ON THE TESTING OF CORRELATED EFFECTS WITH PANEL-DATA

Using orthogonal deviations of the variables, correlated effects biase s are regarded as misspecification biases due to the exclusion of rele vant variables in a standard regression model. Following this approach , Hausman- and Chamberlain-type tests of correlated effects are obtain ed as Wald tests in an extended model estimated by OLS, and robust gen eralisations are suggested. Alternative estimators which introduce res trictions in the regression of the effects on the explanatory variable s are proposed. Finally, the paper extends the results to dynamic mode ls.