In a previous post, I argued that “pay for performance” is not practical in the investment management business because most of the variation in performance over horizons of three years or less is due to random noise, not investing skill. If investors cannot reasonably rely on three-year track records, how then should they choose an investment adviser or an investment strategy? One alternative is to use longer, say 20-year track records. Unfortunately, such data are hard to come by for either investment managers or investment strategies. Furthermore, over such long horizons, things change. Managers turn over and the conditions that produced a successful outcome in the distant past may longer apply.
At Cornell Capital, we have concluded that track records are not an effective tool for making investment decisions. What can be used as an alternative? The main place we turn is to the theoretical soundness of the investment thesis. If fundamental research is involved, as it often is in the cases we consider, we ask why is it reasonable to conclude that a security was mispriced? Aswath’s Damodaran blog is an example of what we have in mind. In a post on September 21, Prof. Damodaran valued Amazon at about $1,250 per share based on a detailed fundamental valuation which he made available as an Excel spreadsheet. In response to that valuation, Prof. Damodaran sold Amazon short at a market price of $1,950. Upon studying Prof. Damodaran’s post, we agreed with him that Amazon was overpriced, though not by as much as he suggested, and responded by taking an effective short options position. The investment was not based on Prof. Damodaran’s track record, but on what we concluded was the soundness of his analysis.