Content area
Full Text
In recent years, there has been considerable discussion relative to the investment benefits of what is often called the endowment model of investing (Sheikh and Sun [2012], Leibowitz, Bova, and Hammond [2010], Keating [2010]). While various definitions exist, a common characteristic (Keating [2010]) is that the endowment model allocates a significant portion of assets to traditional alternatives (e.g., private equity, real estate) as well as modern alternatives (hedge funds, real assets). The increased exposure to less liquid investments is based on the belief and some supporting research (Khandani and Lo [2009]) that there is an excess return to less liquid assets, due in part to the liquidity risk. Of course it is assumed that this liquidity risk may be less for certain endowments and other institutions for which long-term liabilities would enable them to invest in assets with a long investment time horizon and would enable them to bear the liquidity risk of such investments. More generally, the reason for investing in various traditional alternatives such as private equity boils down to anticipated high returns to private equity.1 Investors want to emulate the high returns of the Yale Endowment, which are widely advertised in the press. To illustrate, The Economist, on March 10, 2011, asserted that, for Yale at least, "The university's private-equity assets have produced an annualized return of 30.4% since inception."
What is typically ignored is that this 30.4% number is not a rate of return and cannot be compared to returns of other asset classes. This 30.4% is a so-called since-inception IRR (SI-IRR) and it is the performance metric that is recommended by the GIPS standards.2 This article uses the annual reports of Yale Endowment (from 2000 to 2010) as a case study to show how the GIPS-recommended performance metric may sometimes be uninformative of the true historical returns, display several oddities, and be highly misleading.
Obviously, this article is related to extensive literature on performance measurement, on the debate between time-weighted and cash-weighted returns (e.g., Dichev [2007], Kazemi, Schneeweis, and Szado [2013], and Nesbitt [1995]), geometric versus arithmetic returns (e.g., Jacquier, Kane, and Marcus [2003]), and so forth. But this article makes a much simpler and somewhat tangent point to that literature. It simply shows that using SI-IRR entails misleading...