Exploring the impact of financial literacy on predicting credit default among farmers: An analysis using a hybrid machine learning model

Abstract

This study explores whether financial literacy can enhance the ability to predict credit default by farmers using machine-learning models. It introduces a hybrid model combining k-means clustering and Adaboost to predict loan default using data on 10,396 farmers who obtained credit from Chinese rural commercial banks, including demographics, household finance, credit history, and financial literacy. We systemically compare the results of models with and without financial literacy variables, which indicate significant improvement in the predictive accuracy about credit risk when financial literacy factors are included. Our findings confirm that financial literacy is a crucial indicator of farmers' ability to make informed financial decisions, reducing their likelihood of loan default and suggesting its utility as a screening tool or supplementary credit risk assessment variable. This research has profound implications for financial inclusion and credit risk management, indicating that financial institutions can leverage financial literacy data to evaluate farmers’ creditworthiness and design effective financial education programs. This study enriches the literature on credit risk prediction by introducing financial literacy as a predictor of credit default.