A loan vintage is every loan a lender originated during a defined period — a month, a quarter, or a year — grouped together so their performance can be tracked as a single cohort. The label is fixed to the origination date and never changes: a loan booked in January 2024 belongs to the Q1 2024 vintage for the rest of its life, even as the loan itself ages month by month. That separation between cohort and age is what makes vintage analysis useful, because it lets analysts line up two cohorts at the same point in their life cycle and ask which one is performing better, without the noise of looking at a whole portfolio at once.
Every loan in a vintage faced the same underwriting rules, the same rate environment, and roughly the same economy at the moment the borrower signed. That shared starting line is the whole point. If you want to know whether a lender’s credit standards worked, you have to compare loans that were judged by the same standards.
How Vintages Get Grouped and Sub-Segmented
The mechanics are simple. Each loan gets tagged with its origination date, and performance data rolls up by that shared date. From there, most serious analysis breaks the cohort into sub-segments based on risk characteristics.
Common stratification factors include credit-score bands, loan-to-value ratios, debt-to-income ratios, geographic region, collateral type, and loan term. The OCC’s guidance on allowance methodologies notes that assets within a vintage “can be sub-segmented by a secondary risk characteristic, such as risk rating,” with the loss rate for each sub-segment tracked separately against the original balance.1Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptroller’s Handbook
Some risk factors compound. The FHFA Office of Inspector General has documented that a mortgage with both a high DTI and a high LTV carries what the industry calls “risk layers,” and layered risk is treated as materially more dangerous than any single factor in isolation.2Federal Housing Finance Agency Office of Inspector General. An Overview of Enterprise Debt-to-Income Ratios Stratification within a vintage is how analysts detect whether a lender quietly stacked those layers during a particular origination window.
What Vintage Analysis Reveals
The core purpose is separating cause from effect. When losses climb across a portfolio, the obvious question is whether the economy weakened or the lender’s standards slipped. Vintage analysis answers it by comparing cohorts born under different conditions. If the 2024 vintage is outperforming the 2023 vintage at the same age, the lender probably tightened underwriting between those periods. If every recent vintage is deteriorating at the same pace, the problem is more likely macroeconomic.
Vintage curves also work as an early-warning system. A cohort tracking above its expected loss curve in the first twelve months signals trouble well before those losses hit the income statement at full weight. When a vintage deviates from the benchmark, the follow-up work begins: was the deviation driven by a specific geography, a credit-score band, or a product feature like interest-only payments?
There’s a feedback loop into risk modeling too. If a model predicted a 2 percent lifetime loss rate for a cohort and the actual loss curve is tracking toward 4 percent at the halfway mark, the model’s assumptions about that borrower segment were wrong. The data feeds directly back into recalibration, capital planning, and loss provisioning.
The Metrics Tracked Across a Vintage
A handful of measures do most of the work. They’re usually plotted as curves against months since origination, and the shape of each curve tells a piece of the story.
Cumulative Net Loss Rate
This is the headline number: total dollar losses the cohort has experienced, net of recoveries, divided by the original principal balance. The denominator stays fixed at the original balance, so the rate only moves in one direction. The OCC describes the calculation as tracking “net charge-offs of each vintage divided by the original principal balance, which remains the denominator in each calculation,” producing “a cumulative life-of-loan loss rate based on historic averages.”1Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptroller’s Handbook Two vintages with identical original balances but different cumulative net loss rates at the same age are easy to rank. Lower is better.
Delinquency Rates
Delinquency gives an earlier read on stress than waiting for actual charge-offs. Rates are segmented by severity: 30 days past due, 60 days past due, and 90-plus days past due. The CFPB tracks the 30-to-89-day delinquency rate as “a measure of early stage delinquencies and an early indicator of the mortgage market’s overall health.”3Consumer Financial Protection Bureau. Mortgages 30-89 Days Delinquent A sharp rise in 60-day delinquencies within a vintage usually foreshadows a spike in 90-day delinquencies a few months later.
Roll Rates
Roll-rate analysis quantifies loan movement between delinquency buckets from one period to the next. A roll forward measures the share of loans in a given bucket that moved to a worse status; a roll backward measures the share that cured. A vintage where 30-day delinquent loans roll forward at 25 percent is in a very different place than one rolling forward at 10 percent, even if both vintages currently show the same headline delinquency rate. Roll rates reveal the momentum behind the numbers.
Prepayment Speed
Prepayment speed measures how fast borrowers pay off loans ahead of schedule, through refinancing, sale of collateral, or extra payments. The standard metric is the conditional prepayment rate (CPR), expressed as an annualized percentage of remaining principal expected to prepay. A higher CPR means less interest income for the lender or investor. It also caps late-stage credit risk, because a borrower who refinances at month 18 cannot default at month 36. Vintages originated during falling-rate environments tend to show elevated CPRs, which compresses the window during which losses can accumulate.
Reading a Loss Curve
When cumulative net loss is plotted against months since origination for multiple vintages on the same chart, the curves fan out over time. Each starts at zero, rises as defaults accumulate, and eventually flattens as surviving loans settle into a stable repayment pattern. The point where the curve begins to flatten is called the seasoning point, and it marks the transition from active loss realization to a more stable phase.
