Hedge funds in crisis: not always the safest bet

Hedge funds are often purchased with the intention of hedging an existing investment portfolio against systematic risk. The financial crisis which first hit in 2007 has given such investors an opportunity to find out whether this strategy really works. It appears that many have been disappointed.

Hedge fund returns are not normally distributed

First the easy bit: hedge fund returns are in fact not normally distributed. Even the HFR hedge fund of funds index shows a significant asymmetry, which is characteristic for hedge funds. Specifically, the long-term average for monthly returns is 0.68% (January 1995 to June 2008). There are 89 monthly values above and only 72 below this value. Statisticians describe the resulting distribution curve as left-skewed. It is even more pronounced at the respective ends: extreme values which are more than 1.5 standard deviations from the average can be seen 16 times in the positive area, compared with just nine times in the negative. Which is essentially good news for investors.
The bad news is that four of the nine negative extreme values occurred within the last thirteen months, whereas the last positive runaway occurred in February 2000 (+5.21%). So is there a systemic crisis, after all? Let‘s not throw the baby out with the bath water.

Distribution of returns for Hedge Fund Research Fund of Funds Subindex: clear asymmetry, as is generally typical for hedge funds. Example for the dark blue bar: 28 of the 162 monthly yields (January 1995 to June 2008) were in the range of one half standard deviation σ (1.67%/2 = 0.84%) left and right of the arithmetic mean χ (0.68%) – i.e. between -0.16 and 1.51%.

In the context of the turbulence during the financial crisis, which has affected virtually all of the markets, the average losses of the hedge funds of funds have been comparatively moderate. During the Russian Crisis of 1998, which entailed the spectacular crash of the LTCM hedge fund as well as a financial crisis the effects of which were felt the world over, there was a similar concentration of extreme periods in the HFR funds of funds index and the highest ever monthly loss was recorded (-7.5% in August 1998). Therefore there is little reason to doubt that hedge funds of funds will see a resurgence as investment instruments once things have settled down again. But: in the past they have not offered a reliable hedge in the sense of hedging against systematic risk. During the long bear market between March 2001 and the fall of 2002, hedge funds of funds brought literally nothing (HFR FoF index 03/2001 - 09/2002: -0.4%). Relative to the stock market, this was still an outstanding result, but was of no help to investors who so desperately needed a counterbalance to the stock exposure. The balance since the beginning of the current financial crisis has also been negative (July 2007 to June 2008: -0.3%). And we can also assume that hedge fund portfolios as they are currently structured will show similar performance in future crisis, offering capital protection at best, but little else.

Most hedge funds are stock market-dependent

The reason lies in the essence of the hedge fund universe, which is by nature reflected in portfolios which are formed from it. More than half of the around 2 trillion dollar hedge fund market is to a greater or lesser degree dependent on stock market trends - long-biased long/short equity, the greatest single strategy, particularly strongly, „Event Driven“ or „Emerging Markets“ hedge funds to a lesser extent. The reason is that all of these strategies trade predominantly in stocks or stock derivatives. These hedge fund styles are unable to completely detach themselves from the direction of the market, however skilfully they are managed unless they follow market neutral strategies. This directional dependency is by nature clearly reflected in the various hedge fund indices. The HFR sub-indices „Equity Hedge“, „Event Driven“ and „Emerging Markets“ all correlate highly positively with the MSCI World stock index (between 0.74 and 0.66 in the period 1995-2008). And because the funds which are based on these strategies together account for a large part of the sector, it is not surprising that this stock-dependency also reappears in the broad overall indices. The major hedge fund benchmarks show long-term correlation above +0.5 (again between 1995 and the present). The most clearly correlated is the HFR Fund Weighted Index (+0.76), the lowest is the Credit Suisse/Tremont Hedge Fund Index (+0.52), calculated on the basis of handpicked and comparatively fewer index funds.
The typical hedge fund of funds seems to contain an asset weighted mix of hedge fund styles which corresponds to the asset weights in the overall industry. This theory is underlined by the highly positive correlation between the various funds of funds indices and their overall benchmark. Taking HFR as an example: the HFR Fund of Funds Index shows a long-term correlation coefficient of 0.91 in relation to the HFR Fund Weighted overall index.
To sum up: hedge funds of funds are no more capable than the major hedge fund styles of hedging against the systematic downward risk in the stock segment. Which doesn‘t make their attributes any less valuable - after all, good products are capable of returning yields to match those of stocks at considerably lower risk. We must remember, however, that the ideal deployment of this investment category in a portfolio is all about better diversifying the stock exposure - rather than hedging against it.

