The BNPL abstraction: why "four easy payments" isn't

Cost Me Research Desk · May 28, 2026

A $200 jacket on a credit card asks one question: do you want this? A $200 jacket on Klarna asks a different one: can you spare $50 every two weeks? The second number is smaller, the commitment is the same, and your brain processes them in completely different ways. That gap is the entire business model of Buy Now, Pay Later.

BNPL didn't invent payment decoupling. Credit cards already did most of the work — a 2001 study by Prelec and Simester found that bidders in a sealed-bid auction were willing to pay roughly twice as much when using a credit card versus cash for the same item (Prelec & Simester, 2001).¹ What BNPL did was take that effect and engineer it further, removing the few remaining friction points credit cards still had.

What “abstraction” means in payments

In behavioral economics, an abstraction is anything that separates the moment of consumption from the moment of payment. Cash is the least abstract payment method possible — you hand over physical paper and the loss is immediate, visible, and final. Each layer of abstraction stacked on top of cash mutes that signal a little more:

• Cards turn the loss into a swipe. The money is still real, but you don't feel it leaving.
• Saved cards and one-click checkout remove the typing step. No card number to enter means no pause to reconsider.
• Subscriptions remove the decision entirely. You consumed in March, you paid in March, but you also paid in April, May, and every month since without choosing to.
• BNPL goes further than any of the above. Consumption is right now, the price you see is one-quarter of the actual cost, and the next three payments live in a future that, to the deciding brain, barely exists.

Each layer is a feature for the merchant and a tax on your decision-making. The merchant's revenue per click goes up; your ability to evaluate whether you actually want the thing goes down.

The four-payment trick

The classic BNPL offer — popularized by Afterpay, Klarna, Sezzle, and Affirm's Pay-in-4 — splits a purchase into four equal installments two weeks apart. The first installment is due at checkout, so a $200 purchase shows up as $50 today. The framing matters more than the math.

Three things happen in the buyer's head when the price on the button changes from $200 to $50:

1. The anchor moves. Whatever you were willing to spend on a jacket, you were probably benchmarking against the displayed price. Cut the displayed price by 75% and the threshold to say yes drops with it, even though the total cost didn't change.
2. The pain signal weakens. The fMRI literature on the “pain of paying” shows the insula — the brain region that processes physical pain — fires more strongly the larger the loss feels. A $50 number fires it less than a $200 number, even when the actual loss (eventually) is identical.
3. The future gets discounted. Humans systematically undervalue costs that arrive later. This is called hyperbolic discounting, and it's extreme over short horizons: a payment due in six weeks feels much less than half as real as one due today. BNPL doesn't just split the cost — it pushes 75% of it into the psychologically-weaker future.

The four-payment frame is a behavioral instrument, not an accounting one. The math is the same as a no-interest loan that pays itself off in six weeks. The math is not the point.

What the research actually shows

The Consumer Financial Protection Bureau's 2022 market report — the most comprehensive look at U.S. BNPL data to date — found that loan originations across the five largest BNPL lenders grew from 16.8 million in 2019 to 180 million in 2021 (CFPB, 2022).² Average loan size: about $135. These are not big-ticket purchases. They're the basket an extra $30 dress, the cart-abandoned shoes, the Sephora order.

The CFPB's 2023 follow-up survey of consumers found a more pointed result. Among BNPL users, roughly two in five had spent more than they would have otherwise as a direct consequence of BNPL availability at checkout (CFPB, 2023).³ That's self- reported, after-the-fact awareness — the unconscious effect is presumably larger.

Merchants are not shy about the lift. Public BNPL companies report basket-size increases of 30%–85% on checkouts where the BNPL option is offered versus where it is not. Some of that is selection — customers who would have spent more anyway choose BNPL — but a meaningful chunk is the decoupling effect doing exactly what the Prelec / Loewenstein framework predicts.

The 8% Invisible Surcharge

BNPL providers charge merchants a fee on every transaction — typically in the 4%–8% range, several times higher than the ~2%–3% a credit card costs them. Merchants accept that cost because the basket-size lift makes it pay for itself. But the fee doesn't come out of profit margins. It comes out of prices.

Once a meaningful share of a retailer's checkouts use BNPL, the rational pricing move is to raise base prices to absorb the higher processing cost across the whole catalog. That increase shows up on every shopper — the cash buyer, the debit-card buyer, the credit-card buyer, and the BNPL user alike. Nobody opts out of it because the surcharge isn't labeled. It's just inflation in the listed price.

This is one of the quieter consequences of payment abstraction. The fee that subsidizes the “four installments, zero interest” offer is being collected from everyone walking through the store — including, ironically, the cash-only shopper who would never use BNPL on principle.

