Dynamic heterogeneity in cross-country growth relationships

Empirical literature examining cross-country growth within the simple Solow framework invariably chooses to let the relationships between growth and its determinants stay constant over time. In an era of rapidly changing economies, this may not be wise. We track these correlations from 1982 until 2003. We find that the relationships between the standard determinants and growth are for the most part, quite stable; however, one-quarter of them do change in a significant way. Theoretically, the findings are interesting because of the possible causes for the structural changes, and the fact that three-quarters of the relationships are stable when maybe they shouldn't be. Methodologically, the findings are interesting because modelers using these variables should account for this heterogeneity to reduce bias when drawing inference from data spanning this period.