Private credit’s default rate is 2.51% and 6.0%. Both are correct.

Why do two default rates disagree by a factor of two?

Because they are answering different questions, and only one of them counts a maturity extension. Proskauer’s Private Credit Default Index printed 2.51% for Q2 2026, down from 2.73% in Q1, across 716 loans representing $195.6 billion of original principal (Proskauer Rose, 28 July 2026). Fitch’s US private credit default rate printed 6.0% on a trailing-twelve-month basis at the end of the same quarter, a record, up from 5.7% (Fitch Ratings, 30 July 2026).

Fitch’s own composition explains most of the gap. Of 32 private credit default events in Q2 2026, involving 20 new defaulters across roughly 1,300 tracked US borrowers, more than half were maturity extensions rather than missed payments or interest deferrals.

Moody’s states the mechanism outright rather than picking a side: the 2025 private credit default rate likely ranged between 1.6% and 4.7% depending on whether distressed exchanges are included, and distressed restructurings accounted for approximately 65% of all defaults that year (Moody’s Analytics, 28 April 2026). A three-point spread, created by a definitional choice.

Between them sits the only series that explicitly counts rescues. KBRA’s Middle Market Default Monitor, covering 2,785 sponsored middle-market borrowers and more than $1.2 trillion of direct lending debt, counts payment default, imminent default, and cases where sponsor or lender intervention prevented a payment default. Over the twelve months to 30 June 2026 it reached record highs of 92 borrowers and more than $30 billion of debt: 3.3% by count and a peak 2.4% by debt (KBRA, Q2 2026 Middle Market Compendium, 28 July 2026).

So the answer to which number is right depends on what is being decided. A lender quoting a falling default rate is quoting hard credit events. A rating agency quoting a record is counting every accommodation a creditor was forced to grant. Both are true at once, and the difference between them is the entire subject of this outlook.

Is this a credit event or a liquidity event?

A liquidity and funding event that has not yet become a credit event. Payment defaults are genuinely low and the cures are genuinely expensive, which is a specific and unstable condition rather than a comfortable one.

On the benign side: the leveraged loan payment default rate was 1.34% by issuer count at 30 June 2026, below its 1.51% ten-year average, and 0.97% by par amount, having fallen mainly because one large June 2025 default rolled out of the trailing window (PitchBook LCD, 10 July 2026). Covenant defaults in direct lending were 3.1% in Q1 2026, in line with the average since 2020 (Lincoln International, 7 May 2026).

On the other side, the cures are running hot. Direct-lending amendment activity rose 13% quarter on quarter in Q4 2025, with maturity extensions up 14%, covenant holidays up 14% and sponsor equity infusions up 31% (Lincoln International, 11 February 2026, as at Q4 2025). Payment-in-kind interest was present on 10.6% of all direct-lending loans and represented 8.9% of total interest income at Q1 2026. Loans that carried no PIK at close but carry it today reached 5.9% of all loans, against 2.5% at the end of 2021, and that cohort’s average loan-to-value has risen from 39.4% at inception to 76.1% (Lincoln International, 7 May 2026, as at 31 March 2026). Lincoln’s own description of that metric is a shadow default rate.

The indicators that cannot be amended point the other way. The share of the Morningstar LSTA US Leveraged Loan Index trading below 80 cents was 6.87% in June 2026, against 3.06% a year earlier, having peaked at 7.23% in March 2026, the highest since December 2022. The loan facility downgrade-to-upgrade ratio rose back to 1.25x in June from 1.18x in May (PitchBook LCD, 10 July 2026). B-minus spreads have widened 57 basis points since Q4 2025 to S+411 while BB-minus and B-plus spreads barely moved (PitchBook LCD, Q2 Leveraged Loan Market Wrap, 30 June 2026). The market is not repricing credit generally. It is repricing one rating notch.

The stress is also concentrating in larger credits. Of roughly 5,000 companies held by business development companies at 31 March 2026, 538, or 10.6%, showed signs of credit pressure, a 15% increase in the quarter, while the volume of first-lien and unitranche investments under pressure rose 44% to $35.4 billion. Half of those 538 companies did not use PIK in the preceding twelve months, which means PIK-based early-warning screens miss half the problem (PitchBook LCD, analysis of more than 170 BDCs, 21 July 2026).

One number decides which way this resolves. KBRA reported median interest coverage holding at 1.6x, but the share of borrowers with improving coverage plateaued after more than two years of gains, and median EBITDA growth fell to 24% from 27%, the largest quarter-on-quarter decline in the series (28 July 2026). Earnings growth has been doing the deleveraging work since 2023. That is the engine that just started slowing.

