Warren Buffett on Pensions (1975)
This is the full text letter from Warren Buffett to Katherine Graham discussing pensions, as released by Fortune. I find this easier to read on things like Instapaper than the PDF version.
PENSIONS
There are two aspects of the pension cost problem upon which management can have a significant impact: (1) maintaining rational control over pension plan promises to employees and (2) increasing investment returns on pension plan assets.
The Irreversible Nature of Pension Promises
To control promises rationally, it is necessary to understand the basic arithmetic and practical rules governing pension plans. The first thing to recognize, with every pension benefit decision, is that you almost certainly are playing for keeps and won’t be able to reverse your decision subsequently if it produces subnormal profitability. As a practical matter, it is next to impossible to decrease pension benefits in a large profitable company—or even a large marginal one. The plan may embody language unequivocally declaring the company’s right to terminate at any time and providing that contributions shall be solely at the option of the company. But the law has eroded much of the significance such “out” clauses were presumed to have, and operating practicalities render any residual rights to terminate moot. So, rule number one regarding pension costs has to be to know what you are getting into before signing up. Look before you leap. There probably is more managerial ignorance on pension costs than any other cost item of remotely similar magnitude. And, as will become so expensively clear to citizens in future decades, there has been even greater electorate ignorance of governmental pension costs. Actuarial thinking simply is not intuitive to most minds. The lexicon is arcane, the numbers seem unreal, and making promises never quite triggers the visceral response evoked by writing a check. In no other managerial area can such huge aggregate liabilities—which will be reflected in progressively increasing annual costs and cash requirements—be created so quickly and with so little immediate financial pain. Like pressroom labor practices, small errors will compound. Care and caution are in order.
Deceptive Arithmetic of “Promise Now—Pay Later”
If you promise to pay me $500 per month for life, you have just expended—actuarially but, nevertheless, in a totally tangible sense about $65,000. If you are financially good for such a lifetime promise, you would be better off (if I have an average expectancy regarding longevity for one my age) handing me a check for $50,000. But it wouldn’t feel the same. And, if you promise to pay me 1,000 hamburgers a month for life which, superficially, may sound equivalent to the previous proposition (assuming a present hamburger price of 50¢), you have created an obligation which, in an inflationary world, becomes most difficulty to evaluate. This creates a risk we talk of later as an “earthquake risk”. One thing is certain. You won’t find an insurance company willing to take the 1,000-hamburgers-a-month obligation off your hands for $65,000—or even $130,000. While hamburgers equate to 50¢ now, the promise to pay hamburgers in the future does not equate to the promise to pay fifty-cent pieces in the future. So, before plans are introduced or amended, the financial consequences (particularly in a world of significant inflation which I believe to be close to a certainty)[*] should be clearly understood by you. Consulting actuaries are very good at making calculations. They are frequently terrible at making the assumptions upon which the calculations are based. In fact, they well may be peculiarly ill-equipped to make the most important assumptions if the world is one of economic discontinuities. They are trained to be conventional. Their self-interest in obtaining and retaining business would be ill-served if they were to become more than mildly unconventional. And being conventional on the crucial assumptions basically means accepting historical experience adjusted by a moderate nudge from current events. This works fine in forecasting such factors as mortality and morbidity, works reasonably well on items such as employee turnover, and can be a disaster in estimating, the two most important elements of the pension cost equation, which are fund earnings and salary escalation. [*] My views regarding inflationary possibilities are more extreme than those of most respected observers. It is difficult to substantiate a dogmatic view, since conclusions rest more upon social and political judgments than upon economic training and analysis. You should recognize the subjective nature of the reasoning leading to my pessimistic conclusions regarding inflation over the longer term—and not be reluctant to correct accordingly.
