Regular readers may recall a back-of-the-envelope analysis of hedge fund index dispersion that we ran a few months ago. We found that there was a significant range in the July returns across data from different index providers. This dispersion was modest at the composite level, but was quite significant at the sub-strategy level.
The thumbnail chart at the left contains the two standard deviation ranges of various sub-strategy indexes (one standard deviation on either side of the mean across an average of about 8 index providers per index). If you click on the thumbnail, you’ll see that the highly diversified funds of funds and composite indexes tend to have a minimal dispersion. However, there was significantly less commonality between the equity market neutral, managed futures, short-bias, and event-driven indexes reported that month.
While November’s results begin to roll in, we ran the same test again on October’s returns (which have all now been fully reported by the providers we track). With the absolute value of returns significantly larger in October, one might expect the ranges to be commensuratley larger. As you can see by the chart below, this was largely true (note that we had to set the grid lines at 5% vs. only 1% for July):
In fact, the average standard deviation across all sub-strategy indexes was up 3.5x in October (vs. July). The standard deviation of Convertible Arbitrage indexes went from 0.44% in July to 3.58% in October, an increase of over 8x. Oddly, the standard deviation of Equity Market Neutral indexes was actually down in October (from 2.82% in July to 1.12% in October). In other words, there was more “agreement” between the index providers in October than there was in July.
When you rank the sub-strategies from smallest dispersion to greatest, you can see that the order has also changed. Notably, Event-Driven went from near the bottom to the very top of the list. In other words, most databases reported similar results for this segment in October. Fluke or a common “event-driven” factor in October? Who knows?
Studies have shown than a very small proportion of the world’s hedge funds actually report to 5 or more databases. So it comes as no surprise that index returns differ across providers. As a result, you can’t really interpret any single index as a true representation of a sub-strategy’s performance. Unfortunately, this problem seems to be compounded by the fact that the level of alignment between providers is very dynamic on a sub-strategy basis.