Asymptotics for the Least Squares Estimator in the Classical Regression Model参考.ppt

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Estimating σ2 Measurement Error y = ?x* + ? all of the usual assumptions x = x* + u the true x* is not observed (education vs. years of school) What happens when y is regressed on x? Least squares attenutation: Why Is Least Squares Attenuated? y = ?x* + ? x = x* + u y = ?x + (? - ?u) y = ?x + v, cov(x,v) = - ? var(u) Some of the variation in x is not associated with variation in y. The effect of variation in x on y is dampened by the measurement error. Measurement Error in Multiple Regression Twins Application from the literature: Ashenfelter/Kreuger: A wage equation that includes “schooling.” Orthodoxy A proxy is not an instrumental variable Instrument is a noun, not a verb * * * * * * * * * * * * * * * * * * * * * * * * * * Data Setup Create; G=log(G); Pg=log(PG); y=log(y); pnc=log(pnc); puc=log(puc); ppt=log(ppt); pd=log(pd); pn=log(pn); ps=log(ps); t=year-1960$ Namelist;X=one,y,pg,pnc,puc,ppt,pd,pn,ps,t$ Regress;lhs=g;rhs=X;PrintVC$ Regression Model Based on the gasoline data: The regression equation is g =?1 + ?2y + ?3pg + ?4pnc + ?5puc + ?6ppt + ?7pd + ?8pn + ?9ps + ?10t + ? All variables are logs of the raw variables, so that coefficients are elasticities. The new variable, t, is a time trend, 0,1,…,26, so that ?10 is the autonomous yearly proportional growth in G. Least Squares Results +----------------------------------------------------+ | Ordinary least squares regression | | Model was estimated Sep 19, 2005 at 12:18:37PM | | LHS=G Mean = 5.308616 | | Standard deviation = .2313508 | | Model size Parameters = 10 | | Degrees of freedom = 17 | | Residuals Sum of squares = .3776938E-02 | | Standard error of e = .1490546E-01 | | Fit R-squared = .997