No homeopathic remedy

Admittedly the proportion of the portfolio must be high enough to attain this diversification effect. Looking back over the last ten years, it would have been better to dispense entirely with direct stock investments and instead to put together a portfolio of hedge funds of funds: a passively managed portfolio of European Blue Chips (Dow Jones Stoxx 50) would have delivered a negative return of -1.6% before charges between the end of June 1998 and the end of June 2008 - with a breath­taking annualized volatility of over 19%. In the same period, the HFR Fund of Funds Subindex posted an 89% profit (also before charges) with a volatility of just 4.6%. Each reduction of the stock proportion in favor of hedge funds of funds would therefore have reduced the portfolio risk while at the same time increasing the returns.

HFR „Equity Market Neutral“ sub-index compared with US and European stocks: comparable performance over the years with dramatically different risk.

Even if we were to look at a five-year window, which is less favorable towards the stock market, the result would be the same. Between June 2003 and June 2008, the Stoxx 50 posted a total yield of 19.3% (1.8% annualized) with an annualized volatility of 13.6%. The HFR Fund of Funds Index fared better in every respect: total yield 34.2%, annualized return 3%, volatility 5%.
By contrast, any investor looking for an instrument to hedge a stock portfolio geared towards buy and hold could follow two major strategies:

1.) Short selling
2.) Investing into assets with a strong positive earnings expectation in both cyclical market directions

Learning from hedge funds: hedging through short selling

The most obvious variation of short selling is barely suitable in practice: although options or short certificates may well minimize portfolio fluctuations, the returns would be reduced. In the extreme case of a naive hedge, let’s say by buying a short ETF, the portfolio would make no headway. Which is why professionals turn to other strategies.
The most common of which forms the basis for many hedge funds of the „Equity Market Neutral“ style. Stocks are purchased on one or more markets which the manager believes to be undervalued. As a countermove, overvalued instruments or the entire market (usually in the form of options) are sold short. Consequently, the systematic market risk (beta) should be significantly reduced or even eliminated altogether. The outperformance of the purchased stock portfolio in respect of the market (alpha) is what determines the yield. The relevant sub-index of the HFRI (HFR Equity Market Neutral) shows that this strategy clearly works. In the period December 2007 to July 2008, this benchmark posted a yield (186%) that was almost identical to the US S&P 500 stock index and volatility was so low that bond holders were green with envy. Looking at the top of the chart, investors might be tempted to dispense entirely with classic stock investments, instead opting to dedicate that part of the portfolio entirely to Equity Market Neutral hedge funds. And for long-term as well as cautious investors, this would have made good sense in the past. On the other hand, foregoing the extremes of the stock markets also means foregoing the extra profits during strong bull cycles.
Another disadvantage of this market-neutral hedge fund segment is the - at least in Austria - very narrow choice, as well as the generally limited absorptive capacity of the funds.

Directional opportunities with managed futures

The alternative (or complement) to market-neutral strategies are assets, which show low to zero correlation with the stock market and a generally positive earnings expectation. Ideally, investors would choose instruments capable of performing well in downward markets (we are trying to hedge against the downward risk), but which can also contribute positively to the portfolio in upward stock market periods. Such assets are extremely rare. Commodities, for example, which show little long-term correlation with the stock market, are generally prone to nose-diving along with the stock market whenever a recession is feared - a situation which hit commodities investors hard in the summer of 2008. The only reasonably homogenous class which fits the requirements profile are systematic strategies based on trend-following or mean reversion models like managed futures. They were capable of cushioning a stock portfolio through highly positive yields both during the stock market crisis at the beginning of the millennium and during the current bear market, at the same time showing near-zero correlation with stocks. This „dual effect“ is the result of the directional nature of most trend-following trading models used in conjunction with managed futures. They are positioned according to the long-term direction of the market. Which is why they are on the buy side in sustained upward markets, whereas in bear markets they are on the sell side in order to make money from falling prices. This tends to result in positive correlation in upward markets and negative correlation in downward markets - precisely the behavior one would hope for from a hedging instrument. Of course, these globally-oriented funds trade not only in stocks, but also in currencies, interest and commodities, whereby the bet does not always pay off in all phases. The choice of pro­duct also has to be made very carefully, given that systematic managed futures funds are relatively homogenous in terms of their correlation (i.e. the direction of their returns). The risk-return profiles, however, are very widely distributed. It is easy to find products in this market segment which produce huge volatility but which ultimately deliver only meagre returns.
This includes, particularly in recent years, the „classic“ products of a long-term trend-following approach, which generated considerable returns during the 1980s and 1990s but which is performing less and less well in today‘s market environment. A situation, incidentally, which back in 2005 prompted FTC to shift from a long-term trend-following trading strategy to a diversified multi-style model comprising several low-correlated systems.
Managers who at the other end of the broad spectrum offer consistent long-term performance with proportional volatility are on the rise internationally, although not always easy to find. This makes managed futures indices less suited to gauging the effect of any fund in which you would still be willing to invest today.
To further illustrate the point, we have therefore taken an equally weighted sample from 15 of the biggest managed futures funds worldwide which have been in existence for at least 10 years. The sample includes the flagship funds of the two biggest European managed futures specialists (MAN and Winton), “old boys“ from the USA (Campbell, Chesapeake, EMC and Rabar), active since the 1980s, as well as two Austrian systems (one of which is the FTC Futures Fund Classic). The managers in our sample encompass around half of the estimated current investments in managed futures. The long-term correlation with the asset-weighted managed futures index of CASAM/CISDM as well as that of Credite Suisse/Tremont is over 0.9% in each case - our selection is therefore representative, but also filtered according to quality.
In our 10-year observation window (June 1998 to June 2008), our managed futures sample shows a respectable overall performance of around 242 % (annualized: 13 %) with an annualized volatility of 15.4 %. You don‘t have to be an asset manager to immediately recognize that each combination of this fund package in our stock portfolio, which emulates the STOXX 50, must increase the returns. But the portfolio risk also decreases: the initial risk of the pure block of stocks was 19.9 % (annualized volatility) during the observation period. If the stock quota is gradually reduced by 10 % in favor of our futures funds, then the risk of the mixed portfolio is continually reduced in a ratio of up to 50:50 to 11.3 %. This „minimum variance portfolio“ posted a total yield of 120.3 % (annualized return 8.3 %) and also showed the lowest interim loss (-15.2 %). Greater returns with a moderately increasing risk could have been achieved with an even larger managed futures quota. The best risk-return ratio would have been between 80 and 90 %.

