Track 15 · Applied Statistics for Professionals · Lesson 4 of 12
Default rates in lending
Estimate default rates with confidence intervals, compare them across segments, and interpret logistic-regression predictors of default without causal over-claims.
A consumer-lending credit committee wants to know how often its personal loans go bad, which borrowers are riskiest, and whether its scorecard ranks risk well. Every answer is a proportion with uncertainty — and every answer comes from loans the lender chose to approve.
A default rate is a proportion: loans that defaulted within a fixed window (here 90+ days past due within 24 months) divided by loans that had the chance to default over that window. Fixing the window matters — a portfolio of mostly new loans always looks safe because its loans have not had time to fail.
Included with a Formula Quest membership
It has 8 graded exercises with feedback, hints and worked solutions. After it you will be able to:
- Estimate a default rate over a fixed window with a confidence interval
- Compare segments with both percentage-point differences and relative risks
- Convert logistic coefficients into odds ratios for meaningful changes and read them as adjusted associations
- State the limits of credit data: approved loans only, no causal claims, period-specific rates