[R] plm package twoways effect problem

[yves croissant]

It's difficult to guess what the pb is without the data. 

Try to estimate a 'simple' model, ie without the dummies (something
like :

dsnfi3LP=plm(lnLP~lnC+lnL+lnM
+lnE,data=newdata,effect="twoways",model="within")

Try also pdim(newdata) to check if you don't have 1 observation for some
individuals.