Illustrated Elemental Actuarial Principles
To illustrate a few actuarial principles worth understanding, but without employing the technical jargon and the asterisks, let’s use the example of your household. Assume that you, personally, make irrevocable promises to pay pensions of $300 per month for life after they reach 65 to, say, four household employees. To make it easy, let’s say that they each are 55 years of age at present. If you make that promise today, you have reduced your net worth today by about $70,000. (For simplicity’s sake, I am ignoring some variables such as sex of employees—women live longer and therefore cost more death before 65, etc. In calculating such factors the actuaries shine. ) Why are you immediately $70,000 poorer? Well, if you set aside a $70,000 fund now and invest it at 7% interest—and let all such interest remain in the fund to be compounded—the principal value of the fund will grow to about $140,000 in ten years when your employees reach 65. And to buy them a lifetime annuity of $3,600 per year will then cost about $35,000 each, utilizing the entire accumulated capital of the fund. So if you make the promise and it is binding legally or morally—figure you have spent today $70,000, even though you don’t have to payout a dime of cash for ten years. Now take it one step further and assume that your employees each are earning $600 per month but, instead of promising them $300 per month upon retirement, you promise them 50% of their salary at the time they retire. If their increases run 7.2% per year—and they probably will in the world I foresee—they will be earning $1,200 per month by the time they reach 65. And it will now cost you $70,000 each to purchase annuities for them to fulfill your promise. The actual cost, today, of modifying your promise from 50% of present pay to 50% of terminal pay was to exactly double the fund that needed to be set aside now from $70,000 to $140,000. Many pension plans use final average pay (usually the average of the last five years, or the highest consecutive five years in the last ten years employed) and some use career average pay. I am not arguing here which should be used, but am illustrating the dramatic difference in costs that can occur because of rather minor-appearing changes in wording. Pension costs in a labor intensive business clearly can be of major size and an important variable in the cost picture, particularly in a world characterized by high rates of inflation. I emphasize the latter factor to the point of redundancy because most managements I know—and virtually all elected officials in the case of governmental plans—simply never fully grasp the magnitude of liabilities they are incurring by relatively painless current promises. In many cases in the public area the bill in large part will be handed to the next generation, to be paid by increased taxes or by accelerated use of the printing press. But in a corporation the bill will have to be paid out of current and future revenues—with interest—and frequently with what is, in effect, a cost-of-living escalator.
The “Earthquake Risk”
In Germany, in the great inflation of the early 1920s, the entire Daimler Benz (Mercedes) Automobile Company was selling in the market for the price of 300 motor cars. Almost all past investments were nearly worthless, and current salary levels were astronomical in relation to past history. Under such conditions, or conditions far short of such an extreme, the burden of any pension benefits owed by a business, which are based on current salary levels though related to much earlier service in employment, must be backed almost exclusively by the current value (earning power) of the business. Advance funding simply evaporates. For example, assume that salaries (and the cost of living) are moving upward at 25% per year and the pension fund is earning 10% per year—a set of assumptions not ridiculously different from what exists in England at the moment. Under such conditions, funds put aside for retirement immediately begin to shrink in relation to promised benefits. Every month fewer hamburgers can be purchased from the funds contributed to the pension plan—even after accumulating dividends and interest on the funds. Almost no one chooses to think about this sort of “earthquake risk” in dealing with pension plans, any more than people choose to spend much time thinking about nuclear war. It may be my earlier-mentioned bias, or my mathematical bent, but I believe some “unthinkable” inflation-related calculations should be made—and even considered before any company assumes open-ended pension obligations guaranteeing a large number of persons absolute protection against inflation by gearing benefits without limit to final pay or escalating benefits to persons already retired, based on changes in the cost of living. Thus, the really devastating possibility regarding private pension plans is sustained double-digit inflation. When salaries move ahead at a substantially higher rate than investment returns and benefits are tied to final salaries (or, even more expensively, cost-of living increases after retirement as in recent rubber and aluminum contracts), it is virtually impossible to pre-fund obligations. Like it or not, you become much like the Social Security Fund, absent the power to tax. Should that occur, future purchasers of the products of the company must be willing to do so at prices that reflect not only the wages of current workers, but the promises to past workers. Some businesses will have economic characteristics allowing them to pass along these costs, but others will have major troubles. On balance, I believe we are in relatively favorable businesses under such circumstances. I do not believe this problem can be solved by the investment process. I mention it for completeness, not because I have answers—and to urge caution in making pension benefit promises subject to dramatic escalation through substantial attrition in the purchasing power of money. Now let’s look at funding and investment behavior appropriate to an economic world at least reasonably similar to the past, recognizing that such a world is far from a certainty.
