Publications
Partners
Broad acceptance of hedge funds is remarkable in light of their limited history and lack of transparency. It's tough to understand how returns are earned and difficult to project whether returns - or the funds themselves - are going to persist. In this business, almost anyone can call himself a hedge fund manager. So what accounts for investor fascination with this asset class? The promise of performance.
Assessing the Asset Class
According to TASS, the most comprehensive global hedge fund database available, historical returns have been strong. What we call "directional funds" - a category of funds making some sort of directional bet on the equity market and thus frequently compared with stocks - returned 16,6% pa from 1996 through 2005. "Absolute-return funds" claim to be resistant to the direction of the equity market and, like bonds, aim to provide more stable returns with lower volatility. These returned 12,4% annually. And funds of funds, which reflect a blend of hedge funds, returned 10,6% a year.
These returns are impressive. However, we view them sceptically, as they are inflated by a number of well-documented biases. Two of the most significant are survivorship bias and backfill bias. With the former, a manager stops reporting returns on a fund that underperforms. With the latter, a manager simultaneously launches several funds but delays reporting returns; eventually, he reports the returns of only those that succeed. The winning funds' historical returns are then backfilled into the database. After adjusting returns for these biases, the results were significantly lower. Together, backfill and survivorship biases accounted for between 320 and 720 basis points of reported returns.
Even adjusted for biases, however, hedge funds do have attractive risk/return profiles. For the 1996-2005 period, absolute-return funds still would have compared favourably with bonds, delivering 8,1% annualised, versus 6,2% for the Lehman US Aggregate Index; at lower volatilities.
Directional funds narrowly outperformed stocks, at 9,4% vs 9,1%; but did so with much less volatility: just 9,4% compared with the S&P 500 at 15,6%.
Diversification is another heralded benefit. Absolute-return funds, for example, are only slightly more correlated to stocks than is real estate. And directional funds are only modestly more correlated to stocks, at 0,48. Higher return, lower volatility and greater diversification are all good arguments for an allocation to the asset class.
Beyond Volatility
Unfortunately, this rosy picture fails to fully capture the experience of any institution out to choose a hedge fund. To begin with, those low correlations are not particularly stable. When we divided our return history into two streams - those during rising equity markets and those during falling equity markets - we found that correlations with the stock market increased just when an investor least wanted them to (Graph 1). We suspect this is due to asset liquidity, particularly problematic when the market is stressed.
This supposition is bolstered by an analysis of hedge fund losses. When we viewed the largest peak-to-trough loss experienced by a composite of absolute-return funds over the past decade, we found that, at 10%, the loss was nearly triple that for bonds, despite funds' lower volatility. Interestingly, directional funds did a much better job than stocks at preserving capital: Their worst collective loss of 17% was substantially smaller than the -45% in the S&P 500.
We take less comfort in this than one might expect, because the type of turmoil the market experiences appears to make a meaningful difference in the magnitude of funds' losses. This is evidenced by the variation in return between funds that did and did not exist between August and October 1998 - the culmination of the Asian flu and Russian debt crisis. Directional funds operating during that time saw an average worst loss for any individual fund of -36% between 1996 and 2005; those not around in 1998 saw an average worst loss half that size.
Thus, while hedge funds seem to hold up comparatively well when market trauma is limited to certain sectors (for example, during the bursting of the Internet bubble), they seem not to fare so well when systemic factors cause liquidity to dry up.
One final risk to consider is that of total meltdown: involuntary fund closure. On average, 0,3% of funds close involuntarily each year. From this figure, we calculate the cumulative odds that a portfolio of 10 randomly selected funds will suffer a fund default in a 10-year period to be a staggering one in four. It's therefore important to understand why funds default. A recent study by Capco found that half of all involuntary closures resulted from operational risks, which included risks surrounding trade processing, accounting, valuation and reporting. A little more than 70% of this was traced to behaviour ranging from borderline to squarely criminal, including false or misleading valuation of investments, fraud and misappropriation of funds. Inadequate risk controls that failed to contain style drift and desperate trading accounted for another 15% of operational failures. And inadequate investment in technology or human capital to control risks accounted for the balance.
Diminishing Returns
Thus far, we've tempered expectations based on historical returns and risks, but we also believe there are headwinds that will make it difficult for hedge fund managers to deliver returns on a par with those seen historically.
Among these are narrow risk premiums across the global capital markets. Narrow spreads mean fewer and smaller opportunities for hedge funds to exploit. A few years ago, a great strategy was to buy the cheap shares and sell the expensive ones, benefiting as values converge. But today, the difference between cheap and expensive is much narrower. At the end of 2005, the cheapest 20% of US stocks were trading at 14 times earnings, versus eight times earnings in 2000. Meanwhile, the most expensive were trading at just 18 times earnings in 2005, compared with 48 times in 2000. In fixed income, we've seen similar spread compression. For example, the yield spread between the JP Morgan Emerging Market Bond Index and the 10-year US Treasury bond declined from 650 basis points in 2000 to just 230 basis points at the end of 2005.