Managed futures: pure alpha or independent beta

What is remarkable about this result is not so much the high earnings capacity of managed futures funds, but the profile of these returns. It appears as if the futures returns are entirely independent of stock market trends. Over the ten-year period, the correlation is -0.21, so the beta is also negative and uncorrelated at -0.16. Based on the equation for the regression line (see chart), an annualized alpha (0.0115 * 12/Beta) could be calculated as the excess return ratio of the managed futures portfolio. But: where there is no correlation, the values of beta and alpha are of no significance. For this reason it therefore helps to look at the managed futures component as an independent beta source. This makes all the more sense considering there is a highly positive correlation among the 15 funds, which also show a high correlation with popular managed futures benchmarks (for more information see Futures June 2008).

12-month returns of the Stoxx 50 stock index and the managed futures sample compared with the respective 12-month correlation (gray line): the total yield of managed futures comes from the high number of positive periods rather than individual extreme values.

To understand where the apparently wondrous effect of the managed futures portfolio comes from, we must look more closely at the distribution of the returns. This is best illustrated by a chart with the rolling annual returns of both portfolio components. It is obvious that the high total yield of the futures component does not come from extreme performance peaks. On the contrary: the highest individual performances in any twelve-month window between June 1998 and June 2008 were instead delivered by the stocks of the Stoxx 50. The high overall performance of managed futures is much more a result of predominantly positive individual periods (92 out of a possible 109) and a minority of loss periods. These are also much less pronounced than for the block of stocks: the worst 12-month window for the futures sample delivered a loss of 8.3 % (2/2004 to 2/2005). By comparison, the greatest setback for the Stoxx 50 was 46.2 % (3/2002 - 3/2003). To compensate for such a loss you would need profits of almost 100% - investors holding a Stoxx 50 block of stocks during this period would have had to wait until 2007. Looking at the correlation between the two portfolio components, we can confirm that the „change in direction“ during our observation period actually worked: during the „bull markets“ at the end of the 1990s and after 2003, the 12-month correlations reach temporary positive maximums of around 0.7. During the stock crisis at the start of the millennium, however, we see negative peaks of -0.7. At the end of the observation period we can clearly see that the futures funds were positioned against the stock market in the final months. The highly positive correlation, which was still in evidence at the end of 2007, becomes a negative correlation, so losses on the stock markets stand vis-à-vis current profits for managed futures.

Summing up: what anyone with stocks needs are managed futures

What emerges from our previous deliberations is a confirmation of numerous studies affirming the positive nature of managed futures as part of an investment portfolio. Any investor holding stocks as part of a buy and hold strategy would be well advised to invest in systematic managed futures. Because of the extremely varied risk-return profiles of the individual funds of this asset class, however, the careful choice of manager as well as diversification in the managed futures portfolio are key to the success of such a hedging strategy.