The Investment Management Problem Inherent in All Pension Plans
Once having committed to provide pensions, how do we pay? The law and prudent business practice mandate that we start putting aside funds on a fairly orderly and consistent basis from which we can fulfill our promises. In this manner we pay the employee currently while he is being productive for the company, and we simultaneously set up a savings account (collectively, not individually) which will accumulate at interest so as to purchase an annuity (not actually “purchased” in most plans, of course, just assured by the pension fund) fop him at retirement which will discharge our promise to pay him throughout his non-productive years. Thus, our current cost and current cash requirements (if our estimates are accurate as to what we will earn on the savings account as well as other estimates regarding turnover, salary escalation, longevity, etc.) will reflect his total lifetime employment costs to us spread rather evenly over his productive years. (This advance funding treatment, matching full current costs against current production, contrasts with the Social Security Plan’s program which essentially taxes current producers to pay current non-producers. This simply means moving a portion of current national output of goods and services over from those who produced it to those who are non-producing, and to whom promises have been made. If such a system had been in effect for a very long time, the demographic profile remained fairly constant, the promises remained fairly constant, and there was no inflation, the net effect from such a pay-as-you-go approach would not substantially differ from an advance funding basis. However, these conditions do not exist which may make for various problems—including some that exacerbate inflation and thus have negative fall-out for the economics of private plans.) Because a business corporation, unlike the Federal government, has to create a “savings account”—an investment accumulating and investment management operation of some sort—to properly fund its pension plan, it must make investment management decisions, with respect to pension plan assets.
The History of Corporate Pension Plan Management Act One—The Awakening
A few decades ago pensions were a relatively new thing at most companies, so that the “savings accounts” were in their formative stages and therefore much smaller. Furthermore, promises were fractions of those presently made, so that the amount eventually required in the savings account to purchase the required annuity at retirement was correspondingly smaller. Thus the amounts paid into pension funds (“savings accounts”) were largely forgotten so far as managerial responsibility was concerned until the great awakening of the 1960s. At that time managements noticed:
(1) The funds had grown to the point where they sometimes were 25% to 50% of the net worth of the company—often making the assets employed in the savings account larger than those employed in the company’s largest division. Here are a few figures which I have handy from year-end 1972:
(In millions)
Company
Corporate Net Worth
Pension Fund Assets
A & P
$ 599
$ 236
DuPont
3,268
1,817
Firestone
1,251
423
IBM
7,555
1,023
U. S. Steel
3,577
2,239
So, while U. S. Steel had a visible $3.6 billion in net operating assets which management probably spent 99% plus of their business hours thinking about, they had $2.2 billion in the “bank”, whose economic results would impact future values for the shareholders, dollar for dollar, with the economic results of the steel assets. There literally were years when the savings account earned more than was earned out of all operating assets of the steel business. (In fairness to U. S. Steel, it should be mentioned that they were one of the pioneers in recognizing the importance of pension assets—and have done a better-than-average job through in-house management.) (2) The returns actually realized on the “savings account” had an enormous impact on costs. A sustained 1% change in earning rate could easily swing the annual cost to the contributing company by 15%. (3) During a period when equities had produced fabulous returns, many of the plans had been invested largely in bonds—which not only bore low fixed rates of 3% to 5% in the earlier periods, but also had suffered significant shrinkages in market values as interest rates increased secularly. (If interest rates go up, bond prices must go down—and if the bonds are long-term and the rates rise sharply, prices go into a power dive.) (4) Many managements thus saw their largest division earning dismal rates of return with substantial market value shrinkage in the bond component, while all around them high returns were being realized from stocks with little apparent effort or talent. If a company had $100 million invested in its engine division earning 12% by managerial zeal and ingenuity, why tolerate $100 million in its pension fund “division” poking along at only 4% because of inattention—particularly when increasing the $4 million to some larger figure would have the same impact in future earnings for owners as raising margins on engines. Intensive effort on production, research and sales might only produce an increase from 12% to 13% in the engine area, since decisions already had been so near to optimal, but it was easy to imagine 4% becoming 10% in the pension fund area if just average results were attained in equity investment. And, of course, who would settle for being just average?
