A Global Analytical Comparison of Transfer Pricing Determinants: Tax Rates, R&D, and Predictive Models

This study presents a comprehensive global investigation into the determinants of transfer pricing behavior among multinational enterprises (MNEs), focusing on the roles of statutory corporate income tax (CIT) rates, R&D tax incentives, and macroeconomic variables. Utilizing panel data from 112 countries over the period 2005 to 2023, we apply both fixed-effects regression and advanced machine learning algorithms—Random Forest and XGBoost—to evaluate the impact of these factors on profit shifting.The results show that higher CIT rates significantly discourage reported profits in high-tax jurisdictions, while generous R&D incentives and favorable IP regimes attract disproportionately large income allocations. GDP per capita is positively associated with regulatory enforcement strength. Machine learning models outperformed traditional econometrics, highlighting complex interactions and improving predictive accuracy.This study bridges the gap between economic theory and AI-based modeling, offering novel insights for tax authorities and policymakers seeking to strengthen international tax compliance. It provides evidence supporting reforms such as tightened IP nexus rules and the application of minimum global tax policies under OECD Pillar Two.