14 August 2007
With a traffic jam of potential hedge fund blow-ups developing in Greater Hedgistan (mainly among the large quant funds), we are sure to hear a lot in the coming months about the “high failure rate” of hedge funds in general.
In fact, GMO’s Jeremy Grantham got things started early last week with a prediction that “…within 5 years…up to half the hedge funds…in existence today will have simply ceased to exist.” (Grantham Says Hedge Funds, LBO Funds to Collapse)
Widely-read newsletter author John Mauldin took issue with Grantham’s typically dire prediction by pointing out that this would be in keeping with the attrition rate of typical small businesses (which, of course, most hedge funds are):
“…while a lot of hedge funds in the market today will no longer be here in five years, the real reason is that they simply did not generate enough cash flow for themselves and their investors to survive. You can actually have a profitable year and see your assets under management leave…”
Mauldin is right. In fact, a study updated earlier in the year clearly illustrates how hedge fund attrition stats need to be taken with a grain of salt.
The report by UMass’ Hyuna Park explores the ability of standard deviation, kurtosis and skewness to predict hedge fund failures. But it also aims to calculate a “real failure rate” for hedge funds that ignores the benign closure of funds for a myriad of other “non-failure” reasons. Says Park:
“Defining hedge fund failure is a challenge because it is difficult to obtain detailed information on defunct hedge funds. In addition, liquidation does not necessarily mean failure in the hedge fund universe since successful hedge funds can be liquidated voluntarily due to the market expectation of the managers or some other reasons.”
Park cites earlier studies that find the attrition rate of hedge funds is about 8% based on the number of funds that have ceased to voluntarily report to major industry databases. (For a frame of reference, consider that Grantham’s 50% closure number equates to only 13% per annum - slightly higher when you consider that some closures over the next 5 years will undoubtedly occur in funds that don’t yet exist.)
But Park finds that many of the funds that ceased to report their returns actually kept chugging along - apparently just tired of repeatedly having to mess around with the various databases’ extranets each month, or having merged one fund into another to keep a lid on expenses.
Park defines “real failure” as a fund that ceases to report after a) a negative average rate of return over the 6 previous months and b) a decrease in AUM over the previous 12 months. The “real failure rate” between 1994 and 2004 was actually 3.1% (compared to an ovrall attrition rate of 8.7%). This is even more striking when you consider that Park’s definition of “failure” is a broad one. After all, asset decreases will result directly from poor performance, and many poorly performing funds continue to operate but stop reporting simply because they cease to realize any marketing value from doing so).
As regular readers will know, we’d also suggest that many hedge funds might close (under either benign or under distressed conditions) because traditional long-only players will begin to eat their lunch with “hedge fund light” products like 130/30, portable alpha and market-neutral mutual funds.
The paper contains the following example of two un-named global macros funds that closed in 2000 (Fund “S” and Fund “L”). As the chart below illustrates Fund “S” experienced a drop in AUM in 1998 while Fund “L” experienced major drops in AUM during both 1996 and 2000.
While both closed their doors in 2000, Fund “L” was a big Loser that year:
Apparently Fund ”S” just sailed peacefully off into the sunset while Fund “L” fell off the end of the Earth.
So Grantham’s prediction may turn out to be right - but not necessarily because recent volatility has changed the long-term dynamics of the hedge fund industry. As in the past, many more hedge funds will sail away than will sink.
No comments:
Post a Comment