All 60 of the United States’ leading university endowments underperformed a cheap “tracker fund” over the past decade, according to a study reported in the Financial Times. The culprit is the “Yale Model” of university endowment investment, whereby they focus on private equity, hedge funds and unquoted investments rather than on ordinary stock markets. From this data, the Yale Model clearly has not worked, at least in the last decade. It is worth examining what went wrong.
The Yale Model of endowment investment was developed by David F. Swensen, chief investment officer of the $25 billion Yale endowment since 1985. It consists of dividing the endowment portfolio into roughly five or six equal parts and investing each part in a different asset class. It takes advantage of the fact that college endowments represent very long-term money, so can invest in asset classes such as timberland that would be far too illiquid for normal institutional or retail investors. Swensen’s assumption, backed up in theory by the Capital Asset Pricing Model, was that highly illiquid assets have much higher returns than liquid ones. That assumption is a necessary condition for the Yale Model to be a valid investment strategy. The Yale Model also avoids fixed-income and commodity investments, which are assumed to have too low a return to be on the optimum investment frontier for an endowment.
After Yale enjoyed initial success with the model its use spread to most major college endowments, including the $41 billion Harvard endowment. The results, while initially encouraging, have in the last decade been very questionable. The low-cost Vanguard S&P 500 “tracker fund” of U.S. equities returned 14.7% per annum in the decade, while a balanced indexed portfolio of 60% global equities and 40% bonds returned 9.0% per annum. Of the 60 endowments considered the best, Bowdoin, returned 12% per annum, less than a simple indexed equity portfolio. 18 of the 60 endowment portfolios studied, including Harvard (with an 8.5% per annum return) returned less than the 9.0% per annum of the mixed bonds/equities portfolio. Indeed Harvard, the largest and most expensively run of the endowments, was much closer to the bottom of the rankings than the top, its annual return being only 1.1% better than that of bottom-ranked Southern Methodist University.
To be fair, the last ten years have been an especially bullish period for the U.S. stock market, beginning as they did near the bottom of the market in March 2009 and missing out the large decline in 2007-09, and being fueled by ten years of “funny money.” They have been much less favorable, relatively, for international markets, which is one reason why the 60/40 portfolio, based as it is on a global equity index, has underperformed the S&P 500 index by such a large margin. To the extent that college endowments have invested in international stocks, one of the Yale Model’s six categories, they may have underperformed the U.S. market, but they have given up very little in liquidity. Furthermore, the preceding decade, 1999-2009, saw a substantial outperformance of the U.S. market by international stocks. Since most global markets appear less overvalued than the U.S. market (though developed country markets suffer the same problems of slow growth and “funny money” interest rates as the United States) they may also outperform the U.S. market in 2019-29.
The Yale Model’s disdain for bonds and commodities contravenes the CAPM, which will explain to you that any investment can match the investor’s risk/return desires by the use of leverage. Thus a bond portfolio can achieve a high return by financing it in the short-term market, for example. Commodities also would appear an asset class that Yale Model devotees are missing. Gold, a poor investment over the two decades leading up to the design of the Yale Model, would have returned a perfectly respectable 12% since the turn of the century. That return could have been improved still further by judicious selection of mining stocks, which are leveraged against the gold price, because their operating costs are essentially fixed. The best gold mining stocks are fairly small companies (the big ones indulge in excessive merger and acquisition activity, which is generally very dilutive of shareholder interests). However, the illiquidity caused by their small size should not be a problem for an endowment, for the reasons outlined by Swensen.
It is the other four asset classes of the Yale Model: real estate, timber, private equity and hedge funds, that are on examination problematic for different reasons. Real estate, whether urban office, commercial or industrial buildings, or apartment complexes let out on rental (a business in which endowments have been active since 2009) has been an especially good investment over the last decade. It will have tended to raise endowments’ annual returns above that of the 60/40 portfolio and in some cases above that of the S&P 500. Real estate returns are however extremely dependent on interest rates (one reason why President Trump, a real estate man to the fingertips, wants rates lower). Consequently, many areas of real estate, especially trophy properties in the largest cities, have become very overpriced, and will suffer when rates finally rise. Still, real estate satisfies perfectly the criterion of being highly illiquid, suitable for endowment investors. There is one proviso to this: in severe downturns, such as 2007-08, real estate tends to bleed cash, as vacancy rates rise. Hence funds must be available from the rest of the endowment to subsidize real estate investments during such periods.
Timber and natural resources, on the other hand, are nicely illiquid and in timber’s case yield a good income from cutting once the forest is fully grown. Like real estate, however, timber investments require specialized management generally unavailable within the endowment itself. This immediately identifies a key problem with all these illiquid asset classes: if the endowment hires professional managers to mange the portfolios, it is not itself adding any value, and it is subjecting the investment to two loads of fees.
Timber is also a highly cyclical investment, on a different cycle from the economy as a whole, and there is thus a risk of selling timber investments at the bottom of the market, as statistically-oriented but ignorant investment managers decide the asset class is uninteresting. Harvard, for example, unloaded much of its Latin American timberlands in 2017-18, taking a $1 billion write-down. The Yale Model can be quickly undermined by misguided timing decisions, because the illiquid assets can only be sold at a discount.
The big drag on endowment returns in the last decade has been their investments in private equity and hedge funds. This should not be surprising. The CAPM, which states that the return on any given asset class depends on its risk and its liquidity, assumes that investors are wholly rational in determining how much of their portfolio to allocate to each class. In private equity and hedge funds, however, this assumption has manifestly been untrue. The funds themselves are run by very persuasive salesmen, whose gigantic remuneration depends primarily on how much money they can attract into their funds, rather than on those funds’ returns. Investors thus over-allocate to these asset classes, led by the dulcet tones of the salesmen. Initially, in the 1980s and 1990s, because the amount of money invested in these asset classes was relatively small, their returns were genuinely superior, mostly from scarcity value than from any special skill on the part of the investment managers.
However, since 2000 and more particularly since 2008, the amount of money allocated to these asset classes has been hugely out of proportion to the investment opportunities therein. Initially, this raised the prices of their investments, so that private companies sold at a substantial premium to their public cousins and hedge fund favorites like junk bonds traded as if they had very little risk. However, the We Company failed IPO in September, where a $47 billion valuation collapsed to $8 billion, has cast all “dekacorn” valuations into doubt, since it seems likely that hugely unprofitable companies can only sell at such valuations because foolish investors have bid them up.
With “junk bonds” issued with weaker and weaker covenants, and private companies at unprecedented valuations, the chance for a crash is high, and it will be worse in these asset classes than in conventional equities (and much worse than in international stocks). One of the sales pitches for private equity and hedge funds is their alleged immunity from stock downturns; there will be no such immunity this time. Endowments following the Yale Model will see their performance trashed and their asset base shrink, as the disadvantages of piling gigantic private equity and hedge fund fees on returns no higher than in other areas will become only too apparent.
When the Yale Model has finally been buried, endowments will be able to return to first principles; buying low-risk investments in a variety of areas where the endowment managers themselves have the expertise to manage the investment, and do not pay gigantic fees to gunslinger third parties. The U.S. capital market will be greatly improved by the change – who knows, we may even get a revival of genuine entrepreneurship, with new company formation in decline since 1978.
(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)