Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Theodore Palivos


The first two essays in this dissertation deal with the empirical examination of growth, specifically with the reexamination of convergence. The first essay introduces the issue and provides the preliminary results. Most of the empirical studies on the determinants of economic growth use cross-country analysis, employing long-term averages of relevant variables on each country. In these studies a negative relationship between average growth rates and initial incomes is interpreted as an indication of convergence in cross-country incomes and therefore, presented as evidence in support of neoclassical growth models. Such an analysis, however, ignores dynamic information that can explain part of the variation in growth rates. Furthermore, Quah (1993a) and Evans (1996) show that inferences from cross-country regressions regarding convergence are misleading. An alternative approach, based on unit root tests in panel data, suggests that income differences are persistent even after allowing for country specific differences. The second essay modifies the hypothesis to test the implications of endogenous growth models. Stability tests indicate that performance of developing countries has worsened in the post-1973 period. The results indicate an unambiguous convergence only within OECD. This essay also illustrates that using log income differences rather than levels shifts the distribution toward convergence. The dissertation is completed by a third essay that examines the PPP, UCIP and constancy of real rates of return across a large number of countries using the methodology introduced in the first essay. Findings support all these parities, indicating the failure of earlier studies to confirm these hypotheses is likely the result of low power of standard unit root and cointegration tests.