stata考试do文件.doc

*--------------------------Chapter 4-------------------------- * *------------------4.1 存在异方差时的稳健估计----- ***************************************************** *Example 8.1 Heteroscedastic Regression insheet using tablef9-1.csv,clear * keep if avgexp0 //有平均支出为零的去掉 generate incomesq=income^2 regress avgexp age ownrent income incomesq *绘制残差 rvpplot income,xlabel(0(2)12) xline(2 4 6 8 10,lstyle(grid)) ylabel(-500(500)2000) /// yline(0 500 1000 1500,lstyle(grid)) *------------------------------------------------------------ *Example 8.2 The White Estimator *OLS regress avgexp age ownrent income incomesq *Robust Standard Errors. regress avgexp age ownrent income incomesq,vce(robust) *注意：这里给出的稳健标准误是Davidson/MacKinnon提出的修正1,到这里就可以了。 *要获得最初White稳健标准误将协方差矩阵再乘以(n-K)/n matrix b=e(b) matrix V=e(V)*((72-5)/72) ereturn post b V ereturn display //White Robust Standard Errors *Davidson/MacKinnon修正2 regress avgexp age ownrent income incomesq,vce(hc2) *-------Baum 6.1.2 use /data/imeus/fertil2,clear regress ceb age agefbrth usemeth,vce(robust) regress ceb age agefbrth usemeth,vce(cluster children) *-------------------------------------- *Newey-West *Example 19.4 Autocorrelation Consistent Covariance Estimation insheet using TableF2-2.csv,clear tsset year * generate t=_n generate lnG=log(gasexp/(pop*gasp)) //蓩吖腔詬谢貞私 generate lnY=log(income) generate lnPG=log(gasp) generate lnPNC=log(pnc) generate lnPUC=log(puc) generate lnPPT=log(ppt) generate lnPN=log(pn) generate lnPD=log(pd) generate lnPS=log(ps) * regress lnG lnPG lnY lnPNC lnPUC predict e,resid tsline e,title(Autocorrelated Residuals from Gasoline Regression) *Newey-West estimator newey lnG lnPG lnY lnPNC lnPUC,lag(3) *---------------------------------------------------------------------------- use /data/imeus/ukrates,clear summarize rs r20 tsset month regress D.rs LD.r20 display _N^(1/4) newey D.rs LD.r20,lag(5) *------------------------4.3 异方差的检验----------------------- *Example 8.3 Testing for Heteroscedas