北大高计讲义Lecture 9 PPT.pdfVIP

  • 1
  • 0
  • 约2.78万字
  • 约 5页
  • 2020-10-06 发布于浙江
  • 举报
The Endogenous Explanatory Variable In the classical Gauss-Markov setup: y β +βx +β x +...+β x +u Endogenous Explanatory 0 1 1 2 2 k k Variable and IV estimation an explanatory variable is said to be endogenous if it is correlated with u, i.e. CovCov(( xx , uu)) ≠≠00 i OLS on the above equation will lead to endogeneity bias in the parameter estimates: −1 −1 βk ×1 (X X ) X y β+(X X ) X u −1 E (β) β+(X X ) E (X u ) ≠β The Endogenous Explanatory Variable The Instrumen

文档评论(0)

1亿VIP精品文档

相关文档