Alpha Investing


© 2005 Will Goetzmann and Ming Fang


Tom Griggs works as a quantitative analyst for Cherry Partners a fund company that, among its many products, offers the Cherry Multi-Strategy Hedge Fund.  The fund currently manages $700 MM.  Its target is absolute return.  Its benchmark is 300BP over Treasury Bills.  The fund seeks to achieve this through diversification across hedge fund styles.  Their approach to selecting fund managers is to identify the “best in class”  for each style of manager -- narrow these down to a handful of real stars with quantitative tools and then to do considerable  further due diligence, including on-site visits,  third-party credit review, calls to other clients  and other investigations. 


The Cherry Multi-Strategy Hedge Fund has three long-short equity managers among its 15 hedge funds.  It recently experienced a surprise with one long-short manager – the market exposure of the fund was well over .6 over the course of the last month.  In discussions with the manager about this, she explained that she was relatively bullish on the stock market rebound.  Cherry analysts explained that their goal was to hire a long-short manager who is focused on the relative mispricing of equity and not to keep a systematic bet on the market.  The manager explained that she focused on absolute return and that the beta was Cherry’s concern, not hers.  These discussions eventually led to the termination of the manager.


Now, Tom Griggs is looking for a potential replacement.  He is scanning databases of hedge fund managers.  These databases provide historical performance on the funds as well as analytical measures: Sharpe Ratio, Alpha, Information Ratio, and Drawdown among others.  Typically Alpha Investing looks for managers who can deliver repeatable alpha, where alpha is defined in a multi-factor framework: a market factor, interest rate exposure, exposure to small stocks and exposure to a “value” factor.  Although strict neutrality to all of these is not essential in a long-short fund, consistent exposure to any one of them suggests that the return of the manager is coming from systematic factors as opposed to arbitrage in expectations.   Finally, general measures about drawdown are important – funds with major drawdown events might be “doublers” managers who try to make up a monthly loss by increasing exposure to systematic factors.  Tom has also heard that it is possible to enhance Sharpe Ratios by taking negative skewness.   He also has read that autocorrelation in hedge fund returns is common due to valuation issues.  He wants to check if these funds have this problem and think through whether it is an issue of concern.  


Tom has selected a set of funds that have potential.  He is preparing a report on these managers.  His report will include a discussion of the measures of risk adjusted rates of return, a consideration of whether this alpha is “repeatable” i.e. did it all come from one trade or is it apparent in sub-periods of the data, whether the volatility of the fund changed a lot through time, whether the average exposure was between 0% and 20%.


Tom’s report will also address the question of fees.  Although 1% fixed and 20% incentive fees with a high water mark of Libor are common, he is convinced that a net positive beta fund should not be compensated as an absolute return manager.  Given that Cherry has some capacity to negotiate a different fee structure with the fund,  he wants to think through an propose a variation of the 1 and 20 model that takes into account the systematic exposure of the funds.




Using to historical information about the funds, and the Pertrak database, prepare Tom’s report.


  1. Hedge fund data can be downloaded from Pertrac by opening  “C:\program files\pertrac\universe\long-short (long history).unv.” – click “File->Open New Universe” and select “long-short (long history).unv  If this is difficult, please E-mail Ming Fang to receive a spreadsheet of data.


  1. Return data for the three Fama-French factors can be downloaded from Professor French’s website: