A Rough Start to the Year for Long/Short Funds

A look at systematic drivers of return

Originally published in the April 2008 issue

In this article, we explore what factors have been driving the recent poor performance in long/short funds. We focus on two subgroups of these funds-long bias (directional funds) and no bias (non-directional funds). We find that these funds' underperformance in recent months can, in large part, be attributed to declines in well-known systematic sources of return and risk.

Generally, we find that:

The poor performance of directional funds has been due in part to:
– US stocks with high earnings variability
– European stocks with low yields
– European growth stocks
– European small caps
– Emerging markets

The poor performance of non-directional (or market neutral) funds has been due in part to:
– US stocks with low leverage
– US stocks with high earnings variability
– European stocks with low yields
– European small caps

A look at the recent decline in long/short hedge fund performance

In the last few months, many long/short hedge funds have experienced negative returns, extending a decline which began in the middle of last year. While some fund categories such as convertible arbitrage and event-driven funds have remained resilient in the last year, others have not fared as well. In just the last month, both directional and non-directional (i.e. market-neutral) funds have experienced significant losses. Table 1 puts these declines in context.

Given these declines, what factors appear to be at work? We analyze the performance of a sample of funds using a returns-based model, which maps the returns of individual hedge funds to both well-known systematic factors (including market, industry, and style factors) as well to strategy factors that are relevant for each fund's peer group. The full details of the model are laid out in Alvarez and Levinson (2007)i.

Because hedge funds rarely make available their holdings, hedge fund risk models must rely primarily on returns. As with any returns-based analysis, identifying systematic sources of risk is made more difficult by the presence of noise in the return series. In addition, exposures to sources of risk are rarely constant, in particular, for hedge funds managers who are likely to change their strategy depending on their market outlook.

Lastly, identifying those factors which are relevant for hedge fund returns can be very difficult; especially when hedge funds expand their holdings into quasi-equity and fixed income instruments, such as structured products.

To measure the influence of suspected drivers of recent hedge fund performance, we rely on a dynamic returns-based model. The model is based on the state space (or dynamic linear) modelii that is capable of capturing time-varying exposures to factors driving return over time. Dynamic exposures are warranted in those cases when hedge funds are rapidly changing their factor exposures over time. The recent crisis in equity and bond markets is likely to have caused hedge funds to alter their risk exposures so the following analysis makes use of the state-space modeling methodology.

Directional funds (long bias)

The recent declines in Long Bias funds appear to be driven by biases toward the following Barra factors: positive exposure to US and European markets, positive exposure to the US Earnings Variability Factor, negative exposure to the European Yield Factor, positive exposure to the European Growth Factor, negative exposure to the European Size Factor, and slightly positive exposure to emerging markets. Table 2 shows the factorsiii we identify as important to the significant declines in long bias funds ranked by importance. Average exposures to the factors over the time period are also shown.

What the results suggest is that, on average, long bias funds were taking directional bets towards European markets while only slightly long US markets. They were also, as a group, long European growth stocks and stocks in the US with high earnings variability. These bets constituted a large portion of the poor performance of Long Bias funds overall, though individual funds may have been differently impacted. But depending on the degree of leverage for each fund, the impact from these simultaneous down-movements could be dramatic.

To get a sense for how these exposures shown above evolve over time, we next plot the exposures of the MSCI Long Bias Hedge Fund Index to the relevant factors. This is shown in Figures 1 and 2 beginning in 2007. These figures suggest several possible trends during the observed period:

  • Long bias funds have shown significant and consistent exposure to the European equity market. In addition, the analysis suggests that there was a slight increase in this exposure after the summer crises.
  • Long bias funds have had two significant reductions to US equity market exposure. First, there was a considerable reduction after the moderate, but significant, market turmoil at the end of February 2007. The second significant reduction occurred over the summer crises. It is interesting to note that our estimates show that they may have caught the crises early and lowered their exposure during July and early August.
  • Long bias funds have significantly shifted towards growth in Europe after the summer months.
  • Exposure to stocks with high earnings variability has been significant over the last year and has increased after the summer crises.
  • Beginning in Sept/Oct of 2007, long bias funds have moved from a relatively neutral stance to a strongly negative bias to the European yield and US Size factors.


Non-directional funds (no bias)

Meanwhile, no bias or so-called "market-neutral" funds have been hurt by positive exposure to US and European markets, negative exposure to the US leverage factor, positive exposure to the US earnings variability factor, negative exposure to the European yield factor, and negative exposure to the European size factor. In contrast to long bias funds, non-directional fund exposures to emerging markets and the European Equity Growth factor are quite insignificant. Instead, these funds appear to have been hurt by exposure to US stocks with higher than average leverage. Interestingly, an important factor that has helped offset some of the losses for no bias funds is the US equity short-term momentum factor. The index's overall negative exposure to this factor suggests that the funds may have captured some of the short-term reversal in the market, which was helpful in stemming some of the market losses during the August crises.The fact that no bias fund performance has been eroded by overall exposure to the US and European markets is surprising. On one hand, no bias funds appear to have been less exposed compared to the long bias funds, as we would expect. On the other hand, these exposures are still not indicative of true market neutrality. A number of explanations are possible. For instance, market forces (i.e., the assumptions for correlations and volatilities used to create the hedge) may have unexpectedly moved against these funds. Or alternatively, short positions in Asian stocks may have implied a bias towards European and US markets depending on the portfolio construction method, for instance, global-sector-based hedging programs.

  1. Alvarez, Miguel, and Levinson, Mike (2007), "Hedge Fund Risk Modeling," MSCI Barra Model Insight, April 2007.
  2. The state-space model is used with the Kalman fitler (see Hamilton 1994) to capture time-varying exposures.
  3. Note the following definitions of these factors: The European Equity Growth Factor represents the performance of stocks with higher than average earnings growth, asset growth, sales growth, and/or dividend growth. The European Equity Size Factor represents large caps based on log of market capitalization and total assets. The European Equity Yield Factor captures the performance of stocks with higher than average current yield. All factors can roughly be interpreted as baskets that are long all stocks that have a higher-than-average exposure to the criteria in question and short all stocks with a lower-than-average exposure.
  4. The US Trading Model Size factor is also shown here as it is an important driver in the months leading up to Jan 2008.