As the compensation for taking risk falls, it's bound to become more difficult for hedge funds to earn a premium. At the same time, there are more managers chasing that premium. Indeed, 35% of all stocks have high hedge fund ownership today, compared with just 9% in 1999.
Lastly, high fees are eating into this diminishing potential alpha. It's not uncommon to see some funds charge as much as 2% of assets, and with funds of funds, the second layer of fees can be as large as the first. For investors seeking an 8% after-fee return, once performance-based fees are added in, this fee burden can translate into an additional 300 basis points of return for a single fund and an additional 670 basis points for a fund of funds.
The Manager matters
For traditional stock and bond managers, the overwhelming driver of return is the underlying asset class. Only about 20% of the variance in these managers' returns comes from their decision making; the balance is due to the market itself. With hedge funds, however, the relationship between the stock or bond market and the value added by the manager is turned on its head: with hedge funds, only 20% of return variance traces to market factors, while 80% traces to manager decisions.
The effect can be seen in the dispersion of hedge fund returns over the past 10 years. Graph 2 contrasts after-fee returns for hedge funds to that for active bond and stock portfolios. For absolute-return funds, the dispersion was more than 19%; for directional funds, it approached 39%. Compare this with roughly 1% and 10% for long-only bond and stock managers respectively. This means that, while it's possible to pick a good hedge fund manager, it's also possible to pick a very bad one.
Choosing a Hedge Fund manager
How should an investor select a manager? Performance is an obvious starting point. Unfortunately, we found that top-quartile hedge fund managers fared no better than random relative to other managers.
Still there is evidence that skilful manager selection is possible, even if past performance isn't a good guide. Graph 3 shows that funds of funds have done an impressive job picking hedge funds for investment (though fees typically eat up this premium).
We think a good hedge fund manager will display strength in three areas: a clearly identifiable competitive advantage; a disciplined and sophisticated approach to risk management; and strong infrastructure augmented by independent oversight.
A manager should be willing and able to decompose returns into their elements - market exposure, asset selection and trading - and to separate the effects of leverage from those of skill. A manager's track record should make sense given the historical context - that is, a value manager should outperform when value stocks do. And managers should be candid about the limitations of their style.
Gaining such insights is critical. We isolated the alpha of several types of hedge funds in graph 4 by removing the effects of markets, investment style and company capitalisation. The residuals represented unexplained returns. On the left, we plotted the alphas of convertible arbitrage funds against those of event-driven funds. The former exploit mispricings between convertible bonds and their underlying stocks; the latter include specialists in distressed securities, bankruptcies and merger arbitrage. These funds should have little relationship to one another, but their correlation is fairly strong, suggesting a common ingredient - probably illiquidity. On the right, we charted a similar relationship between convertible arbitrage and long/short equity funds. These mysterious correlations underscore the need for identifying the manager's sources of return.
Sources of return should also be replicable. For example, a major source of long term return for hedge funds was the IPO boom of the late 1990s. But those days are unlikely to be repeated.
Once an investor is convinced a manager can add value, it's time to consider risk management. As with returns, a manager should decompose and explain sources of volatility and how they're managed and should conduct stress testing in up and down markets.
Lastly, because operational risks are a significant hazard, an investor should focus not only on the integrity of the people but on the integrity of technology, systems and information. Due-diligence efforts should be tailored to the fund's investment style, as a generic approach may miss risks unique to investing in certain asset types. If a manager shifts investment style or adds a new asset class, due-diligence efforts should be redoubled; historically, errors are far more likely when managers are working in unfamiliar territory and have adapted legacy systems to cope with new types of investments. And pay attention to complex and/or illiquid assets, which have been the source of most funds' problems.
Evaluating a specific hedge fund is an exceptionally difficult but, in our view, worthwhile task. Hedge funds employ tools that can amplify the results of skilled managers, including the ability to go both long and short, and the use of leverage. Higher potential returns, lower volatility and diversification benefits argue for an allocation to this asset class. When searching for a manager in this murky market, look for transparency, consistency with respect to sources of return and a willingness to maintain an open dialogue. In the end, keeping informed is the best defence.
Joseph Gerard Paul
Paul became Alliance Bernstein's CIO for Small and Mid-Capitalisation equities in 2002, CIO for Advanced Value Fund in 1999 and Co-CIO - Real Estate Investments in 2004. He was named a member of the Institutional Investor All-America Research Team every year from 1991 through 1996.
