|
|
|
|
Current Performance Attribution Methods are Wrong?Damien Laker, CompoundingHappens.com11th February 2008Recently, there has been increasing chatter suggesting that there is something fundamentally wrong about accepted methods for performance attribution. One suggestion used the same heading as this article, except there was no question-mark. It said that attribution calculations should use money-weighted returns (rather than time-weighted returns). Some people say that return attribution is useless, and that a risk-based factor model is the only reasonable solution. Others say that the traditional approach is fatally flawed, because it entails the concept of interaction. It now seems fashionable to rubbish the traditional approach to performance attribution. I think that it’s fine to suggest new approaches or incremental refinements. But when people suggest that traditional performance attribution is worthless or wrong, we are in danger of losing something of great value. I belong to the first generation in Australia where the incidence of polio was almost zero. In the generation before mine, thirty thousand Australian children contracted the disease. Some died, and many were crippled. After routine immunisation began, the incidence of polio plummeted, and now Australia has been declared polio free. Nobody in Australia has contracted polio for years. This is a wonderful triumph of technology. “What does this have to do with performance attribution?”, you may ask. There is a vocal group that opposes the use of immunisation for diseases such as polio. Some of them say that immunisation simply doesn’t work. Others say that the risks of immunisation exceed the benefits. This reminds me very much of what people say about performance attribution. We hear that attribution is fundamentally wrong (i.e. doesn’t work), or that it is too hard to implement (i.e. that the risks exceed the benefits). However, the key difference is that responsible medical practitioners are unanimous in their support for the concept of immunisation, and for its efficient implementation. On the other hand, performance practitioners never seem to have given a clear collective endorsement of performance attribution. Rather, they are the ones who seem to spend much of their time saying that there is something fundamentally wrong with performance attribution. Freedom of speech is a wonderful thing, but if it is exercised unwisely, it can hold back useful technological advancements. The key question, of course, is whether performance attribution is a fundamentally sound technology. In an article of this length, we cannot adduce all the arguments pro and con. My own view is that the traditional approach to performance attribution is perfectly good. But, as they say, “the perfect is the enemy of the good”. Could it be that our industry allocates too much resources to squabbling about a perfect approach to performance attribution, and too little resources to efficiently implementing a good one? |
|
Send mail to
webmaster@compoundinghappens.com with
questions or comments about this web site.
|