The Phantom Debt Fog

Most BNPL plans — particularly the short-tenor Pay-in-4 offers — are not reported to the major credit bureaus. That's historically been pitched as a benefit (a missed payment won't damage your credit) and a feature of the product's small-dollar nature. The second-order effect is what the industry quietly calls the phantom-debt fog: lenders have no visibility into how much short-tenor debt their borrowers are carrying, and borrowers often don't either.

The CFPB report flagged this directly: users stacking multiple BNPL plans across different providers can carry meaningful total obligations that don't show up in any credit file (CFPB, 2022).² A mortgage underwriter looking at a borrower's file in 2026 might see a clean history, a 740 FICO, and a 28% debt-to-income ratio — while the borrower actually has $1,400 in active BNPL plans across Klarna, Afterpay, Affirm, and Sezzle that the file never captured.

The fog cuts both directions. Lenders mis-price risk because they're underwriting against a partial picture. Borrowers mis-budget because the four-, six-, and eight-app sprawl never aggregates into a single “here is what I owe” number. The phantom debt is real money. It just doesn't live anywhere a single screen can show it.

The Anatomy of a Debt Trap

The CFPB's consumer survey also looked at who actually uses BNPL heavily, and the picture isn't reassuring. Compared to non-users, heavy BNPL users were more likely to carry credit-card balances, more likely to be delinquent on other accounts, more likely to have overdrafted their checking account, and more likely to score in the subprime credit range (CFPB, 2023).³ The product that markets itself as a responsible alternative to credit cards is, in practice, most heavily used by exactly the people credit cards already weren't working for.

The trap shape is predictable. The first BNPL plan feels free — no interest, four small payments, the basket shows up at the door. The second one feels free for the same reason. By the fourth or fifth concurrent plan, the biweekly debits start hitting the checking account on the same paydays as rent and utilities. When the timing doesn't work, the BNPL providers attempt the charge anyway, the bank covers it and levies an overdraft fee, or the charge fails and the BNPL provider levies a late fee. Both costs are unlabeled in the original transaction. Neither shows up in the “four easy payments” framing on the checkout button.

The subprime selection effect compounds it: the buyers most sensitive to displayed price (rather than total cost) are exactly the buyers BNPL's checkout-button math is engineered for. The product works on everyone, but it works most aggressively on the segment with the least cushion to absorb its second-order costs.

The FICO 10T Shockwave & Silent Rejections

For most of BNPL's history, the phantom-debt fog protected users in one specific way: it kept the behavior off their bureau file. That insulation is ending. Newer scoring models — FICO 10T and the BNPL-aware update FICO announced in 2025 — are explicitly incorporating BNPL data where bureaus receive it. Affirm began furnishing all loan data to Experian in 2025. Other providers are following.

The bureau-data side is the visible part. The less visible part is that lenders are increasingly using AI-driven underwriting that pulls bank-transaction data directly (via Plaid, MX, and similar) and flags BNPL biweekly debit patterns even when no bureau record exists. A mortgage, auto, or high-limit credit decision in 2026 can land on a decline that the borrower can't reverse-engineer from their credit file, because the signal that drove the decision lives in the transaction data instead.

Practically: BNPL usage in 2022 was nearly invisible to underwriters. BNPL usage in 2026 is increasingly visible — sometimes through the front door of an updated FICO model, sometimes through the side door of cashflow underwriting, sometimes as a silent rejection on a major credit decision without a clear stated cause. The product hasn't changed. The visibility around it has.

The Securitization Time Bomb

The macro story behind BNPL is even quieter than the consumer-level story, and probably matters more in the long run. BNPL providers don't hold the loans they originate. They package the receivables into asset-backed securities and sell them — to hedge funds, pension funds, insurance companies, and the same Wall Street buyers who package auto-loan and student-loan paper. Klarna, Affirm, and the major Pay-in-4 providers have all built ABS programs at scale since 2021.

That model only works if the underlying loans perform. As long as default rates stay low — which has roughly been the case in benign credit environments — the securitization engine keeps spinning and the BNPL companies keep getting cheap capital to originate the next wave of small-dollar loans. When defaults rise, the ABS pricing tightens, the cost of capital climbs, and the providers either pull back on origination or get squeezed on margins.

Two longer-horizon concerns sit behind that mechanic. First, consumer-level: a payment model optimized for maximum frequency of small transactions, financed by institutional capital, structurally pushes the economy toward a “post-ownership” model where households don't own things — they hold ongoing payment obligations on them. A jacket bought in eight micropayments to a financial institution is, in a small way, a different relationship with the object than a jacket bought outright. Stacked across millions of households and decades, the difference compounds.

Second, systemic: phantom-debt fog plus securitized receivables plus AI-driven origination is the same general recipe — different instruments, different decade — that produced the housing-credit crisis. Nobody's claiming BNPL is a 2008-scale risk yet. But the regulators flagging it — CFPB, the Fed, the Bank of England, ASIC in Australia — are flagging the same structural vulnerability: a consumer-debt category growing faster than the data infrastructure tracking it.

