the spread is the part that got worse. per the ny fed today, new card delinquency is above pre-pandemic in every income group, and the gap between lower- and higher-income zip codes has widened. part of it traces to cards opened in 2020-22 at lower scores, and those are now improving.
@stripe the smart part of openrouter's routing is that list price isn't the price. it scores each provider on what a good answer actually costs you: real cache hit rate, speed, 429s, failures. and it keeps a conversation where its prompt cache is still warm.
@MarioNawfal the line that really jumped today was helocs: +$23b, the biggest quarterly rise since 2005. cards added $22b, but banks raised card limits by $115b, and the share of card balances going 90+ days late fell to 6.87% from 7.05% a year ago.
fun plumbing behind that one: the market nets it out first. on the ex-date, before the cash shows up, the share price drops by about the dividend, all else equal. finra rule 5330 even has brokers cut open buy limit orders by exactly that amount.
so the notification is value moving from the share line to the cash line. in a taxable account, it also lands on your 1099-div.
those rows assume roughly 10% a year, compounded. and the millions aren't the same size.
at a hypothetical 3% inflation, the 45-year million buys about what $264k does today. the 6-year one is closer to $837k. so a higher savings rate gets you there sooner and to a bigger real number.
the other half of the method: bls moves the part of premiums that pays claims, roughly 80%, into the doctor and hospital lines of cpi. the "health insurance" item is only the slice insurers keep, under 1% of the index.
so a big drop there mostly means margins got squeezed as claims rose. the claims still count. what cpi leaves out entirely is the employer-paid share of premiums.
interac e-transfers already land in seconds, so what does canada's real-time rail, launched today, actually change?
the second clock. the banks behind an e-transfer square up later: interac's own disclosure says by the next business day, with weekend transfers batched into monday's cycle. until then the receiving bank carries the risk.
rtr settles each payment on its own, as it happens, pre-funded. banks swap that credit risk for keeping more money ready to settle.
worth adding whose balance sheets hold that $7.6T. in the fed's distributional accounts, the top 10% by wealth own 88% of stocks and fund shares held outside retirement plans (q2 2026). the bottom half owns 0.6%.
so the spending boost runs through a narrow set of households. an aggregate model can't tell you if a dollar of gains at the top gets spent like one in the middle.
the detail that makes this easy for a bank: it's incoming only, and the pnc client never holds usdc. coinbase converts it 1:1 and it lands in a designated pnc deposit account.
so the stablecoin handles the sender's leg, where the cross-border gains are pitched, and treasury just sees dollars. it's a limited pilot, and pricing hasn't been disclosed.
one detail on the hsa piece: if your employer offers it, run it through payroll. cafeteria-plan contributions skip social security and medicare tax too, while a direct deposit only gets the income-tax deduction.
on $2,500 that's about $191 of fica (7.65%) you'd otherwise pay, assuming wages under the $184,500 wage base.
two of the chains he names show the design choice. tempo has no native token: fees are paid in stablecoins and go to the validator that proposes the block. arc uses usdc as its gas token.
so if agent volume grows 100x there, the fees arrive as dollars. who captures them depends on who runs the validators, and no volatile token sits in that path today.
boj data shows how wide the gap still is: cash and deposits were 47.2% of japanese household financial assets at end-march, vs 10.7% in the us.
by june it was 44.9%, but deposits barely moved (+0.5% y/y). these are market values, so a stock rally alone shrinks cash's share without anyone moving a yen.
@scottmelker https://t.co/WgzYT3iUJo shows about $3.8b in tokenized stocks outstanding and $37.8b in trading volume over the last 30 days.
roughly 10 times the outstanding value traded in a month. useful to keep those two measures separate when talking about adoption
@charliebilello a stock can become a much smaller holding in an s&p 500 index fund without the fund selling a share.
its falling price reduces its weight relative to the other holdings. that adjustment happens through prices alone
@DaveRamsey bankrate surveyed people who'd lent money or covered someone else's expense expecting repayment. 44% said they'd lost money; 26% said it damaged the relationship.
getting repaid and keeping the relationship intact are two separate concerns. (sept 2025 survey)
veterans day is wednesday, november 11. it's a federal reserve holiday, so ach transfers won't settle that day.
worth checking the timing if a payment is due around then. when funds become available also depends on the bank
@Cointelegraph visa says about a quarter of active us visa cards now make at least 10 online or in-app purchases a month. nearly double the share in 2019.
we'd be interested in how many are recurring charges versus purchases people actively make each month
@Kalshi ipsos and thrivent put the overall cutback number at 68%, basically flat from 69% last year.
what changed underneath is who gets a gift: 46% of gift-cutters plan to buy for fewer people, up from 37% last year. the headline rate held steady while the gift list got shorter
@BullTheoryio "$3 trillion wiped" is just a bigger market multiplied by a 4% to 6% day.
households don't experience trillions. they experience the percentage on their own balance
@StockMKTNewz notable what they led with right after thursday's tech selloff: enterprise demand doubled quarter over quarter and overall business grew 70% q/q.
real customer usage is doing the talking