Invest1 publisher3 min readPublished
Spreads on the weakest US corporate credit widened to 10.53 points from 8.08 a year earlier, but with defaulted bonds returning 29 cents instead of the usual 40, the extra 2.45 points buys a lender less cover than it did.
The Investor · Invest desk

Compiled by The InvestorSomething wrong?How this is made
A spread is a price for a probability, so do the division. Take the 10.53 points [1] and set it against what a lender actually loses when a triple-C bond goes bad: at a 29-cent recovery over the trailing twelve months [5], the loss is 71 cents, and 10.53 divided by 0.71 covers a roughly 14.8% annual default rate [4]. Run the same spread against the 40-cent average of the past 25 years [5] and it covers 17.6% [5]. The spread widened 2.45 points, or about 30% on the 8.08 of a year ago [1][2], and a meaningful part of that went to standing still on severity rather than to buying new protection.
Fifteen percent [3] is a price. It becomes a constraint at the moment a bond matures and has to be reissued, and there is no maturity schedule here for the lowest-rated cohort. That absence means the claim that refinancing costs are now binding on the weakest balance sheets is only an inference, since nothing here measures it directly. What is measured is where the damage sits: most default-related actions occurred at companies at the bottom of the ratings scale [8], and Diamond Hill's Henry Song says firms built in a zero-rate environment "are essentially just hanging on" [9]. Song's comment offers a portfolio manager's characterisation of that stress; it does not itself document a maturity wall.
The policy path pushes the same direction. CME FedWatch, as of 4:35pm Korean time on the 6th, put a 25 basis point increase this month at 59.4%, ahead of the odds of a hold [12], and a December move to a 4.00% to 4.25% target at 36.3%, 50 basis points above the current level [13], which places the current range at 3.50% to 3.75% [6] and triple-C money at about four times the top of it [7]. August payrolls at 162,000 against a 55,000 median Bloomberg forecast [14] came in near three times the estimate [8]. Note also that the 15% is not an independent fact: 4.82 [4] plus 10.53 is 15.35 [11], so most of what the weakest borrowers pay is the Treasury base plus a spread that has to clear a 29-cent recovery.
The reporting carries both readings. Ninety One's John Stopford calls the pattern a possible canary for a wider risk-asset sell-off [16], and 22V Research's John Roque says something always breaks when the 10-year moves this much [17]; against that, cable and satellite recorded the most default-related actions, ahead of industrials and paper and packaging [18], and a levered industry in secular decline defaults at any funding cost. Corbin Capital's John Koch adds that years of aggressive debt restructuring have eroded investor confidence in over-levered borrowers [19], which is also a serviceable explanation for 29 cents, since creditors moved down the priority stack recover less. Composition plus recovery mechanics is the better-supported read; it fails the day default actions show up in a sector with no secular story to blame.
Ranked by verification strength, evidence, and original report placement.
The 10-year US Treasury yield recently rose to 4.82%, its highest since 2023, which the Financial Times attributed to persistent inflation combined with heavy borrowing by governments and artificial intelligence companies.
The CME FedWatch tool, as of 4:35pm on the 6th Korean time, put a 59.4% probability on the Federal Reserve raising its benchmark rate by 25 basis points this month, higher than the odds of a hold.
The same tool put a 36.3% chance on a December increase to a range of 4.00% to 4.25%, which is 50 basis points above the current level.
Spreads on debt issued by companies rated triple-C or lower widened to 10.53 percentage points from 8.08 percentage points a year earlier.
At that level, triple-C or lower rated borrowers must offer an average yield of about 15% to sell bonds.
Recovery rates on defaulted bonds over the past 12 months stood at 29%, well below the 40% average of the past 25 years.
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 providers, one set of eyes
Every figure is checkable in principle and unverified in practice here: the Financial Times for the 4.82% benchmark and the 10.53-point spread, JPMorgan for $40.1bn of default actions, CME FedWatch for the rate odds, the Bureau of Labor Statistics for the 162,000 payroll print. All of it reaches us through one relay, and the softest passage is also the one doing the most narrative work, with the widening gap between rating tiers and the direction of fund flows attributed to analysts who are never named.
No uptake to count
There is nothing to adopt in a spread story, and the closest analogue, where investors are actually putting money, is described only as a direction of travel toward strong credits with no fund flows, issuance volumes or investor counts attached.
Warning quotes run ahead of the tally
The language of detonation, canaries and 'something always breaks' sits in the same piece as $40.1bn of default actions up 9% and investment-grade spreads called stable, which describes damage confined to the bottom of the ratings scale rather than a market coming apart. The number that is undersold is the 29-cent recovery rate: it does more to explain why 10.53 points of spread is thinner protection than it looks than any of the quoted warnings manage.
Buy-side voices, undisclosed books
Four of the five attributed views come from firms that run or sell credit positioning: Corbin Capital Partners, Ninety One, Diamond Hill and 22V Research. A manager holding higher-quality paper that still pays a high coupon does better the longer the flight to quality described here continues, and no position is disclosed for any of them. The default tally comes from a bank's research desk rather than a ratings agency or a regulator.
Internally consistent, externally unchecked
Two things hold this in the middle. The spread, yield and recovery figures are specific and they reconcile, with 4.82% plus 10.53 points arriving at 15.35% against the roughly 15% average yield quoted. Set against that, one outlet at one remove carries the whole story, and the rate section asks a reader to accept a 59.4% chance of a hike this month on futures pricing alone, with no Federal Reserve language and only an August payroll surprise as support.
product
Nebius funds $4.5bn of AI capacity on terms that pay lenders mostly in stock2 publishers
invest
10-year yield has added 80 basis points since the Fed started cutting1 publisher
product
Anthropic put a finished quarter on the table. Procurement should notice why.1 publisher
invest
Nvidia's $500bn GPU pool makes the seller the guarantor of its own demand1 publisher
Publishers with included, body-backed reporting in this cluster.
1 article · September 6, 2026