This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Introduction Properties of OLS Goodness of Fit Linearity Assumption Econ 103, UCLA, Fall 2010 Introduction to Econometrics Lecture 4: Simple Regression (cont.) Sarolta Laczó October 5, 2010 Introduction Properties of OLS Goodness of Fit Linearity Assumption Introduction We consider the linear regression model Y i = β + β 1 X i + u i The OLS estimators are: ˆ β 1 = ∑ i ( X i ¯ X )( Y i ¯ Y ) ∑ i ( X i ¯ X ) 2 = s XY s 2 X ˆ β = ¯ Y ˆ β 1 ¯ X Introduction Properties of OLS Goodness of Fit Linearity Assumption Introduction/2 Outline: Properties of OLS (just like we did for ¯ Y , an estimator of μ Y ) Goodness of fit of the regression Discussion of the linearity assumption Introduction Properties of OLS Goodness of Fit Linearity Assumption Properties of OLS/1 1 The sample regression function obtained through OLS always passes through the sample mean values of X and Y . 2 ¯ ˆ u = P i ˆ u i n = 0 (mean value of residuals is zero) 3 ∑ i ˆ u i X i = 0 ( ˆ u i and X i are uncorrelated) Note: these results hold by construction, without the OLS assumptions. Introduction Properties of OLS Goodness of Fit Linearity Assumption Properties of OLS/2 Under the OLS assumptions (see Lecture 3), we have the following results for ˆ β 1 : 1 E ( ˆ β 1 ) = β 1 . In words, ˆ β 1 is an unbiased estimator of β 1 2 As the sample size n increases, ˆ β 1 gets closer and closer to β 1 , i.e. ˆ β 1 is a consistent estimator of β 1 . 3 If n is large, the distribution of ˆ β 1 is well approximated by a normal. In particular, ˆ β 1 ∼ N β 1 ,V ar ( ˆ β 1 ) , where V ar ( ˆ β 1 ) = ˆ σ 2 u ∑ i ( X i ¯ X ) 2 ˆ σ 2 u = ∑ i ˆ u 2 i n 2 = ∑ i ( Y i ˆ Y i ) 2 n 2 For ˆ β : V ar ( ˆ β ) = ∑ i X 2 i n ˆ σ 2 u ∑ i ( X i ¯ X ) 2 Introduction...
View
Full Document
 Spring '10
 Davis
 Linear Regression, Regression Analysis, OLS, Mean squared error, linearity assumption

Click to edit the document details