Pensions Could Be Better

by K. MINTURN SEDGWICK

I

WHEN a man buys a house or a ear, he knows that he will use it at once. Both he and the architect or engineer know that the good and bad feattures of those things will show up within a matter of months if not weeks. Largely for this reason, Americans are today buying good ears and good housing. In the decade 1950-1960 the American public will spend $100 billion on new cars and a like amount on new houses.

There is another product which has only recently come into equally general demand, and by the end of the decade the public will probably have spent nearly $100 billion on it too. It is an intangible product and, unlike cars and homes, it is not bought by the individual user but is bought for him. It does not have to face the lest of use as a rule until twenty, thirty, or even fifty years after the first down payment is made. Unfortunately, any realistic appraisal of this product point s to the inescapablo conclusion that the public, is getting an extremely poor value for its money, not because il is being fleeced by the unscrupulous, but because neither vendor nor buyer appears to be aware of the probable long-time performance of the product. I mean pension plans.

Any dispassionate study will show that almost without exception every pension plan at present in operation would have failed in its purpose spectacularly in the past half of this century and - unless the planning is changed — will almost surely fail in the future. This failure is clearly evident in the few plans which are more than twenty years old, and the failure has already resulted in inadequate payments to the pensioners despite the fact that the original schemes have usually been supplemented at great additional cost. There are few if any funded pension plans that have pensioners who entered the plan half a century or more ago. If there were more, the failure of the standard type of plan would by now be generally recognized.

Most people will agree that a pension plan should be designed to provide a standard of living for the pensioner and his wife should she survive him, bearing a reasonable relation to that which his compensation permitted during Ids employment. Were living costs unchanging, a pension of somewhere between 40 and 60 per cent of average compensation during the last ten years of employment might well be considered ample. Many present-day plans set out with such a goal in mind; but almost without exception the pension is designated for one life, with an unchanging number of dollars regardless of the cost of living. If the plan has an alternative which will protect a widow, it usually entails a diminution of about one third of the benefit. This diminution is based on t he finding that the survivor of a couple, when the man is 65 and the wife is 62, has a life expectancy of 23 years as against the man s alone of 14 years. It is unfortunate that the pension prospect us generally fails to emphasize that the pension rates quoted are for one life only. Thus a $2400 per year pension shrinks to $1000 if it has to protect a man s wife as well as himself.

Let us turn the clock back and assume that we are starting a standard pension plan in 1900, and then see what happens to its beneficiaries. In 1900 the average annual industrial wage was about $400. Thus a pension of $200 would appear generous. This assumption that living costs would remain fixed was far more reasonable in 1000 than today because the preceding thirty years had seen one of those rare periods in economic history when the price level had failed to rise.

It was reasonable in 1000 to assume that the net average earnings on a fund invested in fixedincome-bearing securities would produce 4 per cent. (At present pension planners usually assume 2.5 per cent.) In 1900 a man (in had a life expectancy of about IB years. Therefore, a capital fund of approximately $2000 would be needed, when a man retired, to pay him about $200 annually for life. Taking the ease of a worker 25 years old in 1900, it was fair to assume that $20 per year contribution over a forty-year period, invested and reinvested at 4 per eent, would produce about $2000, which in turn would provide the $200 annual pension. Such a plan in 1900 would doubtless have seemed incredibly generous; in fact, hardly anyone was bothering about the fate of retired workers at that time.

But when our sample man retired in 1940 with a wife four years his junior, it would have taken some remarkable planning to make $200 cover his annual living expenses. It is obvious that the pension would have failed completely to provide for the worker’s old age, if single. And if there were an option in the plan which would provide for his wife should she survive him, the pension paid would have been reduced to $133. Had it not been for Social Security, this couple would have had to be supported by private or public charity if they had no savings to fall back on.

But let’s be fair; let’s assume that the company would have recognized by 1920 that if its pension plan was to have any meaning it would have to find a capital sum of at least $5000 for each man by 1940 to provide an adequate worker’s pension. Even had they been able to invest money the ensuing twenty years at 4 per cent, the contribution would have had to be increased by roughly $100 a year. Thus the cost would have grown from $20 to $120 per man in that age group. Had there been thousands of employees involved, it is evident that the financial problem produced by increasing the pension plan would have been staggering. When 1940 arrived the net return on fixed investments had fallen to 2.5 per cent, so that $5000, instead of buying a single-life annuity to produce $500 annually, would only have produced about $450 for one life or $300 for two. Long before 1953 arrived, the inadequacy of the pension would have been apparent to all. Yet the cost of the pension during the second half of its funding period was six times what was originalh planned! Thus it is evident that the type of pension which provides a fixed sum annually would have failed to protect the worker even if greatly liberalized and at a cost to the company many times what it had anticipated.

