Function 2 · consumer lending

Is the installment book performing as priced, and what does it expect to lose?

Not quite. The two vintages written under the loosened cutoff reached 10.4% cumulative default by twelve months on book, against 4.0% for the vintages before them. Expected loss on the open installment book is $8,409,265, 5.53% of $152.1M exposure, after an observed to expected overlay that adds $3,484,210 because the scorecard under predicts those vintages. Every portfolio limit is within its threshold.

Expected loss, open installment book
$8,409,265
Before overlay $4,925,055
Loss given default, installment
79.6%
One minus recoveries two years after charge off
CECL style lifetime loss (simplified)
$19,170,851
2.71 times twelve month expected loss
Current to 30 days, monthly
1.46%
Cure from 30 days 42.21%

Cumulative default rate by months on book, one line per origination quarter

0%5%10%15%05101520253035Months on book
  • Every other vintage
  • 2024Q4 (loosened)
  • 2025Q1 (loosened)

Cumulative default rate

The two vintages written under the loosened cutoff (2024Q4 and 2025Q1) reached 10.4% cumulative default by twelve months on book, against 4.0% for the vintages before them. Each curve stops where its youngest loan's history ends; nothing is extrapolated.

Source: generated, Harborline installment loans, seed 20260831, as of 2026-08-31.

Monthly transitions between delinquency buckets

Bucket next monthcurrent306090+charged offpaid offcurrent97.1%1.5%1.4%3042.2%3.2%54.6%0.0%6022.2%6.2%71.6%90+9.0%8.4%82.6%

Rows: bucket this month. The further a cell sits from the background, the larger its share, on a log scale.

A current loan rolls to 30 days past due with probability 1.46% a month; once there it rolls on to 60 with probability 54.6% and cures with 42.2%. The diagonal dominates every row, as it must in a book that mostly pays.

Source: generated, Harborline installment loans, seed 20260831, as of 2026-08-31.

Share of loans not yet defaulted, by score band at application

0%20%40%60%80%100%05101520253035Months since origination640 and over600 to 639560 to 599under 560

Share not yet defaulted

Twelve months after origination, 98.6% of loans scored 640 and over have not defaulted, against 82.7% of those under 560. The Cox model's proportional hazards test is published beside it.

Source: generated, booked installment loans, seed 20260831, as of 2026-08-31.

Expected loss by score band

Probability of default from the scorecard with the vintage overlay, loss given default from recoveries two years after charge off, exposure as the balance at the as of date.

Score bandLoansExposurePD (with overlay)LGDExpected lossEL rate
under 560371$4.6M23.69%79.6%$889,78319.40%
560 to 5997,032$71.2M9.82%79.6%$5,787,8658.13%
600 to 6396,701$62.4M3.16%79.6%$1,625,9882.61%
640 and over1,724$14.0M0.91%79.6%$105,6290.75%

Source: generated, Harborline installment loans, seed 20260831, as of 2026-08-31.

Champion and challenger

The challenger ranks 0.005 AUROC better and agrees with the scorecard's ranking at 0.962. It was not promoted. A gradient boosted model cannot state Regulation B reasons directly, so a decline would be explained by a different model from the one that made it, and the gain is smaller than the promotion gate its report states.

Measure, test monthsScorecard (champion)Monotone GBM (challenger)
AUROC0.6620.667
Gini0.3240.334
KS0.2380.250
Brier score0.06260.0624
Expected calibration error0.02860.0279
Stability, development to test (PSI)0.00320.0035
Validation conclusionapproved with conditionsnot approved

Both reports: pd_scorecard and pd_challenger.

A CECL style illustration, simplified

Twelve month default probability with the overlay, stretched to a lifetime by the ratio of cumulative default at thirty months to twelve on the vintages old enough to show both (2.71 times), then scaled by a macro factor of 0.898 for the supervisory baseline's average unemployment of 4.5%. It is labelled simplified because a real allowance adds qualitative factors, reversion and prepayment this does not.

VintageExposure12 month PDLifetime PDMacro factorLifetime loss
2023Q3$0.8M4.24%11.46%0.90$64,287
2023Q4$2.9M3.33%9.01%0.90$186,901
2024Q1$4.2M2.95%7.98%0.90$241,781
2024Q2$5.2M3.51%9.49%0.90$351,072
2024Q3$6.7M4.83%13.07%0.90$625,941
2024Q4$8.8M9.07%24.55%0.90$1,551,275
2025Q1$11.0M10.04%27.17%0.90$2,132,590
2025Q2$12.5M6.39%17.28%0.90$1,543,531
2025Q3$15.4M4.47%12.10%0.90$1,335,543
2025Q4$18.8M6.56%17.74%0.90$2,383,928
2026Q1$21.2M6.86%18.56%0.90$2,808,148
2026Q2$25.6M6.92%18.72%0.90$3,429,950
2026Q3$19.0M6.85%18.54%0.90$2,515,904

Source: generated, Harborline installment loans, seed 20260831, as of 2026-08-31.

What drives time to default

A Cox model on the same loans, concordance 0.704. A proportional hazards p value under the usual threshold means that covariate's effect changes with loan age and its ratio is an average.

CovariateHazard ratio95% low95% highPH test p value
Score, per 20 points0.5950.5700.6210.990
Loosened vintage1.9601.7612.1820.003
Unemployment at origination, per point1.1201.0431.2010.016

Loans still at risk, by month

Score bandAt risk at 06121824
under 56058742829418977
560 to 5999,8937,2144,8052,9451,538
600 to 6398,7826,4644,3682,7931,616
640 and over2,2161,6561,174756446

Source: generated, booked installment loans, seed 20260831, as of 2026-08-31.

Portfolio limits

LimitMeasuredLimitStatus
Loosened vintages' share of installment balances13.03%35.00%✓ within
Installment balances 30 or more days past due2.58%4.00%✓ within
Nonperforming loans to gross loans0.54%1.00%✓ within
Mortgages' share of gross loans72.02%75.00%✓ within

Source: generated, Harborline installment loans, seed 20260831, as of 2026-08-31.

The overlay, vintage by vintage

Observed default over the scorecard's mean predicted default, by origination quarter. The most recent vintages carry 1.78.

VintageObserved over expectedLoans
2023Q31.15399
2023Q40.951,321
2024Q10.811,455
2024Q20.991,494
2024Q31.341,580
2024Q42.201,897
2025Q12.421,998
2025Q21.721,822
2025Q31.191,222

Method and limitations

  • The loan book is generated, with a stated risk structure the pipeline recovers before it is trusted. It is not any real lender's portfolio.
  • Default is 90 days past due or charge off. A curve stops where its youngest loan's history ends; nothing is extrapolated past the as of date.
  • The CECL figure is an illustration of the method on this book, not an allowance. It has no qualitative factors, no reasonable and supportable forecast period with reversion, and no prepayment model.