Dimitrios Stafylas, of Aston University, Aston Business School, in Birmingham, England, has proposed a “holistic” model of hedge fund performance attribution with three key features:
- Irrespective of underlying fundamentals, it is only during good times that hedge funds as a class deliver significant excess return to their investors;
- Also during these good times, size, age, and redemption restrictions matter. That is, small funds, young funds, and those with restrictions deliver higher alpha than their peers, but
2a. During bad times, small funds suffer more than large funds, and funds that manage to survive without redemption restrictions outperform those that did have such restrictions, although,
2b. during these bad times, the superior performance of young funds over old ones remains.
- “Strategies with specific characteristics can even deliver significant negative alpha, conditional on stressful market conditions.”
Specifically with regard to that third point: Stafylas finds that “directional strategies … with no redemption restrictions present negative alpha to investors.” Non-directional strategies in young funds do likewise. Unfortunately (for the unwary investors), “it is possible to pay high fees for negative alpha during stressful market conditions.”
Building the Model
In building this model, Stafylas focused mostly on the North American market, which as he says is the investment focus for 72% of the worldwide total of hedge funds, with $1.9 trillion in assets under management. He uses data going back to the beginning of 1990, and up through the first quarter of 2014. This data covers three turns of the U.S. business cycle. A footnote indicates that he plans another paper to treat with the same multi-factorial approach of funds that do not have direct exposure to the North American region.
Within that data base, the “good times” referenced above consisted of the first seven months of 1990, the period from April 1991 to March 2001, again from December 2001 to December 2007, and from July 2009 to March 2014.
For purposes of the above observations, a fund is “young” so long as it is less than 62 months (five years and two months) old.
The decision whether to treat a fund as having redemption restrictions is treated for purposes of this model as a binary one, lock-up or not, although Stafylas acknowledges that there are “soft” restrictions, such as restrictions on redemption frequency, short of a lock-up. Those restrictions would have to be treated in a more full account of the subject matter.
Size and Strategy
Stafylas explains the greater success of smaller funds in good times thus: the most talented fund managers are incubated by the larger funds, then, once they have built experience, go off on their own, thus putting their acquired expertise on the small fund side of this divide. Other considerations may be involved. For example (Stafylas mentions) small funds pay a higher price for failure, so that “higher pressure” produces better performance. Also, it is possible that the bigger the operation, the further the talented top managers are from analysis of investments at the security level, and this layering of analysis becomes counter-productive.
Stafylas found also that, as intuition suggests, directional strategies are more volatile in their returns than non-directional strategies. Here he isn’t as binary in approach as he is with redemptions. He uses three classifications: directional, semi-directional, and non-directional. He regards short bias, long only, sector, and long/short strategies as directional; event driven, multi-strategy, others, and global macro as semi-directional. That leaves relative value, market neutral, and CTA for the non-directional class.
Stafylas acknowledges certain limitations in his study. For example, he did not consider leverage as a factor in performance, since in his database there was inadequate time-series information for that.
Related Work
This is not the first time Stafylas has visited such questions. He was also the lead author of a 2015 article on “issues in hedge fund index engineering” with Keith P. Anderson of The York Management School and Muhammad Moshfique Uddin of the University of Leeds, Division of Management.