Bayes estimation of short-run coefficients in dynamic panel data models

Abstract

Introduction This chapter considers the estimation of the mean coefficients of dynamic panel data models in the presence of coefficient heterogeneity across cross-sectional units. It is well known that when the regressors are exogenous, the estimators based on the sampling approach such as fixed and random effects yield consistent estimates of the mean coefficients when the number of cross-sectional units approaches infinity (see, for example, Zellner (1969)). However, Pesaran and Smith (1995) have demonstrated that the same results do not carry over to dynamic models. The neglect of coefficient heterogeneity in dynamic models creates correlation between the regressors and the error term as well as causing serially correlated disturbances, thus rendering the within estimators biased. Moreover, the bias of the within estimators does not disappear even asymptotically. Neither do the usual remedies such as instrumental variables estimation technique or differencing the variables work in this case. The asymptotic bias of the usual within estimator is a function of the degree of coefficient heterogeneity and the extent of serial correlation in the regressors. While the inconsistency of traditional pooled panel estimators have been well established for dynamic random coefficient models, it is hard to say that there is similar clarity in the literature about the appropriate estimation technique that needs to be used (see Maddala et al . (1997) for a review).

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 220 citations exist upstream.