To view the full discussion, click here.
As you've been in the business for a long time, you know we all talk about avoiding data mining. But by virtue of the fact of your experience, you're constantly in the process of data mining because your experience tells you what the answer is before you start on a lot of research projects.
It's fairly easy to get seduced by a new idea that matches what your bias is. As a firm, I'd say the primary lesson that we've learned is that if someone has an idea, it does not start at the computer. It actually starts with a piece of paper.
They must write down what they believe behaviorally they're extracting from the market. We are not data miners and we want to avoid being data miners. Nor are we quants that are looking for the data to tell us the answer. We actually want behavioral finance to tell us the data, which requires a researcher to be with us a fair amount of time before he is expressing something that he's seen in the markets.
We have all of our researchers trade and lose money; get caught in a cotton-up move or whatever that might require them to stay up all night with the position that they feel terrible about themselves. We're looking for a researcher to say, "I think the market does X for the following reasons because this is how people behave."
We believe firmly that every investor is rational; it's just that their objective function is different than another. And when we can describe a rational decision that creates a price anomaly, that's what we're looking to extract.
I would say the primary lesson I've learned is to make certain that you start with a behavioral concept that you believe makes sense in the market.