It is well known that databases of historical hedge fund returns suffer from a range of biases - chiefly ‘survivor bias’. The worst funds cease reporting their results, sometimes simply because they go out of business, and some of the best stop reporting when they no longer need to raise assets.
But new research presented at the fourth Lyxor/NYSE Liffe Hedge Fund Research Conference in Paris by George Aragon of Arizona State University suggests there are also significant biases introduced by delays to and clustering of performance reports to databases.
In a paper co-authored with Vikram Nanda of the Georgia Institute of Technology, Aragon examines daily changes to a closely followed database since January 2009. He finds that reporting delays are longer when performance is worse, and that those who delay often report two or three periods in clusters - which, on average, show poorer results being reversed by later, better ones.
“This is an important issue for performance measurement,” Aragon said. “Some data vendors provide the precise date on which a fund began reporting. Likewise, it would also be useful for data providers to include the precise date of reporting, and possibly also the archive they may have of historical reporting dates.”
He added that there could be perfectly valid reasons for delays. Reporting early could involve information leakage that impairs competitiveness, for example, and marking assets to market can be a challenge during liquidity shocks. In difficult markets, reporting to a database may simply not be a top priority.
However, while they cannot rule-out these more ‘innocent’ explanations, Aragon and Nanda explicitly suggest that these are “strategic” decisions. Correlation of poor performance and delayed reporting remains significant across all funds - which all face the same market conditions at the same time: one might assume that a good manager would be just as consumed by tricky markets as a bad one, but the good managers still manage to report results earlier.
Transparency is usually considered to be an unalloyed positive when it comes to hedge fund investing and reporting. But another empirical study presented at the conference showed that greater regulatory transparency from individual hedge funds can lead to lower returns, higher fees and greater correlation with other funds.
Zhen Shi of Georgia State University examined the characteristics of a sub-group of funds before they began filing 13-F portfolio disclosure to the US regulator, the SEC, and after, for the period 1994-2010. The sub-group were funds with assets under management at or around $100m (€76m), the threshold over which funds must begin disclosure, to control for the effects of size on performance.
The research found that the returns to the funds fell by 4% annually on average after they began reporting - and recovered for the test group of funds that went the other way, from reporting to not reporting.
Shi did not present any conclusions as to why this effect was evident, but suggested that funds took more risk before reporting and became more conservative once they started.
Investors may like to think the returns hedge funds report to databases reflect the actual returns from assets they hold - or at least that, once the final audit is in, reported numbers will be revised promptly and finally. But that is not the finding of a third study from the conference by Tarun Ramadorai and Michael Streatfield of the University of Oxford, together with Andrew Patton of Duke University.
Their paper considers a series of monthly downloads from five databases, from 2007 to 2011, covering performance histories stretching back to 1994. While the majority - 60% of funds - had been subject to no revisions, 40% had revised at least one past record by at least 1 basis point, 32% had revised an old record by as many as 10bps, and 22% had made revisions of as many as 50bps.
Fund managers revised reported returns for months as far back as the late 1990s, and that revising activity is strongly predictive of performance - in other words, dud managers tinker more with their records.