Which comes first – urbanization or economic growth? Evidence from heterogeneous panel causality tests

Abstract

Heterogeneous panel causality tests are employed to consider the relationship between urbanization change and economic growth. Urbanization causes economic growth in high-income countries, but noncausality could not be rejected for both middle-income and Latin American countries. A bi-directional, equilibrium relationship is observed for low-income, predominately African countries where economic growth has a positive, causal effect on urbanization, but where urbanization has a negative, causal effect on economic growth. Hence, urbanization and economic growth either co-evolve in low-income/African and high-income countries, or else the two processes are decoupled for middle-income and Latin American countries.

Keywords:

• heterogeneous panel causality
• economic growth
• urbanization

JEL Classification:

• C23
• O18
• O54
• O55

Acknowledgements

We thank J. Carrion-i-Silvestre and S. Fachin for providing their GAUSS codes and T. Malinen for providing the GAUSS code for the Banerjee and Carrion-i-Silvestre (2006) test. The system-GMM estimator was developed for STATA by David Roodman.

Notes

¹ A panel made-up of nonhigh-income Asian countries would contain only 10 countries, which we judged too small to be worthwhile/insightful.

² MATLAB code for the Dumitrescu and Hurlin (2012) test was retrieved from: www.runmycode.org/CompanionSite/site.do?siteId=51

³ Dumitrescu and Hurlin’s code does not allow for the lag structure to vary by cross section.