Idea Transcript
Michigan State University Department of Economics
Jeffrey M. Wooldridge
ECONOMETRIC METHODS WITH CENSORED DATA Helsinki, Finland, June 4-8, 2007 Overview This course will cover models and estimation methods for censored data, broadly speaking. We will consider cross-sectional and panel data applications. For responses that are corner solutions – that is, outcomes that can occur at endpoints with positive probability – we discuss standard methods, such as Tobit models, but also two-part models. For true data censoring (such as top coding or duration censoring), we will cover standard censored regression methods and also less parametric approaches. Text J.M. Wooldridge, Econometric Analysis of Cross Section and Panel Data, MIT Press, 2002. Outline and Reading Assignments The unlabelled readings are from my text listed above. My lectures will tend to follow the text, although some of my more recent work is available through working papers that can be obtained from my web site http://www.msu.edu/~ec/faculty/wooldridge/wooldridge.html Course Outline (Approximate) Monday, June 4, 2007 9.00 – 13.00 1. Introduction: Corner Solutions and Data Censoring: 16.1 2. Corner Solution Responses a. Quantities of Interest: 16.2, 16.3 b. Tobit Models: 16.4-16.6 c. Two-Part Models: 16.7 plus notes d. Two-Corner Models: notes e. Linear Model with a Tobit-like Endogenous Variable: notes 3. Conditional Mean Models for Corner Solutions: 19.2, 19.4.2 L.E. Papke and J.M. Wooldridge (1996), Journal of Applied Econometrics.
Tuesday, June 5, 2007 9.00 – 13.00 4. Censored Outcomes and Censored Regression a. Censored Normal Regression: 16.1-16.3 b. Censored Least Absolute Deviations: 16.6.4 5. Sample Selection a. Exogenous Sample Selection: 17.1, 17.2 b. Truncated Regression: 17.3 Wednesday, June 6, 2007 9.00 – 13.00 5. Sample Selection (continued) c. Endogenous Sample Selection and Heckman’s Method: 17.4.1 d. Endogenous Sample Selection and IV: 17.4.2 e. Sample Selection with Binary Response Models: 17.4.3 6. Panel Data Models with Strictly Exogenous Explanatory Variables a. Unobserved Effects Tobit Models: 16.8.1, 16.8.2 b. Sample Selection: 17.7.1, 17.7.2 Thursday, June 7, 2007 9.00 – 13.00 7. Panel Data Models without Strict Exogeneity a. Tobit Models with Endogenous Explanatory Variables: notes b. Dynamic Tobit Models: 16.8.3 J.M. Wooldridge (2005), “Simple Solutions to the Initial Conditions Problem for Dynamic, Nonlinear Panel Data Models with Unobserved Heterogeneity,” Journal of Applied Econometrics 20, 39-54, January 2005. c. Models for Multiple Corners Papke and Wooldridge (2007), “Panel Data Methods for Fractional Response Variables with an Application to Test Pass Rates,” mimeo, Michigan State University Department of Economics. Friday, June 8, 2007 8. Empirical Exercises a. Modeling Charitable Contributions b. Estimating a Duration Model for Recidivism c. Panel Data Models for Women’s Labor Supply and Wages
2