Matt King calls his framework “The Great Deferral.” The former head of credit strategy at Citi, who now runs Satori Insights, sees strong corporate profits and persistent investor inflows keeping markets resilient—even as fiscal pressure builds, investment commitments grow and some borrowers postpone confronting their problems.

“The risk hasn’t disappeared, it’s been pushed into the tails,” Matt King (Satori Research).

That does not mean every tight credit spread is wrong. One of the more interesting parts of our conversation is King’s willingness to defend much of the corporate market’s pricing. His concern is what happens when risks outside the corporate balance sheet become risks to the corporate balance sheet—and whether investors will get much warning.

In this episode of Fixed + Floating, we discuss France, government borrowing, the financing of AI investment, private credit and what happens when money rotating between assets becomes money leaving risk altogether.

WHICH MARKET IS WRONG?

We open with a contradiction: French five-year CDS at about 80 basis points, OAT spreads to Bunds at their widest since 2011, and European corporate credit still near its tights. Those are the market levels discussed at the time of recording. Which market is wrong?

“I’m not sure either is wrong,” King says. “But at a minimum, there’s an underlying tension here.”

For much of corporate credit, he argues, profits remain strong enough to support tight spreads. That distinguishes the current market from earlier periods when he thought credit prices had become disconnected from deteriorating fundamentals.

“Mostly, I’d say credit has been following its fundamentals and more so than previously”, Matt King (Satori Research).

The more obvious concerns sit in CCCs, parts of private credit and particular pockets of the market—not uniformly across the corporate universe.

The problem is that the risks investors worry about most are increasingly sovereign, political and macroeconomic. Those risks do not necessarily feed into every corporate bond immediately. But there comes a point when the distinction stops holding.

“It’s that that gives rise to this underlying tension of, gosh, my spreads are really tight, but I still have this macro world where I’m worried about everything.”

THE BORROWER OF CONCERN HAS MOVED

King jokes that he has spent his career looking at bubbles—not through choice, but because the source of concern keeps moving.

Around the dot-com period, it was companies. Before the financial crisis, it was households, mortgages and the institutions financing them. Increasingly, it is governments.

“It’s governments where we worry about reaching a tipping point and that risk spilling over,” Matt King (Satori Research).

The uncomfortable part is the scale. Each successive source of concern involves a larger pool of debt. Sovereign distress can reach much further than problems confined to one segment of corporate borrowing.

Investors have also become accustomed to increasingly fast rescues.

King recalls about five weeks before aggressive easing and liquidity support around LTCM, about three weeks before measures such as TARP after Lehman, about five days during the UK gilt crisis, and a weekend for the SVB rescue package. These are his recollections, rather than a standardized comparison of policy-response times.

“We’ve kind of become addicted to government rescues and central bank rescues,” Matt King (Satori Research).

Meanwhile, governments struggle to take measures that improve long-term debt sustainability but impose short-term costs.

“The moment you try and do the right thing for the long term, some special interest group cries that they’ve been wronged, and then the governments end up backtracking.”

His conclusion: “This cycle will not be broken until the bond market shows who’s boss,” Matt King (Satori Research).

FRANCE AND THE KINK IN THE CURVE

For France, King sees a political constraint as much as a financial one.

“Voters have no appetite for seriousness, and any attempt at seriousness causes the incumbent to be thrown out”

He does not simply blame the governments under President Macron. He argues that they have attempted modest pension reforms and other measures, only to face opposition and retreat. Nor does he think lower energy prices alone would resolve the underlying fiscal problem.

For credit investors, the question is when sovereign pressure starts overriding company-specific fundamentals.

King points to work from the 2011–12 sovereign-debt crisis. At lower sovereign-CDS levels, corporate spreads could trade primarily on their own fundamentals. Around 120 basis points, the relationship changed: companies—and especially banks—became more closely correlated with the sovereign.