The History of Corporate Pension Plan Management Act Two—The Great Leap
And so the hunt was on. Wall Street abhors a commercial vacuum. If the will to believe stirs within the customer, the merchandise will be supplied—without warranty. When franchise companies are wanted by investors, franchise companies will be found—and recommended by the underwriters. If there are none to be found, they will be created. Similarly, if above-average investment performance is sought, it will be promised in abundance—and at least the illusion will be produced. Initially those who know better will resist promising the impossible. As the clientele first begins to drain away, advisors will argue the unsoundness of the new trend and the strengths of the old methods. But when the trickle gives signs of turning into a flood, business Darwinism will prevail and most organizations will adapt. This is what happened in the money management field. The banks had traditionally been the major money managers (leaving aside insured plans) and, by and large, their investing as well as their communication had been lackluster. They felt obliged to seek improvement, or at least the appearance of improvement, as corporate managers searched for yardsticks by which to make their decisions as to whom care of this newly discovered giant “division” should be granted. The corporate managers naturally looked for groups with impressive organizational charts, lots of young talent, hungry but appropriately conscious of responsibility, (heavy on MBAs from good schools), a capacity for speed in decision making and action—in short, organizations that looked something like they perceived themselves. And they looked for a record of recent performance. Unfortunately, they found both. A little thought, of course, would convince anyone that the composite area of professionally managed money can’t perform above average. It simply is too large a portion of the entire investment universe. Estimates are that now about 70% of stock market trading is accounted for by professionally managed money. Any thought that 70% of the environment is going to substantially out-perform the total environment is analogous to the fellow sitting down with his friends at the poker table and announcing: “Well, fellows, if we all play carefully tonight, we all should be able to win a little.”[†] [†] An interesting example of this line of thinking (sub-species: wishful) occurs in the April, 1975, Conference Board “Trends in Pension Fund Administration” article which Marty sent me the other day. In a carefully written “Investment Guidelines” statement by a manufacturing company, it is announced: “We believe it is reasonable to expect long-term results superior to the usual market indexes, and the S & P 500 in particular. Specifically, we look for performance better than this index in all types of market environments.” And yes, Virginia, maybe every football team can have a winning season this year. So, clearly the almost universal expectations of above-average performance in pension fund management were doomed to disappointment. These disappointments were certain to be amplified by a corollary affliction that frequently accompanies pressure for investment performance—higher turnover rates. It is difficult to measure turnover costs with precision, but they certainly must run at least 2% on average when applied to the round trip of purchase and sale. If an investment manager, striving for not only acceptable quarterly performance but also for the appearance of behaving as other highly-thought-of managers are known to be behaving,[‡] moves aggressively to keep his portfolio in the “right” stocks, he easily can average turnover rates of 25% per year. When the performance rage peaked, drastically higher turnover figures were recorded with some managers. [‡] In the short term, it frequently is better to look smart than to be smart, particularly if your employment is to be decided by a rather brief interview. If the fans are going to decide your hiring status based on only a few swings, it is prudent to develop a batting style that will remind them of Joe DiMaggio or Ted Williams, even if long-range your percentage of solid hits with that style is small and you know you obtain better results batting cross-handed. In any event, a 25% turnover rate among professionals as a group, with 2% costs attached to such turnover, reduces group performance by 1/2 of 1% per annum ($1½ billion per year on $300 billion of assets). This means that, instead of chance dispersal of results causing half of all managers to fall above the unmanaged performance level (which has no transaction costs) and half to be below, the frictional drag of turnover costs causes well over half to perform worse than what “average” might be assumed to be. For the reasons set forth above, almost all recent investment management performance by pension funds of large corporations has been fair to poor. Specifically, the Becker study (most comprehensive of all pension investment measuring services) reports the following:
Overall Annual Return
S & P 500
Becker Median Result *
Last 3 market cycles, (5/30/52 to 12/31/74)
5.3%
4.1%
Last 2 market cycles, (9/30/55 to 12/31/74)
2.1%
0.4%
Last single market cycle, (5/30/70 to 12/31/74)
2.2%
(0.3%)
*Excludes bond segment of portfolios so that equity management only is measured against the equity yardstick.
