The Consumer Credit Bubble: How ‘Buy Now, Pay Later’ Pushed the Lower Middle Class to the Brink

The number on the screen changes at a specific point during the checkout process, which occurs in a split second, barely registering consciously. not the entire amount. Just the amount you’ll pay today. A pair of sneakers that cost $200 is reduced to $50. A laptop that costs $1,200 is reduced to $300. The thing is the same. The debt remains the same. However, there appears to be something about the way those smaller numbers are presented that eliminates any hesitation that might have otherwise caused someone to pause. In a very literal sense, the entire Buy Now, Pay Later business model is based on this psychological trickery. As a result, about 30% of American consumers are now enrolled in a credit system that, for many of them, isn’t functioning as advertised.

In 2023, Americans financed more than $100 billion in retail transactions using BNPL services. Globally, the market is expected to grow to $560 billion. What started out as a practical substitute for credit cards for online purchases has developed into something far more intricate, financing not only apparel and electronics but also, increasingly, groceries. 14% of BNPL customers divided their grocery purchases into installments in 2024. That percentage increased to 25% by 2025. Approximately one in three Gen Z users specifically acknowledge using BNPL to pay for food. The nature of a financial product has significantly changed when it is used to spread the cost of milk and bread over a six-week payment schedule. Originally, the product was marketed as a convenience tool for discretionary purchases.

BNPL’s mechanics are fairly simple. A customer chooses Klarna, Afterpay, Affirm, or one of a few rivals at the checkout, which is nearly always done online and increasingly done in-app. Usually charging a fee of two to eight percent, the provider pays the merchant up front. After paying the first installment of 25% at checkout, the customer is left with three additional payments due at intervals of two weeks. No hard credit inquiry. Unless the loan defaults, there will be no reporting to credit bureaus. There is no complete picture of the number of these arrangements that a single customer has operating concurrently across various platforms. It is a parallel credit infrastructure designed to fill in the gaps left by conventional regulation.

These gaps in regulations weren’t coincidental. The Truth in Lending Act’s Regulation Z, which was written decades ago and exempts credit arrangements from disclosure requirements if they involve four or fewer installments and charge no finance fees, is largely responsible for the existence of the modern BNPL industry. Fintech companies saw this “rule of four” as a basis for a completely new product category, even though it was created for a different financial era. By operating within that exemption and providing products that function similarly to credit but are not subject to the same consumer protection regulations as credit cards, Klarna, Afterpay, and Affirm created billion-dollar companies. The outcome is what researchers have begun to refer to as a “shadow credit system,” in which a user can concurrently accrue significant debt across five different platforms without any of it showing up on any credit report until something goes wrong.

Researchers Marco Di Maggio and Emily Williams of Harvard Business School documented the downstream effects in a study that ought to get more public attention than it has. According to their research, customers who use BNPL spend between 10 and 40 percent more on the same purchase than they would if they used a credit card. Users with lower incomes are most affected. Customers who make between $25,000 and $45,000 a year are the most likely to use BNPL; as a result, 20% of them pay overdraft fees and 17% pay low-balance fees. The mechanism is what Di Maggio referred to as the flypaper effect: the behavioral brake that might otherwise slow the purchase releases when the immediate cost of something seems low. Up until the point at which the account balance is depleted, the automatic payment that is taken out of the account two weeks later, four weeks later, and six weeks later feels unrelated to the initial choice.

It’s difficult to ignore the people this system is most actively pursuing. The Consumer Financial Protection Bureau’s own data indicates that the demographic profile of BNPL users is skewed toward women, Black and Hispanic adults, and adults under sixty without college degrees—exactly the groups most vulnerable to economic shocks and least protected against payment disruptions. Generally speaking, these users aren’t using BNPL to purchase upscale goods they couldn’t otherwise afford. They use it to control the flow of money between paychecks and make everyday household expenses seem doable on tight budgets. The 36% of users who say that their main motivation for using BNPL is to manage cash flow aren’t going overboard. They’re extending.

The data on late payments has been trending in a single direction. In 2025, 41% of BNPL users had at least one late payment, compared to 34% the previous year. In absolute terms, the BNPL charge-off rate is still lower than that of credit cards, at about 2% as opposed to about 4%. However, analysts have noted that this comparison is a little deceptive. BNPL loans are small-dollar, short-term loans. Compared to revolving credit card debt, their default rate ought to be significantly lower. In a product category where the average loan size is about $131, the fact that it is running at 2 percent and growing says something specific about the financial situation of those who are not making those payments.

There has been tension within the industry itself. With great fanfare, Apple introduced its own BNPL product in 2023, but it was shut down in 2024 before it had even reached its first birthday. Due to rising interest rates that upset the economics of interest-free lending, Klarna’s valuation dropped from about $45.6 billion in 2021 to about $14.6 billion by late 2024. In order to finance their purchases, providers who promised customers a zero-interest product still need to borrow money, and the math becomes more difficult to maintain as borrowing costs increase. The business model relies on merchant fees, scale, and a consistent flow of late fees from the percentage of users who fail to make payments, more than the promotional materials usually highlight. At the very least, the incentive structure created by that final revenue line is awkward.

The fact that many BNPL users do not perceive these products as credit at all adds to the regulatory picture’s complexity. According to survey data, a sizable percentage of users do not characterize their BNPL arrangements as debt, despite the fact that they are by any reasonable definition precisely that. Instead of feeling like a loan, the products are made to feel like a payment tool. The phrase “split your payment” is used instead of “take out a loan.” The user interface is smooth. The approval happens right away. When the outcome materializes, it may come as a surprise. Regulators in the United States and Europe are still debating whether better disclosure, stricter regulations, or a fundamental rethinking of how these products are sold should be used to close the gap between perception and reality. However, the window for anticipating the issue is closing more quickly than the legislative calendars appear to acknowledge.