Monday, 27 June 2016
Last updated 1 hour ago
Nov 16 2007 | 11:32am ET
by Daniel Karp, Liberty Gateway Asset Management
Do Absolute Return funds produce returns that are independent of the markets? Should we expect return profiles to be about the same in bull, bear or sluggish environments? Many people think so. For example an article about Harvard and Yale endowments in the September 2007 issue of Smart Money magazine (“A League of Their Own”) defines the category this way:
“Absolute return consists of assets that are expected to perform well in good and bad markets and aren’t correlated to broad market averages.”
The article goes on to say that:
“Most hedge funds fall into this category, including so-called long/short funds (which try to profit from both rising and falling markets), merger arbitrage and distressed securities.”
Even a casual glance at the data reveals that this is not the case. The table on the right compares HFR Index returns for some major strategies for the bear market period of 2000-2002, against the bull market period of January 2003 through June 2007. The Hedged Equity, Event-Driven, Distressed and (disturbingly) Global Macro categories clearly perform much better in the bull market. Some of the other strategies do appear to be essentially invariant, but the returns are low, most notably–and disappointingly–in the Market Neutral category. Convertible Arbitrage seems to be counter-cyclical, which would ordinarily be appealing, but in fact the strategy exhibited marked instability over the period. The aggregate, represented by the Fund-Weighted Composite, appears to be heavily
When one considers the true nature of the strategies, the cyclical bias should not be surprising. Shorting securities is difficult and hedging is costly. One would therefore expect the preponderance of strategies to be “long-biased”, and for the market-neutral strategies to post either lower or more erratic returns. This, by and large, is what we observe.
Another pro-cyclical bias in absolute return strategies is more subtle, more insidious. That is the tendency for “convergence” strategies – which include most “value” and arbitrage strategies, to diverge and post losses during unexpectedly volatile periods. This can be devastating for those strategies that are highly leveraged. It can be especially devastating for a portfolio of diverse strategies, normally uncorrelated, which suddenly move together in the wrong direction. That is what we saw this past August. While such periods tend to be infrequent, they are also unpredictable and their impact can be large–one ignores them at one’s peril. The main point here is that these periods are not cyclically neutral. They do not normally accompany bull markets.
Implications for Absolute Return Portfolios
Just because the majority of strategies are not independent of the markets does not mean that structured portfolios cannot be. There is no need to replicate the structure of the industry. Should not one limit choices to those strategies which are truly independent of the markets? One often hears an additional argument for true market-uncorrelated portfolios: “Why should we pay those fees for beta (the market component of return), which one can have for extremely low fees via index funds?” Only alpha–the market-independent component-could ever merit 2% management and 20% incentive.
Other things being equal, the “pure alpha” argument is compelling. Whether other things are in fact equal depends on the opportunity set available. Here is the opportunity set we see: market-independent funds include some asset-backed lending, some trading, various market-neutral equity and some arbitrage. While we value such strategies and strive to include them in portfolios, it is taking things too far to limit portfolios to only these strategies - much too far. Such portfolios are, variously, too conservative, too erratic, too illiquid or inadequately diversified. Meanwhile one is forced to pass over a large array of investment talent merely because the “alpha” is accompanied by some “beta”. We do not see how this could be optimal. Our main point here is that such a restriction is simply not necessary.
A less draconian approach is warranted. We consider sensitivity to market movements – “beta” – as a force to be tamed and kept within bounds, not squelched. A more reasoned approach is to construct a highly diversified portfolio which in the aggregate aims for an expected return of Alpha + [Beta X 0.25]. Here we use the S&P 500 return as our beta proxy. (This is a simplification, as the returns of the underlying strategies are potentially affected by numerous equity and credit markets, but the main point is unaffected.) For the purpose of this discussion, we assumed a reasonable alpha target of 0.9% per month, or just over 11% annually. Although this figure is significantly below what one sees in historical simulations, we recognize that it is, in practice, on the ambitious side. As for [Beta X 0.25], it is derived from historical data, from “due diligence” observations and from collective information on the various strategies. One auxiliary benefit of extensive manager diversification is that the inevitable measurement errors on the individual managers tend to cancel out in the aggregate.
For simulation purposes, 2xs portfolio leverage is employed. The expected return is therefore doubled, and a funding cost subtracted. It should be noted that the beta factor is now 0.50. The following table displays return expectations for various S&P 500 movements. Funding costs are assumed to be 8%, somewhat higher than the present level.