Defining the Determinants of Corporate Financial Performance: A Machine Learning Approach

Corporate Financial Effectiveness Ensemble Machine Learning Geopolitical Risk Sanctions Resource Dependence Theory Resource-Based View (RBV) Shapley Additive Explanations (SHAP) Russian Enterprises Emerging Markets

Authors

Downloads

This study investigates the determinants of corporate financial performance (CFP) among Russian enterprises (2012–2023) through the lens of geopolitical disruptions, employing ensemble machine learning (ML) to address methodological gaps in modeling non-linear institutional interactions. Using data from 25 large non-financial firms, we analyze sectoral, organizational, and strategic drivers, integrating train-test splits (75%/25%) and 10-fold cross-validation to mitigate overfitting. Results reveal that industry affiliation, initially dominant (28% explanatory power pre-2022), declined sharply post-sanctions (15%), reflecting vulnerabilities in globally integrated sectors like manufacturing and extractives. Organizational size exhibited a nonlinear relationship with CFP, favoring comparatively smaller firms’ agility over larger enterprises’ rigidity, consistent with transaction cost economics. Strategic investments in corporate social responsibility (CSR) and research and development (R&D) diminished post-2022 as firms prioritized liquidity and operational stability, aligning with resource-based view principles. Methodologically, Shapley Additive Explanations (SHAP) clarified threshold effects in CSR returns and innovation’s reduced role under sanctions. The study innovates by applying ensemble machine learning to sanction-affected emerging markets, challenging linear econometric assumptions and advancing institutional theory through a crisis-contextualized framework of resource dependence and stakeholder salience. Findings underscore the fragility of intangible assets under systemic shocks and advocate adaptive resource allocation frameworks to balance short-term survival with long-term resilience. This work provides policymakers and managers actionable insights for fostering operational agility and strategic foresight in volatile institutional environments.