First payment default meaning, in the words a lender's own staff use: an early payment default (EPD) is a loan that becomes 30 or more days past due on any of its first three payments. The borrower never made the habit of paying, or could not, or was never going to — and the loan says so within weeks of funding, long before a vintage chart would flag anything. Botminds' Early Payment Default Review exists for exactly this moment: it opens a case the day a loan crosses that line, re-reads the whole file, and proposes why the loan failed — fraud, a credit miss, an onboarding slip or hardship — for an analyst to confirm. That makes the EPD not just a count but an early-warning system, if the definition is written down and the cause is found.
The worked examples on this page come from a fictional sample bank running the solution — Lanternfield Bank, whose August vintage of 4,812 personal loans produced 87 early defaults. Every figure drawn from that vintage is sample data on invented people and partners, not a measured result, and it is labelled as such where it appears. The research figures come from regulators and peer-reviewed studies, linked where they appear.
The EPD definition in two real settings, and the scale it can hide
60days
late on the first, second or third payment — the GSE servicing-released definition of an “EPD Mortgage Loan”
30days
the early-default line in this solution’s default setting: any of the first three payments, configurable per lender
$212B
in identity-related suspicious activity in FinCEN’s 2021 BSA data — one thing an early default can be the first visible trace of
What is an early payment default (EPD)? The definition lenders actually use
An early payment default is a delinquency, not a loss and not a charge-off: the borrower has missed enough of the first payments, by enough days, for the loan to count as an early default under the lender's own definition. Three clauses make the definition, and each one is a choice that has to be written down before the count means anything.
Anatomy of the EPD definition: three clauses, each one a decision
| Clause | What it fixes | The GSE servicing-released wording | The sample solution’s setting |
|---|---|---|---|
| Which payment | Does the first missed payment count, or any of the first three? | “The first, second or third monthly payment due Servicer” | Any of the first 3 payments (options include the first payment only) |
| How many days late | 30 days? 60? More? | “Becomes sixty (60) or more days delinquent” | 30+ days past due (option: 60+ days in the first 6 months) |
| How it is measured | Days past due from the contractual due date, per payment | Each such loan is an “EPD Mortgage Loan” with reimbursement consequences for the seller | Case opened automatically on the servicing feed, on the definition in Settings |
Notice what the definition does not say: why. An EPD clause names a payment and a day count; it says nothing about fraud, credit quality or errors. That silence is the whole problem this series keeps returning to — the definition that triggers the review is not the one that explains the result.
Where the “30 days” and “60 days” thresholds come from, and why they differ by lender
The most-cited EPD threshold in US mortgage lending is 60 days, and it comes from a specific place: the GSE servicing-released agreements. Under Fannie Mae's Exhibit E-2-05, where a servicer has purchased the servicing rights, if "the first, second or third monthly payment due Servicer becomes sixty (60) or more days delinquent", the loan is an "EPD Mortgage Loan" and the seller who delivered it owes reimbursement (Selling Guide, Exhibit E-2-05). That is a contract term about who bears the cost of a bad delivery — written for one product, one channel and one set of parties.
A consumer lender running its own book typically draws the line earlier, at 30 days past due on any of the first three payments, because the point of watching early defaults is to act while there is still something to act on: a ring to close, a rule to fix, a borrower to re-enrol. Some definitions go further out — 60 or more days past due within the first six months — to catch loans that pay once or twice and then stop. None of these is "the" definition. Each is a setting, and Early Payment Default Review treats it as one: the early-default line, the review clock and the risk appetite are all configured per lender, versioned, and recorded on the audit trail. What matters is that the line is written down before the vintage is counted, so that this month's 1.81% means the same thing as last month's 1.06%.
Early payment default vs first payment default vs early payoff — the differences that matter
Three terms sound alike and get used interchangeably in hallway conversation; only two of them are problems. A first payment default is the narrow case of an EPD where the very first due payment is missed — the loudest version, because a borrower in genuine trouble almost always pays once before slipping. An early payoff is not a default at all: the borrower pays the loan off ahead of schedule, which costs the lender expected interest but says nothing about credit quality. Mixing the three up skews the count a lender reports to its own committee.
EPD, first payment default and early payoff: same weeks, opposite meanings
What it is: Any of the first payments goes 30+ (or 60+) days past due, under the lender’s written definition.
