Fund indices shed more light
Hedge fund indices are supposed to provide a window on the world of hedge funds; a way for investors to see how hedge funds are performing against other asset classes and which particular strategies are winning or losing.
In practice hedge fund indices have not yet achieved the transparency that benchmarks for traditional asset classes provide. They provide a kaleidoscopic view of activity that has become further fragmented as the number of hedge fund indices has grown.
A comparison of the main hedge fund indices, carried out by Professor Narayan Naik, director of the Centre for Hedge Fund Research and Education at the London Business School last year, found wide divergences in the key performance statistics. For most years, there was no consensus across indices for the best performing investment strategy for example.
Naik concluded that the indices then available did not provide a true benchmark of hedge fund performance. This is not disputed by index providers, who concede that an index is effectively a fund of funds, somebody’s view of what an index should be.
Since Naik’s study, the universe of hedge fund indices has expanded with the arrival of a hedge fund index from MSCI. Perhaps more significant, two investable indices: Standard & Poor’s S&P Hedge Fund Index in November and HFR’s new index, the HFRX in March.
The trend towards indices that strike a balance between representativeness (what the hedge fund world is like) and investability (how much of that world the investor can buy into) is producing a kind of consensus in the indexing business that database indexes, with their vast but inaccessible universes are not enough.
However, there are different routes to the same goal of investability, and variables that will affect the overall performance of investable indices. The key variable is the choice of whether to evenly (fund) weight or asset (capital market) weight the index.
The largest single reason for the differences between hedge fund indices performance data is the system of weighting they choose, says Justin Dew, director of portfolio services at Standard & Poor’s. “If you’re an asset-weighted index you are going to live or die by the performance of one or two sectors, such as long/short equity and macro. That is going to be a very significant proportion of your return.
“When you’re an evenly weighted index you’re going to have a bit less volatility and slightly more consistent returns. The reason is that when long/shorts are not performing maybe CTAs are.”
The S & P HFI gives equal weightings to each major style while the HFRX has an asset-weighted global index. “We felt that being evenly weighted was more advantageous because we avoided the super Nasdaq effect of the asset-weighted alternative, as opposed to some of our peers who chose the other.” Different systems of weighting have accounted for variations in reported performance over the past five years, says Dew.
Another variable that can cause differences in index performance is the use of cluster analysis and stratified sampling. HFR pioneered cluster analysis to identify managers that correlate most closely with the various strategies. It also helps eliminate the “outlier” managers whose performance is out of the ordinary and inexplicable.
Joseph Nicholas, founder and chief executive officer of HFR explains: “In the broad index you don’t have certainty about the composition. So you’re making up for that lack of certainty with diversification. You diversify away the outliers. In the investable index, cluster analysis eliminates outliers.”
Stratified sampling and cluster analysis can ensure that that the managers selected for certain strategies are representative of their particular space. For example, they can help identify institutional quality managers that are representative of commodity trades advisors space and therefore highly correlated to the managed future sector. However, there are strategies, such as macro, where cluster analysis will not help, says Dew.
Hedge fund indices, like the funds they report, will always have an individual tilt. This will mitigate against the kind of consensus possible for traditional asset class indices. However, the growth of investable indices, which is largely driven by the institutional investor, suggests greater convergence in future.