Author :
Abstract
This study examines the usefulness of alternative data sources, including mobile payment activity, e-commerce spending, psychometric traits, and digital footprints, in assessing creditworthiness when traditional credit bureau scores are unavailable or incomplete. The study addresses the issue of financial exclusion in developing economies by evaluating whether these non-traditional indicators can effectively predict loan repayment behavior. Primary data were collected from 400 respondents in India using a structured questionnaire covering demographic characteristics, financial literacy, digital transaction behavior, psychometric attributes, and loan performance. Logistic regression analysis was employed to estimate the relationship between alternative credit indicators and the probability of default. Descriptive statistics, correlation analysis, and odds ratios were used to interpret the results, and traditional credit scores were incorporated where available for comparison. The results indicate that timely bill payments, higher frequency of UPI transactions, greater conscientiousness, and stronger digital footprint scores are significantly associated with a lower probability of default. In contrast, a larger number of existing loans and higher interest rates increase repayment risk. Although
traditional credit bureau scores improve the predictive accuracy of the model, nearly half of the respondents lacked such scores, highlighting a substantial financial inclusion gap. The findings suggest that alternative data can serve as reliable indicators of creditworthiness and offer financial institutions an effective approach to extending credit access to underserved populations
excluded from conventional credit evaluation systems.
Keywords
Abstract
This study examines the usefulness of alternative data sources, including mobile payment activity, e-commerce spending, psychometric traits, and digital footprints, in assessing creditworthiness when traditional credit bureau scores are unavailable or incomplete. The study addresses the issue of financial exclusion in developing economies by evaluating whether these non-traditional indicators can effectively predict loan repayment behavior. Primary data were collected from 400 respondents in India using a structured questionnaire covering demographic characteristics, financial literacy, digital transaction behavior, psychometric attributes, and loan performance. Logistic regression analysis was employed to estimate the relationship between alternative credit indicators and the probability of default. Descriptive statistics, correlation analysis, and odds ratios were used to interpret the results, and traditional credit scores were incorporated where available for comparison. The results indicate that timely bill payments, higher frequency of UPI transactions, greater conscientiousness, and stronger digital footprint scores are significantly associated with a lower probability of default. In contrast, a larger number of existing loans and higher interest rates increase repayment risk. Although
traditional credit bureau scores improve the predictive accuracy of the model, nearly half of the respondents lacked such scores, highlighting a substantial financial inclusion gap. The findings suggest that alternative data can serve as reliable indicators of creditworthiness and offer financial institutions an effective approach to extending credit access to underserved populations
excluded from conventional credit evaluation systems.
