Loading today’s stories
Invest1 distinct publisher3 min readPublished
Fed, Equifax and Fitch data for Q2 show delinquency falling across prime and subprime. The 90-plus rate that argued the opposite was stale charge-offs, according to the New York Fed.
The Investor · Invest desk
Compiled by The InvestorSomething wrong?How this is made
A stock delinquency rate and a flow delinquency rate are not the same instrument. The 90-plus figure that carried the consumer-cracking argument through the past year [8] counts what is sitting in the bucket on the measurement date. If lenders leave the same charged-off accounts on credit reports for longer, the ratio climbs even when no additional borrower goes bad. The New York Fed's own reading, published with its Q2 household debt report [5], is that the increase came from "a pool of stale, charged-off debts that lenders have been reporting for longer durations, rather than a fundamental worsening in the incidence of delinquency" [6]. Before the pandemic those balances came off reports sooner and dropped out of the Equifax-derived rate, so the longer reporting is the new behaviour [7].
Take that artifact out and the earlier buckets, which respond to fresh distress rather than to filing habits, agree with each other. The Fed's 30-plus rate for all commercial banks is down 19 basis points in a year and 37 over two [1][14]. The Equifax 60-plus rate, which includes store cards and subprime accounts, is down 18 basis points and 35 [2][15]. Fitch's prime card ABS series sits at 0.84%, about a third of the all-card rate [4][16]. The subprime-inclusive number is the one that matters here, because prime cardholders were never the argument.
Two caveats belong on the table. The Equifax public series only starts in June 2020 [3], so "lowest since the free money era" describes a six-year window rather than a cycle. And the Fed's 2.85% is seasonally adjusted while the Equifax 2.69% is not [1][2], so the three-year-low claim properly belongs to the bank data.
The balance side gives the caution thesis even less to work with. Statement balances of $1.26 trillion, up 4.5% year over year [9], are measured before payments are made, and most of that spending is cleared by the due date without accruing interest [13]. Set the growth rate against throughput: Nilson put 2025 card payment growth at 6.1% [11], and Visa reported US cardholder volume up 9% through Q1 2026 after 6.8% the year before [12]. Balances are growing roughly 1.6 points slower than the money crossing the same rails [17]. Against Wolf Richter's $6.9 trillion estimate for 2025 volume [11], which follows a measured $6.51 trillion in 2024 [10], outstanding balances work out to about 18% of annual volume, or 2.2 months of card spending [18]. Richter also reckons cards are $4.3 trillion short of tapped out on unused limits [19].
None of this says household finances are comfortable, and one improving quarter across three data providers is not a cycle call. It does say that the series most often used to argue deterioration was tracking how long lenders keep dead accounts on a credit file, and that every card measure of new distress moved the other way in the same quarter. An argument for a cracking consumer now has to be built somewhere other than card performance.
Ranked by verification strength, evidence, and original report placement.
The 30-plus days delinquency rate on credit cards issued by all commercial banks declined to 2.85% in Q2, seasonally adjusted, the lowest since Q2 2023, down from 3.04% a year earlier and 3.22% two years earlier, according to Federal Reserve data based on regulatory reports filed by all commercial banks.
For prime-rated cardholders, the 60-plus days delinquency rate declined to 0.84%, the lowest since the free-money era and well below any time before it, according to Fitch Ratings data on ABS backed by prime credit card balances.
The New York Fed released its Q2 Households Debt and Credit Report earlier in August and addressed the 90-plus day delinquency question in an accompanying blogpost.
The New York Fed wrote: "We find that the stock delinquency rate is rising because of a pool of stale, charged-off debts that lenders have been reporting for longer durations, rather than a fundamental worsening in the incidence of delinquency."
In pre-pandemic years banks removed stale charged-off debts from credit reporting sooner, so those old debts disappeared from the Equifax 90-plus day delinquency rate; keeping them on reports for longer is a new trend.
The 60-plus days delinquency rate across all credit cards, including private label store cards and subprime cards, declined to 2.69% at the end of Q2, down from 2.87% a year earlier and 3.04% two years earlier, according to Equifax (not seasonally adjusted).
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Named institutional data behind the core numbers, one publisher relaying it
The quantitative core is attributed to identifiable primary sources — Fed regulatory filings for the 2.85% bank 30-plus rate, Equifax for the 2.69% all-card 60-plus rate, Fitch prime ABS for 0.84%, a direct New York Fed quote for the stale-charge-off finding, and Fed/Visa disclosures for payment volume. That is strong sourcing. It is capped below the high band because every figure reaches us through a single publisher with no independent corroboration in the cluster, the supplied body is truncated before the credit-limit and debt-to-income evidence, and the interpretive claims (revolving share, '$4.3 trillion of headroom', how widely the 90-plus chart was cited) carry no figures at all.
Card usage disclosed at scale by both a regulator study and a network
Real-world usage underlying the story is directly disclosed rather than inferred: the Fed payments study measured $6.51 trillion of 2024 US card payment volume (+11.7% over two years) and Visa separately reported US cardholder volume up 9% year over year through Q1 2026, while the NY Fed's Q2 report puts statement balances at $1.26 trillion. That is broad, independently disclosed usage. It is not higher because the 2025 volume figure central to the article's framing is an extrapolation of a third-party estimate rather than a measured release.
Data claims land; the headline reassurance runs ahead of what is shown
Mildly overstated. The delinquency and payment-volume claims are specific and properly attributed, and the New York Fed quote genuinely undercuts the 90-plus day chart, so the core is not hyped. The gap comes from the framing layered on top: the lede's '$4.3 trillion away from tapped out' has no supporting credit-limit arithmetic in the supplied text, the 'balances measure spending, not borrowing' conclusion depends on an unquantified assumption about how much revolves, and the claim that the 90-plus chart was 'endlessly cited' is asserted without a single example. Those push the reassurance further than the presented evidence supports, while the underlying quarterly improvement is real.
No disclosure of publisher funding or positions in the supplied material
The cluster contains one self-published commentary piece and no information about its monetization, sponsorships, subscription model, or the author's positions in consumer-credit or payments assets. The cited data providers (Fed, Equifax, Fitch, Nilson, Visa) also have commercial relationships to the credit and payments industries that the supplied text does not characterize. Rather than infer motives from tone or headline style, this dimension is left unmeasured.
Confident on the quarterly data, guarded on the conclusions drawn from it
High confidence that the reported Q2 delinquency rates, balance figures and the New York Fed's stale-charge-off explanation are as described, since each is attributed to a named institution and the NY Fed language is quoted directly. Confidence is held to the middle band by structural limits: one publisher, no independent corroboration, a truncated body that cuts off the credit-limit and debt-to-income evidence, an unverified third-party estimate carrying the 2025 volume framing, and the absence of any revolving-share or flow-based delinquency figure that would test the 'consumer is fine' reading.
invest
The card networks just picked the referee for agent checkout, and it looks like EMVCo2 distinct publishers
invest
Treasury moved $742 billion in a week. The price was a 5.216% thirty-year.1 distinct publisher
invest
1,051 trades in one month, and a $185 million window where the number should be1 distinct publisher
invest
Central banks have sold $233 billion of Treasuries since February. Private money did not blink1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.