The Determinants of Democracy Revisited: An Instrumental Variable Bayesian Model Averaging Approach

TL;DR

This paper uses an advanced Bayesian approach with instrumental variables to identify key determinants of democracy, highlighting arable land, youth, life expectancy, and GDP as significant factors across different country groups.

Contribution

It introduces the use of IV Bayesian Model Averaging to robustly identify democracy determinants while explicitly addressing endogeneity issues.

Findings

Arable land is the most persistent predictor of democracy.

Youth population, life expectancy, and GDP per capita are also significant.

State fragility influences democracy in developing countries.

Abstract

Identifying the real causes of democracy is an ongoing debate. We contribute to the literature by examining the robustness of a comprehensive list of 42 potential determinants of democracy. We take a step forward and employ Instrumental Variable Bayesian Model Averaging (IVBMA) method to tackle endogeneity explicitly. Using the data of 111 countries, our IVBMA results mark arable land as the most persistent predictor of democracy with a posterior inclusion probability (PIP) of 0.961. Youth population (PIP: 0.893), life expectancy (PIP: 0.839), and GDP per capita (PIP: 0.758) are the next critical independent variables. In a subsample of 80 developing countries, in addition to arable land (PIP: 0.919), state fragility proves to be a significant determinant of democracy (PIP: 0.779).