What it tells the lender: The loudest early signal: fraud, a credit miss or an onboarding slip, all still invisible at this point.
What it is: The first due payment is missed — the narrowest case of an EPD.
What it tells the lender: The loudest version of the signal; a ring’s loans often never pay at all.
What it is: The borrower repays early, in full, with no delinquency.
What it tells the lender: An interest-cost event, not a credit event. It never belongs in the EPD count.
The distinction pays off in practice. On the sample vintage, every one of the 19 ring loans missed every payment from the first due date — first payment defaults in the strictest sense — while the 13 borrowers whose autopay the bank's own app never saved missed a payment they had authorised. Same tree, different meaning; what fraud looks like next to a credit miss is the next page's subject.
What makes EPD “the loudest signal a lender gets”
A loan that misses one of its first payments is telling you, within weeks, one of three things: the borrower was never going to pay, the credit policy let in a borrower it should not have, or the bank broke something between approval and the first due date. No other early indicator compresses so much into so little time. A FICO score is a forecast; an EPD is an event. And the event arrives while the evidence is still fresh — the application, the verification results, the device and account data, the decision record, all weeks old rather than months.
The research says the signal is worth listening to at face value. In a National Bureau of Economic Research study of the non-agency mortgage market, Piskorski, Seru and Witkin compared what intermediaries disclosed about loans at sale with the loans' actual characteristics in credit-bureau data, and found that "about one out of every ten loans has one of these misrepresentations" — misreported occupancy status or misreported second liens (NBER Working Paper 18843). More to the point for an early-default desk: "ex post delinquencies of such loans are more than 60% higher when compared with otherwise similar loans." A loan whose file was flattered defaults far more often than an honest loan with the same headline numbers. The early default is often the first visible crack in a file that looked fine at approval.
The first 90 days of an early default: from a missed payment to a confirmed cause
- First payment missedday 1
- 30+ days past due: the EPD linedays 30–31
- File re-read and cause proposed5 business days
- Cause confirmed, fix routedby ~day 45
- Charge-off risk without actionthe ring stays open
Why a vintage default rate cannot tell you what caused it
A vintage rate is a numerator over a denominator: 87 early defaults over 4,812 August loans is 1.81%, against 1.06% for the six vintages before. That number is necessary and almost useless on its own, because it averages three very different failures into one. In the sample vintage, the 1.81% decomposes into a fraud ring of 19 loans arriving through one partner sub-source, 21 credit misses admitted by a July rule change, 13 borrowers an app defect made late, and a residual of honest hardship. Strip out the 19 ring loans and the 16 onboarding slips the bank caused itself, and 52 early defaults remain: 1.08% of the vintage, inside the 1.25% appetite. The vintage chart cannot see that decomposition; only the files can.
This is why an EPD rate, reported alone, tends to produce the wrong meeting. It rises, the committee asks "credit or fraud?", and the honest answer from the number itself is "yes". The rate's proper job is to trigger the review, not to substitute for it. FinCEN's scale figure gives a reason to take the trigger seriously: in the agency's 2021 Bank Secrecy Act data, "approximately 1.6 million, or 42% of around 3.8 million total BSA reports, equivalent to $212 billion in suspicious activity, related to identity" (FinCEN, Identity-Related Suspicious Activity, January 2024). Some of that activity first surfaces to a lender as a loan that never paid — one row in someone's vintage chart.
Does an EPD always mean the loan gets sold back to an investor?
No. Review and repurchase are different events, and conflating them makes desks either complacent or panicky. What an EPD reliably triggers on the mortgage side is a quality-control look: "Fannie Mae may review mortgage loans it has purchased or securitized (including those with early payment defaults, those that have been foreclosed, as well as any other mortgage loan)" (Servicing Guide, A2-4-01). The same guide describes how the review population is built — a statistically valid random sample, "augmented with targeted, discretionary sampling" — which is the regulated version of a simple idea: early defaults get looked at more closely than the average loan, because they are more likely to have something wrong with the file.
What follows a review is a separate question with a separate standard of proof. The academic measure of how much misrepresentation was circulating gives the scale the contract fights are about: the NBER authors estimated that misrepresentation on just the two dimensions they could quantify "could result in forced repurchases of mortgages by intermediaries up to $160 billion" (NBER Working Paper 18843). For a lender's own book — especially unsecured consumer lending, where there is no investor to sell back to — the practical consequence of an EPD is earlier and quieter: find the cause while the evidence is fresh, and route the fix to the team that owns it. What those fixes cost, and what they save, is worked through in what an unexplained early default actually costs.
