格林 面板数据讲义-4-mde.pptVIP

Econometric Analysis of Panel Data Econometric Analysis of Panel Data 4-A. Minimum Distance Estimation Chamberlain’s Model Chamberlain (1984) “Panel Data,” Handbook of Econometrics Innovation: treat the panel as a system of equations: SUR Models, See Wooldridge, Ch. 7 through p. 172. Assumptions: Balanced panel Minimal restrictions on variances and covariances of disturbances (zero means, finite fourth moments) Model the correlation between effects and regressors Chamberlain (2) Chamberlain (3) - Data Chamberlain (4) Model Chamberlain (5) SUR Model Chamberlain (6) Chamberlain (7) Estimation of Σ Chamberlain (8) Estimation of Π FGLS. Use the usual two step GLS estimator. OLS. System has an unrestricted covariance matrix and the same regressors in every equation. GLS = FGLS = equation by equation OLS. Denote the T OLS coefficient vectors as P = [p1, p2, p3 …, pT]. Unconstrained OLS will be consistent. Plim pt = πt, t=1,…,T OLS is inefficient. There are T(T-1) different estimates of ? in P and T-1 estimates of each δt. Chamberlain Estimator: Application Chamberlain Estimator Efficient Estimation of Π Minimum Distance Estimation: Chamberlain (1984). (See Wooldridge, pp. 442-446.) Asymptotically efficient Assumes only finite fourth moments of vit Maximum likelihood Estimation: Joreskog (1981), Greene (1981,2008) Add normality assumption Identical asymptotic properties as MDE (!) Which is more convenient? MDE-1 MDE-2 MDE-3 MDE-4 MDE-5 MDE-6 MDE-7 MDE-8 MDE-9 Minimum Distance Estimation Carey Hospital Cost Model Multiple Estimates (25) of 10 Structural Parameters MDE (2) MDE (3) Maximum Likelihood Estimation MLE (2) Rearrange the Panel Data Generalized Regression Model Least Squares GLS and FGLS * * William Greene Department of Economics Stern School of Business Cornwell and Rupert: Lwageit = αi + β1Expit + β2Expit2 + β3Wksit + εit αi projected onto all 7 periods of Exp, Exp2 and Wks. For each of the 7 years, we regress Lwageit on a constant and the three va