In two previous posts (Identifying Alternative Risk Premium Factors and Alternative Risk Premium Factors), we outlined how Quantitative Investment Strategy indices (QIS) can be used to construct common alternative risk factors that contain largely orthogonal information. These risk factors are based on systematic investment strategies across asset classes. Moreover, we highlighted the type of information conveyed by the various alternative risk premium factors. In this paper, we analyze how these factors can help decompose hedge fund index returns and compare them to traditional factors commonly used to describe hedge fund returns.

Data And Methodology

We focus our analysis on a period covering the last 15 years of monthly data, thus covering the period from 2007-05-31 to 2022-04-31. We construct the RC factors as outlined in Identifying Alternative Risk Premium Factors and Alternative Risk Premium Factors. These factors mostly mimic systematic trading strategies, such as those potentially invested in by some hedge funds. In addition, we add 12 market factors (“beta”) as described in Alternative Risk Premium Factors which gives us a total of 58 factors. In the later analysis, all of these factors serve as potential explanatory variables and can be easily identified as they begin with the two letters “RC”.

In addition to alternative risk premium factors, we add 11 factors that are more common in hedge fund analysis to our set of potential explanatory variables. These may consist of market indices or Barra risk factors. It is worth noting that all alternative risk premium (“RC”) factors are directly tradable.

In the following we analyze two aspects:

  1. How well is our set of explanatory variables suited to describing hedge fund index returns?
  2. What factors (by asset class or premium) primarily help explain hedge fund index returns? Are these predominantly classic risk factors or alternative risk premium factors?

We analyze both questions simultaneously. Since the set of potential explanatory variables is very large, we first select the best candidates for each index by performing a LASSO. We then regress the index returns on the selected variables. Using the regression coefficients, we can then decompose each annualized index return into the contribution explained by the factors and an unexplained portion, the alpha contribution.

Fund-of-Funds And Hedge Funds

HFRI Fund of Fund

Figure 1: Return Decomposition of HFRI Fund of Fund Index. Source: Resonanz Capital

Over the 15-year period, the annualized average returns of fund-of-hedge funds (HFRI FoF) can be explained by 7 different factors, plus cash and a marginal alpha contribution. Given the negligible alpha, our set of explanatory variables describes the HFRI FoF index very well. Almost all of the selected factors are predominantly equity-based, with RC Commodity Beta as the only exception. Apart from cash, global equity market beta (RC Equities Beta) is the largest contributor (0.95%), while commodities beta and equity momentum detract from the annualized return (-0.11% and -0.46%, respectively). In addition, the group of RC factors is the largest overall contributor to the annualized return, while Barra’s momentum factor and the MSCI Emerging Markets Index are the only classical factors selected in LASSO.

HFRI Fund Weighted Composite

Figure 2: Return Decomposition of HFRI Fund weighted Composite Index. Source: Resonanz Capital

The annulized return of the average hedge fund (HFRI FWC) provides a much higher alpha contribution (1.26%) over the last 15 years than the average fund-of-hedge-fund. Still, the LASSO identifiey 4 factors, of which global equity market beta has the overall higest contribution to annulized returns (1.52%). This comes as little surprise, as the HFRI FWC has historically been dominated by Equity Hedge, a hedge fund style known for higher equity market exposure. In addition, returns rely on a broader range of asset classes that include equities, credit and commodities. However, all factors are beta-driven and only one classical factor (MSCI Emerging Markets Index) is selected, while the remaining belong to the group of “RC” factors.

Convergence

HFRI Relative Value

Figure 3: Return Decomposition of HFRI Relative Value Index. Source: Resonanz Capital

Over the 15-year period, the HFRI Relative Value Index has a relatively high alpha contribution (+2.49%), which represents the highest of all the indices analyzed. Of the 6 factors selected, 5 represent beta, while only RC Multi-Asset Low Volatility represents a systematic strategy. 2 classical factors and 4 “RC” factors were selected. Among the factors, Credit contributes the most overall (0.36% and 0.36%). This is most likely due to the relatively high proportion of FI RV funds in the index. The remaining contributors are equities, multi-asset and commodities.

