On the testing of heterogeneity effects in dynamic unbalanced panel data models

This article analyses the behaviour of the Holtz-Eakin test for the presence of individual heterogeneity effects in dynamic small-T unbalanced panel data models in the presence of endogenous but predetermined regressors. The test behaves correctly for a moderate autoregressive coefficient. However, when this coefficient approaches unity, the presence of an additional regressor sharply affects both the power and the size of test.