I am esimating a panel model using variables already in log differences. In order to chech for serial correlation in these equation should I put these variables as arguments in the xtserial or just the log-values?
Thanx

Piero Esposito

University of Rome "La Sapienza"
Department of Economic Science