Country Risk, Financial Development, And Renewable Energy Consumption: Evidence from G20 Countries

Abstract

Background:  The global energy transition toward renewable energy sources remains constrained by complex institutional, financial, and geopolitical barriers, particularly among G20 economies that collectively contribute more than 75% of global CO₂ emissions. Despite recent growth in renewable energy investment, the share of renewable energy in total final energy consumption across G20 countries remains heterogeneous and unevenly distributed, reflecting structural differences in financial system maturity, political risk environments, and fossil energy path dependencies. Objective: This study analyzes the effects of country risk measured through the International Country Risk Guide (ICRG) composite index and financial development on renewable energy consumption across 18 G20 member countries over 1992–2021 (540 observations). Method: Employing fixed-effects panel regression as a baseline and Mean Group Estimator (MG) as the primary estimator, the study yields three key findings. Results: First, financial development exerts a significant negative effect on renewable energy consumption (β = −4.433**), contrary to conventional expectations, confirming the dominance of the financial incumbency hypothesis in the G20 context an entrenched financial system perpetuates fossil fuel financing orientation. Second, composite risk is insignificant, confirming aggregation bias in composite risk indices that conceals individual risk dimension effects. Third, CO2 emissions exert a highly significant negative effect (β = −1.403***) as the only variable consistent across all estimations, confirming carbon lock-in as a universal barrier to energy transition. Conclusion: The reversal of the financial development coefficient from a statistically insignificant positive estimate in OLS FE (β = 8.328, t = 1.32) to a significant negative estimate in MG (−4.433**) empirically demonstrates the aggregation bias consequences concealed in homogeneous panel estimators.