IMPACT OF GEOPOLITICAL RISK ON THE CAPITAL STRUCTURE DECISIONS OF SMALL OIL COMPANIES IN EMERGING MARKET ECONOMIES.

ABSTRACT
The impact of geopolitical risk on capital structure decisions of 77 small oil and gas firms operating in emerging economies is studied over the period 2009–2024. Geopolitical Risk (GPR) Index developed by Caldara and Iacoviello (2022) is used to measure geopolitical uncertainty. The primary measure of leverage in this research is Debt-to-Equity Ratio (DER). Firm-specific financial indicators such as , net profit margin (NPM), return on investment (ROI), liquidity (current ratio), firm size (total asset), firm age, tangibility, and EBITDA, and other important external (macroeconomic) variables, such as oil price volatility (OPV) and interest rates are used.
The methodology used in this study is multi-model quantitative framework. The choice of this rigorous methodology is to show that under uncertainty capital structure behavior is regime-dependent, driven by multiple interactions and nonlinear effects. The estimation techniques adopted for the study are complementary as well as reliable, namely Ordinary Least Squares (OLS), Generalized Additive Models (GAM), and Random Forest regression. Whereas the OLS established baseline linear relationships and provided estimation for conventional statistical inference, the Generalized Additive Models (GAM) captured smooth nonlinear effects and threshold behavior in leverage responses. Random Forest regression, with its machine-learning approach, identifies complex interactions between variables that are nonlinear while ranking the relative importance of both macroeconomic and firm-level determinants. Based on approximately 1,600 firm-year observations, the empirical analysis combines accurate data pre-handling and logarithmic transformation of firm size, harmonizing extreme values, and mean imputation for limited missing observations, to ensure efficient and robust results.
The updated empirical findings show that the expanded OLS specification provides substantial explanatory power, with a Multiple R² of 0.6113 and an Adjusted R² of 0.5662. The overall model is statistically significant (F-statistic = 13.56, p-value < 2.2e-16), indicating that the included firm-level and macroeconomic variables jointly explain a significant proportion of leverage variation. The GAM improves nonlinear detection with an adjusted R² of approximately 0.251 and deviance explained of about 28.3%, confirming the presence of smooth nonlinear patterns and threshold responses in capital structure adjustments. The Random Forest model demonstrates strong nonlinear predictive performance, achieving an in-sample R² of 0.9056, an Out-of-Bag R² of 0.5035, and an out-of-sample test R² of 0.5615, indicating robust predictive structure even on unseen data.
 
The results of the empirical and statistical data review therefore indicate that geopolitical risk holds an impact on leverage at a statistically significant level and economically exerts a meaningful effect. Leverage adjustments become more pronounced when geopolitical uncertainty intensifies, reinforcing the argument that capital structure decisions in emerging-market oil and gas firms are shaped by dynamic risk conditions rather than stable linear relationships.
Furthermore, the variable-importance assessment shows that macroeconomic and sector-level risks—particularly oil price volatility and interest rates—play a dominant role in influencing leverage decisions, while firm-level characteristics such as liquidity and size function as stabilizing cushions that moderate exposure to geopolitical shocks. Profitability indicators exhibit comparatively weaker independent effects once macro-financial conditions are accounted for.
The discussion integrates Trade-Off Theory, Pecking Order Theory, Resource Dependence Theory (RDT), and contemporary financial frameworks to conclude that capital structure decisions for oil and gas firms in emerging markets primarily reflect strategic risk-management responses to uncertainty rather than static optimization choices.
Generally, this research contributes to capital structure literature by demonstrating that geopolitical risk remains a critical determinant of leverage dynamics for small oil and gas companies in emerging economies, while also showing that nonlinear econometric and machine-learning methods are essential for accurately capturing financing behaviour under conditions of political and economic volatility.

Keywords: Capital structure; Debt-to-Equity Ratio; emerging economies; leverage, oil and gas firms; geopolitical risk; oil price volatility; interest rates; OLS; Generalized Additive Models; Random Forest; BRICS+.