A new white paper from New York Life looks at the role of alternatives in portfolio construction and argues that usual risk-return based approaches can underestimate risk and lead to over-allocation to the alternatives. The paper, by Amit Soni, an NYL portfolio manager, proposes a new method "to quantify performance dispersion risk and incorporate it in the portfolio construction process."
The long bull market, especially in equities, has pushed valuations to levels that make many investors nervous. Also, (and closely related) the long period when central banks have worked to keep money cheap and carried bloated balance sheets in order to achieve that, is at or near an end. As Soni observes, these conditions also have investors worried about diversification. Having both stocks and bonds may not make one adequately diversified, given "the fear that a sharp rise in rates might trigger a sell-off in equities as well."
The Underestimation of Risk
This means that investors are looking to alternatives, but they should look warily. Because, Soni writes, an "underestimation of risk" comes about if one uses the best-known alternative indexes-those "which are derived from aggregating or pooling returns provided by hedge fund or private equity managers." There are two big problems with these indexes. First, they "disguise the significant dispersion of returns for managers within the index." Second, they are "subject to a smoothing effect that comes with illiquidity."
Soni writes that there has already been a lot of research addressing the second of those problems, the smoothing issue. This white paper is designed to contribute to solving the other problem - that which arises from the dispersion of returns.
In a few words, that problem is this: if one looks at large-cap equities, or for that matter, intermediate-term fixed-income instruments, one finds that some do better than others, but that this dispersion is small. On the other hand, the performance dispersion of PE or of US-domiciled hedge funds is much larger. Thus, an investor does himself considerable harm from selecting an underperforming manager in an alternatives space, and this risk is "not captured by the volatility of the index, but should be taken into consideration," Soni says.
The Simple Approach is the Right One
A simple approach to this is to treat the risk from volatility of the index, ?i, and the risk of performance dispersion by the underlyings, ?pd, as uncorrelated. This implies that the total risk is the square root of the sum of the squares of those two risk components.
Soni's argument is that this simple approach is the right one. One problem with application is that the data may not be available for the entire cross-sectional distribution for ?pd. Another problem is that of an outlier, a "severely underperforming fund with minimal assets," which "likely had no implications for external investors in general," but which would lead to a distorted standard deviation number.
The Real PE Risk
But quintile performance numbers can serve as a surrogate for performance dispersion. They can be tertile, quartile, or decile. "In most cases," Soni says, "this approximation of the variance or standard deviation is close to the actual variance or standard deviation of the entire distribution, and, therefore, a reasonable approach to compute ?pd with easily accessible data."
The performance dispersion for private equity serves as an example. "For private equity, the dispersion between the 25th and the 75th percentile … performance is 11.6%." Soni asks us to assume a normal distribution of this dispersion. One can argue that this assumption is a dubious one, in an age where "fat tails" are seen as routine, where thousand-year floods come along every five years or so - still, given a default assumption that this dispersion is normally distributed, Soni calculates that the total risk of an investment in private equity is 12.8%. He also says that if one ignores the issue of dispersion of results one will erroneously conclude that the risk of such an investment is only 9.6%.
The difference between a 9.6% risk and a 12.8% risk doesn't lead Soni to infer that institutional investors such as New York Life ought to leave PE and other alternative investment asset classes alone. But it does and should lead to an emphasis on "robust manager selection process with due diligence on investment process and risk management."