The 2.9% Number Is Lying

In 1979, a man handed a credit card to a clerk at a retail counter and someone from U.S. News & World Report photographed the moment. That's all: a card, a counter, a transaction. The photograph is remarkable now only because it captures the system at its point of ordinariness. Before $1.28 trillion in outstanding balances (NY Fed, Q4 2025), before interest charges reached $160 billion in a single year, before the average APR climbed past 21%. The man in the photograph was entering a system. He didn't know what it would become. Neither did the banks.

Fifty-five years later, the system publishes a number every quarter that is supposed to tell you whether American borrowers are managing. The delinquency rate on credit card loans, reported by the Federal Reserve, stood at 2.94% in Q4 2025. It has fallen for six consecutive quarters. The chart looks like recovery, but the chart is lying.

Where the Average Breaks

It lies the way a national temperature average lies about Phoenix in July: by distributing an extreme across a landmass large enough to absorb it. The all-bank delinquency rate blends every credit card portfolio in the country, JPMorgan Chase's and the community bank in Yakima's, into a single figure. The blending is where the story disappears.

Banks outside the top 100 by asset size, the institutions that serve the borrowers the largest banks increasingly don't want, report a credit card delinquency rate of 6.62% as of Q4 2025. More than double the headline. It peaked at 7.87% in Q4 2023, the highest the series has ever recorded, and has come down only modestly since. If you bank with a community lender or a mid-tier regional, you live in a different credit economy than the one the aggregate describes.

The split between the headline rate and the small-bank rate isn't a footnote. It is the entire story. The 2.94% is the average of a system where the dominant players have spent three years cutting their subprime origination share from 23.3% to 16.4%, pushing riskier borrowers to exactly the institutions least capitalized to absorb their losses. The number didn't fall because American borrowers got healthier. It fell because the biggest banks got pickier about who they'd lend to.

The St. Louis Fed confirmed the geographic version of the same fracture: from Q2 2021 through Q1 2025, credit card delinquency rose 63% in the lowest-income 10% of ZIP codes and 44% in the wealthiest 10% (May 2025). Both numbers are bad. But the gap between them is the mechanism: the average improves when the improving segment is larger and growing faster than the deteriorating one. The people inside the number don't experience the average. They experience their rate.

The Machinery Behind the Split

The spread between a prime borrower's rate and a subprime borrower's rate is itself a form of default prediction, priced in advance and extracted monthly. Underneath the headline aggregate sits a relationship between who borrows, who lends, and what each pays. The average has no language for it.

The commercial bank interest rate on credit card plans reached 21.76% in Q3 2024, the highest ever recorded in the FRED series, and has barely retreated since, standing at 21.00% as of Q1 2026. But even that figure is an average. The Consumer Financial Protection Bureau's 2025 market report found the average APR on general purpose cards hit 25.2% in 2024, and 31.3% on private label cards. The store-branded cards are disproportionately held by lower-income borrowers. They pay roughly six percentage points more for the same access to credit.

Matt Schulz, chief consumer finance analyst at LendingTree, put it plainly: "High interest rates are murder on those with credit card debt." Murder is not an economist's word. It's the right one. When the cost of carrying a balance rises 52% in two years while wages don't, the math becomes a mechanism of attrition. Borrowers don't default all at once. They make minimum payments. The Philadelphia Fed found the share of active large-bank accounts making only the minimum hit a 12-year series high in Q4 2024 before retreating. And they hold. Until they can't.

The lag is where the pipeline reveals itself. The all-bank charge-off rate on credit cards stands at 4.11% as of Q4 2025. It peaked at 4.69% in Q3 2024, well after the delinquency rate had already started falling. A borrower misses payments (delinquency), holds in that state for 120 to 180 days, and then the bank writes the balance off its books (charge-off). The delinquency rate is the entrance to the tunnel. The charge-off rate is the exit. They don't move in sync. When one falls and the other stays elevated, the system is processing the casualties of the previous quarter's stress. The chart improves. The person doesn't.

The Oldest Story in Commerce

David Graeber argued in Debt: The First 5,000 Years that credit preceded money. Mesopotamian merchants stamped clay tablets as IOUs long before anyone minted coins. The credit card isn't a modern invention so much as an acceleration of a pattern as old as trade itself: the gap between owing and paying is where the power lives. What changed is the scale, the rate, and who absorbs the cost when the gap widens.

