In recent years the structural approach to econometrics compared to reduced form econometrics has become more popular. This involves tight combination of theoretical economic models and statistics in order to estimate parameters of interest. Imposing more theoretical structure in the way that we use data and statistical methods is meant to provide guidance and sometimes can even uncover parameters that are not easily estimable with reduced form methods. Even for non-econometricians this can potentially be interesting because simulation and sampling can be an important part in structural estimation and the techniques are well applicable in other social sciences.

This branch of econometrics, as a branch of statistics, does not seem to have any introductory textbooks so far. I have only found more advanced material like Structural Econometric Models by Choo and Shum (2013) or the survey chapter by Reiss and Wolak.

Could someone point me towards a set of lectures or perhaps even a book (that I just haven't found yet) which would provide an introduction to structural econometrics? Ideally this would be based on examples with different approaches including code or a guide on how to replicate these examples for better understanding.

I am aware of several research papers especially in industrial organization

  • modeling of state dependence (Rust, 1987)
  • demand estimation (Berry, 1994; Berry, Levinson, and Pakes, 1995)
  • estimation of productivity (Olley and Pakes, 1996)
  • estimation of market power (Nevo, 2005; Sovinsky, 2008)

but most of them are difficult to follow. So if someone knows about a more gentle introduction this would be of great help.


3 Answers 3


I am not aware of anything like this. Paarsch and Hong's An Introduction to the Structural Econometrics of Auction Data and Ada and Cooper's Dynamic Economics come closest.

The usual classroom approach is to read classic papers and perhaps replicate one along the way. Here's one example (Jean-Marc Robin). Here's are more labor oriented lecture notes (Chris Taber).

  • $\begingroup$ Thanks for your answer Dimitriy. I started a bounty in order to draw a bit more attention to this question because I am very interested in this topic. Perhaps you have further suggestions with some applied examples? I recently bought Wolpin's "The Limits of Inference without Theory" which gives good theoretical examples; now I would look also for more application. $\endgroup$
    – Andy
    Oct 26, 2014 at 20:12

If you also consider structural methods for macroeconomics then perhaps the book Structural Macroeconometrics by DeJong and Dave will be interesting.

Mathias Andre has some structural econometrics problem sets on his website. The questions are similar to the example in chapter one of Wolpin's book you mentioned in the comment to the previous answer and there are also examples of estimating value functions. There are some corrections to Wolpin's book here.

Victor Aguirregabiria makes data and code to several estimation procedures available on his website. So does Aviv Nevo.

Abbring and Klein have published Matlab code for dynamic discrete choice model.

Simon Quinn has lecture notes that start off at the basics and then go on telling you how to use maximum likelihood in estimating utility functions. Data and code should also be on his website. For this purpose the book Maximum Likelihood Estimation by William Gould and his co-authors is valuable because maximum likelihood and simulated maximum likelihood are often used tools in structural econometrics.

  • $\begingroup$ I tried to find a compromise among the good answers here. Dimitriy got the accepted one so +2.5 for the quickest answer with good references. The bounty goes to user45086 for the answer that best fit the description of the problem, +6. And j-kahn's answer got +1 from me for the excellent reference to Kenneth Train's book. Thanks for your answers and I hope some more might come in the future. $\endgroup$
    – Andy
    Nov 2, 2014 at 15:56

One problem is that structural estimation varies a lot depending on what field you are in, as the models used vary a lot. To me, at least, structural estimation in labor and marketing looks wildly different from finance and macro. It might be helpful to separate estimation methods (method of simulated moments, simulated maximum likelihood, Bayesian filtering) from model computation (dynamic programming and value function iteration). For estimation methods Gourieroux and Monfort are a good in-depth survey. For model solution Miranda and Fackler are another good reference, though you can also check out any number of other books on dynamic economics. In combining model solution and estimation, for finance this survey is a very good place to start, for Macro you've already been pointed to DeJong and Dave, and I would add Rust (1994) to an already long list of IO/marketing surveys you're looking at, as well as Kenneth Train's textbook.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.