(January 22, 2013) – It’s become a truism of institutional investing: major asset owners with low payouts and long-term liabilities can absorb—and make a premium on—illiquid assets.
But whether it’s a forest or venture capital stake, financial economist Andrew Ang argues that premium is probably not all that it seems. Ang’s latest paper, “Illiquid Asset Investing,” is both a qualitative and quantitative take-down of the alternatives-saturated endowment model.
He contends that the vast majority of assets are in fact illiquid—the US residential real estate market alone is worth about the same as the total market capitalization of the NYSE and NASDAQ combined. Furthermore, liquid assets dry up when investors need them most, such as during the financial crises when the market on commercial paper froze along with many other markets. These kinds of events happen regularly, Ang argues, and investors ought to consider illiquidity risk alongside other exposures, like inflation and interest rates. But if investors tend to be bad at estimating illiquidity risk, they are at least as unreliable forecasting reward.
Ang quotes from a memo sent out to Harvard’s Council of Deans in December 2008 to illustrate the fallibility of reported returns on illiquid assets. The note asked each school to cut expenses, and cautioned that as bad as the endowment’s reported losses were, the true losses were even worse: “‘Even the sobering figure is unlikely to capture the full extent of actual losses for this period, because it does not reflect fully updated valuations in certain managed asset classes, most notably private equity and real estate.’”
The memo gestures to one of the three key factors that, according to Ang, cause people to both inflate expected returns and understate risk.
1. Survivorship Bias: Those who win on illiquid investments stick around and participate in the market, while losers close up shop and tell no tales. (Or, rather, many fewer tales, particularly since reporting results is voluntary.)
2. Infrequent Trading: Risk models lowball estimates of volatility, correlations, and betas when their calculations are based on thin sets of return data. Furthermore, smoothing of sporadic returns—those in real estate, for example—hides the range of the data set with a mean. Ang unsmoothed real estate return data from the early 1990s, and found the lowest reported return dropped from -5.3% to -22.6%.
3. Selection Bias: Investors tend to buy high and sell high. Real estate turns over much more frequently when prices are up, and buyout funds take companies public only in bullish stock markets. Likewise, venture capitalists tend to sell start-ups when markets are hot.
“Treat reported illiquid asset returns very carefully,” Ang advises. “Survivors having above-average returns, infrequent observations, and the tendency of illiquid asset returns to be reported only when underlying valuations are high will produce return estimates that are overly optimistic and risk estimates that are biased downwards. Put simply, reported returns of illiquid assets are too good to be true.”
Read Andrew Ang’s entire paper here.