Consider the trajectory. Total credit card balances bottomed at $659 billion in Q1 2014, rose to roughly $881 billion by 2019, then compressed during the pandemic. Stimulus checks, reduced spending, forbearance. Balances fell to $841 billion in Q1 2022. Since then, the rubber band released. Balances reached $1.28 trillion in Q4 2025, an all-time record, up $436 billion in under four years. Americans now carry 45% more credit card debt than the 2019 average.

The Great Financial Crisis provides the scale marker. The all-bank delinquency rate peaked at 6.77% in Q2 2009. The charge-off rate peaked at 10.54% in Q4 2009. Today's figures, 2.94% and 4.11%, are less than half those levels. True and important. But the 2009 crisis was a synchronized event. Unemployment spiked. Housing collapsed. Credit froze. The current stress is not a crisis. It is a sorting. The aggregate numbers are moderate precisely because the system has bifurcated. The large banks have de-risked their portfolios. The small banks have absorbed what was shed. The national average reflects the weighted outcome of two populations living in different credit economies.

A Federal Reserve FEDS Note published in November 2025 confirmed this interpretation, finding that "the slowdown in credit card borrowing since early 2024, as well as the lagged effects of past tightening in bank credit lending standards, likely have contributed to the recent stabilization of the credit card delinquency rate." The rate stabilized not because borrowers recovered, but because banks tightened who they'd lend to. The denominator shifted. The average fell. The people who needed credit the most were pushed to the institutions least able to price it fairly.

The Counterforce Nobody Expected

There are real signs that the worst of the post-pandemic adjustment may be behind us. The Boston Fed found that even among the lowest FICO quintile, the share of accounts becoming newly delinquent peaked in December 2024 at 1.95% and has since declined. The all-bank delinquency rate has fallen for six straight quarters. The charge-off rate ticked down in Q4 2025. These are not trivial movements.

The genuine open question is whether stabilization at an elevated level is good news or a new floor. The 2019 average delinquency rate was 2.58%. Today's 2.94% is 35 basis points above that. The charge-off rate, at 4.11%, is 40 basis points above its 2019 average of 3.71%. These are not GFC numbers. But they represent a system that has settled into a higher-stress equilibrium while APRs remain near record highs. The Q1 2026 figure of 21.00% sits almost six full percentage points above the 2019 average of 15.05%. The cost of treading water has increased even as the headline risk metrics moderate. Whether that equilibrium holds through a softening labor market remains the open variable.

In January 2026, President Trump proposed a 10% cap on credit card interest rates. As of April 2026, the proposal has not advanced legislatively, but it acknowledged something the Fed data has been saying quietly: the pricing of consumer credit has become a structural burden, not a cyclical one. "A credit card rate cap is enormously popular with Americans," Matt Schulz told CNBC. "That's why we've seen big names on both sides of the aisle propose credit card rate caps in recent years." The proposal may go nowhere. But the fact that it surfaced tells you where the pressure is building. Not in the aggregate. In the lived experience of every household carrying a share of the $1.28 trillion in outstanding balances.

What the Number Hides

The man in the 1979 photograph was entering a system that would, over the next half-century, grow to hold more outstanding debt than the GDP of all but a handful of nations. The quarterly number published to describe that system's health, the 2.94% all-bank delinquency rate, performs the same averaging trick Americans have seen in school district ratings that mask building-level collapse and in national crime statistics that distribute a few safe neighborhoods across an entire city. It takes a true figure and uses it to tell a false story. The indicators worth watching now, the small-bank delinquency rate at 6.62%, the charge-off rate at 4.11% still processing last year's stress, the credit card APR at 21.00% holding near its all-time peak, and total balances at a record $1.28 trillion, are components of an American Distress Index sitting at 64.0 (Elevated), where debt stress accounts for 41.6% of the composite weight. If the split between the headline rate and the small-bank rate continues to widen through 2026 rather than converge, it will mean the bifurcation isn't a post-pandemic adjustment but a permanent feature of American consumer credit. Two systems running under one number, one of them recovering, the other still paying the cost of a recovery it was never included in, and the average is still lying about both.