Hedge fund statistics?
Hedge funds in "aggregate" lost money last month. Many good hedge funds gave some of their year to date returns, while others were revealed as beta mutton dressed up as alpha lamb. Skill is rare so "average" returns are always going to be poor. However some hedge funds did well recently. Unfortunately, it will take decades to be STATISTICALLY confident whether a manager made money due to skill or luck.
But do we need to wait decades for quantitative studies to confirm common sense? If you give a basketball to Kobe Bryant or a soccer ball to Ronaldo we already know they will be able to do more with it than a random person. Similarly give money to hedge fund managers run by Paul Tudor Jones or Warren Buffett, for example, and the chances are high they are going to do a better, PERSISTENT, job than a random investor. Why in sports is it obvious that skill exists, but, even today, many people doubt the existence of investment skill? The market is supposedly "efficient" so there is no alpha to be captured!
"Sell in May and go away" is a market adage with a surprising level of historical accuracy. Anyone who followed it this year looks good right now. I am not a big fan of seasonal trades, primarily because of the paucity of data - only 110 "Mays" (1896-2005) for the Dow, for example and much shorter for most other benchmarks. But sometimes the old rules work even without sufficient quantitative evidence. Sadly the long only crowd MUST obey their lemming like "mandate" to be fully invested at all times, unlike good hedge funds that manage risk, go short or to cash when necessary.
Limited timeframes flaw most hedge fund survivorship studies. Despite thorough data scrubbing and a plethora of statistics, they use JUST a decade or so of monthly returns, ie only around 120 data points per fund, if available, which it often isn't. The databases that these studies use include lots of managers, some of whom are hedge funds and many others who say they are, but cannot escape the inherent temporal bias and sample size inadequacy of reaching conclusions based solely on a mid 1990s to mid 2000s data set.
But do we need to wait decades for quantitative studies to confirm common sense? If you give a basketball to Kobe Bryant or a soccer ball to Ronaldo we already know they will be able to do more with it than a random person. Similarly give money to hedge fund managers run by Paul Tudor Jones or Warren Buffett, for example, and the chances are high they are going to do a better, PERSISTENT, job than a random investor. Why in sports is it obvious that skill exists, but, even today, many people doubt the existence of investment skill? The market is supposedly "efficient" so there is no alpha to be captured!
"Sell in May and go away" is a market adage with a surprising level of historical accuracy. Anyone who followed it this year looks good right now. I am not a big fan of seasonal trades, primarily because of the paucity of data - only 110 "Mays" (1896-2005) for the Dow, for example and much shorter for most other benchmarks. But sometimes the old rules work even without sufficient quantitative evidence. Sadly the long only crowd MUST obey their lemming like "mandate" to be fully invested at all times, unlike good hedge funds that manage risk, go short or to cash when necessary.
Limited timeframes flaw most hedge fund survivorship studies. Despite thorough data scrubbing and a plethora of statistics, they use JUST a decade or so of monthly returns, ie only around 120 data points per fund, if available, which it often isn't. The databases that these studies use include lots of managers, some of whom are hedge funds and many others who say they are, but cannot escape the inherent temporal bias and sample size inadequacy of reaching conclusions based solely on a mid 1990s to mid 2000s data set.
© posted by Hedge Fund on 3:20 AM
<< Home