Where does the schedule actually fall?

In 2028 and 2029, and it was put there deliberately in the first half of 2026. The near-term cliff has been cleared; the cost is that the restructuring cycle has been rescheduled to years whose base rates and spreads nobody can currently forecast.

The Morningstar LSTA US Leveraged Loan Index stood at $1.57 trillion outstanding at 30 June 2026. Loan volume maturing through the end of 2027 narrowed to $32 billion from $62 billion at the end of 2025, while loans maturing in 2029 and beyond grew by $129 billion over the same six months (PitchBook LCD, 17 July 2026). The exception is the sector that needs it most: the roughly $40 billion 2028 software maturity wall has barely moved this year.

The instrument doing the work is amend-and-extend. There was $106 billion of it in H1 2026 against roughly $84 billion in H1 2025, an increase of 26%, with $27 billion in June alone across 24 transactions, and institutional volume of $39 billion in Q2 2026, the strongest quarter in the recent series (PitchBook LCD, 17 July 2026 and 30 June 2026). LCD’s own framing is that sponsors raced to chip away at the 2028 wall before the window closes.

Sponsors are doing it in defence rather than in growth. Private-equity-backed borrowers drove 74% of institutional maturity-extension amendments in Q2 2026 while accounting for just 44% of new-money activity, against a five-year average nearer 70% (PitchBook LCD, 30 June 2026). The practical inference for a borrower is to assume no new sponsor equity unless it is explicitly committed in writing.

The most under-read number in the market is who is being extended. Of 2026 amendments, 30% went to issuers rated BB-minus or higher, up from 11% in 2025, while the B-minus share fell to 27% from 44% (PitchBook LCD, 17 July 2026). Record volume and a collapsing B-minus share together mean the weakest cohort is being cut out of the extension market, not served by it.

Private credit’s wall sits in the same two years. A review of SEC filings from 74 business development companies found only about $15 billion of $84 billion of analyzed assets maturing in 2026, with the bulk peaking in 2028 and 2029 (Reuters, 1 May 2026, on Q1 2026 filings). In the middle market, maturities through 2026 fell to 6% of borrowers by count and about 2% of debt (KBRA, 28 July 2026), which is a good statistic for the market and a bad one to be inside: the straightforward extensions have already been done.

And extension is not repeatable by default. Lincoln International estimates that 30% to 40% of direct lending deals maturing in the next two years have already extended once, and frames the alternatives as an incremental extension or a restructuring. Leverage has risen roughly half a turn from inception across all vintages, and closer to a full turn for surviving 2019 and 2020 vintages (Lincoln International, 11 February 2026).

What happens when an extension is not available?

Increasingly, a change of control rather than a paper exercise. The out-of-court tool that defined the last three years is being used far less, and creditors are taking assets instead.

Sixteen index issuers conducted liability management exercises in the twelve months to 30 June 2026, against 36 in the twelve months to June 2025, a 56% decline and the lowest count since August 2023, with none at all in June 2026. Their share of the trailing default landscape fell to 52% from a peak of 73% in July 2025 (PitchBook LCD, 10 July 2026). What is left has turned consensual: of eleven exercises that closed in Q1 2026, two were pro rata, two were drop-downs, four were deals away and three were amend-and-extends, with no classic uptiers (9fin, Q1 2026 LME update).

The candidate pool has not shrunk with the completions. LevFin Insights counted 50 issuers at near-term risk of a liability management exercise in January 2026, against 51 in January 2025 (CreditSights and Covenant Review, quarterly update through Q4 2025, 14 January 2026). Deferred, not avoided, and that deferral is 2027 supply.

The reason it matters is the re-default rate. Covenant Review and LevFin Insights have tracked 37 bankruptcies, 25%, ultimately resulting from 148 liability management transactions, without double-counting companies that ran more than one before filing (as at 31 December 2025). One in four of these ends in Chapter 11 anyway, and practitioners describe the 2021 and 2022 vintage transactions now defaulting a second time.

Meanwhile lenders are simply taking the keys. Direct lenders foreclosed on $24.2 billion of principal in 2025 and a further $15.2 billion in the first quarter of 2026 alone, against $13.6 billion across the preceding three years combined, with nearly 75% of those change-of-control transactions relating to 2021 and 2022 vintage deals (Lincoln International, 7 May 2026, as at 31 March 2026). A single quarter at $15.2 billion annualizes to about two and a half times the whole of 2025.