Banks are particularly exposed through government-bond holdings and the possibility of deposits leaving the country. Non-financial companies can remain more insulated, but not indefinitely.

“When sovereigns get into difficulty, they do change taxes, they do change laws, they do resort to sudden nationalizations.”

The roughly 120-basis-point level is a historical reference, not a universal trigger. King nevertheless thinks it helps explain why France’s sovereign stress has not yet translated into a much broader corporate-bond repricing.

His presentation describes France as still “left of the kink—for now.”

THE BORROWERS KEEP ISSUING. THE BUYERS HAVE A CHOICE.

France is part of a wider change in the bond market: the largest borrowers are not responding to rising yields in the way investors might expect.

“Neither governments nor hyperscalers, so the big borrowers in fixed income, are yield sensitive for now, or sufficiently yield sensitive for now,” Matt King (Satori Research).

At the same time, buyers are becoming more price-sensitive.

Pension funds, insurers and central banks previously supplied substantial demand. Increasingly, the marginal buyer has more choice over where to allocate capital—and requires a higher yield.

“But the new buyers have a choice, and as a result, they demand a higher yield”

For governments, the pressure takes time to become visible because the interest bill adjusts as debt matures and is refinanced. But the mismatch between borrowers that keep issuing and buyers that demand more compensation can compound.

For hyperscalers, the mechanism is different. Expectations of future profits matter more than a modest increase in financing costs.

“What’s different about the hyperscalers is that, as the spreads and the yields have gone up, they haven’t cared”, Matt King (Satori Research).

His deck describes the twist in this corporate cycle as “the capex, not the debt”: profits still outpace debt, while the investment is unprecedented.

This is not simply a story of already weak borrowers stretching their balance sheets. Strong issuers can sustain a borrowing cycle for longer—even while their issuance puts pressure on the rest of the market.

THE AI BORROWING BOOM HAS BARELY STARTED

How much more debt might the hyperscalers issue?

King cites consensus estimates of roughly $1 trillion of further issuance by 2028, and $2 trillion-type figures from Epoch AI. These are estimates he references, not a forecast developed in the interview.

The more important question is what would make the borrowing stop.

In what Silicon Valley views as a winner-takes-all environment, he expects the weight of issuance to continue until something breaks. His most likely candidate is confidence on the equity side.

Elevated equity valuations can make growing debt appear manageable. Investment spending also supports revenues and earnings elsewhere in the economy, extending the apparent strength beyond the technology companies themselves.

But free cash flow and off-balance-sheet commitments create a different picture for a credit investor.

“Suddenly you look at all of the debt that was issued, and then you realize it’s not as sustainable as you thought previously,” King says of the moment equity confidence changes.

He repeatedly stresses that we are not there yet.

I push back on the idea that less hyperscaler issuance must be bad for bondholders. If issuance falls because investment slows, reduced supply and lower real yields could support the bonds of stronger issuers.

A slowdown in financing demand and a collapse in confidence about the cash flows supporting the debt are not the same outcome. That distinction matters when thinking about who wins and loses if the AI investment cycle turns.

MONEY ROTATION, NOT JUST MONEY CREATION

If central banks have been draining liquidity, why have markets continued to rise?

King separates two questions: how much money or credit is being created, and where investors choose to put it.

Since 2024, he thinks the break from central-bank liquidity has been most obvious in equities, less pronounced in high yield and barely apparent in investment grade.

Aggregate credit creation is not exceptionally strong by historical standards, he argues. The more extraordinary feature is investors’ willingness to allocate money to risk assets.

“That exuberance or willingness to take risk is just off the charts,” he says. Investors have been “educated to buy dips” and to expect central banks to intervene if necessary.

This does not mean money creation has become irrelevant. King explicitly argues that fiscal activity has been an important source of credit creation, especially in the US. He also describes money-market funds financing government borrowing and supporting the expansion of repo and the Treasury basis trade.

He puts US repo at about $7 trillion, more than double its level in 2022.

