Why Retailers Love Buy Now, Pay Later

Clicking “Buy Now, Pay Later” at checkout feels like a small convenience: take the product home today, pay for it later, usually with no interest. But behind that click is an economic mechanism quietly reshaping online retail—and a puzzle about why stores are so eager to lend you money for free.

A new study co-authored* by Manju Puri, the J. B. Fuqua Professor at Duke University’s Fuqua School of Business, sets out to solve it. The paper, “The Economics of Buy Now, Pay Later: A Merchant’s Perspective,” published in the Journal of Financial Economics, focuses primarily on why merchants offer Buy-Now-Pay-Later (BNPL). After all, the store usually foots the bill, sometimes even paying an outside provider like Klarna or Affirm to run the loan.

Why would a seller voluntarily finance its customers’ borrowing?

A booming market built on free credit

BNPL has gone from niche to ubiquitous. Worldwide, it has grown into a roughly $500 billion market, up from about $50 billion in 2019. At its core, BNPL is a short-term, unsecured consumer credit offered right at checkout, letting shoppers split a purchase into a few installments (“Pay in four,” common in the U.S. and U.K.) or defer it in a single later payment (“Pay in 30 days,” popular in Europe), typically at zero interest.

For the shopper, it’s close to free money. For the merchant, it’s a cost: the store absorbs the financing, whether it lends in-house or pays a provider to do it. That’s the opposite of a credit card, where the customer pays the interest and fees.

So why would the merchant volunteer to pick up the tab?

Price discrimination in disguise

The study’s central insight is that BNPL is a clever form of price discrimination—charging different customers different effective prices for the same product. A merchant can’t simply post a lower price for cash-strapped shoppers; everyone would claim the discount. But by bundling the product with a subsidized zero-interest loan, the store can quietly lower the effective price for the people who value that loan most—those with limited access to cheap credit.

In the process, the merchant captures a larger share of consumers that a single sticker price would leave on the table.

“If you offer the product at the regular price, but you bundle it with a zero-interest loan, the result is a lower effective price of the product,” Puri said. “This typically benefits less wealthy consumers who value getting a subsidized loan more.”

A “Robin Hood” effect—with a catch

This flips the usual script. Because lower-creditworthiness customers face the steepest borrowing costs elsewhere, they gain the most from a free loan, so BNPL effectively redistributes value toward the shoppers who need it most.

The authors call this a “Robin Hood effect,” the opposite of the “Reverse Robin Hood” pattern earlier research has documented for credit cards, where reward-card perks for the affluent are effectively subsidized by everyone else.

“BNPL brings new buyers into the market—people who want to purchase but might hesitate because of liquidity constraints,” Puri said. “Credit cards work the other way: the fees are paid by the customers, so the loan is not subsidized, and the interest rates offered vary with customers’ credit scores—lower-credit customers face higher interest rates.”

There’s an important catch, though, Puri noted. The Robin Hood benefit holds only if shoppers behave rationally. If they’re prone to “present bias”—overweighting the purchase today and underweighting the bill that comes later—the same interest-free financing can tip into overborrowing, and the net effect on those customers becomes ambiguous.

That tension is one reason BNPL has drawn growing scrutiny from regulators worried about household debt, Puri said.

Proof from a real store

To test the theory, the researchers went inside a large German online furniture retailer that ran BNPL in-house, giving them a rare view of both the benefits and the costs.

BNPL was wildly popular there—the single most-used payment method, chosen for just over half of all transactions.

The decisive evidence came from a randomized controlled trial: the retailer showed BNPL at checkout to some customers and withheld it from others. The results showed that offering BNPL lifted sales by about 20%, mostly by converting browsers into buyers, with an added bump from shoppers choosing pricier items once the option was available.

And the effect landed exactly where the theory predicted—among lower-credit shoppers. Their purchases were two to three times more responsive to BNPL than those of high-credit customers, evidence that the free loan was working as a targeted price cut for the people most constrained by cost.

That doesn’t mean the store says yes to everyone, Puri said. 

BNPL is a credit decision, and the retailer managed its risk carefully, rejecting roughly one in eight applicants.

Consistent with what expected, approvals skewed toward customers with higher credit scores and toward higher-margin products—the cases where extra sales were most likely to outweigh the cost of the occasional default.

The economics: an inverted U

So does it pay off? Yes, but unevenly, Puri said. Lending isn’t free: at this retailer in 2021, about 2.5% of BNPL transactions ended in default, though after collection efforts the actual revenue loss was closer to 1.17%.

Because BNPL increases both sales and default risk at once, its profitability traces an inverted U across the credit spectrum:

• For the most creditworthy customers, BNPL barely moves sales, but it’s cheap to offer because they rarely default.
• For the riskiest customers, the additional sales can be swamped by default losses.
• The sweet spot is the middle—shoppers too constrained to lean on premium credit cards, but reliable enough to pay the bill.

Why it works better online than in the aisles

None of this would be possible without technology, Puri noted. Online, a merchant can assess a shopper’s risk in real time from their digital footprint—the device they’re using, the contents of their cart—and approve or decline a loan before checkout, no lengthy application required.

That such signals can gauge credit risk is itself a finding of Puri’s earlier research. In “On the Rise of FinTechs: Credit Scoring Using Digital Footprints,” published in the Review of Financial Studies, she and her co-authors showed that a few simple digital traces—the device and operating system a shopper uses (an iPhone versus an Android, say), the time of day they buy, even how they found the site—can predict default about as well as a traditional credit score. The digital footprint complements credit scores, and it can open the door to credit for people credit agencies barely see.

That’s also why BNPL is far more common online than in brick-and-mortar stores, where screening a customer on the spot is harder to pull off.

This showed up inside the German furniture store, which offered BNPL on its website but not in its physical shops, just as the model predicts.

“Technology has made small-scale lending feasible at scale,” Puri said. “That’s the real innovation—the ability to make a credit decision in seconds, not days.”

How far does it generalize?

Although the evidence comes from a single retailer, what gives the results broader reach is the logic behind them, Puri said. The mechanism needs only one ingredient: that the merchant, not the customer, pays for the loan.

And that’s precisely how the major providers operate, she said, as Klarna draws the overwhelming majority of its core revenue from retailers rather than shoppers, and Affirm openly pitches merchant-subsidized 0% financing as a sales tool, an alternative to cutting prices.

The same incentives that show up at one store, in short, are present across much of the industry.

Tailoring payment to different customers

At its core, BNPL is a tool for tailored pricing: lower effective prices for customers with little access to credit, and more revenue from those who can pay upfront—all while widening the pool of people able to buy in the first place.

“You’re tailoring the payment structure to different kinds of customers,” Puri said, “even for something as small as a $400 chair.”

“The appeal is clear,” she added. “BNPL gives consumers with little access to traditional credit the chance to participate in the market while helping merchants increase sales and profitability.”

*Co-authors:

Tobias Berg (Goethe University Frankfurt), Valentin Burg (Humboldt-Universität zu Berlin), Jan Keil (Indian School of Business)