Payrolls beat lifts Fed hike odds to 59%
A 162,000 print against a 53,000 forecast put a September hike back in play while the thirty-year barely moved. The same backdrop is reshaping corporate supply, data centre paper, credit’s public scoreboard and who owns your order data.
Payrolls beat lifts hike odds while the long end stalls
August non-farm payrolls came in at 162,000 against a consensus of about 53,000 — Reuters' poll had 56,000 — the strongest month since March. Unemployment held at 4.1%, average hourly earnings rose 0.3% on the month to $37.75 for annual growth of 3.1%, and July's initially reported 23,000 decline was revised to a 21,000 gain. Futures pricing for a hike at the 15-16 September meeting moved from 52% before the release to 59% after. The curve did almost all of that at the front: two-year yields rose 5.3bp to 4.39%, ten-year 1.8bp to 4.782%, and the thirty-year half a basis point, to 5.2475%. The long bond arrived at this print already stretched — it touched 5.34% in mid-August, the highest since 2007, and has spent 55 days above 5% this year, the most in any year since 2006 — and Treasury's expanded buybacks, at least $4bn per operation across the 10-to-20 and 20-to-30 year buckets, begin on 9 September and run to 4 November. John Briggs of Natixis called those operations "a drop in the bucket"; Gregory Faranello, quoted by Bloomberg, put the alternative plainly: "If you want to get the long end down, you tighten rates." Options pricing implies a thirty-year near 5.7% by late November.
The buyback question has moved twice in three weeks — first size, then funding — and this week the market answered both by pricing a rate hike instead: the front end did more than ten times the work of the long bond on a number that came in at three times consensus. That is the market saying the long end is a term-premium problem no published operation is targeting, which leaves anyone hedging long-dated liabilities carrying a risk the front end does not proxy, in a bear-flattener where the cheap hedge and the crowded hedge are now the same trade.
Reuters via Yahoo Finance · InvestmentNews · Bloomberg via Advisor Perspectives · US Treasury
Issuers pull 2027 funding forward into a record September
Dealers expect US high-grade supply of as much as $250bn this month against last September's record $207.5bn, with Bank of America's strategists at roughly $190bn — still the second-highest month on record — and other estimates around $215bn. Last week produced $53bn of US investment grade issuance, with up to $60bn guided for the current week; year-to-date supply stands at $1.3trn, up 4%, against $212bn of US high yield, up 5%. Demand has not thinned yet: IG funds took $3.8bn of inflows on the week and $49bn year to date, while high yield saw $250m out. Europe has run above €40bn for three consecutive weeks, taking euro IG to €521bn, up 12%. The motivation is defensive rather than opportunistic. Tom Murphy of Columbia Threadneedle: "If I was a CFO or treasurer and had something to do in 2027, I'd probably pull it forward into 2026." Moshe Tomkiewicz, head of investment-grade DCM at Mizuho Americas, called it a "'devil you know' type of situation where you'd rather issue now with credit spreads still relatively tight." Lesya Paisley of MacKay Shields countered that "caution and discipline in terms of being selective and demanding proper compensation are warranted."
Pre-funding on this scale means the index is being reshaped by treasurers' hedging decisions rather than by anything the market is signalling — duration, sector weights and issuer concentration all drift on a calendar set outside the portfolio. The operational consequence is sharper than the analytical one: at $53bn a week against $3.8bn of fund inflows, primary allocation is where risk is cheapest to source and secondary is where it is dearest, so the binding constraint becomes how much new issue a desk can actually process in a session — credit approvals, limit checks, allocation reconciliation — and that capacity is a headcount and workflow number nobody budgeted for a record month.
Bloomberg via Advisor Perspectives · The DESK
Data centre paper meets its first real buyer resistance
Data centre ABS has passed $10bn of issuance in 2026, the leading esoteric asset class, after $15.5bn in 2025 against $2.4bn in 2020. Two deals this week showed the bid thinning. GlobalCapital reported on 3 September that spreads have widened and that GI Partners' $375m Chicago data centre CMBS, guided on 28 August, "is moving slowly"; the same day, Vantage gave investors an early look at an ABS ahead of an anticipated October repayment date on a 2021 deal. Blackstone's QTS showed a $280m BBB- tranche of a $1.47bn securitisation to selected accounts roughly two weeks ahead of syndication, and investors told GlobalCapital the long lead time was recognition that the deal "may have got a rough reception if they had just gone straight to syndication." All of this follows the SEC staff position in August that data centre securities are not asset-backed securities — removing Reg AB 15-g disclosure filings and the five-business-day waiting period, on a petition from Latham & Watkins — after which Morgan Stanley research described data centre ABS spreads as "still near YTD tights despite the move in corporate markets."
