Published Research

Credit Without Credit Histories

Lessons from U.S. securitized auto loans for consumer lending in African markets. A loan-level empirical study of what actually predicts repayment when formal credit histories are thin or absent.

ABSTRACT

What the paper set out to test.

This paper examines how consumer credit can be efficiently extended in environments where formal credit histories are limited or incomplete. Using loan-level data on U.S. securitized auto loans disclosed under Regulation AB II and accessed via Wharton Research Data Services, it studies repayment performance, delinquency transitions, and default outcomes in a market where lenders combine collateral values, borrower characteristics, and contract design rather than relying solely on credit scores. The analytic sample comprises 239,374 unique loans drawn from 154 trusts issued by 18 sponsors, yielding just over 4 million loan-month observations between November 2016 and December 2024. Estimating discrete-time hazard, logit, and linear-probability models with trust-clustered standard errors, the paper finds that loan-to-value ratio, vehicle age, and contract maturity retain economically and statistically significant predictive power after conditioning on credit scores, and that removing score information from the model reduces explanatory power only modestly. The results are then calibrated to African consumer-lending conditions, showing that conservative loan-to-value limits and faster amortization can materially reduce projected default. The paper positions collateral-backed consumer credit as a distinct informational architecture for extending responsible credit where histories are thin.

0
Unique loans in the analytic sample, drawn from 154 securitization trusts issued by 18 sponsors
4.0M
Loan-month observations, November 2016 through December 2024
0.087→0.059
Pseudo-R² before and after removing credit score entirely from the hazard model
1.84x
Odds ratio: a delinquent loan in negative equity is roughly 1.84 times as likely to roll into default
KEY FINDINGS

What the data actually showed.

Finding
01
Credit scores are not doing as much work as most lenders assume.
Removing credit score entirely from the delinquency-hazard model drops explanatory power only modestly, pseudo-R² falls from 0.0874 to 0.0591. Loan-to-value, vehicle age, contract term, and payment burden each retain independent, statistically significant predictive power (p<0.001) after conditioning on score. Collateral and contract design are not a fallback when scores are missing; they carry real information on their own.
Finding
02
Negative equity is the strongest late-stage trigger of default.
Once a loan is delinquent, being in negative equity, where the amortized balance exceeds depreciated collateral value, raises the odds of rolling into full default by a factor of 1.84. The effect is largest early in a loan's life, when balances are high relative to depreciated collateral, and fades as equity builds, a pattern consistent with borrowers' incentive to walk away from underwater collateral rather than a story about unobserved borrower quality.
Finding
03
Under macroeconomic stress, credit scores lose roughly half their predictive power, right when it matters most.
During the COVID-19 disruption and the 2022–2023 inflation and rate-tightening episode, the marginal explanatory power that credit score adds to the model fell from 0.0416 to 0.0236, close to a halving. Leverage and amortization structure became more predictive over the same period. This is the paper's most consequential finding for African lending markets: asset-based underwriting signals hold up, and become relatively more useful, precisely during the stress periods when thin-file borrowers are hardest to assess.
Finding
04
Calibrated to African lending conditions, underwriting design choices move projected default more than any single input.
Holding the estimated hazard coefficients fixed and shifting the covariate distribution to reflect thinner credit files, higher payment burden, and weaker recovery conditions, the calibration projects an 83.8% twelve-month cumulative default rate in a base thin-file scenario. Imposing a conservative loan-to-value limit brings that down to 49.8%. Faster amortization alone brings it to 56.6%. Elevated income volatility pushes it up to 93.0%. The message for lenders: leverage discipline and amortization speed are levers that can be pulled even when borrower history is unavailable.
CALIBRATION TABLE

Projected 12-month cumulative default by lending scenario.

The estimated hazard applied to alternative covariate profiles. These are calibrations, not forecasts, the policy-relevant signal is the gradient across design choices, not the absolute level.

01
US benchmark, prime borrowers
Reference
13.9% projected 12-month default
02
US benchmark, subprime borrowers
Reference
77.9% projected 12-month default
03
African base case, thin credit file
Calibration
83.8% projected 12-month default
04
African + conservative LTV limit
Calibration
49.8% projected 12-month default
05
African + faster amortization
Calibration
56.6% projected 12-month default
06
African + elevated income volatility
Calibration
93.0% projected 12-month default
METHOD

How the estimates were built.

DATA

Regulation AB II loan-level disclosures

Monthly asset-level performance data on securitized auto loans, accessed through Wharton Research Data Services, deduplicated to one observation per loan-month across amended filings.

MODEL

Discrete-time hazard, logit, and linear probability

A complementary log-log hazard for delinquency onset, a logit for the delinquent-to-default transition, and a two-way fixed-effects linear-probability model to benchmark explained variance, all with trust-clustered standard errors.

IDENTIFICATION

Credit-score attenuation

Performance is estimated with the continuous score, with score collapsed into three coarse bins, and with score excluded entirely, isolating exactly how much predictive power sits outside formal credit histories.

ROBUSTNESS

Competing-risk and stress-period checks

Results hold after treating prepayment as a competing exit rather than censoring it away, and the stress-period interaction isolates how the same underwriting signals behave under the 2020 and 2022–2023 macro shocks.