Is There Hope? Can a Wise Corporation Assure Superior Investment Performance for its Pension Plan?
If above-average performance is to be their yardstick, the vast majority of investment managers must fail. Will a few succeed—due either to chance or skill? Of course. For some intermediate period of years a few are bound to look better than average due to chance just as would be the case if 1,000 “coin managers” engaged in a coin-flipping contest. There would be some “winners” over a 5 or 10-flip measurement cycle. (After five flips, you would expect to have 31 with uniformly “successful” records—who, with their oracular abilities confirmed in the crucible of the marketplace, would author pedantic essays on subjects such as pensions.) In addition to the ones benefitting from short-term luck, I believe it possible that a few will succeed—in a modest way—because of skill. I do not believe they can be identified solely by a study of their past record. They may be operating with a coin that they know favors heads, and be calling heads each time, but their bare statistical record will not be distinguishable from the larger group who have been calling flips indiscriminately and have been lucky—so far. It may be possible, if you know a good deal about investments as well as human personality, to talk with a manager who has a decent record and find that he is using methods which really give an advantage over other investors, and which appear to be likely to provide continued superiority in the future. This requires a very wise and informed client—and even then is not free from pitfalls. For openers, there is one huge, obvious pitfall. I am virtually certain that above-average performance cannot be maintained with large sums of managed money. It is nice to think that $20 billion managed under one roof will produce financial resources which can hire some of the world’s most effective investment talent. After all, doesn’t the big money at Las Vegas attract the most effective entertainers to its stages? Surely $50 million annually of fees on $20 billion of managed assets will allow (a) an array of industry specialists covering minute-by-minute developments affecting companies within their purview; (b) top-flight economists to study the movement of the tides; and (c) nimble, decisive portfolio managers to translate this wealth of information into appropriate market action. It just doesn’t work that way. Down the street there is another $20 billion getting the same input. Each such organization has its own group of bridge experts cooperating on identical hands—and they all have read the same book and consulted the same computers. Furthermore, you just don I t move $20 billion or any significant fraction around easily or inexpensively—particularly not when all eyes tend to be focused on the same current investment problems and opportunities. An increase in funds managed’ ‘dramatically reduces the number of investment opportunities, since only companies of very large size can be of any real use in filling portfolios. More money means fewer choices—and the restriction of those choices to exactly the same bill of fare offered to others with ravenous financial appetites. In short, the rational expectation of assuring above average pension fund management is very close to nil.
Illustrations of Reality in Pension Fund Investment Management
I recently received some interesting figures from a pension fund involving about $250 million of assets. The 9-1/2 year record through June 30, 1975 of the three major banks involved (managing $20—$50 million per bank) follows, compared to “average” as defined in two ways, the S & P 500 and the D-J Industrials:
Annual Compounded Rate of Return (including Dividends)
Dow-Jones Industrials
+2.8%
S & P 500
+3.8%
Continental Illinois
+1.8%
First National City
+1.0%
Morgan Guaranty
+3.4%
Until recently, Chase had managed a comparable segment of the funds but was terminated because of poor performance—probably worse than the other banks shown above. Each percentage point plus or minus affected this fund by about $2 to $4 million over the time period measured. A further interesting calculation was made. On June 30, of the portfolio held by Continental, 64% of the securities held, measured by market value, also were in the Morgan portfolio and 44% were duplicated in the First National City portfolio. Similar overlap was demonstrated in other ways. I am familiar enough with the record of this fund to state that no unusual constraints have been placed on the managers, and no special factors have to be recognized to interpret the numbers. I believe that Morgan’s above-average record here, compared to that of other large banks, is typical of its relative performance at other pension funds. I also feel the predictive value of that relative standing to be low. Pensions & Investments magazine recently rated the 35 largest banks for which five-year performance records of commingled equity funds were available. Ten did as well as or better than the S &P, consistent with what might be expected according to probability theory if everyone were operating solely according to chance and there was a modest drag on performance caused by transaction costs. More importantly, I know of no set of statistics which would demonstrate any opposite view relating to managers handling large amounts of money. The evidence all seems to confirm that it is unwise to expect above-average investment results from a corporate pension plan, conventionally managed.
