How Splitting a Payment Became a Shadow Credit System
An old idea, moving faster
For most of modern retail history, the question that stood between a person and a purchase was simple: can I afford this? Now a second question has taken its place at the checkout: can I afford the monthly payment?
GlobalData’s Buy Now Pay Later – Thematic Research report put global BNPL transaction value at $120 billion in 2021, on track to reach $576 billion by 2026. Other research firms measure the market differently, by originations rather than gross transaction volume, or by US activity alone, and land on smaller figures, so the exact number depends on what’s being counted. What’s not in dispute is the direction: this went from a niche checkout option to a mainstream one in well under a decade.
Paying in installments itself is not new. Layaway let Depression-era shoppers reserve an item and pay it off before it ever left the store. Hire purchase did the reverse, handing over furniture or appliances first and collecting the payments after, and became a fixture of postwar British and Australian retail. Store credit generalized both into an account a shopper could keep drawing on. What’s different about BNPL isn’t the installment. It’s the speed: financing that used to require a separate application and a wait now appears as a single tap at checkout, underwritten in the seconds before the sale rather than arranged in advance of it.
The price that stopped feeling like a price
A phone that costs six hundred dollars is a big decision. The same phone offered at thirty dollars a month barely registers as one. A car that costs forty thousand dollars gets cross-examined by a whole household. Offered at five hundred and forty-nine dollars a month, it gets test-driven on a Saturday and driven home the same afternoon.
Nothing about the phone or the car has changed. The total owed hasn’t changed either. What has changed is the unit the brain uses to judge the cost. People rarely multiply a monthly payment across its full term and compare it to the total price. They compare the installment, not the bill, because the installment is the number sitting in front of them at the moment of the decision.
This works on sellers as much as buyers. When a business wants something to look more affordable, it now has two ways to do it: lower the price, or lower the size of the payment. The second option is cheaper for the business and just as effective at the register, because the number the shopper is weighing is rarely the full price anymore. It’s the payment. That logic reaches back into product design itself. A phone maker or furniture brand doesn’t only ask whether people will pay six hundred dollars for something. It asks whether the price can be held under a thirty-dollar monthly figure, because that’s the number that will actually move units at checkout. Storage tiers, trim levels, and financing terms get built around psychologically comfortable monthly amounts as much as around cost, which means the sticker price stops being the thing anyone is actually designing toward.
How the loan disappears
Ask someone who financed a laptop through a credit card whether they took on debt, and most will say yes without hesitation. Ask someone who split the same laptop into four BNPL payments, and the answer often gets softer. They didn’t borrow, they just split the payment.
That difference in language does real work. Debt signals risk and consequence. Splitting a payment sounds like a checkout preference rather than a loan. The mechanics are the same either way; what’s changed is the word attached to them, and the word is doing the job the interest rate used to do, deciding how seriously a person takes what they’ve just agreed to.
The checkout page has been redesigned around that same softer framing. It used to do one job, processing payment. Now it often offers financing, warranties, and sometimes insurance in the same few taps, before the transaction is confirmed, each one presented as a feature of checkout rather than a separate financial product being sold alongside it. Retailers do this because it pays off. Merchants who accept BNPL pay the provider a transaction fee well above what they pay on a standard card. A 2026 Federal Reserve research note put the gap at five to eight percent for BNPL against two to three percent for cards, citing academic research on the sector; broader industry surveys report a wider two-to-eight-percent range for BNPL, averaging closer to four to six percent, depending on the provider and the merchant’s negotiating leverage.
Retailers accept the higher cost because BNPL tends to lift conversion rates and order sizes enough to make the math work. Those lift figures come mostly from provider and industry case studies rather than independent audits, vary by merchant and category, and are best read as directional rather than precise; with that caveat, they consistently put the conversion lift somewhere in the twenty to thirty percent range, with average order value rising twenty to forty percent. Enough retailers have seen a large enough lift that some have built recent sales growth and pricing decisions around BNPL remaining available, which makes them more exposed than they might realize if approval rates tighten or a provider’s terms change.
Borrowing tomorrow’s demand, today
BNPL doesn’t necessarily create more demand. Often, it just moves demand forward in time. A television that would have been bought in December gets bought in August instead. For the retailer, the sale shows up today. For the household, a slice of next month’s income is already spoken for before it’s been earned.
The fair counter-case deserves a real hearing here. For a large share of users, BNPL isn’t replacing a cash purchase they’d have skipped otherwise, it’s replacing a credit card purchase they’d have made anyway, at a lower cost. Paying in four installments at zero percent is a better deal than carrying the same balance on a card at twenty percent or more, provided every payment lands on time. Some of what looks like “borrowed demand” in the aggregate is really just demand that was always going to be financed, moving from a more expensive form of credit to a cheaper one. That’s a genuine benefit, and it complicates any claim that BNPL is purely inflating spending rather than partly just repricing it.
What that counter-case doesn’t fully answer is the aggregate picture. When a large number of households pull spending forward at the same time, whatever the individual reasoning, consumer spending in that period can look stronger than the underlying economy actually is. A rising share of retail activity becomes financed by expectations of future income rather than income already in hand, and that’s true whether each individual decision was a smart trade-down from a credit card or not. That distortion is easy to miss in monthly retail sales figures, and it tends to unwind quickly once conditions tighten.
