Invest Skeptically

Here are two great articles from the Zephyr Associates Inc. website on Mutual Fund Analysis and Peer Group Analysis. They discuss the curses of endpoint bias and survivor bias that skews the analysis of mutual funds.

Endpoint bias can be seen in typical mutual fund ads that compare, for example, 1-year, 3-year, and 5-year returns versus some index or benchmark. However, the cutoff date for these statistics can drastically change the attractiveness of the fund — that’s the effect of endpoint bias. The Zephyr paper shows that, by just changing the endpoint (the cutoff date), you can turn a mutual fund that looks like it’s consistently outperforming into one that looks like it’s consistently under-performing the market index. This statistical trick is useful in mutual fund sales where marketers tend to “sell last year’s winners”.

Survivor bias can be seen in typical mutual fund fact sheets where they try to show what quartile (i.e. how they rank) the fund is in relative to their “peers”. The problem with this is two-fold: (1) the universe of “peers” is usually selected based on broad mandate, or type of fund, rather than the style of the managers; (2) poor performing funds get dropped out of the universe as they get wound-down or amalgamated into other funds — that’s the effect of survivor bias. This is a game that is often played by marketers because no one wants to be the fund that’s below the median. As an aside, always be wary of these presentations if they exclude the relevant index/benchmark. Your fund may be in the top quartile, but who cares if the market is in the top decile?

The conclusion: be skeptical of what you see in marketing materials and ask for more details. What details? Stay tuned for Part 2…