Major Options in Pension Fund Management
With respect to pension fund investments in securities, several paths can be followed, singly or in combination: (1) One or more large conventional money managers may be retained, with the expectation that performance will be slightly poorer than “average” because of costs involved. If the manager is very large and conventional—any major bank will do—directors of the corporation should run no risk of imprudent behavior as fiduciaries under the new ERISA statute. Conventional behavior is safe—even when the potential litigants have the benefit of hindsight. If we should go this route, I see no reason to change from Morgan. (2) Management in security selection can be, in essence, abandoned by simply creating a portfolio which is equivalent to “average”. This minimizes turnover and management costs, and probably accomplishes about the same result as (1)—perhaps with just a touch more potential liability for results, since a tacit admission has been made that no effort is being expended to “manage”. Several funds have been established fairly recently to duplicate the averages, quite explicitly embodying the principle that no management is cheaper, and slightly better than average paid management after transaction costs. This policy could be implemented by participating in such funds, or on a do-it-yourself basis. (3) A manager can be found handling smaller amounts whose record has been good for the right reasons. Then hope that no one else finds him. Good records of any type usually have attracted massive money flows—whether the record was based on unusual skill, luck, or even, occasionally, semi-fraudulent activity which has “manufactured” the record. Even those records, which I believe to have been based at least partially on skill have wilted when subjected to torrents of money. A further problem is that in no case were the superior records I have observed based upon institutional skills, which could be maintained despite changes in the faces. Rather, the good results have been accomplished by a single individual or, at most, a few, working in fairly specialized areas in which the great bulk of investment money simply had no interest. It has been very difficult to out-think the pack on General Motors, IBM, Sears, etc. Rather, the unusual records—and there have been few that have been maintained—have been achieved by those who have worked relatively neglected fields in,which competition was light.[§] [§] Your win-loss percentage in tennis will not be determined by the absolute level of ability that you possess. Rather, it will be determined by your ability to select inferior opponents. If you select with care it will be quite easy to attain a winning percentage higher than, say, Cliff Richey while he is playing on the tour. Application of this principle is the key element in bridge, poker, or investments. (Harder to apply in the latter, however—it is easier to identify a couple of palookas at the bridge table.) Many pension funds, including the fund to which I referred earlier, have attempted to find superior but relatively unknown managers still working with small sums. This often involves dozens of interviews and usually comes down to the past record, particularly the recent past, plus an articulate practitioner who “looks” reasonable and respectable. In my opinion, based only on impressions, the overall record from this selection process has been poor and very likely worse than the mildly below-average record of the major money manager. The reason isn’t too hard to see (particularly with a rear view mirror). Much of Wall Street is a succession of fashions. Obviously some individuals will have hit the most recent fashion, and their record will look correspondingly good maybe sensationally good if they have a reckless streak and have played a particular trend very hard. But fashion-hitting never has been successfully maintained, to my knowledge. And the manager primarily selected for recent hot performance knows what he is expected to do. He is to perform—and quickly. So the new small manager’s decisions frequently are characterized by high turnover, major mistakes, even more furious activity to catch up, etc. It has not happened every time, but my hunch is that the sum total performance of the relatively unknown go-go managers for pension funds has been worse than the lumbering, stiff-legged minuet performed by the major banks. (4) A fixed income (bonds) investment strategy can be followed, which presently allows returns of about 9% per annum that can be locked in for some time. This option currently is becoming somewhat more popular, should be quite defensible under ERISA, and will look good or bad, depending upon what returns from equity holdings develop over the next decade or two. This essentially is the same decision that was made by default several decades ago by many companies, except the rates then were 4% instead of 9%. I do not recommend an attempt by us to go back and forth from bonds to stocks. This is a skill possessed by few, if any—and certainly not by a group. If we could master this particular form of alchemy, there would be little reason to do anything else. The comfort level produced by this option is likely to be high under any conditions except high rates of inflation—which will produce distress under all of the options discussed herein. (5) My final option—and the one to which I lean, although not at anything like a 45-degree angle—is mildly unconventional, thereby causing somewhat more legal risk for directors.