Who Benefits from Alternative Data for Credit Scoring? Evidence from Pe

The World Bank reports that as many as 1.4 billion individuals worldwide remain unbanked. These customers typically do not have credit scores, which results in a lack of access to credit. In this paper, we document how retail data can be used to construct a credit score for these customers and in-turn offer credit. Our study relies on unique data that was acquired in partnership with a conglomerate in Peru. We merge data from the Peruvian financial system, which provides a detailed record of every citizens financial history, with customer loyalty data, and credit card payment data. We use these data sources to construct credit scores for customers with and without a financial history. Our simulations show that approval rates increase from 15% to between 31% and 47% for customers without a financial history. For customers with a financial history, there is very little change in the approval rate of 87%. We explore why alternative data may benefit customers without a financial history. We conclude with a discussion of implications for consumers, firms and policy makers if this type of credit scoring methodology is adopted.