EFFECTIVENESS OF FISCAL POLICY USING DOUBLE MACHINE LEARNING: GCC AND HIGH-INCOME ECONOMIES
EFFECTIVENESS OF FISCAL POLICY USING DOUBLE MACHINE LEARNING: GCC AND HIGH-INCOME ECONOMIES

Summary/Abstract: This study examines the causal effects of fiscal policy on macroeconomic stability and non-oil GDP growth in hydrocarbon-dependent GCC countries and diversified high-income economies. Using a balanced panel of 17 countries from 2000 to 2024, comprising 425 country-year observations, it applies Double Machine Learning to address endogeneity, nonlinearities, and high-dimensional confounding. The analysis focuses on fiscal balance, government expenditure, and tax revenue as shares of GDP, while controlling for institutional and structural factors such as fiscal transparency, sovereign wealth fund governance, public investment efficiency, and economic complexity. The model combines Lasso, Random Forest, and Gradient Boosting with cross-fitting, instrumental-variable checks, fixed-effects OLS, mediation analysis, and heterogeneity tests. The results show significant positive effects of fiscal balance (0.052), government expenditure (0.038), and tax revenue (0.033) on GDP growth (p < 0.01). These effects are partly mediated by public investment efficiency and financial development and are stronger in diversified economies with better institutions. The findings suggest that GCC countries can improve fiscal policy effectiveness through institutional reform and economic diversification.

• Details
• Contents