Will hedge funds regress towards index-like products?
By: William Fung, London Business School & David Hsieh, Duke University
Published: January 2007
Will hedge funds succumb to the same fate as large chunks of the mutual fund industry: commoditization through ETFs and index funds? Merrill Lynch seemed to think so last October. And now, so do academics William Fung and David Hsieh. This research paper provides a “tool kit” to identify alternative beta, distinguish it from its hard-to-copy cousin, alternative alpha, and replicate it using basic trading rules that the authors say “capture the essence of hedge fund strategies”. These trading rules explain what they call an “accidental alpha” produced by traditional (linear) factor regressions.
The paper draws close parallels between mutual funds and hedge funds:
“There is a sense of deja vu among hedge fund investors that many hedge fund managers are beginning to resemble active managers in the mutual fund industry of the pastâ€”failing to deliver returns commensurate to the fees and expenses they imposed on investors. History tells us that over-priced active managers will be replaced by low-cost passive index-liked alternatives. Could the same process be taking place in the hedge fund industry?”
William Sharpe’s technique for delineating alpha from beta in mutual funds in 1992 assumed constant exposure the various underlying factors. But Fung & Hsieh say that hedge funds’ propensity to dynamically allocate capital means that traditional factor exposures cannot properly explain hedge fund returns. Ergo, passively investing in factors alone cannot replicate hedge fund returns.
So Fung & Hsieh introduce “primitive trading strategies” (PTSs) to the equation. An example of a PTS might be a merger arbitrage strategy that goes long the target and short the acquirer, or simply an S&P500 index put. Unlike traditional factors, PTSs are non-linear. However, a hedge fund can still be represented (and replicated) by a linear combination of (non-linear) PTSs.
Hedge fund returns, they say, can be represented by a weighted average – not just of factor returns – but also of various PTS returns. Omitting these PTSs from a regression will yield the “accidental alpha”.
Introducing PTSs into a regression can have the effect of removing any non-linearity from other (traditional) factors and improving the model. But the trick, they say, is to assemble a comprehensive PTS “toolkit” to explain all non-linearity, leaving the hedge fund portfolio as static linear combination of traditional factors and/or PTSs.
In conclusion, the paper makes several prescriptions:
“…investors may wish to gain direct exposure to specific PTSs. Some PTSs may add diversification to the standard stock and bond portfolios while delivering positive risk premia commensurate for placing capital at risk. For example, over the 2000-5 period, the average return of the market factor (RMRF) was -11 bp per month. In comparison, SMB, HML and the currency lookback straddles, respectively, had average returns of 67 bp, 114 bp and 34 bp per month. They also had low correlation to RMRF: 0.26 for SMB, -0.50 for HML, and -0.12 for the currency lookback straddles…These same characteristics are often used to illustrate the diversification benefits of certain hedge fund stylesâ€”typically proxied by peer group averages of hedge funds.”
So managing a hedge fund portfolio may not be all about alpha after all. It may be all about dynamic allocations to non-linear risks – from which a fair premium can be earned.
But where can we get these “primitive trading strategies” and how much do they cost? Says the paper:
“As most PTSs can be implemented at a fraction of the cost of prevailing hedge fund products, these PTSs are natural low-cost alternatives to hedge fund products based on similar strategies. Research over the last decade has created a library of rule-based, executable, PTS-like hedge fund replication strategies.”
And finally, Fung & Hsieh return to their original question: “Will hedge funds regress towards index-like products?”
“The existence of these index-like hedge fund products can also act as catalysts to improve the price discovery process in the hedge fund industryâ€”more efficient fee structure with equitable risk-return sharing between investors and managers. This is in fact a healthy development for the hedge fund industry, one where alpha producers with limited capacity can be sufficiently compensated for their skills and beta-only products will regress to being index-like alternatives at lower fees.
“The success of lowâ€“cost synthetic hedge funds will inevitably lead to an improvement in the return quality (better performance at lower fees) of the surviving hedge funds. However to replicate these better performing hedge funds, some of which will exhibit skill-based alternative alpha, will require new technological innovations that are likely to come at ever increasing price tags and replication risk. Ultimately the question as to whether hedge funds will become index-like products, will depend on the answer to the fundamental question that precipitated this process, but with a qualification â€”namely do hedge funds add value (have alpha) that cannot be replicated at a lower cost? The answer to this question will no doubt emerge over time. In the meantime, low-cost transparent synthetic hedge funds that offer exposures to specific PTSs are likely to become the, index-like, vehicle of choice for delivering the returns of maturing hedge fund strategies. Efficiently priced, dynamically managed combinations of these investable PTSs will challenge inefficient portfolio products such as some over-priced investable hedge fund indexes and funds-of-hedge funds.”