Evaluation of Credit Scoring Models in Consumer Lending by Commercial Banks in Anambra State

Credit scoring models are increasingly adopted by Nigerian commercial banks to improve consumer lending decisions, yet their application and effectiveness in South East markets such as Anambra State require empirical assessment. This study evaluates the credit scoring models used in consumer lending by commercial banks in Anambra State, examining their predictive accuracy, fairness, and practical limitations. A mixed-methods design was adopted, combining structured interviews with 30 credit risk officers from six commercial banks in Onitsha, Awka, and Nnewi with documentary analysis of credit scoring framework documents. Predictive accuracy was measured by the Gini coefficient and Kolmogorov-Smirnov statistics derived from historical loan performance data. Fairness was assessed by examining differential scoring outcomes across gender and occupation categories. Thematic analysis was applied to qualitative interview data. Results indicated that the Gini coefficients for deployed scoring models ranged from 0.34 to 0.51, indicating moderate to good discriminatory power. However, scoring models exhibited systematic underscoring of informal sector workers, who constitute 65% of loan applicants in Anambra State. Female applicants faced a 12% lower average credit score under models weighted toward formal employment income, raising fairness concerns. The study concludes that credit scoring models in Anambra State commercial banks lack contextual calibration for informal economy participants and recommends recalibration using alternative data sources such as mobile transaction histories and utility payment records.

Keywords: credit scoring, consumer lending, commercial banks, Anambra State, informal economy