Models.Behaving.Badly

Emanuel Derman’s new book, Models.Behaving.Badly, is a cautionary tale.  Derman, a former Goldman Sachs quant, examines why confusing illusion with reality can lead to disaster on Wall Street and in life.  You can read a bit about Derman (in his own words) here and here; his occasional blog is here. He is currently a professor at Columbia University, where he runs the financial engineering program, and is also a principal at Prisma Capital Partners, where he co-heads risk management.

As Derman puts it, the book is “about metaphors and analogies, about the nature of modeling and theorizing, the difference between them, why financial models will intrinsically and always at best be very limited approximations to reality, and what to do as a consequence.”

I had this general point driven home to me over the week-end in a surprising way.  On Friday, my better half and I went to see Moneyball.  I have been a big fan of the Michael Lewis book since it first came out back in 2003 (see my comments here) and was looking forward to the movie.  I was not disappointed.  It’s a great yarn with interesting applications to business and to life.

A small but significant role in the film was the Yankee cast-off, David Justice, played by Stephen Bishop (the actor, not the musician), a former minor league baseball player turned actor.  Bishop and Justice look a lot alike (see below) and Bishop had spent many hours as a kid idolizing Justice and mimicking his swing.   

Since Bishop and Justice had a bit of a relationship before the movie, Bishop drew on that experience and called on Justice for help.  No less an authority than Justice’s wife (not ex-wife Halle Berry) approved of the result, acknowledging that Bishop “nailed it.”

Happily, Moneyball the movie doesn’t deviate from reality nearly as much as the typical “based on a true story” Hollywood blockbuster.  But it is still a far from perfect representation of reality.  Obviously, movie-makers have much more of an interest in telling a good story than in scrupulously sticking to the facts.  But the limits of film vis-à-vis reality are far greater than those relating to telling a good story. 

On Saturday, I went to a youth football game near my home.  One of my wife’s students had written the best persuasive essay in her class.  Since the essay argued why his teacher should go watch him play football, we were in the stands.  David Justice was one of the coaches roaming the sidelines. I’m not interested in violating anyone’s privacy here, so suffice it to say that the real David Justice was significantly different from the David Justice character I had seen portrayed on screen the night before, at least in that instance.*  I shouldn’t have been surprised by that, but I was.  

Derman argues that in finance, models can only hope to provide a simplistic and very limited approximation to reality.  Movies — as in the portrayal of David Justice — cannot aspire to more and do not.  Indeed, they do not even aspire to that much. We would all do well constantly to bear in mind the limits of financial modeling in general as well as the limits of any specific model.  They aren’t representational; at best they are illustrative.

The resemblance of the Moneyball David Justice to the real David Justice is more than coincidental.  But it was still a long ways from reality — just like economic models.

___________

* Obviously, people are more variable and complex than any model, which means that it would be easy to push this metaphor too far.  However, one instance of inaccuracy is sufficient to falsify a model, yet it is not enough to conclude that it is altogether useless.  Indeed, it might be highly accurate overall, just mistaken in the particular instance I witnessed.  That said, based upon what I saw, “highly accurate overall” seems unlikely.