To pool or not to pool : revisited

This paper provides a new comparative analysis of pooled least squares and fixed effects estimators of the slope coefficients in the case of panel data models when the time dimension (T) is fixed while the cross section dimension (N) is allowed to increase without bounds. The individual effects are allowed to be correlated with the regressors, and the comparison is carried out in terms of an exponent coefficient, delta, which measures the degree of pervasiveness of the fixed effects in the panel. It is shown that the pooled estimator remains consistent so long as delta < 1, and is asymptotically normally distributed if delta < 1/2, for a fixed T and as N -> ∞. It is further shown that when delta < 1/2, the pooled estimator is more efficient than the fixed effects estimator. Monte Carlo evidence provided supports the main theoretical findings and gives some indications of gains to be made from pooling when delta < 1/2. The problem of how to estimate delta in short T panels is not considered in this paper.