HFRI Equity Market Neutral

Figure 4: Return Decomposition of HFRI EH Equity Market Neutral Index. Source: Resonanz Capital

In contrast to Relative Value, Equity Market Neutral (EMN) has only a marginal Alpah contribution of 0.19%. In addition, 9 factors contribut relatively evenly to the annulized return of the HFRI EMN Index. This could be due to the relatively hedged approach of the funds in this index. Of the 9 factors selected, only Barra’s US Momentum factor is not a “RC” factor, while the remaining 8 factors are all from the Alternative Risk Premium (“RC”) group. Thus, the “RC” factors explain the returns of the Equity Market Neutral Index very well.

Value

HFRI Equity Hedge

Figure 5: Return Decomposition of HFRI Equity Hedge Index. Source: Resonanz Capital

 

Similar to Equity Market Neutral, Equity Hedge (EH) had only a marginal alpha contribution during the 15-year observation period. While 9 factors were selected for EMN, however, EH can be explained very well by only two factors: Globals and Ermerging Markets Beta, with the first factor being by far the largest contributor to annulized returns (2.33%).

HFRI Event Driven

Figure 6: Return Decomposition of HFRI Event Driven Index. Source: Resonanz Capital

After Reletive Value, the HFRI Event Driven Index has the second highest alpha contribution (1.49%) among all indices. Nevertheless, the LASSO selects 4 “RC” factors, of which Global Equity Beta is the largest contributor (1.45%), followed by US High Yield Credit Beta, Multi-Asset Volatility and Commodities Beta. All in all, only about 34% of the 4.33% annulized return of the index can not be explained by factor exposure or cash return, so the “RC” factors perform quite well in explaining the return of an average Event Driven fund.

Divergence

HFRI Macro Systematic Diversified

Figure 7: Return Decomposition of HFRI Macro Systematic Diversified Index. Source: Resonanz Capital

The HFRI Macro: Systematic Diversified Index has the second highest detraction from alpha among the indices (-1.73%). Thus, the “RC” factors can explain these returns very well. Among the 4 “RC” factors selected, there are three that incorporated Momentum premiums for different asset classes. A fact that makes intuitive sense given that the index is mainly composed of CTAs that tend to follow trend-following strategies. Since the RC factors are directly tradable, the negative alpha can be interpreted as fees and costs associated with an average CTA investment.

SG Trend

 

Figure 8: Return Decomposition of SG Trend Index. Source: Resonanz Capital

The SG Trend Index is comparable to the HFRI Macro: Systematic Diversified Index. It had the largest alpha detraction (-2.09%) among the indices for the 15-year sample period. Unlike the HFRI Macro: Systematic Diversified Index, only two “RC” Momentum factors (Multi-Asset and Commodities) are sufficient to explain the average CTA return. Considering that these CTAs usually follow systematic trading rules, the “RC” factors seem to reflect these strategies particularly well. A fact that is not very surprising given that all “RC” factors are constructed using systematic investment strategies.

Conclusion

QIS indices are systematic investment strategies that span multiple asset classes and are directly investable by entering into swap contracts. We construct 58 different factors from a universe of more than 1000 such indices. In previous posts (Identifying Alternative Risk Premium Factors and Alternative Risk Premium Factors) we have shown that these factors contain largely orthogonal information. In this post, we have highlighted that these factors help in explaining the returns of various hedge fund indices. The higher the degree of systematic strategies within these hedge fund indices, the better the “RC” factors are at explaining their returns. Moreover, we show in a side-by-side comparison that these “RC” factors are in most cases better suited to describe the returns of hedge fund indices than classical factors/indices.

Subscribe to our Newsletter