A reinvestigation of the existence of multiple cross-country growth regimes

Abstract

The purpose of this study is to replicate and extend a part of the study ‘’ Multiple regimes and cross-country growth behavior’’ carried out by Steven N. Durlauf and Paul A. Johnson (1995). Despite the dominant theory of a single-regime model existence in cross-country growth theory, they have found strong evidence of multiple-regimes. Firstly, we are exploring the literature that are directly or indirectly related to this topic. On the second part, we present in detail the S.N. Durlauf and P.A. Johnson (1995) theory framework. On the third part, we follow the reasoning of the authors and empirically try to reassess their method, focusing on estimating the unconstrained regressions accompanied with the relative Wald tests. These estimations cover the 1960-1985 (96 countries), 1960-2015 (96 countries) and 1960-2015 (108 countries) time frames. Our findings are not always consistent to the main study findings, due to mixed evidence regarding the regimes in the growth process. Finally, we draw the conclusion based on the most important results of our empirical approach. Overall, we point out that: i) for the 1960-1985 time frame, the results are mixed, regarding the multiple regime existence, as we reject the single regime specification for some cases (under literacy rate in two-way split and when the variables interact) but we do not for some others (under the output per capita in the two-way split and in all cases in the three-way split) ii) for the expanded set of countries and span of time, we still produce mixed results (rejection of the single regime only when under output per capita we break the sample at three parts and when we let the variables interact with each other), showing a contradiction between the control variables and iii) for the 1960-2015 using the original sample, we find exclusively evidence of a single regime existence.