His concern is less that every leveraged position is inherently unstable than that funding pressure and correlated positioning could cause investors to move together.

“The main risk that they point to is a liquidity shock more than a credit shock.”

DEFAULTS DEFERRED, NOT AVOIDED

The other part of the deferral occurs within individual credits.

Weak covenants, refinancing and liability-management exercises can postpone a conventional default without preserving the value of the original creditor’s claim.

King argues that default measures need to account for LMEs, including transactions described as voluntary that nevertheless impose significant losses on lenders. He also questions the incentives of private-equity-owned borrowers that use improving profits for dividend recaps rather than sustained deleveraging.

When problems are postponed, eventual recoveries can be worse.

“We’ve abandoned the covenant, so nothing defaults, but at the same time, when we get the defaults, the recovery rates are really low,” Matt King (Satori Research).

Abundant liquidity can reinforce the process. An analyst may consider a credit expensive, while a portfolio manager receiving inflows still needs to deploy capital.

“What that gives you is stuff that should have been defaulting priced to par, and then when you do get the default, yeah, the cliff is all the steeper.”

But there is an important qualification.

I ask whether the public market already recognizes the weak borrowers: CCCs trade much wider while the rest of credit remains tight.

King agrees—to a point. There is not just a cliff in spreads. There is also a cliff in fundamentals. BB and B profits remain strong.

For those liquid credits, he says, “if there’s a bubble, it’s a bubble in earnings and not just in markets being distorted.”

PRIVATE CREDIT: THE GATES ARE HOLDING

In private credit, King sees a different reason problems can take longer to surface: investors cannot necessarily withdraw their money.

Redemption gates reduce the immediate risk of a run and allow time for assets to be refinanced or problems to be worked through. They do not establish that the assets are correctly valued.

“It does leave you with the lingering feeling, is this really marked correctly?”

One possible transmission channel is that investors sell what they can, rather than what they would prefer to sell.

“It could get to the point where you pull your money out of the liquid fund because you can’t pull it out of the illiquid fund,” King says. “But we’re not there yet.”

That qualification runs through the conversation: postponement can buy time, but it can also leave the eventual adjustment concentrated in a smaller number of more severe outcomes.

WHEN ROTATION TURNS TO EXIT

What would change King’s view?

“Outflows. That’s easy.”

Higher rates should affect both borrowing decisions and asset allocation. It is one thing to buy credit or equities when cash earns nothing; another when a bank account or money-market fund offers 4–5%, as in the comparison he makes during the interview.

Yet inflows have remained remarkably resilient.

King’s concern is a feedback loop: investors withdraw money, prices weaken, financing conditions and fundamentals deteriorate, and cash becomes more attractive—prompting further withdrawals.

That self-reinforcing exit is what he thinks has largely been missing from this cycle.

He uses two images to describe a market that absorbs pressure until it does not: a ball in a bowl, and a brick attached to elastic.

“You pull and pull and pull and nothing happens, and then suddenly it comes and hits you in the face.”

MATT KING ON…

JAPANESE SELLING

King does not dismiss Japan’s influence, but he thinks direct overseas-bond selling is too small to carry the whole explanation.

“We’re talking tens of billions, we’re not talking hundreds of billions.”

The broader influence of rising Japanese yields, term premia and repatriation concerns may matter more than the direct flows.

REAL RATES

King is skeptical of the reassuring explanation that rising real rates are simply a sign of stronger growth.

He emphasizes renewed demand for capital: government spending, infrastructure, defense, supply-chain diversification and hyperscaler investment. The private sector increasingly wants physical capacity, rather than the asset-light model rewarded in earlier decades.

THE COST OF BEING RESPONSIBLE

Protecting a portfolio against an abrupt adjustment is difficult when the market continues rewarding momentum and dip buying.

“Responsible investors everywhere are going out of business,” King says, describing the pressure on investors who have worried about the macro backdrop and underperformed those taking more risk.

