German researchers wishing to find the genetic determinants to stock picking skills recently played an interesting trick on their test subjects. (Hauke R. Heekeren , et all 2009).1 The game rewarded contestants with points for discerning patterns and making the best choice from several options. The fun part started when a player chose the best answer seven times in a row, at which point, the researchers suddenly changed the pattern.
Subjects in the lab experiment scored highest when they could quickly adjust to outcomes that didn’t match their expectations. This is an excellent model of how the stock market works. Once a pattern is widely recognized, it changes. We all know this, so to prevent those with natural stock-picking talent from having an unfair advantage and to keep all the mediocre fund managers gainfully employed, we designed the concept of investment styles.
Academic research demonstrates that none of the standard investment styles ever add any value (Laurent Barras, et all, 2009)2, Thank goodness. There are just too many people out there beating their benchmarks every year! Fortunately the winning funds used methodologies that only worked extremely well for a limited period of time, and then, for no reason at all, God, or somebody, just stepped in and changed the rules. The purpose of the investment style, of course, is to prevent any of the really clever fund managers from adapting to the change, and for the most part it seems to work. Of course, there will always be some fund managers who try to cheat on this system by engaging in what is known as style drift. These fund managers will only get caught if they perform abnormally well, in which case, the simple solution is to fire them.
Style Counsel
The problem for anyone trying to integrate a group of effective industry analysts into a cohesive team when you have already promissed your sponsors that you will follow a given investment style is that the factors that contribute to success in any one sector are almost always different from any other sector. An overly zealous adherence to investment style can have a devastating effect on the performance of individual analysts within a firm. Let’s start off with a very simple example. The Low Price book anomaly has been known to add very persistent excess return to both Long only and Long/Short books in Japanese Equities. (Please see figure 1.) At least it used to, but let’s say you’re a Value Fund that has gotten religion about this anomaly and then you hire a retail analyst with a reputation for very sound valuation-driven research. Now you are in trouble. Low Price/Book strategies are about the surest way to lose money in Japanese retail stocks that was ever invented. (Please see figure 2).
In my view, the best way to deal with this is to hire an analyst who has enough confidence to tell his boss to fuck off whenever she feels its appropriate. Very painstaking and scientific research demonstrates that it takes about 7 years for an analyst to develop this level of confidence, so never hire anyone with less than 7 years of experience.
Managing Angels
Fortunately for those of us who need to live with them, experienced analysts can still make horrible mistakes. As a manager, you need to be able to distinguish between routine mistakes that happen every day and are best used for opportunities to build character, and those mistakes that prove we have somehow lost the plot. Figure 3 shows the performance of low EV/EBITDA vs high EV/EBITDA within the technology sector. Notice how outstanding this simple metric performed between 200 and 2006. Then it just failed. This is a potential game changer. In my view, a good analyst will usually pick up on this change even faster than a trader or a PM, but no one can be a genius at every turn. Even Michael Jordan needed a coach once in a while. The worst thing you can do in such situations is to pretend that you know more than the analyst. Do you imagine that Phil Jackson ever really tried to teach Michael Jordan anything about how to play Basketball? That probably wouldn’t have worked so well.
What you can and must do in such situations is to show that you are monitoring the performance fairly, and that you are measuring the underlying attributions of that performance in a way that is at least careful and hopefully very clever. If you have good attribution data, make it available to the analysts on an internal website. Good analysts will learn to use it themselves without being prompted. If not, you can ask them, “Why isn’t EV/EBITDA working anymore?” If they don’t have an answer, tell them that this is information that you need to have. You’ve made your point without making it personal and without assigning any blame to anyone. If you don’t have the precise attribution data, you will have nothing constructive to offer.
Style Creeps
Don’t get me wrong. There are certain types of style drift that no sensible naked analyst would approve of. If you sell yourself as an expert on US Health Care stocks, for example, please don’t invest my money in Bolivian Bonds. But as long as you stick to what you know, there’s never any need to set yourself up to win this year’s Darwin awards. Remember, it’s not the strongest or the smartest who survive. It’s the ones who adapt.
2. Most recently, Barras, et all, in a study of 2,076 funds, found that just 0.6% had alphas exceeding zero. Barras, Laurent, Scaillet , O. and Wermers, Russ R., False Discoveries in Mutual Fund Performance: Measuring Luck in Estimated Alphas (April 20, 2009). Journal of Finance, Forthcoming; Swiss Finance Institute Research Paper No. 08-18.
Low EV/EBITDA models that strip out a company’s chunky equity holdings destructed when the market (correctly) started discounting forward lower stock market values. Also, the demise of “activists” got people to realize that much touted “hidden value” was never, ever going to be realized. Actually it was worse than that because as the activists with big holdings in low EV/EBITDA stocks ran into trouble, people (correctly) anticipated large blocks of those stocks becoming available at disadvantageous prices…..
Bull Dog Sauce defeated Steel Partners in the courts in June of 2007 and activism seems to have died on that day, but it is difficult to argue that markets anticipated this, since Bull-Dog’s stock price plunged 68% relative to topix during the subsequent two weeks, and another 45% in the next six months. So how does one explain that the EV/EBITDA model actually died in early 2007 – 6 months before the BullDog ruling? My view is that rather than one causing the other, both were correlated to a third factor, most likely the rush into and back out of Japanese stocks by foreign investors who had no idea what they were doing.