计量经济学讲义1.pptVIP

Econometric Theory Lecturer: Dr.Jingtao Yi Room 715 Business School, RUC Lecture 1: Two-Variable Linear Regression I 1. Introduction; 2. Bivariate relationships; 3. The two-variable linear regression model. Introduction Structure of Economic Data Cross-Sectional Data Time Series Data Panel Data Cross-Sectional Data (wage and education) observation no. wage education female 1 3.1 11 1 2 3.2 12 1 3 6.2 16 0 . . . . . . . . 499 11.5 9 1 500 3.5 12 0 Time Series Data (min wage and unemployment) observation no. year average min wage unemployment 1 1950 0.2 15.4 2 1951 0.3 16.0 3 1952 0.4 14.8 . . . . . . . . 44 1993 3.4 19.8 45 1994 3.5 16.8 Panel Data (city crime) obs. no. city year murders population police 1 1 1991 5 35000 440 2 1 1992 8 35900 471 3 2 1991 2 64300 75 4 2 1992 1 65100 75 . . . . . . . . . . . . 297 149 1991 10 260700 286 298 149 1992 6 245000 334 299 150 1991 25 543000 520 300 150 1992 32 546200 493 Bivariate Relationships Examples Covariance and Correlation Coefficient Bivariate Relationships The limits of r: Proof: Nonsense/Spurious correlation Two-variable Linear Regression Model Conditional Expectation Case 1: discrete bivariate probability distribution Case 2: bivariate normal distribution Two-variable Linear Regression Model Two-variable Linear Regression Model Conditional Variance (bivariate normal) Homoscedasticity: the conditional variance of Y is invariant with X. Two-variable Linear Regression Model A Conditional Model If X, Y follow bivariate normal distribution, then g(X) is linear in X as: Two-variable Linear Regression Model For individual i, Classical assumptions for disturbance: Two-variable Linear Regression Model Then, this is to assume that The two-variable linear regression model is specified as: Model Estimation Estimators and Estimates Estimators for the bivariate linear regression model To estimate the model, we have to estimate parameters as follows: Model Estimation The procedure to estimate the