The most frequently cited number on American consumer distress is that 12.8% of credit card balances are now 90 or more days past due — up from 7.6% in late 2022. New York Fed researchers published a paper on Tuesday arguing that the number is real, the alarm is misplaced, and the gap is almost entirely a bookkeeping story.
The finding accompanied the New York Fed's Quarterly Report on Household Debt and Credit for the second quarter of 2026, released August 11. Total household debt edged down by $13 billion, or 0.1%, to $18.8 trillion. The share of all outstanding balances delinquent at least 30 days fell to 4.7% from 4.8% the prior quarter.
Stock Versus Flow
The companion Liberty Street Economics post — "How Distressed Are Consumers? Reconciling Diverging Credit Card Delinquency Measures," by Donghoon Lee, Daniel Mangrum, Joelle W. Scally, Tejas Sinha and Wilbert van der Klaauw — separates two measures that are routinely treated as one.
The stock delinquency rate counts every balance currently sitting 90+ days past due. That is the 12.8% figure. The flow rate counts balances newly transitioning into delinquency each quarter, and it has run near 7% and, in the authors' words, "remained relatively stable for almost two years."
The two diverged because, as the authors put it, "the stock kept rising as charged-off debts accumulated" after new delinquencies stabilized in early 2024. The mechanism is a change in lender reporting behavior: between 2004 and 2012, only about 40% of charged-off debts were still being reported a year later. By 2024 that had doubled to 80%. Balances lenders have already written off increasingly stay on credit reports, inflating the stock measure without a single additional household missing a payment.
Strip those balances out and, the researchers found, "all measures of credit card delinquency remain stable after 2024." Their conclusion: "when the question is 'how are households doing right now?' the flow delinquency rates ... provide a more accurate view of current consumer repayment behavior."
Why Bank Data Says 2.9%
The reporting quirk also explains a discrepancy that has puzzled analysts. The Federal Reserve Board's data from commercial bank Call Reports puts credit card delinquency near 2.9% — roughly ten percentage points below the credit bureau figure. That gap is mechanical. At 180 days past due, an account is charged off and leaves the bank's delinquency denominator entirely, whether or not the borrower's situation improved. The bureau data keeps it. Neither measure is wrong; they are answering different questions.
What Actually Moved in Q2
Beneath the headline decline, the composition was mixed. Credit card balances rose $21 billion, or 1.7%, to $1.26 trillion. Auto loan originations hit $211 billion, a record in nominal terms. Home equity balances added $19 billion, extending a four-year climb.
The overall drop was driven by a $74 billion fall in mortgage balances, which the New York Fed attributed to a "servicer transfer gap" — delayed credit reporting when a mortgage moves between servicers — and expects to reverse next quarter.
Spending, meanwhile, has not retreated. Personal consumption expenditures rose 3.2% in the second quarter, and Bank of America Institute data showed credit card spending excluding gasoline up 4.3% in July.
The Divide That Is Real
None of this makes the consumer picture uniformly healthy. New York Fed researchers still describe a K-shaped economy, noting that "there are a lot of households that live paycheck to paycheck." Student loan balances 90+ days delinquent stood at 10.3% as of the first quarter, and auto loans at 5.6% — both well above the credit card flow rate.
The distinction matters for anyone reading consumer credit as a signal on the economy. A stock delinquency rate climbing on stale charge-offs is a lagging indicator of stress that already happened. A flow rate holding flat is a real-time reading, and right now it says elevated but not deteriorating. With July CPI due Wednesday and the September Fed decision still contested, the difference between those two readings is the difference between a consumer cracking and a consumer merely strained.
Sources: Federal Reserve Bank of New York (Quarterly Report on Household Debt and Credit, Q2 2026), Liberty Street Economics, CNBC, Bloomberg, CNN Business, Federal Reserve Board of Governors, Bank of America Institute