Empirical evidence on the two-gap hypothesis: A revised analysis

This paper reconsiders the problem, analyzed by Weisskopf (1972a), of empirically classifying a set of low-income countries as having had their growth performance primarily constrained by the availability of foreign exchange or of domestic savings. Ambiguities in Weisskopf's procedure are noted, and an alternative, based on the extension of classical likelihood-ratio methods to the case of regression parameters subject to linear inequality constraints, is suggested. Reclassifying a subset of countries from Weisskopf's sample, we find stronger evidence against the importance of an independent foreign exchange constraint than indicated by Weisskopf's original results.