Risk Management and Fund Sizing

Fund sizing is a part of risk management that you don’t hear much about. A manager with a stellar track record who does not manage AUM — perhaps closing the fund to new investments or even returning capital — often will enjoy a storied career of destroying value even as he maintains that stellar — though somewhat waning — inception-to-date performance.

People flood into successful funds. But there are things that can go wrong, and eventually do. The manager might not be able to put a larger amount of capital to work. The idea might run its course. The manager might start spending more time counting his money and enjoying his fame.

I first appreciated this point watching the tail spin of Tiger Management in the 1990's. The New York Times had this to say as Tiger imploded: Still, Mr. Robertson’s long-term record is a stand-out. Mr. Robertson’s 25 percent annualized return since his fund’s inception in 1980 handily beats the 17.5 percent average annual gains in the Standard & Poor’s 500-stock index.

That stand-out record didn’t matter for most of Tiger’s investment dollars. If you got into the fund in the mid-1990s, what you saw was a slow motion train wreck. At its peak Tiger managed $21 billion. Robertson could not put the large capital to work on individual stocks, so he moved from stock picking to macro, and in the fall of 1998 when Russia defaulted on its debt, Tiger lost $1.6 billion on bad currency bets, ending down 4 percent for the year. In 1999, value shares were steamrolled by technology, and Tiger’s returns were down 19 percent. Then in 2000, down another 14 percent before calling it quits.

Things don’t have to be so spectacular to make the point. It can be a slow leak rather than a blow out. Forget inception-to-date returns, and ask if the fund has added value on an asset-weighted basis. Are investors worse off overall than if they had simply held a passive index. Has the manager enjoyed a career of losing money for clients, and been paid billions and built up a fanbase in the process.