His discussion of positioning centers on cheap convexity and exposures that could respond to sovereign stress. But professional investors face two risks: losing money in a downturn, and losing assets under management by being too cautious during the continued upturn.

ABOUT THE GUEST

Matt King is the founder of Satori Insights and a global markets strategist. Previously Global Markets Strategist (and previously Global Head of Credit Strategy at Citi and Head of European Credit Strategy at JPMorgan.

His research connects credit markets, liquidity, investor flows and the wider macroeconomic environment. In this episode, he presents his October 2026 framework, “The Great Deferral.”

https://www.linkedin.com/in/matt-king-0147194/

https://satoriinsights.com/

https://www.linkedin.com/company/satori-insights/

SUPPORT MATERIAL

E23 Matt King Transcript

124KB ∙ PDF file

Download

Download

The Great Deferral Satori Insights Oct 2026 V2

2.68MB ∙ PDF file

Download

Download

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe to Fixed + Floating on YouTube · Spotify · Apple Podcasts

Follow Fixed + Floating on X · Fixed + Floating on LinkedIn

Follow Josef on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

My interpretation is narrower: the divergence reflects, in part, the repricing of a segment of leveraged finance whose debt is no longer sustainable at the current cost of capital.

An increase in financing costs that forces a borrower levered at 7x into a restructuring may remain manageable for an otherwise comparable borrower levered at 3x. With public CCCs representing a small share of high yield, concentrated stress in that segment need not translate into comparable widening across the broader market.

Higher interest rates affect highly leveraged capital structures through two channels:

1. A higher discount rate reduces enterprise value, holding projected cash flows constant.

2. A higher all-in financing rate increases cash interest expense when floating-rate debt resets or fixed-rate debt is refinanced, reducing free cash flow.

Lower enterprise value erodes asset coverage. Higher interest expense reduces the amount of debt the business can sustainably service. The impact depends on leverage, cash generation and the timing of refinancing—not the rating alone (e.g. Virgin Media).

A highly leveraged borrower with negligible or negative levered free cash flow can require a debt reduction after a relatively small increase in financing costs. A borrower with lower leverage and stronger cash generation can absorb the same increase. My argument is that these rate-sensitive capital structures are concentrated in the CCC bucket, not confined to it.

Sector-specific operating pressures and liability-management exercises also matter. They can interact with the financing-cost mechanism described here, but this piece does not attempt to quantify their separate contributions to the CCC differential.

In XAIA’s Q1 2024 quarterly letter, which I co-authored, we argued that higher rates would keep the CCC differential elevated. The letter identified both valuation losses and rising financing costs as pressures on highly leveraged borrowers. The following piece further develops those mechanisms and their implications for credit analysis.²

For whom the spread tolls — some background on CCCs

CCCs sit near the bottom of the high-yield ratings scale. A CCC rating can reflect operating weakness, excessive leverage or subordination (subordinated instrument rated CCC and more senior instrument rated BB).

The CCC bucket is also much smaller than it was. Its share of the US high-yield index fell from 19% in January 2014 to 8% at the end of September 2026, while its spread widened. These are market-value weights: part of the decline reflects lower bond prices rather than fewer CCC issuers.¹

Historically, CCC spreads often widened when government yields fell because both reflected weaker growth. Falling earnings increased default risk, while government bonds rallied as investors sought safety and priced in monetary easing. Stronger growth could produce the reverse: rising yields alongside tighter credit spreads.

The highly leveraged structures built during the zero interest rate period now face an additional constraint. Their debt was sized against lower financing costs. Higher rates reduce the amount of debt their cash flows can sustainably service, even if operating performance holds up.²

These borrowers remain exposed to weaker growth. But stronger growth provides less protection if the improvement in cash generation is insufficient to offset higher financing costs induced by overall higher interest rates.

The favorable outcome is disinflation with resilient growth: financing costs decline without a material reduction in operating cash flow.

