Buy Now, Pay Later and the Phantom-Debt Problem

Buy Now, Pay Later didn't just add a checkout button — it created a parallel layer of consumer debt that, for much of its rise, was invisible to the credit system that prices nearly everything. The debate over BNPL is usually framed around its size. The more important question is its visibility: what happens when a fast-growing slice of household borrowing can't be seen by the models that decide who gets a mortgage, a car loan, or a card.

The core thesis

The central argument of this report is that BNPL's defining feature is not its growth rate or its default rate but its invisibility to the credit reporting system. A credit score works because it aggregates a borrower's obligations into one picture; BNPL, for much of its history, sat outside that picture entirely. That creates a specific and underappreciated risk: not that BNPL is inherently dangerous, but that debt the system cannot see is debt the system cannot price. When a meaningful and fragmented layer of borrowing is invisible, every downstream risk model — for mortgages, auto loans, and cards — is working from an incomplete portrait.

This reframing matters because it separates two questions that are usually conflated. Is BNPL a systemic threat today? Probably not, given its current scale. Is BNPL a source of hidden, mispriced risk and a symptom of deeper strain? Yes — and that is the more useful lens for anyone who lends, invests, or regulates.

The danger of BNPL isn't mainly its size or its defaults. It's that a fast-growing, fragmented layer of consumer debt has been invisible to the models that price all other credit. You cannot price what you cannot see.

What BNPL actually is

Buy Now, Pay Later is point-of-sale installment financing: at checkout, a shopper splits a purchase into several payments rather than paying all at once. The dominant form is the "pay-in-4" plan — four installments over roughly six weeks, frequently interest-free — offered by providers like Affirm, Klarna, and Afterpay and, increasingly, embedded directly into the checkout of major retailers both online and in stores. Its appeal is straightforward: it converts an uncomfortable price into a manageable schedule, without the interest of a revolving card, and it's approved in seconds. For merchants, it lifts conversion and average order value, which is why adoption has spread so quickly.

The scale — read carefully

Here discipline is required, because headline figures for BNPL vary wildly depending on what's being measured — global versus US, gross merchandise volume versus outstanding debt, all providers versus pay-in-4 only. The most careful US estimate comes from the Federal Reserve Bank of Richmond, which put the real transaction value of BNPL at roughly $70 billion in 2025 — about 1.1% of total credit card spending — growing around 20% per year since 2021. Global gross-merchandise-volume estimates run far higher, into the hundreds of billions, but those measure a different, broader thing.

The honest read: in the United States, BNPL is a meaningful and fast-growing channel, but still small relative to credit cards. That scale is precisely why its direct financial-stability footprint looks limited today — and precisely why the visibility problem, not the size, is the story worth watching.

The phantom-debt problem

This is the heart of the matter. Unlike credit cards, which report balances and payment history to the bureaus, the majority of BNPL loans — especially pay-in-4 — have historically gone unreported. The Richmond Fed and others have adopted a vivid term for the result: phantom debt. Because these obligations don't appear on credit reports, a lender assessing a mortgage, an auto loan, or a new card cannot see them — and therefore cannot accurately gauge the borrower's true debt load.

Two features make this worse than a simple data gap. First, loan stacking: the CFPB has found that a large share of BNPL users — roughly two-thirds in its research — carry multiple BNPL loans at once, often across different providers, with no centralized system tracking the total. A borrower who looks unlevered on paper may be juggling several simultaneous plans. Second, the invisibility runs in one direction that matters most: it hides liabilities. An underwriter approving a car loan may have no idea the applicant already owes across several BNPL plans — making the lending decision, in the words of practitioners, a matter of "flying blind." Multiply that across millions of transactions and the credit system's core function — aggregating a borrower's obligations into one honest picture — is quietly undermined.

The slow move to reporting

The obvious fix is to bring BNPL into the credit reporting system — and that is happening, but slowly and unevenly. Some providers, including Affirm and Klarna, have begun reporting to bureaus such as Experian and TransUnion, and FICO has introduced ways to incorporate BNPL data into scores. In principle, this closes the gap: on-time BNPL payments could help a borrower's score (a real benefit for thin-file consumers), while missed payments would appear like any other delinquency.

But integration is fragmented and incomplete. Reporting practices differ by provider and bureau; the short, unusual structure of pay-in-4 loans doesn't fit neatly into scoring models built for monthly installment debt; and providers have shown genuine ambivalence about handing their proprietary repayment data to the bureaus, wary of what they give away competitively. The result is a patchwork in which some BNPL activity is visible and much still isn't — which, from a risk perspective, can be the worst of both worlds, because partial data can mislead as easily as no data.

Who uses it — and the risk

BNPL's user base is the key to understanding its risk profile. Usage skews younger and lower-income — precisely the households with thinner cash cushions and greater sensitivity to economic downturns, and precisely the cohort showing the most strain elsewhere in consumer credit. This overlaps directly with the bifurcation we document in the consumer credit cycle: the same stressed lower cohort driving the K-shaped divergence is the one leaning most on BNPL.

The delinquency data must be read with the same care as the size data. Surveys point to a rising share of BNPL users missing payments — one industry survey found roughly 47% reported a late payment in the past year, up from prior years, while Federal Reserve survey work has put the figure lower, around a quarter. Methodologies differ, so the level is uncertain, but the direction is a warning. Compounding this, loss rates on standard pay-in-4 products have historically stayed under 1%, but the loans are so short — often six weeks or less — that delinquencies can spike quickly during stress, giving servicers little time to react. Short duration is a feature in good times and a vulnerability in bad ones.

The private-credit connection

BNPL debt doesn't stay with the providers — and this is where the phantom-debt problem meets the institutional capital markets. Private credit's appetite for consumer receivables, including BNPL, has surged: purchases of consumer loans reportedly jumped roughly 14-fold to around $136 billion in 2025, channeled through forward-flow agreements, warehouse facilities, and asset-backed securities. In other words, invisible consumer debt is being packaged and sold to institutional investors — a direct link to the growth of asset-based finance and the broader private credit boom.

The concern this raises is one of compounded opacity: if the underlying loans are inconsistently reported and the borrowers' total leverage is hard to see, then the risk embedded in the securities built on top of them is correspondingly hard to assess — for originators and institutional buyers alike. It is a reminder that data gaps at the consumer level don't stay at the consumer level; they propagate upward into the structured products that fund the system.

The stability question

So where does this leave us? The measured conclusion — consistent with the Richmond Fed's analysis — is that BNPL's direct impact on financial stability appears limited at present, given its scale relative to credit cards and the historically low loss rates on pay-in-4 products. There is, as of early 2026, no clear evidence of a BNPL-driven crisis. This should temper the more breathless "ticking time bomb" framing.

But limited direct impact is not the same as no concern, and the structural issues are real: fragmented, partly invisible debt; concentration among financially fragile users; short terms that permit rapid delinquency spikes; and a distribution channel that pushes the risk into structured products. The right posture is neither alarm nor complacency but vigilance — treating BNPL as a warning light on the dashboard of consumer credit rather than a fire. The deeper lesson points beyond BNPL itself: it is a case study in why the credit system's visibility is a form of infrastructure, and why lenders increasingly need cash-flow and alternative data to see what the bureaus miss — the shift we examine in the new architecture of credit. In a world where debt can hide, the ability to see the whole borrower becomes a decisive advantage.

Frequently asked questions

---

This report is for general information only and does not constitute financial, legal, or investment advice. Figures vary widely by definition and source; where possible, this piece anchors to Federal Reserve and CFPB analysis, which readers should consult directly.