Pay in 4 is a spending button, not a payment method

2026-10-21 · spending, credit · about the paper: Marco Di Maggio et al., 2022

The most dangerous button at checkout doesn't charge interest. It splits your total into four painless pieces, promises zero fees if you're on time, and feels like a payment method. A working paper from Di Maggio, Williams and Katz suggests it behaves much more like a spending prompt, one that rewrites your budget for months after the purchase you actually meant to make.

(Di Maggio, Williams and Katz, "Buy Now, Pay Later Credit", NBER Working Paper 30508)

TL;DR: First BNPL use raises total spending by about $130 on impact and, by the paper's cleanest estimate, about $60 per week permanently. That's roughly $3,100 a year of extra spending from a "0% interest" button. Overdraft-fee odds rise 20% relative. Spending rises even for people with credit cards and savings buffers.

The data

The researchers had de-identified bank and credit card transactions: 400,000 users and 39 million transactions, half who used BNPL and half who never did. They compared everyone's spending before and after first BNPL use, then tightened it up with a second design that used retailers' adoption of BNPL as the shock, so it wasn't just that big spenders discovered installments.

What happens the week you first use it

Total spending jumps about $130 in the week of first use. Only about $30 of that is the BNPL purchase itself. The rest is spillover into regular retail, essentials and discretionary stuff, the button didn't just finance a purchase, it loosened the grip on the whole budget. The elevated spending persists across all 24 weeks the study observes.

The spending basket itself shifts, about 12% toward retail on impact, still 1.5% elevated weeks later.

The number to remember

The cleanest estimate in the paper: +$60 per week of permanent extra spending, almost none of it flowing through BNPL itself. Call it $3,100 a year of behaviour change from a button that markets itself as a convenience.

Taylor checks out with pay-in-4, Riley pays in full. Nothing about their incomes or intentions differs, just the checkout button. If Taylor's spending drifts toward the study's estimate, that's $60 a week finding its way out the door. Redirected into an index fund at a 6% average return instead, $260 a month compounds to roughly $180,000 over 25 years. Any compound interest calculator will confirm the math, which is the least arguable part of this post.

The fee spillover nobody advertises

BNPL is marketed as the friendly alternative to credit cards. The transaction data shows what happens to the rest of the account: after adopting BNPL, the probability of an overdraft fee rose 0.47 percentage points (a 20% relative increase) and low-balance fees rose 0.48 points (17% relative). The installment schedule drains the chequing account between paycheques, and the old banking fee machine picks up the difference. Four interest-free payments, plus whatever the overdraft costs.

It's not just broke people

Here's what rules out the comfortable explanation. The spending surge shows up even for users with liquidity buffers and credit cards, people who could pay full price and knew it. Standard economic models of borrowing constraints can't produce spending bumps this size from loans that are mostly repaid within six weeks. The authors' term is a "liquidity flypaper effect": the credit sticks where it lands and spending follows it, for reasons that are psychological rather than mathematical.

The survey data agrees. Among users, 38.4% said the purchase "wouldn't fit the budget", 24.7% wanted to borrow without a credit check, 14% couldn't get approved for a credit card at all. The button is doing exactly what it was designed to do.

The honest bit

One real positive: for lower-income users, BNPL smoothed consumption, it reduced how sharply spending had to swing with each paycheque. That's not nothing. The caveats are real too: this is a single bank's data, the window overlaps COVID-era spending weirdness, it's a working paper rather than peer-reviewed, and no observer can see debt stacking across multiple BNPL apps, which is the industry's favourite blind spot.

Canadian checkouts now serve up the same buttons from Klarna, Afterpay, Sezzle and friends. This study is US data, so the direction likely travels and the exact dollars may not.

What to actually do

• Treat the pay-in-4 button as a spending decision today, because the data says that's what it is, whatever the checkout copy says about flexibility.
• If you do use it, log the payment schedule somewhere you'll actually look. The overdraft finding means the cost shows up in a completely different account than the purchase.
• Turn installments off at checkout where your browser or the retailer allows it. Friction in reverse.
• Pick one thing this month to skip or delay, then send the $60 to an automatic contribution instead, and let compounding make the trade visible.

How strong is the evidence?

Scale: Strong = replicated across multiple peer-reviewed studies · Moderate = one solid peer-reviewed study or consistent findings in limited settings · Suggestive = working paper, small samples, or not yet replicated · Thin = non-peer-reviewed source or conflicting evidence.

Since this post was written, the BNPL collection has grown around it — the synthesis of all five studies separates what's replicated from what isn't.

Next steps

• Doing money beats knowing money — automating around willpower, which is what checkout buttons are built to exploit.
• The skim view: digest + full PDF — every claim above, one tap from the source.
• The paper itself — Di Maggio, Williams and Katz, NBER Working Paper 30508.