Forecasting Illicit Practice Impacts on Economies of Conflict Countries: XGBoost Approach

The article examines the impact of illegal practices on economic growth in countries experiencing armed conflict in the context of increasing global geopolitical instability. The study aims to develop and test a predictive approach to quantitatively assess the impact of corruption, money laundering, and cyber risks on the economic growth dynamics of conflict countries using modern machine learning methods. The empirical basis of the study is panel data for 83 countries during 2012–2024. The dependent variable is the logarithmic growth rate of GDP per capita, and the key influencing factors are indicators of corruption, money laundering risk, cybersecurity, and conflict intensity. The methodological basis is the extreme gradient boosting (XGBoost) algorithm, supplemented by the construction of lag variables, interactions for conflict countries, feature importance and SHAP interpretation methods, and scenario analysis. The scientific novelty of the study lies in combining machine learning with cross-country panel data to forecast economic growth in conflict conditions, as well as in identifying nonlinear, asymmetric, and delayed effects of illegal practices. Unlike traditional econometric approaches, the proposed model allows for complex interrelationships among institutional, financial, and security factors, as well as the heterogeneity in responses across countries’ economies. The results show that all XGBoost models achieve stable, high-quality forecasting, with the most accurate results from a comprehensive model that considers all types of illegal practices. It was found that economic growth in conflict countries exhibits significant inertia, while the effects of corruption, money laundering, and cybersecurity manifest with a time lag. SHAP analysis revealed contradictory, context-dependent effects: corruption and money laundering can partially compensate for institutional weakness in the short term, while cybersecurity deterioration mostly has a negative impact on economic dynamics. The practical significance of the results lies in the possibility of using the proposed approach to support decisions in the areas of anti-corruption policy, financial monitoring, and cybersecurity, as well as formulating strategies for economic stabilization and post-conflict recovery, considering the specifics of individual countries and development scenarios.