Working Paper: NBER ID: w14659
Authors: Sumit Agarwal; Paige M. Skiba; Jeremy Tobacman
Abstract: Using a unique dataset matched at the individual level from two administrative sources, we examine household choices between liabilities and assess the informational content of prime and subprime credit scores in the consumer credit market. First, more specifically, we assess consumers' effectiveness at prioritizing use of their lowest-cost credit option. We find that most borrowers from one payday lender who also have a credit card from a major credit card issuer have substantial credit card liquidity on the days they take out their payday loans. This is costly because payday loans have annualized interest rates of at least several hundred percent, though perhaps partly explained by the fact that borrowers have experienced substantial declines in credit card liquidity in the year leading up to the payday loan. Second, we show that FICO scores and Teletrack scores have independent information and are specialized for the types of lending where they are used. Teletrack scores have eight times the predictive power for payday loan default as FICO scores. We also show that prime lenders should value information about their borrowers' subprime activity. Taking out a payday loan predicts nearly a doubling in the probability of serious credit card delinquency over the next year.
Keywords: No keywords provided
JEL Codes: D03; D14; D82
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Taking out a payday loan (G51) | Increases the probability of serious credit card delinquency (G51) |
| Prior credit card liquidity (G21) | Decision to borrow from payday lenders (G51) |
| Teletrack scores (G51) | Predictive power for payday loan defaults (G51) |
| Credit card liquidity (G21) | Taking out a payday loan (G51) |