Attention Capital Allocators

I’ve been reading a number of papers from Michael Mauboussin of Legg Mason recently, he looks in depth at the role of luck and skill in many “probabilistic fields”. A probabilistic field is any activity where the outcome is the result of a probability of a win and where the Law of Large Numbers dictates the outcome over many samples. So for example, many sports games, gambling, and obviously investing. We never quite know for sure if our next trade will be a winner or a loser but we know over a long period of time that if we have a positive expectancy, we should make money.

Michael Lewis wrote the following in his book MoneyBall:

Over a long season the luck evens out, and skill shines through. But in a series of three out of five, or even four out of seven, anything can happen. In a five-game series, the worst team in baseball will beat the best about 15% of the time. Baseball science may still give a team a slight edge, but that edge is overwhelmed by chance.

Goyal & Wahal (2008) did a study of 3,500 plan sponsors on data of returns from 1994 to 2003. Plan sponsors include major capital allocators, pensions funds, unions, endowments, charities, etc, representing about $636 Billion dollars in allocated assets and $108 Billion in redeemed assets. As an aside that represents ~$9.5 Billion & ~$1.6 Billion in management fees assuming 1.5% charge. What they find is allocators tend to invest in the money managers with above average gains the previous three years and fire those managers with the worst performance over those three years. Now that may not be surprising, but what is surprising is the future performance AFTER the 3 year window. Those that had above average performance, and hence allocated capital, tend to have LOWER returns the next 3 years while those that were fired, and hence capital redeemed, tended to have ABOVE average returns the next 3 years.

So the allocators have it backwards and simply do not understand reversion to the mean. If we take Mauboussins’ argument that investing, due to the high level of competition, has a large amount of luck and therefore high tendency to mean-revert, then this makes a ton of sense. Any streak of good performance is due to luck and therefore should revert back to the long run average.

So how does one find a “skilled” money manager. Another interesting study by Cremers and Petajisto (2009) identify a new method of predicting performance based on the activity and make-up of a money managers portfolio. The method is called “active share” and it simply measures the fraction of the portfolio that is different from the benchmark index. High Active Share, that is low composition of the portfolio versus the index, correlates well with excess returns.

So why don’t more managers do this?? Simply boils down to career and business risk. It is much safer and more profitable to join the herd in the investment management business. Though the Steve Cohens and Jim Simmons of the world get a lot of our attention from their ridiculous performance fees, the management fees keeps the gas in the Bentley’s and the Nanny on overtime for most money managers. Nice little profession i’nit?

So to the capital allocators of all shapes and sizes, learn about reversion to the mean, how to differentiate “skill” versus “luck” when it comes to investing. The material is out there and a little dose of common sense helps out a lot. Quantitative work can be a great tool in this regards but one should also have a thorough understanding of the underlying dynamics involved with money manager selection. Hopefully more on that later…

~ by largecaptrader on July 28, 2010.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: