Who Benefits from Alternative Data for Credit Scoring? Evidence from Peru.

This article examines how retail transaction data can be used as alternative data to construct credit scores for individuals lacking traditional credit histories, focusing on a unique dataset from a Peruvian company that integrates loyalty card transactions, utility payment–based scores, and Peru's administrative credit registry (RCC). The study finds that incorporating retail data substantially improves credit risk prediction and increases credit approval rates for applicants without a credit history—from 16% to between 31% and 48%—while having minimal impact on those with established credit histories. Retail data enhances risk differentiation particularly for no-history applicants by introducing greater heterogeneity in predicted risk scores, effectively serving as a substitute for formal credit registry information. The authors discuss implications for financial inclusion, lender strategies, and fairness concerns, emphasizing the need to balance credit access expansion with risks related to consumer welfare, privacy, and potential biases in alternative data usage.