Alternative Viewpoints: Due to funds’ lack of persistence, the Sharpe ratio has no validity as an investment decision tool

29 Oct 2009

There have been many studies on hedge fund manager return “persistence”.  Persistence, after all, is a necessary precondition for the existence of alpha.  Like alpha itself, you might expect that the persistence of a good Sharpe ratio may be possible in less mature (more informationally inefficient) markets.  But a new study by Siewling Lay, CAIA, finds that this intuition might be wrong.

Special to AllAboutAlpha.com by: SiewLing Lay, CAIA, senior analyst, GFIA

SL_LayMany investors use the Sharpe ratio conveniently to categorize the risk-adjusted return profile of a hedge fund.  Implicit in its use is the assumption that the fund’s Sharpe ratio is somehow persistent over time – that a good fund manager will stay “good”.  As a result, many investors look to the Sharpe ratio as an indication of how a manager might perform in the future.  If investors decide to include it in their assessments of a fund’s attractiveness for investment, its persistence and reliability would clearly be important.

You might expect that good managers are able to persist in less efficient markets such as emerging markets.  To explore this, my GFIA colleagues and I tested whether in fact Sharpe ratios of Asian hedge funds persisted on a multi-year time frame.  What we discovered might come as a surprise.

Firstly, to ensure that no single fund benefitted from a certain market environment, we examined hedge fund performance over a common timeframe: July 2007 to July 2009 (i.e. not since the inception of each fund).

As you can see from the table below from our report, funds that fall below the 25th percentile show little consistency on a year on year basis.  In fact, only 28% of funds in the top quartile in 2007 actually remained there in 2008:

nopersistence

Amazingly, we found that a greater percentage of bottom quartile funds in 2007 (31%) ended up in the top quartile in 2008.

nopersistence2

No more than 50% of all funds in the same category remained in that category the subsequent year. In fact, the highest percentage (46%) is represented by funds which were doing better than the 75th percentile in 2008 but which in the following year were among the worst performing funds as measured by the Sharpe ratio!…

As we move into this year, 2007’s winners staged a modest comeback and were the first category of 2007 finishers to recoup their losses (red line in chart below).  But the comeback is modest at best since all 4 quartiles were mere percentage points apart by mid-2009:

nosharpeshirlock

So it would appear that investing based on Sharpe ratios back in 2007 would have produced dismal results on a multi-year view.

But some studies have shown that alpha might be more persistent than raw returns since raw returns are often dominated by various market betas.    So to remove the different tailwinds experienced by different sectors, we compared each fund against only those in its own sector (Asian equity ex. Japan, Asian equity inc. Japan, Chinese equity, Japanese equity, Macro/Multi-strategy).

With the exception of the macro/multi-strategy funds, Sharpe ratios of funds in 2008 give a better indication of 2009 Sharpe ratios for the best and worst performing funds.   The persistence of the ratio is particularly high for the Asia equity ex Japan funds, which saw nearly half of the worst performing funds in 2008 based on Sharpe ratio remaining at the bottom of the pack in 2009, and 38% of the best performing funds still at the top of the pack in 2009 (see chart below).

somepersistenceAlthough this suggests past performance may, in fact indicate future performance for certain funds in certain sectors, in general we find no persistency in using the Sharpe ratio as a tool for portfolio construction in the hedge fund universe.   In other words, the Sharpe ratio has no validity as a forward-looking investment decision tool.

Be Sociable, Share!

2 Comments

  1. Peter Urbani
    October 29, 2009 at 9:18 pm

    We find similar results in this study http://www.infiniti-analytics.com/kb/kb/article/infinitisfascorearapm

    Interestingly we find that the Sortino ratio does worse than expected due to the fact that, because of the high quality of the 36 underlying funds, it is possible for the computer to find a set of weights that gives an in in-sample portfolio with no downside deviation. When we repeat the study using Indices instead of individual funds the Sortino ratio out-performs the Sharpe ratio.


  2. Ranjan Bhaduri
    November 2, 2009 at 1:47 am

    Nice post. Yes, the Sharpe Ratio is not at all stable. Dr. Galen Burghardt et al, also proved this in a paper (http://www.intercontilimited.com/mfutsarchive/correlation_duration.pdf). It’s reckless and not prudent to use the Sharpe Ratio as an important metric given that it’s not stable, doesn’t take into account the higher statistical moments, that it can be gamed by smoothing for illiquid hedge funds, and that it penalizes for upside volatility.


Leave A Reply

← Gone in 45 days: Who said hedge funds were illiquid? One reason why equity allocations may never fully recover from recent injuries →