Static pool calculator: cumulative default by cohort
Paste a loan tape summarised by cohort and month. You get each cohort’s cumulative default curve by months on book, the same curves a fund draws before it lends, and the average weighted by what each cohort originated.
The file is read in your browser and is not sent anywhere.
The denominator never moves, so a growing book cannot hide its losses
Each cohort is the loans originated in one month. Its curve divides the principal that has ever been 90 or more days past due by the principal the cohort originated, and reads that share at every month the cohort was observed. The denominator is fixed on the day the cohort is closed, which is the point of a static pool: new lending can grow the book, but it cannot dilute a single cohort’s losses.
Months a cohort was not observed stay empty. The calculator does not fill a gap or stretch a young cohort to the length of an old one, because an extrapolated curve is a forecast, and a fund reading the tape wants to see where the data stops.
For what the curves are for and how a rating agency reads them, see what a static pool is and why a fund asks for one.
Questions this raises
- What file does the static pool calculator need?
- A CSV with four columns: cohort (the month the loans were originated, like 2024-01), originated (the principal that cohort lent), month (the month observed) and defaulted (principal of that cohort that has ever been 90 or more days past due by then, written-off amounts included). One row per cohort per month observed. Commas or semicolons both work, and the column order does not matter.
- What counts as defaulted?
- Principal that has reached 90 days past due at any point up to the month observed, including what was later written off. Loans that cured stay counted, because the curve is cumulative: it should rise or stay flat. If a cohort's curve falls, the calculator still draws it and says where, since a fall usually means cured loans were netted out of the tape.
- Why is the average weighted by what each cohort originated?
- Because a large cohort and a small one do not carry the same risk to the book. Each month on book is averaged over the cohorts actually observed at that month, weighted by their originated principal. A young cohort simply drops out of the months it has not reached yet, and nothing is extrapolated.
- Is my loan tape uploaded anywhere?
- No. The file is read and calculated in your browser, and the page sends no data from it. If you accepted analytics, we count that a calculation ran and how many rows and cohorts it had, nothing from the rows themselves.