Model Uncertainty and the Robust Determinants of World Happiness: An Extended Bayesian Model Averaging Analysis

Abstract

This paper provides an extended analysis of the determinants of cross-country happiness using Bayesian Model Averaging (BMA). Building on the 2024World Happiness Report dataset, we explicitly account for model uncertainty in evaluating the relative importance of economic, social, and institutional factors. The results confirm that social support, freedom to make life choices, and perceived corruption are the most robust predictors of happiness. While GDP per capita remains positively associated with well-being, its importance diminishes once model uncertainty is accounted for. These findings highlight the multidimensional nature of happiness and underscore the importance of institutional and social conditions in shaping subjective wellbeing.