Cross-country growth regressions: problems of heterogeneity, stability andinterpretation

The paper discusses the issues of heterogeneity and stability of cross-coun try growth regressions that have been used to study the problem of converge nce. Almost all studies use pooled regressions. The paper considers the iss ue of pooling under heterogeneity using a hierarchical Bayesian method and estimates growth regressions for different panels studied in earlier papers , and different regimes. The conclusion is that the convergence rates are h igher than those obtained from pooled regressions under the assumption of h omogeneity and that there is instability over time in the relationships.