Below the rated universe the stress has already converted into filings. There were 4,589 commercial Chapter 11 filings in H1 2026, up 28% from 3,595 in H1 2025, and 1,663 Subchapter V elections, up 50% from 1,107 (Epiq AACER and the American Bankruptcy Institute, 8 July 2026). Subchapter V is the lane for businesses under $7.5 million of debt. Commercial Chapter 11 filings rose 28% in the same period that the leveraged loan payment default rate fell.

Fitch expects the published rates to catch up. Its July monitor states that the H1 decline in US leveraged loan and high-yield default rates was driven mainly by base effects rather than credit improvement, and that default activity is expected to accelerate in the second half of 2026 (Fitch Ratings, 17 July 2026). Its full-year 2026 forecasts of 4.5% to 5.0% for leveraged loans and 2.5% to 3.0% for high yield were reaffirmed unchanged on 16 July 2026.

What is a first lien worth if it goes wrong?

Somewhere between roughly 50 cents and 90 cents, and which end applies is knowable in advance from the facts of the credit rather than from an average. The two most current readings disagree by forty points and both are current, so they are worth holding side by side rather than blending.

Octus analysis of private credit restructurings puts average recoveries at approximately 50 cents on the dollar, well below the 70% many managers had cited, and names the cause as follow-on restructurings compounding the deterioration of recovery values on original principal (Octus, 11 May 2026, on the Q4 2025 reporting period). Fitch, measuring its own privately monitored ratings portfolio, described realized lender losses as minimal in 2025, with recoveries in several cases estimated at 70% to 90%, against a 9.2% default rate in that portfolio (Fitch Ratings, 6 March 2026, full-year 2025 as trend context).

The two are reconcilable. Consensual, first-time, single-lender-group restructurings in undisrupted sectors still clear at the top of that range: Q2 2026 dental restructurings produced 45% to 65% first-lien debt reduction with 70% to 90% first-lien recoveries per BDC filings (Octus). Repeat restructurings and disrupted sectors drag the average to the bottom of it. Which one a company is determines the recovery, and that classification is usually visible before the negotiation starts.

Sector makes the difference stark. Excluding asset-based recoveries, par-weighted first-lien recoveries on claims against retail companies with at least $100 million of first-lien debt averaged approximately 49% from 2016 to 2025, and term loan lenders recovered nothing at all in three named retail bankruptcies, with nearly half of retail filings ending in liquidation (Fitch Ratings, 10 June 2026). Fitch has separately reported that US first-lien recoveries fell to a ten-year low in 2025, with liability management transactions supplying new financing ahead of existing first-lien debt as the mechanism. The precise figure sits behind a paywall and is not stated here; the direction is confirmed across three independent reports.

One more variable belongs in the calculation, and it is about the lender rather than the borrower. Non-traded business development company redemption gates were still binding in Q2 2026. Two large vehicles fielded $4.7 billion of tender requests in the quarter, down 13% on Q1; where requests reached 18.8% and 38.1% of shares outstanding against a 5% quarterly cap, roughly 27% and 13% of each shareholder’s request was honoured (AltsWire, 6 July 2026). A fund managing quarterly gates has structurally less appetite to fund a delayed draw, a covenant holiday or a rescue tranche, whatever the credit merits.

What should an owner or sponsor do about a 2028 maturity?

Start the work now, and start it as a management exercise rather than a financing one. The schedule is known, which is unusual, and it removes the excuse for being surprised by it.

Three things are worth doing before any notice period. First, establish where the credit sits in the extension queue: rating notch, coverage, whether it has already extended once, and whether the sector is one lenders are currently declining. Those four facts predict the answer better than any relationship does. Second, obtain every lender’s mark. Octus documents pricing gaps of nearly forty points on the same widely held stressed loan between funds holding it, which means two lenders in the same structure can hold irreconcilable views of what a business is worth before a proposal is even tabled (11 May 2026).

Third, build the operating plan the creditor will ask for. Restructuring practitioners report that private credit lenders now commission carve-out-style hundred-day transition plans before taking control (Octus, 29 July 2026). That document is a description of how a business runs without its current owner, and it takes longer to produce well than a notice period allows.

The macro will not do the work. The Federal Reserve held at 3.50% to 3.75% on 29 July 2026 by nine votes to three, with all three dissents in favour of raising by 25 basis points. Any plan whose trigger is a lower base rate is a plan without a date.