Does Financial Inclusion and Digitalization Reduce Corruption? Cross-Country Evidence and Sectoral Insights

Abstract

This study investigates the impact of key economic and financial variables on the Corruption Perception Index (CPI) across countries. The objective is to identify the most significant predictors of perceived corruption. The analysis employs a variety of quantitative methods, including Linear Regression, Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Attention-RNN, and Random Forest models. The results demonstrate that GDP per capita is the most influential factor in explaining variations in corruption perceptions. Among financial sector indicators, variables reflecting active banking operations (such as outstanding loans and deposits) showed stronger predictive power than digitalization metrics (e.g., number of bank branches and ATMs). This suggests that the intensity of financial activity plays a more critical role than the mere availability of banking infrastructure. The study concludes that fostering economic growth and stimulating active engagement with the formal financial system are effective strategies for reducing corruption perceptions. Policy recommendations include addressing unemployment and promoting the formalization of remittance flows.