[**] It may differ from other common stock programs, more in attitude than in appearance, or even results. It involves treating portfolio management decisions much like business acquisition decisions by corporate managers. [**] I am leaving most of the legal discussion out of this memo—partly due to space considerations and partly because it is not my field. In general, the safest course is to behave as most others are behaving—and to trust in the expertise of large, well-regarded conventional organizations. I also believe it is defensible—although not as automatically defensible—to diversify among a group of well-financed entrenched businesses purchased at reasonable valuations of earnings and assets. The directors and officers of the company consider themselves to be quite capable of making business decisions, including decisions regarding the long-term attractiveness of specific business operations purchased at specific prices. We have made decisions to purchase several television businesses, a newspaper business, etc. And in other relationships we have made such judgments covering a much wider spectrum of business operations. Negotiated prices for such purchases of entire businesses often are dramatically higher than stock market valuations attributable to well-managed similar operations. Longer term, rewards to owners in both cases will flow from such investments proportional to the economic results of the business. By buying small pieces of businesses through the stock market rather than entire businesses through negotiation, several disadvantages occur: (a) the right to manage, or select managers, is forfeited; (b) the right to determine dividend policy or direct the areas of internal reinvestment is absent; (c) ability to borrow long-term against the business assets (versus against the stock position) is greatly reduced; and (d) the opportunity to sell the business on a full-value, private-owner basis is forfeited. These are important negative factors but, if a group of investments are carefully chosen at a bargain price, it can minimize the impact of a single bad experience in, say, the management area, which cannot be corrected. And occasionally there is an offsetting advantage which can be of very substantial value—but for which nothing should be paid at the time of purchase. That relates to the periodic tendency of stock markets to experience excesses which cause businesses—when changing hands in small pieces through stock transactions—to sell at prices significantly above privately-determined negotiated values. At such times, holdings may be liquidated at better prices than if the whole business were owned—and, due to the impersonal nature of securities markets, no moral stigma need be attached to dealing with such unwitting buyers. Stock market prices may bounce wildly and irrationally but, if decisions regarding internal rates of return of the business are reasonably correct—and a small portion of the business is bought at a fraction of its private-owner value—a good return for the fund should be assured over the time span against which pension fund results should be measured. It might be asked what the difference is between this approach and simply picking stocks a la Morgan, Scudder, Stevens, etc. It is, in large part, a matter of attitude, whereby the results of the business become the standard against which measurements are made rather than quarterly stock prices. It embodies a long time span for judgment confirmation, just as does an investment by a corporation in a major new division, plant or product. It treats stock ownership as business ownership with the corresponding adjustment in mental set. And it demands an excess of value over price paid, not merely a favorable short-term earnings or stock market outlook. General stock market considerations simply don’t enter into the purchase decision. Finally, it rests on a belief, which both seems logical and which has been borne out historically in securities markets, that intrinsic business value is the eventual prime determinant of stock prices. In the words of my former boss: “In the short run the market is a voting machine, but in the long run it is a weighing machine.” Specifically, it probably is possible to invest the $12 million in our pension fund in a dozen businesses (maybe more; ERISA emphasizes diversification) with current intrinsic value (measured by private-owner valuations and transactions) attributable to our investment of, say, $20 million and current earnings of at least $1.5 million. The portion of such earnings paid out to us clearly is worth 100 cents on the dollar, and a reasonable batting average by the managements of companies in which we invest should result in the portion reinvested having similar value. If this is the case, such a “pension division” operation will produce better returns than bond investment at current rates. The main long-term risk would be that future returns