Ah, the good old days – when the hedge fund “secret sauce” was revealed and suppliers set about developing the products that would bring it to investors the world over under the moniker “hedge fund cloning”. But as Swiss researchers Erik Wallerstein, Nils Tuchschmid and Sassan Zakerc note in a recent paper called “How do hedge fund clones manage the real world?”:
“Some years ago hedge fund replication was a much discussed topic on the hedge fund horizon. A credit crunch and some hedge fund Ponzi schemes later, the attention has turned elsewhere. 2008 performance of broad hedge fund indices where dismal at best. This did not bode well for selling pitches to persuade investors to turn to funds which replicate this performance.”
However, the trio goes on to argue that the $2 billion hedge fund replication business is far from dead. Hedge fund replicas, they say, “…have several unique and interesting features, many which where attractive during the crises of 2008.”
They analyze the recent performance of 21 hedge fund clones from 17 companies covering the full spectrum of replication techniques from factor replication to distributional replication to mechanical replication. (see the “Alternative Beta and Hedge Fund Replication” category at the right side of this page for extensive coverage of these topics).
Here’s how the 21 fund stack up from March 2008 to May 2009 (chart based on data in paper)…
At first glance, you might be impressed that some replication products actually managed to squeeze out a positive return last year. If so, you’d be disappointed to learn that the biggest winners were actually short versions of hedge fund replicators – designed to take advantage of downturns in the hedge fund industry. However, there were some replicators, notably the Aquila Capital Statistical Market Neutral Arbitrage Fund (distributional approach – see related AllAboutAlpha.com post), the Desjardins Global Asset Management Synthetic Alternative Investment Fund (also a distributional approach – see related AllAboutAlpha.com post) and the Fulcrum Alternative Beta Fund (a factor-based approach – see related AllAboutAlpha.com post).
Overall, the replicators performed admirably. The blue diamond above represents the S&P 500 over the same period and the green diamond represents the HFRI fund-weighted composite index.
As you can see from the chart below from the paper, 100% of the replicators beat the S&P 500 (in blue) during the time period analyzed (March 2008 to May 2009.)
Interestingly, many hedge fund replicators avoided the deep drawdown experienced by the HFRI (in black), but have underperformed the index since last 2008 as the HFRI has come roaring back.
The authors conclude the report with a comment on the state of the hedge fund replication business and a word of warning about one of the value propositions of hedge fund replicas:
“Investors should also question the promise of better transparency in replication products. The trend seems to be that replication models are becoming increasingly complex and it is necessarily a need to also understand why models allocate to certain assets. The distribution approach is a case in point. While the products in this survey indeed have generated the best performance it is not in our view straight forward to understand under which market conditions this method will deliver high returns.”
As we have argued before on these pages, the true objective of hedge fund replication products should, in theory, be to replicate the hedge fund industry (e.g. the HFRI), not to beat it. On those grounds, most “replicas” seem to fail.
However, this assumes that the hedge fund index adequately represents the performance of hedge fund strategies. But as noted academic Bill Fung has pointed out (see related AllAboutAlpha.com post), hedge funds often succumb to extra-investment factors such as liquidity and panic selling that their strategies do not necessarily capture. If these issues were not adequately captured by replicas of hedge fund trading models, then this could explain the under-performance of hedge funds vs. their theoretical potential.