事件研究法stata

事件研究法stata产生估计窗口和事件窗口，为了保证观察值数目，要count，果然有，所以删除usey1.dta,cleargenes_period=1ifdif>=-100&dif<=-30genev_period=1ifdif>=-5&dif<=5bysortcode:egenobs=count(es_period)bysortcode:egeneobs=count(ev_period)*先用市场调整法计算argenar1=ret-mretifev_period==1bysortcode:gencar1=sum(ar1)ifev_period==1bysortcode:egencar1_11=sum(ar1)ifev_period==1*用市场模型法计算arsortcodeegencode1=group(code)tabstatretmretifes_period==1,s(nmeansdminmax)c(s)setmoreoffgenpret=.forvaluesi=1/381{listcodeifcode1==`i'&dif==0regretmretifcode1==`i'&es_period==1predictpifcode1==`i'replacepret=pifcode1==`i'&ev_period==1dropp}genar2=ret-pretifev_period==1bysortcode:egencar2_11=sum(ar2)ifev_period==1