Treasury doubles long-end buybacks

Treasury doubles long-end buybacks and the market erases it

Treasury announced on 19 August that the maximum size of nominal long-end liquidity support buybacks rises from $2bn to at least $4bn per operation in the 10-to-20-year and 20-to-30-year buckets, effective 9 September and running through 4 November, citing "consistent strong sponsorship from market participants" and the "significant volume of high-quality offers" it routinely receives in longer-dated operations. The 30-year fell roughly 10bp on the announcement, which Gennadiy Goldberg of TD Securities called "effectively the equivalent of verbal intervention from the U.S. Treasury." By the next session the whole move was gone, the 30-year back at 5.25% and the 10-year at 4.708%, and on Friday Scott Bessent told CNBC that "we are going to make a market in these" and that operations "could be more than $4 billion per issue." Wednesday's $16bn 20-year auction stopped at 5.204%, against 5.163% in July. Separately, Coalition Greenwich research published this week put electronic trading at 57% of US rates in July on $1.085tn average daily notional, and found that after the July FOMC announcement dealer-to-customer trades did not account for half of volume until thirteen minutes had passed.

Two things are now true at the long end at once: the marginal buyer of duration is a government programme with a published calendar, and the first quarter of an hour of price discovery around a scheduled event does not run through dealer balance sheet at all. Execution assumptions built on dealer liquidity around events — arrival-price benchmarks, event-window TCA, working a large duration hedge into a print — are measuring a market that has moved to ATSs and central limit order books, and the one dependable long-end bid is now the one every counterparty can see coming.

US Treasury · Axios · Yahoo Finance · Bloomberg · 20-year auction result · The DESK

Private credit's retail channel jams as non-accruals climb

LCD data published on 19 August put non-accruals at the ten largest BDCs at 3.95% of debt at cost, up 20bp on the quarter; across all 213 registered US BDCs, covering roughly $516bn of debt, the adjusted measure that captures every tranche of a borrower with one non-accrual reached 3.3%, up 116bp, with $17.3bn of adjusted non-accrual debt and around $772m of interest income at risk — about 204bp of total cash interest income. The Financial Times reported the median at the twenty largest listed BDCs rising to 2.8% from 2.0%, with FS KKR at 7.1% and Blackstone marking its Medallia loan below 50 cents from 60 cents three months earlier, while Fitch has the private credit default rate at a record 6.1% against 3.8% for leveraged loans and 2.7% for high yield. The wealth channel has responded faster than the credit has deteriorated: non-traded BDCs raised about $2bn in the second quarter, down 82% year on year, against a record $23bn of withdrawal requests, with many funds capping redemptions at 5%. PIMCO's Christian Stracke said wealth distributors "do not want to and cannot sell the direct lending private credit retail vehicles any longer" and that most BDCs have "a queue of around 15% of assets under management lined up to exit." Blue Owl cut direct lending to roughly 35% of AUM in July from about 50% two years ago.

The credit numbers are bad but not catastrophic; what broke is the willingness of a distribution channel to hold an asset whose price is an opinion, and a 5% gate is simply the product design becoming visible. Anyone holding or building evergreen private credit exposure is now running a liquidity line item that is a queue rather than a redemption, with no observable series to test a mark against — PIMCO's Lotfi Karoui told Benzinga that scepticism over reported NAVs is "unlikely to ease without a better mechanism for price discovery," and the BDC spread premium over broadly syndicated loans has compressed from over 300bp in 2017-18 to under 100bp, which is thin compensation for valuation you cannot verify.

