Graduate Panel Data Econometrics

General Info

Econ 220C develops the econometric methods used to analyze panel data, where the same units are observed over several periods, in four parts. The first part studies the static linear panel model, starting from pooled OLS and the random-effects estimator and then turning to the fixed-effects model, its identification and consistency, instrumental variables, and clustered inference with time effects. The second part moves to dynamic panel models, where past outcomes influence current ones, and to modern difference-in-differences designs with staggered adoption. The third part develops the theory of extremum estimation, covering the general consistency theorem and the uniform law of large numbers, which prepares the ground for the nonlinear models that follow. The last part applies this framework to limited and qualitative dependent variables, including binary choice and the probit model, censored and truncated regression, and sample selection. The asymptotic theory of extremum estimators, with the method of moments as a special case, provides the unifying thread throughout.

Discussion section: Tuesday from 5:00 to 6:30 pm, SSB 107. Office hour: Tuesday from 6:30 to 7:00 pm, SSB 107.

Material

  1. Introduction to static panel data: data structure, pooled OLS, and the random-effects model. (Slides)
  2. Fixed-effects model: consistency, identification, instrumental variables, and clustering. (Slides)
  3. Clustering and time fixed effects: clustered inference and adding time effects to the within model. (Slides)
  4. Dynamic panel data and staggered DID: persistence in panels and modern difference-in-differences. (Slides)
  5. Practice test: static and dynamic panel data. (Slides)
  6. M-estimation: the general consistency theorem, the uniform law of large numbers, and exercises on NLS and QMLE. (Slides)
  7. Probit model: linear projection, maximum likelihood, instrumental variables, and the uniform law of large numbers. (Slides)
  8. Practice final: binary outcomes with endogeneity, the control function, and discrete choice. (Slides)
  9. Practice final: truncated and censored (Tobit) regression and sample selection. (Slides)

Evaluation

Instructional Assistant Student Evaluation of Teaching available here.