Addendum to my first post: Since I wish to understand what plm does to my data, I tried to manually calculate the demeaned values and use OLS. See below how far I got with the Grunfeld data; formula's are based on Greene's Econometric Analysis.
Obviously, I am missing at least one important step as my results differ from both the plm and the LSDV estimations. Transferring the model to my original dataset, I i) also get different results for the lm vs. the plm models, however with the "manually" demeaned dataset, I do not end up with incredibly small standard deviations in the demeaned values. That is, the step that I am missing when reproducing the plm results is highly probable to cause the error in my analysis. ############### library(plm) library(lmtest) data(Grunfeld) head(Grunfeld) pdata<-pdata.frame(Grunfeld,index=c("firm","year")) head(pdata) fm<-inv~value+capital mod<-plm(fm, pdata, effect="twoways", model="within") y<-Grunfeld$inv x1<-Grunfeld$value x2<-Grunfeld$capital y_i<-c() x1_i<-c() x2_i<-c() for (i in unique(Grunfeld$firm)){ y_i<-c(y_i,rep(mean(Grunfeld$inv[Grunfeld$firm==i],na.rm=TRUE),nrow(Grunfeld))) x1_i<-c(x1_i,rep(mean(Grunfeld$value[Grunfeld$firm==i],na.rm=TRUE),nrow(Grunfeld))) x2_i<-c(x2_i,rep(mean(Grunfeld$capital[Grunfeld$firm==i],na.rm=TRUE),nrow(Grunfeld))) } y_t<-rep(NA,nrow(Grunfeld)) x1_t<-rep(NA,nrow(Grunfeld)) x2_t<-rep(NA,nrow(Grunfeld)) for (t in unique(Grunfeld$year)){ y_t[Grunfeld$year==t]<-rep(mean(Grunfeld$inv[Grunfeld$year==t],na.rm=TRUE),sum(Grunfeld$year==t)) x1_t[Grunfeld$year==t]<-rep(mean(Grunfeld$value[Grunfeld$year==t],na.rm=TRUE),sum(Grunfeld$year==t)) x2_t[Grunfeld$year==t]<-rep(mean(Grunfeld$capital[Grunfeld$year==t],na.rm=TRUE),sum(Grunfeld$year==t)) } y_it<-rep(mean(Grunfeld$inv,na.rm=TRUE),nrow(Grunfeld)) x1_it<-rep(mean(Grunfeld$value,na.rm=TRUE),nrow(Grunfeld)) x2_it<-rep(mean(Grunfeld$capital,na.rm=TRUE),nrow(Grunfeld)) y_star<-y-y_i-y_t+y_it x1_star<-x1-x1_i-x1_t+x1_it x2_star<-x2-x2_i-x2_t+x2_it mod_lm<-lm(y_star~x1_star+x2_star-1,na.action=na.omit) mod_dumm<-lm(Grunfeld$inv~Grunfeld$value+Grunfeld$capital+as.factor(Grunfeld$year)+as.factor(Grunfeld$firm)-1,na.action=na.omit) coeftest(mod) coeftest(mod_lm) coeftest(mod_dumm) ############## Again, any help is highly appreciated. Best wishes, Philipp ----- ____________________________________ EBS Universitaet fuer Wirtschaft und Recht FARE Department Wiesbaden/ Germany http://www.ebs.edu/index.php?id=finacc&L=0 -- View this message in context: http://r.789695.n4.nabble.com/error-in-La-svd-Lapack-routine-dgesdd-tp2125052p4604714.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.