+7 votes
asked ago in General Economics Questions by (200 points)
Hi all,

My econometric training has been in applied or reduced form (RF).
However, in my current occupation I read and need to implement plenty of structural estimation (SE).

I'm looking for a high-quality through book or online course on this - that will take me through the hoops. How do you approach a research question with SE? what should you model first? how does data relate to the model? how do you estimate such models?

It seems to me that RF is very-straight forward. the emphasis is on the design itself, and if we have RCT (or similar), then a simple liner regression of Y on treatment yields the estimates of interest. SE seems much more complicated, with many steps from "question" to "answer".
Though I have both a BA and an MA in Economics and Statistics, Reading SE papers without the proper background on the matter gets me nowhere, fast.

Any help would be highly appreciated.

3 Answers

+1 vote
answered ago by (280 points)
0 votes
answered ago by (150 points)
Highly recommend this book by Ken Wolpin : https://mitpress.mit.edu/books/limits-inference-without-theory
+1 vote
answered ago by (3.3k points)
What field are you interested in? SE can vary across fields. I would find some SE papers in your field that you can replicate. Search for either older seminal papers that are "relatively" simpler and have a lot written about them OR search for more recent papers where you have access to their data and code. For SE, I think learning by doing is key.