particularly vulnerable. We question whether this is “eminently wise” under present conditions unless a substantial degree of inflation is predicted.

On the potential importance of investment of pension funds in real estate through leasebacks, and optimal decision rules regarding such investments, see Ricks R. Bruce , Recent Trends in Institutional Real Estate Investment, Research Report 23, Center for Real Estate and Urban Economics, Institute of Urban and Regional Development, University of California (Berkeley: 1964).

We feel justified in saying this, despite the fact that, in our test, the deflationary variant of the decision rule resulted in a fund of $32.8 million whereas the dollar-averaging approach produced a $34.9 million fund, because it seems most unlikely that any manager who was as deflation-minded as was assumed in our deflationary variant would have been willing to dollar average on a 50:50 basis throughout the period.