on capital obtained by our “division managers” would turn sour. In some individual case that undoubtedly would prove true. It might well be that there also would be a favorable surprise or two. That is true of any acquisition program. The one substantial advantage that this “division” would have would be attractive purchase prices—far below those available for purchasers of entire businesses. If purchases could not be made on such a “bargain” basis, we simply would wait until they could. A second advantage would be the flexibility provided by a public stock market, allowing portions of business operations to be acquired or divested much more easily than entire operating businesses. It must be emphasized that measurement of this program would have to be based on the progress in our share of earnings and assets of the constituent companies—not short-term market movements. We would expect substantial increases in earnings, dividends and asset values over, say, a decade just as we expect with our operating divisions at Washington Post. We would, in mental approach, have created a new diversified division, currently capitalized at about $12 million, which is expected to earn a better-than-bond rate with which to eventually pay retired employees. In a sense, this is what was done in the profit-sharing fund when it accumulated Washington Post stock throughout earlier years—basically a wise decision because the business earned very well internally and thereby grew in value at a substantial rate. It was an attractive partially-owned business, acquired at an attractive price. Until market prices were introduced, which complicated things, the funds built up through this self-owning division were significantly larger than would have been achieved by buying bonds or the general run of common stocks. If it should be decided to implement such a policy, recommendations should be obtained from qualified analysts who clearly understand our rather unusual selection criteria. This could be handled on a quite infrequent basis. As mentioned earlier, a policy also could be followed combining more than one option. Most combinations have been (1) and (3), or (1) and (4), but (4) and (5) could well be more logical.
Pension Asset Management in the Scale of Business Priorities
Of course, the desirability of even considering something non-conventional (not to be equated with more risk; the exclusive benefit provision of ERISA requires fiduciaries to discharge their duties “solely in the interests of the participants and beneficiaries”) relates to the relative impact pension results can have on overall earnings. One of my friends always reminds me: “If a thing’s not worth doing at all, it’s not worth doing well.” If our pension assets were $1 million and likely to grow only a few hundred thousand dollars a year, it would be ridiculous to think of any activity that might incur even unwarranted criticism, let alone litigation. On the other hand, if we had U. S. Steel’s pension situation, we might want to make a truly top priority project out of pension asset management (as they have), even if we might slightly increase personal risks to directors by so doing. We fall in between although, if you look through to the plans to which we contribute but don’t manage (Guild, mechanical, etc.)[††] we are moving more in the direction of U. S. Steel. It is likely that within ten years we will have $40 million in our direct pension “division” and further substantial sums in other plans presently on an indirect basis. Earnings of businesses purchased through this “division” should run at least $5 million per year. [††] In prior years it was felt that by limiting contributions to union plans to so much per shift, or per hour, management was relieved of responsibility for any specific level of benefits, as well as pension fund administration. The shield of this “hands off” position is eroding and we should recognize that we may well have responsibility at some future date for benefits promised under these plans, which we originally thought involved only defined contributions. With this in prospect, we may wish to gain at least partial command over fund investments in this area. Summing up: (1) If the economic world turns out to be one of sustained double-digit inflation—probably still unlikely but not unthinkable—among the carnage will be private pension plans. The investment process can do little to modify that disaster. Hope lies mainly in the care with which past promises have been made, and the ownership of a business whose economic characteristics allow pass-through pricing which includes a large part of past labor costs, as well as full current costs. (2) In a more orderly world, the Care with which promises have been worded remains important, but on a scale that diminishes as inflation moderates. Conventional approaches to money management should not be expected to produce above average results. But average will be perfectly acceptable at low inflation rates. (3) A mildly non-conventional investment approach, emphasizing a business approach to security selection, gives some opportunity for long-term results slightly above average without corresponding increase in investment risk.
Warren E. Buffett
October 14, 1975