The Lumpy Earnings Model
Board: Berkshire Hathaway
If one has found a model that can systematically generate unusual profits the correct behavior is to keep quiet about the existence of the model and to put it into practice on a scale small enough to not disturb the markets in question. Greed comes into play for obvious reasons: If a little bit of a good thing exists, the temptation is to increase the scale.
Completely agree. I think I can shorten your reply into saying that, a successful system, once it is discovered by others, will shortly be arbitraged away.
Here is the exception: such a system, if it is based on deep value contrarianism, will not be arbitraged into failure. The reason for this is that such a system will necessarily experience periods of 3-4 years of underperformance.
Greenblatt points out that in any ten year period, all of the top money managers over that time will spend a year or two in the BOTTOM 10%. Look at Fairholme: Morningstar Fund Manager of the Decade, then, in the 11th year, it sinks to the bottom 10%. Berkowitz coat tailers are not going to arbitrage away his results after that performance. Greenblatt has the same experience lately: he started taking public money under professional management from individuals in 2010. He mostly lagged for 2010, 2011, and 2012. He experienced a LOT of redemptions. I know, because I sold a lot of Berkshire to put it into Formula Investing in 2010, and I talk with the head of trading there. What happened? 2013 made up for 2010, 2011, and 2012, and then some. 50% gains for the year in the managed accounts. Why did it work? Precisely because it is a good system, and people cannot stomach the long three year underperformance.
Why are we all here? Because Buffett/Graham preach a deep value, contrarian philosophy. That philosophy is (from the OP) "a model that can systematically generate unusual profits." Buffett can't be arbitraged away. Neither can Berkowitz or Greenblatt, who have both subsequently bounced back strongly. They can't be arbitraged because the variables are time and patience and confidence. According to Greenblatt, the reigning parameter in meeting performance goals is quarterly. Miss that mark for a couple of quarters in a row, you're likely gone, fired, kaput. Self employed, running a mutual fund? Fairholme had redemptions of something like 50%. Just imagine the freedom enjoyed by being able to wait 2-3 years for a payoff. Flashback: when Buffett bought BNSF, he was roundly criticized, he had lost it again. Flashforward: once more, he is a genius.
So, how do you safeguard "a model that can systematically generate unusual profits?" Answer: through lumpy earnings. Most people can't handle them. Make those earnings lumpy enough, and if you have a system that can generate unusual profits like Buffett, Berkowitz, and Greenblatt have, and those profits, in the long run, will be safeguarded.
One MIGHT think that people would catch on. Not to worry. I remember hearing Clark Howard the financial advisor on the radio a few years back. He said he checked his investment results once every year or so. Clark is a smart guy. Contrast that with eTrade or some such that advertises the first 50 trades are free. Or the baby in the crib in the Super Bowl ads, making stock trades. The guy in the soft chair at Starbucks checking his stocks every 20 minutes. Now, translate that into the 20 or 30 something guy/gal running the average mutual fund. The time horizon is not getting longer to accommodate a systematic model, it is actually getting shorter and shorter all the time. And, paradoxically, the shorter the time horizon gets for the professionals, the better it gets for us contrarians. To meet quarterly metrics, they absolutely have to sell something. Some stocks with the worst short term outlooks. Thing is, some of those stocks still have very good longer term fundamentals.
So the next time you see someone walking down the street staring intently into a small plastic screen, think: he just might be on the other side of the trade.