Preface

In 2010 I was invited to attended my first Economics seminar. I was an honors student at the time, and wasn’t quite sure what I had got myself into. Glenn Harrison stood up the front of the room and presented some work on estimating models of discounting behavior. I don’t know about you, but for me the idea that you can estimate the fundamental parameters of an economic model has always struck me as pretty amazing: we can go all the way from an economic model of how people might behave in an environment, to an econometric model of the data-generating process of our experiment. This is such a powerful tool, both because it provides us with a lens for evaluating our economic model when we get some data, but it also allows us to comment on and estimate the parameters and quantities that our model says are important.

So that’s how I came across structural, but why Bayesian techniques? As much as I think we should learn from our experiments the way mathematics tells us we should (i.e. use Bayes’ rule), I am primarily here because it does the things I want to do better than existing Frequentist techniques. By that I mean in experimental and behavioral economics we have to take heterogeneity seriously. Bayesian hierarchical modeling is a great way to deal with that. Furthermore, in structural applications we often want to comment on transformations of our parameters. These are handled really easily within the Bayesian framework. For me, Bayes is not the right tool if you just want to draw a straight line through your data. Then, I’m quite happy to run a linear regression because it’s easy and fast. However the models we have in mind when designing and analyzing our experiments are rarely linear or homogeneous.

What I have put together here is a collection of notes on how I think about going from an economic model to an econometric model, how to estimate these econometric models, and how to communicate their results.

Throughout this book, my goal is to equip you with some tools that I have used or developed. This is not a book on econometric theory, although I will certainly mention it when we can take advantage of it. Instead, I want it to be a practical “how to” guide for anyone who wants to use these techniques. As such, there will be at least one example in each chapter, taking you from an economic model to an econometric model, and then showing you how to estimate it. Along the way, I will also show you some computational techniques that might be useful for solving some of the more persnickety economic models.