There’s a second layer to this with less to do with money and more to do with information. Every BNPL transaction hands the provider a live feed of what people buy, when, how reliably they repay, and how spending shifts around payday. That behavioral data has become a business of its own, feeding underwriting models and shaping which offers a shopper sees next. The loan is the visible product. The data is often the more durable asset underneath it.
The debt that stacks invisibly
A single BNPL plan is easy to manage. The risk was never the first one. It’s the fifth, taken out to cover a shortfall left by the first four, at a moment when it looks less like a new debt and more like a way out of the old one.
Because each plan is small and often issued by a different provider with no shared view of a person’s other obligations, debt can build up without appearing in one place. LendingTree’s 2026 Buy Now, Pay Later Report, based on a March 2026 survey, found that 47 percent of BNPL users had paid late at least once in the past year, up from 41 percent in 2025 and 34 percent in 2024. The same report found that 25 percent of users were carrying three or more BNPL loans at once, up from 23 percent the year before. That’s loan stacking in practice: a single missed paycheck turning into several missed payments across lenders that don’t talk to each other.
This gap is closing, slowly. Affirm now reports its loans to Experian and TransUnion, Klarna reports to TransUnion, and in 2025 FICO built its first scoring models designed to factor BNPL activity into a traditional credit score. Until that becomes standard across the industry, a meaningful share of household debt sits outside the view of the institutions meant to track it, and outside the borrower’s own running total too.
For a large number of young adults, a BNPL plan is also the first borrowing experience they have, ahead of a student loan, a credit card, or a bank loan. First experiences with credit tend to set expectations for how debt is supposed to feel, and a generation whose first loan felt like a checkout convenience is learning a different baseline than one whose first loan came with a bank’s name on it and a signature required.
What the stress signals actually show
The honest picture sits between two things that are both true at once.
On one side: BNPL has grown up during low interest rates and a resilient job market, and the last two years have offered a preview of stress rather than a full test. Klarna’s IPO priced in September 2025 at forty dollars a share; by mid-2026 the stock traded well below that level, partly on concerns about rising late payments. Affirm’s own disclosures show its allowance for credit losses climbing to roughly six percent of loans held for investment even as its headline delinquency rate stayed close to flat, a sign lenders are bracing for more strain than has shown up yet. None of this is a collapse. But rising late payments, more loans stacked per person, and lenders building bigger reserves is the kind of pattern that tends to widen fast once unemployment rises or credit tightens elsewhere.
On the other side: used occasionally and paid off on schedule, BNPL can be a genuinely free short-term loan, useful for smoothing an unexpected expense or making a large purchase without reaching for a costlier card. It has given people with thin or no credit history a way to buy something they’d otherwise need to save toward for months. The concern isn’t that the tool exists. It’s what happens when something built for occasional use becomes the default way to pay for groceries, everyday clothing, and small recurring costs that used to simply come out of a paycheck. A useful product turns into a dependency at exactly the point where it stops being occasional, and that point rarely announces itself.
Catching up to it
Regulators have moved in fits and starts. In 2024, the CFPB issued an interpretive rule treating BNPL providers as credit card issuers under the Truth in Lending Act, which would have required billing statement disclosures and dispute rights similar to a card. In 2025, under new leadership, the CFPB withdrew that rule and said it wouldn’t reissue a replacement, arguing that open-end credit rules were a poor fit for BNPL’s typically closed-end loans. Consumer advocates have noted that the withdrawal doesn’t settle the underlying legal question, only postpones it.
Into that gap, states have started to step in. New York enacted a licensing law for BNPL lenders in May 2025, as part of that year’s state budget, requiring providers to register with the state’s Department of Financial Services in the absence of clear federal rules. More than a year later, the regulator is still finalizing the rules needed to actually enforce it: the licensing requirement doesn’t take effect until 180 days after those implementing regulations are adopted, and as of mid-2026 they still hadn’t been. Even the one state that moved fastest is still catching up to itself.
The credit bureaus and FICO, meanwhile, have kept moving on their own timeline, building the infrastructure to fold BNPL activity into mainstream credit scores regardless of what happens in Washington or Albany. The pattern across all of it is the same: a product that outgrew the rulebook, followed by regulators, states, and credit bureaus each catching up in their own lane, on their own schedule.
The real shift
Buy Now, Pay Later is really just the first widely adopted expression of something bigger: the moment of borrowing merging into the moment of spending, until they stop being two separate decisions at all. One-click ordering removed the effort of buying. Same-day delivery removed the wait. What BNPL removes is the pause, the moment someone might otherwise stop and ask whether they can really afford this. As that pause keeps shrinking, alongside the broader shift toward subscriptions and upgrade programs where fewer things are ever fully paid off, borrowing stops feeling like a financial decision and starts feeling like just another way to pay.
There’s one more piece worth naming, and it’s the most concrete evidence that the shift is already real rather than theoretical. Millions of households now reach payday with part of that income already committed to payments taken out weeks earlier, for things already used up, worn, or eaten. The paycheck isn’t entirely future income anymore. A slice of it belongs to purchases already made, spent before it arrived.
Every generation has borrowed. Previous generations mostly borrowed for milestones, a home, a car, an education, decisions large enough to justify the weight of debt. What’s changed is that borrowing has become the default way to buy ordinary life, spread thin enough across small purchases that it stops feeling exceptional. And once it stops feeling exceptional, consumer spending, retail growth, and even measures of economic strength start to rest, in part, on people spending income they haven’t earned yet. Klarna, Affirm, and Afterpay are simply today’s names for that shift. Whatever replaces them will do the same thing, only faster and further out of view.
