Upload the loan files your core already exports. Vintage measures what your book has actually done, and shows exactly where every number came from.
Your core produces the record every month: balances, rates, charge-offs, payoffs, the full history of what every loan did. Years of it, sitting in exports that nobody can turn into one picture.
So loss and prepayment assumptions get borrowed from peer data and national proxies. New loans get priced off competitor rate sheets and judgment. Then the exam comes, and the request is the one it always is: support these assumptions with your own experience.
That experience is in the exports. Vintage is what reads them.
Cumulative loss curves by loan age, built vintage by vintage from your own charge-offs. Exposure-weighted, with a lifetime loss rate read at your book's own term. Where the curve runs past your data it is marked as a projection, never blended into the measured figure.
Why it mattersA CECL input you measured yourself, and loss experience you can put in front of an examiner instead of a peer proxy.
Monthly SMM and annualized CPR measured from your own payoffs and prepayments, a remaining-balance curve against the contractual schedule, and the expected life of the book.
Why it mattersALM and cash-flow assumptions that reflect how your borrowers actually behave, not a national average.
The rate a new loan needs to clear your hurdles, built up line by line: funding, expected credit loss, operating cost, fees, and your ROA or ROE target. Enter a proposed rate to see the gap in basis points, then follow the loan month by month to break-even.
Why it mattersPricing you can defend at ALCO, with your own loss and prepayment experience standing behind it rather than a competitor's rate sheet.
Export loan-level files from your core: monthly snapshots, originations, transactions. CSV or Excel, in whatever shape they come out. There is no template to match, and personal information never leaves your browser.
Vintage recognizes your columns, asks about the ones it cannot, and joins every upload into a single loan-level history. The same loan is recognized across files even when the format drifts between years.
Loss curves, prepayment speeds, and pricing measured from your own book. Slice by origination year, product, balance, or any field you uploaded, and every number follows the slice.
Snapshots, origination files, and transaction files, in any combination, with whatever columns your core exports. CSV or Excel, nothing renamed or reformatted first.
Fiserv, Jack Henry, Symitar, Corelation, FIS. There is no integration to build and nothing to install, because Vintage reads the file you already pull, in the format it already comes in.
Vintage recognizes standard fields by name and asks you to confirm the rest. Your status codes are classified once, in your own words, and remembered for every upload after that.
An upload does not have to be complete to be useful. Before anything commits, Vintage tells you what your data supports today and names the exact fields that would extend it.
When someone asks where a number came from, you can show them. Vintage uses historical cohort averaging: a life-table of per-age rates, weighted by exposure. No machine learning, no fitted distributions, no hidden parameters, so a banker or an examiner can reproduce any figure on the screen from the same files.
2018 origination cohort, exposure-weighted
The solid curve runs where your charge-off history reaches. Every point is exposure-weighted.
The bars are the loans behind each age. Fewer remain as the book seasons, and the chart shows it.
Roughly 95% confidence over your data. It widens as fewer loans remain to measure.
Past your history the curve is a labeled projection, drawn apart and never blended into the observed figure.
Sensitive columns are removed and loan and tax identifiers are replaced with one-way scrambled values before anything is uploaded. Vintage's servers never receive raw personal information.
Bring the files you already have. Vintage will show you what your book can measure today, and name the fields that would extend it.