For many consumer loan products, the steepest part of the curve falls within the first 18 to 24 months. After that, the rate of new losses slows. Where a vintage seasons depends on the product: auto loans season faster than 30-year mortgages, and unsecured personal loans often show their losses earlier than either. A vintage with 24 months of history has revealed most of its credit risk. A vintage with only six months is still in the acceleration phase, and projections carry real uncertainty.
The insight comes from comparing curves. If the Q1 2025 curve sits above the Q1 2024 curve at every point, the later vintage is weaker. If two curves start on similar paths but diverge after month 12, something changed in the economy or the borrower mix that only became visible after the initial honeymoon period. That divergence pattern is often the first sign that a downturn is hitting a portfolio unevenly across cohorts.
What Drives Differences Between Vintages
Performance gaps come from two sources: the conditions loans were born into, and the decisions the lender made when booking them. Teasing those apart is the central challenge.
External Forces
The rate environment at origination shapes a vintage in more than one way. Adjustable-rate mortgages originated in a low-rate period face payment shock if rates rise sharply, pushing defaults higher. Fixed-rate loans originated at high rates may see elevated prepayments if rates later fall. General economic conditions, especially unemployment, exert direct pressure on a cohort’s ability to perform. Loans booked just before a spike in job losses will almost always show higher cumulative losses than loans booked during a strong labor market.
Regulatory changes can reset the baseline for an entire generation of vintages. The CFPB’s Ability-to-Repay rule, at 12 CFR 1026.43, requires mortgage lenders to make a reasonable, good-faith determination that a borrower can actually repay before closing, evaluating income, employment, monthly debt obligations, debt-to-income ratio, and credit history.4eCFR. 12 CFR 1026.43 – Minimum Standards for Transactions Secured by a Dwelling Vintages originated after that rule took effect should, in theory, carry lower default risk than pre-rule vintages.
Internal Decisions
Changes in a lender’s own underwriting are the most direct cause of performance shifts. Lowering the minimum credit score, raising the maximum DTI, or allowing a higher LTV all introduce measurably more risk into the cohort. Fannie Mae and Freddie Mac have both acknowledged that as DTI increases, “the level of risk also tends to increase,” and that higher DTI “increases the probability a borrower may be unable to meet all their obligations at some point in the future.”2Federal Housing Finance Agency Office of Inspector General. An Overview of Enterprise Debt-to-Income Ratios
Shifts in target borrower demographics leave fingerprints too. A lender expanding into a new geographic market or moving down the credit-score ladder is building a vintage with a different risk profile even if every other parameter holds. Product design changes matter as well: longer terms, higher advance rates, or interest-only structures all push expected losses higher. Good vintage analysis separates those internal choices from the external backdrop, so the lender can tell whether a weak cohort was bad luck or bad judgment.
Where Vintage Analysis Shows Up in Regulation
Vintage analysis is not only an internal tool. It sits inside two significant regulatory frameworks.
The Current Expected Credit Losses (CECL) standard under FASB ASC Topic 326 requires lenders to estimate lifetime expected credit losses at origination rather than waiting for losses to become probable. The 2023 interagency policy statement from the OCC, Federal Reserve, FDIC, and NCUA explicitly identifies “vintage analysis” as one of the acceptable loss-rate methods for estimating expected credit losses under CECL, alongside the weighted-average remaining maturity method, the probability-of-default/loss-given-default method, the roll-rate method, and others. The same statement notes that vintage itself is one of the risk characteristics institutions may use to segment financial assets for collective loss evaluation.5Federal Register. Interagency Policy Statement on Allowances for Credit Losses Revised April 2023 The OCC’s Comptroller’s Handbook describes the vintage method as “a closed pool method focusing on the origination period” that “can reflect changes in underwriting, regulations, or economic conditions during a particular year, quarter, month, or another length of time,” and notes it “is best suited for portfolios that have large data sets and predictable loss patterns.”1Office of the Comptroller of the Currency. Allowances for Credit Losses – Comptroller’s Handbook
The other place vintage data lives is in asset-backed securities disclosures, where it’s usually called “static pool” data. The SEC’s Regulation AB, at Item 1105, requires ABS issuers to provide static pool information on delinquencies, cumulative losses, and prepayments for prior securitized pools. If the sponsor has less than three years of securitization experience for that asset type, the rule calls for the same data organized “by vintage origination years.” The disclosure must cover at least five years of prior pools or vintages (or as long as the sponsor has been active, if shorter), updated to within 135 days of the prospectus date.6eCFR. 17 CFR 229.1105 – Item 1105 Static Pool Information Item 1105 also requires a narrative description of how the static pool differs from the pool backing the current offering, including differences in underwriting criteria, loan terms, and risk tolerances. For anyone evaluating an ABS deal, that section of the prospectus is where vintage analysis lives, and it is often the single most useful tool for judging how the sponsor’s recent originations stack up against its track record.