The Overreliance Myth vs. Incentive Systems

Sarita Yardi tweeted earlier today about Sinan Aral’s Tweet about a blog post from Jason Gots about the dangers of exclusive reliance on mathematical models to explain human behavior. Got’s post is interesting to me only because it reinforces the myth that Wall Street relies on these models and this reliance was, in large part, causal for the 2008 financial crash.

The overreliance myth is propagated as an explanation instead of the much more difficult to prove reality that people use the models to justify whatever position they want to take. If you try to convince me that many of the intellectual elite on Wall Street do not know the limitations of their own modeling, I will simply nod and assume that you have not yet seen ‘inside’ the beast. On the inside, we know that management will tell the workers to ‘make the model work’ because, as Jon Kleinberg has so eloquently pointed out with a proof, we always can.

Statistical models have great value for examining behavior. But incentive systems that propagate over-reliance on model outcomes are the real problem. The incentive system on Wall Street perpetuates an environment that allows modelers to win, so they do. Modelers work on both the long and short side of the business and they win because they have an environment that wants to fight in this space (because it excludes the dumb money whose pockets both sides want to pick). Similarly, teachers may help students cheat on standardized tests because they have incentives for their students to do well.

The debate should be about the incentive systems first and the models second. That’s what we’ve learned from watching 30 years of model use in science.

12 notes

Show

  1. alexainslie reblogged this from spurpura
  2. spurpura posted this

Blog comments powered by Disqus