Measuring political rivalry and estimating its effect on economic growth

Abstract

In this paper we construct a composite indicator of political rivalry using factor analysis and then build a panel dataset of political rivalry levels for 125 countries during the 1984-2012 period. According to the factor analysis results, while specific institutional quality aspects are fundamental for defining the degree of political rivalry, political regime specific variables and natural resources rents do not appear significant. A preliminary analysis of the constructed indicator shows that political rivalry is clearly inversely related to the level of development, and that there are significant differences in political rivalry levels among countries, depending on their income and geographical location. The results of system-GMM estimations of the effect of political rivalry on economic growth clearly indicate a negative effect, which is equally maintained when other political and institutional variables are simultaneously considered, and when the model is tested on a number of restricted samples, thus confirming the robustness of the baseline results. Additionally, regression results for the restricted samples suggest that the negative effect of political rivalry on growth weakens as the development level increases.