Traditional credit scoring models predict the risk associated with a new loan using a potential borrower’s history with previous loans from data in credit bureau files. In recent years, lenders have tested alternative data such as rent and utility bill payments as supplemental information when a consumer’s credit file is too thin to be useful for automated scoring. That practice is gaining acceptance even if there is no consensus regarding how predictive bill payment history is for repayment of a loan. These facts have contributed to a growing belief that risk management can be aided by de...