Faster to market with less to disclose has arrived at precisely the moment the marginal buyer started asking for more time, and pre-marketing a mezzanine tranche a fortnight early is what an issuer does when the structured bid needs convincing, not when the document set is sufficient. The relief covered here in August has now had its first pricing test, and it lands on the buyer as work: with no standard disclosure package, cross-issuer relative value in the sector depends on diligence a manager performs itself — power contracts, tenant concentration, residual value on the equipment — which is analyst headcount and a data build rather than a screen, and it is being asked for exactly as spread stops paying for it.
GlobalCapital · GlobalCapital · GlobalCapital · InvestmentNews
August reporting leaves credit with one public scoreboard
Tradeweb reported August average daily volume of $2.8trn on $61.2trn of total volume, up 13.7% year on year. Fully electronic US credit ADV was $8.9bn, up 32.5%, taking a 19.0% share of fully electronic US high grade TRACE volume and 26.8% including total volumes; high yield ran at 8.3% fully electronic and 10.0% total. Global cash credit portfolio trading ADV rose 50.4%, with non-competitive portfolio trading up 113.1%. On the rates side, US government bond ADV was $282.5bn, up 28.9%, repo $836.3bn, up 11.5%, credit derivatives $17.7bn, up 51.4%, and municipals $523m, down 2.8%. MarketAxess has published no monthly statistics since the ICE acquisition was agreed; the last comparable public read, from The DESK on 13 August, put Tradeweb at 17.99% of TRACE in June against MarketAxess at 16%. Trumid, which released July figures, has not yet published August. CME, by contrast, reported its second-highest August on record at 29.7m contracts a day, up 6%.
What these releases share is the reporting cycle, and what they show is that it now produces a complete public series for futures and a one-sided one for credit — the missing half being the venue that was acquired. Every best-execution and venue-selection framework built on platform share is now comparing one participant's self-reported numerator against TRACE while its closest competitor has stopped answering, and the monthly releases carry no fee data at all, so the price half of that comparison cannot be made from anything published this week. Rebuilding the denominator from raw TRACE is a normalisation project with a real cost, and it is the kind of spend that only gets approved after a committee asks a share question nobody on the desk can answer.
Tradeweb August 2026 volumes · The DESK · CME Group
Two vendors buy the conversation layer, terms undisclosed
Sense Street said on 1 September that it has integrated with S&P Global Market Intelligence's IssueBook to turn unstructured investor-to-salesperson conversation into structured order data for bookbuilding, pulling order intent and state changes — increases, reductions, withdrawals — out of Bloomberg IB Chat, ICE and Refinitiv messaging and feeding interpreted orders into the book with human validation retained before execution. Chief executive Haroon Hassan framed the problem as "some very highly paid salespeople taking orders from clients in conversation and physically transcribing them into bookbuilding software," and was explicit that "you cannot take a human out of the loop." Pilots have run with one unnamed bank; no customers and no pricing were disclosed. The next day Nasdaq completed its acquisition of Dasseti, announced on 23 July, on undisclosed terms, folding AI-driven due diligence questionnaires, RFPs and ongoing manager monitoring into eVestment. Dasseti covers 17,000 managers and general partners representing $34trn; eVestment carries 4,800 contributing managers, more than 1,200 asset owners and intermediaries, and over 112,000 products.
Both deals take the part of the investment process that has always lived in prose — an order given in a chat window, a manager's answer to a questionnaire — and convert it into a structured record held on somebody else's platform. The closest comparison runs the other way: DirectBooks attacked the same primary workflow from the structured side, wiring 48 underwriters and more than 1,000 investor accounts into Bloomberg's TSOX in June, and Preqin already sits inside eVestment as a data partner, so the divergence here is that both of this week's moves go after the unstructured input rather than the message standard — the harder half, and the one with no public price on it anywhere. Neither disclosed terms, so explicit cost is unknown; the total cost of ownership is the sentence that matters, because the structured history of your own orders and your own diligence answers accrues to the vendor, and it gets priced at renewal, once the alternative is going back to transcription.
A-Team Insight · Markets Media · Nasdaq · Bloomberg and DirectBooks