Why “no interest” is doing real work

Classic credit card debt at least carries an emotional label. Most people, when they think the word “debt,” feel something. BNPL's zero-interest framing strips that label off.

From a strict accounting view, that's defensible — if you pay on time, the four-payment BNPL plans don't charge interest. From a behavioral view, removing the word “debt” from the experience removes one of the few remaining cognitive checks consumers have on serial small borrowing. You can be carrying $1,200 in active BNPL commitments across five apps and not feel like you're in debt, because nothing called itself that.

The CFPB data suggests this is not a hypothetical. Heavy BNPL users in their survey were more likely than non-users to carry credit-card balances, be delinquent on other accounts, and overdraft their checking accounts (CFPB, 2023).³ The four-payment frame doesn't replace traditional debt for these users — it sits on top of it, with the extra feature that it never registers as debt at all.

The opportunity cost the four payments hide

Set aside the consumer-debt question entirely. Even when BNPL is used responsibly and paid off perfectly on schedule, every $200 spent through it is $200 that didn't go into an index fund. At the S&P 500's long-run average return of roughly 10%, $200 invested today compounds to about $3,500 over 30 years. The jacket lasts a few seasons. The forgone return lasts forever.

BNPL doesn't change the opportunity-cost math — the jacket would cost the same in forgone returns on a credit card. What BNPL changes is the probability you ever ran the math. By replacing $200 with $50 on the checkout button, it moves the decision out of System 2 (deliberate, comparative, future-oriented) and into System 1 (reflexive, immediate, anchored on the displayed number).

The four-payment trick works because it's designed to bypass exactly the cognitive process that would otherwise ask the opportunity-cost question.

How to think clearly when BNPL is on offer

The countermeasures are not about willpower. They're about restoring the information the abstraction stripped out:

• Translate the BNPL price back to the real price before you decide. The button says $50, but you're committing to $200. Mentally restore the full number, then decide.
• Ask the opportunity-cost question at the real price. Would you pay $200 today, in cash, for this item, if you knew that same $200 would compound to roughly $3,500 over 30 years? If yes, buy it. If no, the BNPL split didn't change anything that matters.
• Treat “no interest” as a marketing word, not a financial one. The cost of a purchase isn't the interest you pay on it. It's the purchase itself plus the return you forgo on the money spent.
• Audit your active BNPL commitments monthly. If you can't name every open plan from memory, you have more BNPL debt than your brain is tracking. That gap is where the trouble starts.
• Use a 48-hour rule for any non-essential BNPL purchase. The cooldown reintroduces the deliberation the four-payment split was engineered to remove.

The point is not to demonize a payment method

BNPL can be a useful tool — for a known, planned purchase where the four-payment cadence matches your income timing and you'd have made the buy anyway. For furniture, a textbook, a needed appliance, the math is fine and the lift is small.

The problem is that BNPL's entire optimization function — the reason the option is offered, the way the button is styled, the size of the displayed number — is built for the exact opposite case. It's built to convert hesitations into purchases. And the data, both academic and corporate, agrees that's what it does.

The point isn't to never use it. The point is to know that when you do, the screen is showing you a number that was chosen specifically because it bypasses the calculation you would otherwise run.

How Cost Me reframes the BNPL question

Cost Me is, structurally, a tool for putting the full price back into view before the decision happens. Type the real amount — $200, not $50 — and the app shows the 30-year opportunity cost of that money. The brain that was about to process a $50 commitment instead processes the actual one: a $200 spend, with a $3,500 future-value tag on it.

That's not designed to make you feel bad about a jacket. It's designed to give the same brain that was about to decide on incomplete information the missing piece. A purchase made at $200, on full data, is a real decision. A purchase made at $50, anchored to a hidden $200 and an even more hidden $3,500, is a marketing outcome.

Restoring the full number is the smallest possible intervention that still tilts the calculation back toward the buyer.

References

1. Prelec, D., & Simester, D. (2001). Always leave home without it: A further investigation of the credit-card effect on willingness to pay. Marketing Letters, 12(1), 5–12. https://doi.org/10.1023/A:1008196717017
2. Consumer Financial Protection Bureau. (2022). Buy Now, Pay Later: Market Trends and Consumer Impacts. Washington, D.C. consumerfinance.gov
3. Consumer Financial Protection Bureau, Office of Research. (2023). Consumer Use of Buy Now, Pay Later: Insights from the CFPB Making Ends Meet Survey. consumerfinance.gov
4. Prelec, D., & Loewenstein, G. (1998). The red and the black: Mental accounting of savings and debt. Marketing Science, 17(1), 4–28. https://doi.org/10.1287/mksc.17.1.4

Related reading: why paying with cash hurts more than tapping a card, why your brain says yes to the impulse buy, and the 48-hour rule. Back to costme.io.