2

ANOTHER and more expensive, contributory form of pension plan is frequently provided for teachers, clergy, doctors, and welfare workers. When it first came into vogue, about the time of World War I, it was supposed that an annual contribution of 7.5 per cent of salary, if started when a man was in his twenties, would ensure a capital fund sufficient on retirement for an adequate pension. Actual experience has already shown this to be inadequate even when the contribution was stopped up to 15 per cent. Most of such plans today assume earnings of less than 3 per cent, although one which is offered by a leading university close to Boston practically guarantees a 3.5 per cent return and is based on an 11 per cent contribution. The prospectus of the plan shows how a full professor could expect to retire at sixty-seven on a pension which, combined with Social Security, would provide roughly half salary for two lives. This appears to be adequate by today’s standards, provided there is no increase in the cost of living in the next fifty or sixty years. About two thirds of this combined pension is supplied by the university’s scheme.

Let us see how this same plan would have worked if initiated in 1900, and let us assume, since there was no Social Security in 1900, that the contribution was 15 per cent instead of 11 per cent. At the turn of the century the following university salary scale was common : —

The pension built up by this compensation would have been about $1420 annually for two lives, or just under 50 per cent of the final salary.

Conditions changed; teachers’ salaries went up, so did expenses; and when 1940 came, the professor found he had actually lived by the following salary patlern:—

This would have resulted in a pension for him and his wife of about $1910. Although this was materially larger in dollars than he had anticipated, it was not adequate in 1940. Imagine the straits in which the poor professor would have found himself in 1953, unless the university decided to make supplementary grants. In other words, this plan too would have failed, but not as dramatically as the industrial pension we studied earlier, for the simple reason that its cost was three or four times as much in relation to payroll.

As the cost of living has mounted so rapidly in the last few years in this country — though far, far less than in most countries of the free world—more concern is at last being given to our pensions. A not infrequent provision is that a man shall have a percentage of he av erage of his final ten years of salary. This is a helpful step but it is likely to be both expensive and disappointing. Had the professor whose salary and pension we have just studied had an agreement that he should retire on 40 per cent of final salary for a two-life pension, he would have been entitled to $2400.

To provide for this in the last fifteen years of his service would have necessitated an additional annual contribution of about $370, thus increasing the cost of the pension to about 22 per cent of average salary. Few if any institutions have resources adequate to finance such an increase in pension costs; yet even had this heroic remedy been undertaken, the total pension of $2400 would have hardly proved adequate as the sole support of an elderly retired professor and his wife in 1953.

Brief as this review has been, it leaves little doubt, as to the three major weaknesses of present-day pension plans. The first is that they assume no responsibility for widows —or if provision is made, it is at the cost of one third of an already inadequate pension. The question, Why should a pension be expected to support a widow? can best be answered by another: Why should wages or salary be expected to support a wife or family? The second is that these pension plans completely ignore the steady upward trend of real wages for manual workers due to greater productivity. The third fault is that an unvarying price level is anticipated, when economic history provides few instances of such stability and many of a sharply rising trend. The reader may well ask, Are all pension plans foredoomed to failure, despite periodic revision at ruinous expense? This is a serious question in view of the very large amounts of money currently going into pension plans, and it cannot be answered either reassuringly or categorically. However, it can be said that a plan can be devised today that would have worked successfully not only during the past half century of rising prices but also during the previous thirty years, when the price level failed to rise. The final section of this paper will describe it.

3

Two years ago I was asked to divise a pension plan whose assets were to be invested entirely in common stocks. It seemed to me an impossible task because, like everyone else, I had supposed that if a pension plan were to be valid it had to promise a certain dollar income at a future date, and I felt that only through fixed-income-bearing securities was it possible to guarantee a fixed number of dollars thirty, forty, fifty, or even sixty years hence. After studying the problem, I came to realize that a promise of a certain number of dollars half a century hence was meaningless as no one could tell what the dollar would buy. A study of the purchasing power of money over several centuries pointed to the conclusion that at the end of any half century any currency was likely to buy less rather than more goods and services than at the start. A rapid review of the fate of pensions based on such famous currencies as the French franc, the German mark, and even the pound sterling, left me with the feeling that to believe there was any long-term security in the promise of a certain number of dollars required an unusual feat of optimism.

Without the necessity to assure a fixed number of future dollars we are able to use an investment medium which will produce the maximum results for a given investment. The Cowles Commission, composed largely of members of the faculty of the University of Chicago and financed by Alfred Cowles, found that the average annual income of substantially all the industrial common stocks listed on the New York Stock Exchange was 6 per cent from 1871 through 1938. It found also that during the same period the average annual appreciation was 2.8 per cent or a total annual average investment performance of 8.8 per cent. Subsequent studies, although less complete, clearly indicate that the record has been even better from 1938 to 1953.

Thus we have an investment medium which, over an eighty-year period, had an average total investment performance of 9 per cent. Assuming that the operating costs of an actual fund were 1 per cent per annum, we might expect to find an average accomplishment of an all-common-stock investment fund of about 8 per cent. As a matter of fact this is the case. The three oldest of the large investment trusts, each having assets in excess of $100 million, date back to 1925. All have been invested mainly or wholly in common stocks. Had $100 been invested annually in each of these funds since 1925, the least successful would have enjoyed an investment accomplishment of 8 per cent per annum and the most successful better than 12 percent.