PitchBook LCD · Private Equity Wire · Bloomberg via Advisor Perspectives · Private Equity Wire · Benzinga

Tricolor charges show warehouse collateral nobody could verify

The SEC on 18 August charged former Tricolor chief executive Daniel Chu, former chief financial officer Jerome Kollar and former senior finance director Ameryn Seibold over what it describes as a "multi-year scheme to defraud investors by double pledging hundreds of millions of dollars of subprime auto loans." Tricolor raised $1.9bn through ABS offerings and had $945m of principal outstanding when it filed for Chapter 7 in September 2025. David Woodcock, director of the SEC's Division of Enforcement, said the defendants "defrauded investors based on bogus collateral and violated the integrity of our private credit markets." Asset Securitization Report followed on 20 August with the scale: one lender's analysis identified roughly $365.5m of double-pledged principal across seven outstanding bond deals, and warehouse lenders have set aside more than $500m in losses. Kollar and Seibold pleaded guilty in December 2025; Chu goes to trial in October 2026 and his attorney, Matthew Schwartz, says many of the allegations "are inaccurate."

Strip out the fraud and what is left is an infrastructure fact: the same loan sat in multiple securitisations and multiple warehouse lines for years because no participant could see the other participants' collateral, and reconciliation ran on files the originator produced. The explicit cost of fixing that — independent verification agents, loan-level data feeds, custodian attestation — is a few basis points on a deal; the total cost of not fixing it is $500m of warehouse losses now repricing advance rates and diligence requirements for every non-bank originator that had nothing to do with Tricolor.

SEC · Asset Securitization Report

T. Rowe buys the fixed income wrapper it lacked

T. Rowe Price agreed on 20 August to acquire F/m Investments, which managed about $19bn as of 31 July across ETFs, institutional separate accounts, and taxable and municipal SMAs. The deal adds roughly 9% to T. Rowe's fixed income AUM and more than doubles its fixed income ETF assets, against a firm total of $1.87tn. Terms were not disclosed and the transaction is expected to close in early 2027. F/m chief executive Alexander Morris, who will report to T. Rowe global fixed income head Arif Husain, said the firm "needed a partner with relevant expertise, deep resources, and a shared vision" to "continue to innovate and provide client value at scale."

A $1.87tn manager does not buy $19bn of AUM for the alpha; what is being bought is the wrapper and the plumbing underneath it — ETF share-class machinery and SMA infrastructure that rebalances thousands of individual bond accounts against tax lots and client constraints, which is a multi-year data and operations build with headcount attached rather than a product decision. No price was disclosed, so the multiple on that capability cannot be established, and that is the useful part: the buy-versus-build question on fixed income SMA infrastructure is being answered with acquisitions whose cost nobody has to defend in public.

T. Rowe Price · InvestmentNews

Record muni supply meets disclosure that cannot price it

First-half municipal issuance reached $294.884bn, up 5.8% on the same period in 2025, with tax-exempt supply of $269.519bn up 8% and the year tracking towards a record near $600bn; J.P. Morgan now projects roughly $57bn of tax-exempt issuance in August alone, potentially a top-three month on record. It is being absorbed: muni funds took $759m in the week before last, a seventeenth consecutive week of inflows, with $64.7bn year to date, the second-highest pace on record behind 2021. Supply still bit mid-week, with yields cut up to 7bp against Treasury gains of up to 2bp as Los Angeles Department of Airports priced $2.69bn. Against that, Ceres published a review on 21 August of 60 recent offerings across 20 US metro areas with high climate exposure: one-third contained no mention of climate or extreme-weather risk at all, only 12% included quantitative climate metrics or targets, and only 15% disclosed internal accountability for managing the risk. Steven Rothstein of Ceres put it as "enormous climate risk built into the $4 trillion bond market for investors."

Record supply meeting record retail flow means paper clears on concession and structure rather than credit differentiation, and the Ceres numbers explain why it has to: where physical risk is most material the risk factor is often absent from the document entirely, so there is nothing to test against spreads even when the research process wants to. That gap is where the vendor line grows — overlay datasets bought to supply what official disclosure does not, then reconciled onto thousands of CUSIPs whose own offering documents say nothing about it, which is paying twice for one field, while San Francisco PUC's Nikolai Sklaroff points at the mirror-image cost, that the market "has historically failed to provide issuers with credit for resilience investments."

The Bond Buyer · The Bond Buyer · Octus · The Bond Buyer · The Bond Buyer