The Equity Risk Premium Puzzle-Who are the Long-term Holders?

One of the unsolved mysteries of modern finance theory is the "Equity Risk Premium" puzzle. In a sentence, the puzzle is why equities have historically outperformed bonds (in real, dividend-adjusted terms) by such a large margin? For an elegant illustration of the concept, see Brad DeLong's 2006 blog post here.

Of course, stocks should deliver a larger return since they are riskier claims than bonds; and the variance of stock returns -- the measure of risk in modern finance theory -- is manifestly greater than the variance of bond returns. But the magnitude of stocks' outperformance is the puzzle, especially in light of the well-known fact that the variance of stock returns can be mitigated by long holding periods. (note 1)

This latter notion is most famously articulated in Jeremy Siegel's "Stocks for the Long Run" and more infamously in Glassman & Hassett's "Dow 36,000".

Barring a quick rebound in the US stock markets, the Equity Risk Premium puzzle is likely to be a hot topic in finance departments over the next few years. And if the US experiences another decade of meandering stock price performance to match the current Japanese experience, it may be decided that there's no puzzle at all. The major Japanese market indices are hovering at the same levels they first reached twenty years ago and are roughly 80% below their all-time highs. The current generation of Japanese investors are probably not scratching their heads over the inexplicably high historical returns of common stocks.

Recent news events may shed some light on the issue as well. Like many universities, Harvard has recently reported that it's endowment shrunk 22%, or $8 billion, in the last half of 2008 and that it may shrink by as much as 30% when illiquid assets such as private equity and venture capital positions are marked to market. Geraldine Fabrikant has covered this story for the New York Times here

"Harvard Endowment Loses 22%"

and here.

"Endowment Director is on Harvard Hot Seat"

What I find interesting in this news is the fact that "Harvard depends on its endowment for about 35 percent of its operating budget..." representing $1.4 billion of endowment income contributed to annual operating expenses according to this letter from Drew Gilpin Faust, the university's president.

"Letter to the Community - February 18, 2009"

Now, $1.4 billion per year equals 3.8% of Harvard's endowment of $36.9 billion as it stood on June 30, 2008. But $1.4 billion per year comes to 5.4% of the $25.8 billion that Harvard's endowment will shrink to if realized losses are in fact 30%. At current prices, I suppose it's possible to construct a diversified portfolio of this size that would generate a pre-tax return of 5.4% without touching the endowment principal; but it wouldn't be easy and, more importantly, the income stream likely would not grow as fast as the operating expenses it funds.

Hence, Harvard is selling off a number of its endowment positions in public and private equity and borrowing in the debt markets to provide short-term cash. On the operating front the university is tightening its belt, suspending some ambitious construction plans and instituting certain pay freezes. The university will undoubtedly be reaching out to the alumni base for increased giving as it celebrates its 373rd anniversary this year.

But if Harvard, whose endowment is managed with a multi-century time horizon, finds its liquidity affected by the current crash in financial asset prices, who exactly are the long-term holders who can ignore the current volatility in stock prices and sit tight for the long term? Insurance companies like, for example, AIG?

There may simply be an insufficient amount of investment capital with a multi-generational investment horizon that is indifferent to market volatility like we're experiencing today. This alone could explain a good bit of the Equity Premium Puzzle. Moreover, the NYT "Hot Seat" article on Jane Mendillo, Harvard's new endowment manager, suggests a related agency issue. Even if Harvard University can take a truly long-term view, ultimately the investment decisions are made by individuals (or committees) whose career horizons are quite a bit shorter than Harvard's. Even if one truly, truly believes that in the long-run stocks will deliver premium risk-adjusted returns, it's no fun to report a 30% decline on your watch.

So for all those young investment bankers who intend to wait out the current recession in business school for the next two years, start reading up on the Equity Risk Premium Puzzle and the Liquidity Preference Function. They should be hot topics for the next couple of years.

**********

Some more comments by RHH on this topic over at Seeking Alpha

Annals of Rank Hubris, Larry Summers Edition

Larry Summers's Billion Dollar Harvard Gamble



(note 1)
Much, though not all, of the data used to analyze the Equity Premium Puzzle comes from US stock returns from the mid-1920's, or in some cases stretching back to the Civil War. This is largely due to data availability and reliability. Some commentators on the Equity Risk Premium Puzzle have noted that returns for US equities over this period may be related to country-specific factors, specifically the evolution of the US from an agrarian, emerging economy into a political and economic super-power. As such, US returns may reflect a "success bias" making them unrepresentative of global equity returns over the same periods. For a good discussion of global equity returns, see "Global Evidence on the Equity Risk Premium" (pdf).

It should also be noted that even 160 years of stock price data represents only eight non-overlapping 20-year observation periods, a relatively small data sample. Likewise, the commonly used U.S. data back to 1925 represents less than five non-overlapping 20-year periods. This means that new, extreme data points can dramatically change our view on the odds of their occurring. This time last year, one could say that a 50% peak-to-trough decline in the S&P 500 had occurred only once in 88 years, and that was associated with the Great Depression. As of today, it's occurred twice.