Two Credit-Card Delinquency Numbers Disagree. The Gap Is Debt Lenders Already Wrote Off.

Headlines about record credit-card delinquency and headlines about steady credit-card delinquency have both been accurate this year. They are describing different statistics. A post published on August 11 by researchers at the Federal Reserve Bank of New York works out where the two diverge, and the answer is a reporting convention rather than a change in household behaviour.

Two ways to count a delinquency

• A stock measure asks what share of outstanding balances is currently 90 or more days past due. The New York Fed's Consumer Credit Panel, built from Equifax credit reports, produces one.
• A flow measure asks what share of balances newly moved into 90-plus-day delinquency in the quarter. The same panel produces this one too.
• A third measure comes from banks' own Call Reports, filed with the Board of Governors, and counts balances 30 or more days past due on lenders' balance sheets.

Where they part company

The stock measure rose from 7.6% in the third quarter of 2022 to 12.8% in the first quarter of 2026, a steep and continuous climb. The flow measure has been broadly flat for close to two years. The Call Report measure has also been level recently, in line with the flow rate. Only the stock measure is rising.

The charge-off mechanism

When a credit-card balance reaches roughly 120 to 180 days past due, the lender charges it off. On the lender's books the balance then disappears from both the numerator and the denominator of any delinquency ratio, which is why bank-sourced measures do not accumulate old bad debt. On a credit report the balance does not disappear, because the borrower still owes it and the lender keeps furnishing the record to the bureaus. The Consumer Credit Panel therefore keeps counting it, quarter after quarter, as a delinquent balance.

That alone would produce a gap. What made the gap grow is a change in how diligently lenders keep reporting charged-off accounts. Between 2004 and 2012, about 40% of borrowers' charged-off debts were still being reported one year later. By 2024 the share was about 80%. Twice as much old delinquency now stays visible in the data.

What happens when you take it out

The authors recalculate the stock rate excluding charged-off balances. On that basis all three measures line up, and delinquency rates are stable after 2024. Their conclusion is that flow rates give the more accurate view of current repayment behaviour, because the stock rate is partly a record of past defaults that have not yet aged off credit reports.

The quarterly report underneath

The same day, the New York Fed published its Household Debt and Credit report for the second quarter of 2026. Total household debt was $18.8 trillion, down $13 billion on the quarter and up $383 billion on the year.

• Mortgage balances: $13.117 trillion, down $74 billion on the quarter
• Credit cards: $1.263 trillion, up $21 billion on the quarter and $54 billion on the year
• Auto loans: $1.713 trillion, up $28 billion on the quarter
• Student loans: $1.651 trillion, down $7 billion on the quarter
• Home equity lines of credit: $459 billion, up $13 billion on the quarter

Transitions into serious delinquency were 6.97% for credit cards against 6.93% a year earlier, 3.00% for auto loans against 2.93%, and 7.83% for student loans against 12.88%. The share of all outstanding debt in serious delinquency was 2.57%, down from 2.91% a year earlier.

Delinquency rates across most products have held steady over the past two years. Still, new delinquencies for auto loans and credit cards remain elevated.

Why it matters

The practical lesson is about reading, not about credit. When a delinquency figure is quoted, the question to ask is whether it counts balances that are currently going bad or balances that went bad at some point and remain on file. The two answer different questions, and since 2022 they have pointed in different directions.