It will be recalled that in the pension plan we studied earlier in this paper, 3.5 and 4 per cent net earnings were assumed. The following facts show the overwhelming advantage of a higher rate of return. One hundred dollars invested annually at 4 per cent produces, over a forty-year period, $9882, while at 7 per cent it produces $21,360. If the $9882 is used to buy a standard two-life annuity (man 65, wife 62), assuming 4 per cent basic earnings, the annual payment will be only $660. Had the interest rate been 3 percent throughout, the annuity would have been but $472. If 7 percent is charged annually against the $21,360 which is invested in assets earning 7 per cent, the average payment will be $1495.20. Furthermore, if the basic earnings continue at that rate, the fund will not be reduced even though payments continue indefinitely. So much for the theory on which a common-stock pension can be based.

Everyone familiar with common stocks is aware of the wide fluctuations which occur in their price, and this might seem a bar to their use in a pension plan. Let us see, therefore, how the pension plans and pensioners we studied from 1900 on would have actually fared had the annual contribution been invested and reinvested in the Cowles Commission Industrial Common Stocks, less 1 per cent for operating costs up to 1929. From 1929 an actual wellknown common stock fund will be used as the investment vehicle. It was selected for two reasons: first, its results were almost identical with that of the Cowles Commission series during ihe period in which they overlapped, 1929-1938. Second, since 1938 its performance was in line with a general index of common stocks.

It will be recalled that in the first plan we studied, the company decided to place $20 annually per employee in a pension fund and thus produced an annual pension for two of $133 — or would have if orthodox pension investments had earned 4 per cent as late as 1940. Had the $20 been placed annually in a common-stock fund such as we have described, a capital of about $3960 would have been available by 1940. If this had been charged at the rate of 7 per cent annually, it would have provided average payments of $326, or two and a half times the orthodox pension; and even in the worst year, 1942, it would have paid $235, and this would have risen to $427 in 1952. Though not an adequate pension, it would have averaged larger than the expanded orthodox plan of $300 which we saw would have cost $120 annually during the last twenty years of the man’s employment as against the $20 per year of the common-stock plan. This is not the place to describe the mechanics of accumulation and payments, but thanks to modern mechanical bookkeeping the cost would be negligible.

Now let us turn to the type of pension in which the pension contribution is a percentage of salary. We saw in the university plan that 15 per cent might have been placed annually in a pension fund.

Had 15 per cent of compensation been put each year into the common-stock investment medium, the professor would have built up a capital of about $48,000 which would have provided an average retirement income of $3959. The minimum in 1942 would have been $2852, and by 1952 it would have climbed to $5175 instead of the $1910 which the university’s plan secured.

Furthermore, in the case of both of these examples there would have been a relatively large block of capital on January 1, 1953. In the case of the worker’s pension it would have been $6268, and in the case of the professor about $75,900. It is reasonable to suppose that ten years hence these would be at least the same size as at present, and very possibly considerably larger.

Skeptics may argue that the period studied was selected to show either the maximum pension under the proposed plan or the maximum accumulation of capital. As a matter of fact, had these two common-stock pension plans been started any time after 1870, the pensioners would have enjoyed equal or larger average benefits and the capital available on the death of the pensioners would have been about the same as was available in 1953. Only in the year 1932 would the pensions have gone below the 1942 minimum, and then by about 25 per cent.

In the case of a charity it could be arranged to have the capital revert to the institution after the beneficiaries died. Imagine what a boon it would be for a university to receive something in the neighborhood of $50,000 to $75,000 on the death of a retired professor and his wife.

We have seen that owing to the fact that pensions are normally figured on one life, though they actually should provide for two, they at best furnish only two thirds of the dollar benefit originally anticipated, and at worst leave the pensioner’s widow dependent on charity. The average rise in the cost of living since 1900 has been 2 per cent per annum. This creeping long-term inflation lias made our present conventional pensions grossly inadequate by any reasonable standard. When common stocks are used the pensions average more than twice as large and the capital, as measured in dollars, on which they rest remains intact or grows, leaving large sums available either to the beneficiaries’ heirs or to the institution.

What of the future? No man can answer this with assurance. For the pessimists the British experience is encouraging. Since World War I, British industry has endured nearly all the ills that might overtake business in the United States in the next half century. These include tragic unemployment, Steady depletion of basic raw material reserves, ruinous taxation, total war, and socialist government. Nevertheless, a study by the Economist of London showed that a representative list of industrial stocks in Great Britain turned in an average annual investment accomplishment of belter than 8 per cenl during the period 1919—1953. If British business could obtain such a performance during the last third of a century, is it not possible that we in America may do as well in the next half century? Even if stock investment results drop to as little as 5 per cent, if used in the manner described they will produce far better pensions than the standard plans and will accumulate a large amount of capital instead of using it up.

It is heartening to know that practically every pension plan now in existence can be modified along the lines set forth in this article, to take advantage of the long-term benefits of common stocks. In view of the vast sums being spent by business and charitable institutions in pension plans, it is important not alone for the beneficiaries, but for the national economy, that the money be wisely used. Let us hope that those who control the use of these funds will have the courage and foresight to realize that unless pension plans are changed, they will be likely to prove as expensive, as wasteful, and as disappointing on a large scale in the future as they have on a small scale in the past.