Investigate a Quarterly Increase in Loan Delinquency
Company: Affirm
Role: Data Analyst
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Technical Screen
# Investigate a Quarterly Increase in Loan Delinquency
A lending business sees its delinquency rate rise in the past quarter. Explain how you would investigate the increase and distinguish measurement changes, portfolio-mix changes, loan seasoning, and deterioration within comparable borrowers or loans. Do not assume a particular delinquency threshold or portfolio dataset; state what you would request and clarify.
### What a Strong Answer Covers
- A precise delinquency event, denominator, observation date, and reporting convention.
- Data and servicing checks before interpreting the change as worsening credit performance.
- Vintage, loan-age, risk, and product comparisons that separate mix from within-group changes.
- Evidence-based hypotheses and a response that acknowledges causal and data limitations.
```hint Compare loans at similar ages
A growing young portfolio and a maturing older portfolio can have different aggregate delinquency even under unchanged credit quality.
```
### Follow-up Questions
- Why can a loan-count delinquency rate and a balance-weighted rate move differently?
- What would you examine if the rise is concentrated in one origination vintage?
Overview: Investigate rising delinquency through rate definitions, reporting checks, loan vintages, seasoning, portfolio mix, and within-segment deterioration.
Investigate a Quarterly Increase in Loan Delinquency
A lending business sees its delinquency rate rise in the past quarter. Explain how you would investigate the increase and distinguish measurement changes, portfolio-mix changes, loan seasoning, and deterioration within comparable borrowers or loans. Do not assume a particular delinquency threshold or portfolio dataset; state what you would request and clarify.
What a Strong Answer Covers Guidance
A precise delinquency event, denominator, observation date, and reporting convention.
Data and servicing checks before interpreting the change as worsening credit performance.
Vintage, loan-age, risk, and product comparisons that separate mix from within-group changes.
Evidence-based hypotheses and a response that acknowledges causal and data limitations.
Follow-up Questions Guidance
Why can a loan-count delinquency rate and a balance-weighted rate move differently?
What would you examine if the rise is concentrated in one origination vintage?