How is an EPD rate calculated, and what counts as “early”?
The calculation is a fraction with two judgment calls baked in. The numerator is the count of loans that crossed the EPD line — under the written definition, so that the count is reproducible. The denominator is the vintage's funded loans. On the sample bank's August vintage: 87 early defaults ÷ 4,812 funded loans = 1.81%, read at the first-payment milestone against a 1.25% risk appetite. The same arithmetic read at 60 days, or on a rolling quarter rather than a calendar month, would give a different, equally honest number; the definition decides which one the committee sees.
"Early" itself is relative to the loan's own clock, not the calendar: any of the first three payments is the standard window, because a borrower who pays three times has usually formed the habit or shown the intent. Product choices matter here — a lender funding through lead partners may want the partner-level read as well, since one sub-source can carry 5.14% against a 3.0% contract trigger while the book as a whole passes. A rate that cannot be cut by channel, partner and exception is a rate that hides its own explanation.
What should a lender do in the first 30 days after an early default?
The first 30 days after a loan crosses the EPD line are when the file is still warm: the application, the verification results, the decision record, the fraud checks and every payment attempt are all in reach, and the people who touched the loan remember it. The OCC's Retail Lending handbook states the general discipline: "Well-run banks generally handle exceptions in a structured and disciplined manner to promote sound operations as well as compliance with applicable laws and regulations, including those related to consumer protection and fair lending" (Comptroller's Handbook, Retail Lending). An early default is an exception by any reading, and the structured response is to treat it as a question with an answer in the file.
Before a loan counts as “explained”, confirm each of these
- The file is complete
Application, documents, verification results, decision record, fraud checks, channel and every payment attempt — assembled before anyone re-reads, so the read is of the file and not of memory.
- The definition that fired is the written one
Which payment, how many days, measured how. If the case opened on a different line than the policy states, the count is wrong before the cause is found.
- The cause names the evidence, not the vibe
Each proposed cause points to the document, record or check it rests on — the pay stub, the device match, the rule version, the missing debit authorisation.
- The runner-up cause was considered
Fraud over credit is the common wrong call; the file should say what would distinguish them.
- Someone with authority confirms it
Agents propose; an analyst confirms or changes every cause. Nothing counts as explained on a proposal alone.
In the sample run, that discipline is what turns 87 unexplained loans into a signed report in two days — 79 causes confirmed together at the configured threshold, the remaining 8 read one by one. The definition started the clock; the confirmation made the number mean something. To see the whole series from the question of meaning to the question of money, start from Early Payment Default Review.
Questions people ask
What does first payment default mean?
A loan whose very first due payment is missed — the narrowest form of an early payment default. Lenders watch it closely because borrowers in genuine difficulty usually manage one payment before slipping; a loan that never pays once is a different kind of warning.
What is the early payment default definition Fannie Mae uses?
In the servicing-released context, Fannie Mae's Exhibit E-2-05 defines an "EPD Mortgage Loan" as one where "the first, second or third monthly payment due Servicer becomes sixty (60) or more days delinquent" — a contract term about reimbursement for bad deliveries, not a universal definition for every lender or product.
How is a first payment default rate calculated?
Loans that crossed the EPD line divided by loans funded in the vintage, under the written definition. On the sample bank's August vintage, 87 of 4,812 loans is 1.81%, read against a 1.25% appetite. The rate is only reproducible if the definition — which payment, how many days — is fixed first.
Is an early payment default the same as a charge-off?
No. An EPD is a delinquency event measured in days past due; a charge-off is an accounting decision taken later, when the lender concludes the debt is uncollectible. Most EPDs are resolved, cured or explained long before any charge-off question arises.
Does an early payment default always mean fraud?
No — and treating it as if it does is the most common expensive mistake. On the sample vintage the early defaults split four ways: fraud, a credit miss, an onboarding slip the bank caused itself, and genuine hardship. The rate alone cannot say which; only reading the file can.
Can an early payment default hurt a borrower who did nothing wrong?
Yes, and that is the quietest cost. In the sample vintage, 13 borrowers chose autopay and the bank's app never saved the instruction; they were recorded late through no fault of their own. A review that finds the lender's own error can waive the fee and correct the late mark with the bureaus.