If rates instead fall because growth deteriorates sharply, weaker earnings and wider credit spreads can offset the benefit. The old pattern—falling government yields and widening CCC spreads—can reappear.

The spread–yield relationship also changed across the two samples. Before 2022, a 100bp increase in the 10-year Treasury yield was associated with an estimated 118bp narrowing of the CCC-minus-B differential. Since 2022, it was associated with 34bp of widening.³

These are regressions of monthly changes, not isolated rate shocks. The rolling estimates also turn negative during part of the later period (tariff shock). The result is consistent with financing costs becoming more important for highly leveraged borrowers; it does not establish that rates caused the CCC widening. The rating buckets are not fixed portfolios: rating migration, defaults and changes in issuer and sector weights can also affect the observed differential.

Altice illustrates how legacy financing can preserve near-term liquidity while leaving the long-term debt burden unresolved.

Die another day — Altice’s liquidity buys time, not deleveraging

Altice USA illustrates how a borrower can address near-term maturities without resolving an unsustainable capital structure.

In February 2024, S&P assessed Altice’s liquidity as adequate and noted that roughly 86% of its debt was fixed-rate, including floating-to-fixed swaps. Its assessment identified no substantial maturity wall until 2027. Legacy financing limited the immediate increase in interest expense and provided time to execute the turnaround.⁴

But liquidity did not establish debt sustainability. In May, S&P downgraded Altice to CCC+ on capital-structure sustainability concerns. On approximately $25 billion of debt, an increase in the average interest rate from roughly 7% to 9% would add approximately $500 million of annual interest expense. That would absorb approximately $500 million of estimated free operating cash flow—even under an assumption that capex was reduced to about $1.2 billion.⁶

Average yields on traded instruments reflect the gap between the existing interest cost and the return investors required. S&P cited yields of approximately 12–16% on guaranteed notes and 23–27% on unsecured notes. These were not executable refinancing offers, but they showed how far market pricing had moved above the existing blended cost of debt.⁶

Altice had addressed its near-term maturities. However, it cannot refinance the whole capital structure at a sustainable cost for the time being.

Altice also faced declining earnings, competition and substantial capital expenditure. Higher financing costs compounded those operating pressures: the business had little room to absorb either weaker cash generation or a higher interest bill.

The capital-structure problem remained unresolved. In June 2026, the renamed Optimum Communications announced plans to pursue a comprehensive restructuring of CSC Holdings’ debt.¹⁸ S&P subsequently downgraded the company to CCC, citing approximately $6.2 billion of 2027 maturities and the likelihood of restructuring.⁷

The distinction is between postponing repayment and restoring debt capacity. Legacy financing and maturity extensions can preserve liquidity, but if the business cannot service the existing debt at its replacement cost, time alone does not close the gap.

Honey, I shrunk the debt capacity — higher refinancing rates and cash-flow coverage

A central task in corporate credit analysis is to estimate how much debt a business can sustainably service over the life of the loan. The first channel through which higher interest rates affect credit is a reduction in debt capacity: with cash generation unchanged, the business can service less debt.

Every capital structure is different. For the calculations below, we simplify things and use one class of debt, with all creditors ranking equally.

Consider a company with EBITDA of 100 and gross debt of 600 and no cash: 6.0x leverage.

After capital expenditure, working-capital requirements and cash taxes, assume it generates 60m cash before cash interest. We hold cash taxes constant and exclude scheduled amortisation, financing fees and the incremental interest tax shield.

For this analysis, cash-flow interest coverage is:

The numerator deducts reinvestment and taxes before assessing interest coverage. This is more useful for our purpose than EBITDA alone, particularly for capital-intensive borrowers. Cash interest-coverage definitions vary, but i stick to this version to account for varying capital structures.⁹

At an 8.5% all-in cash interest rate, annual interest on 600 of debt is 51. Coverage is 1.18x and post-interest cash flow is 9.

At 11%, interest rises to 66. Coverage falls to 0.91x and post-interest cash flow becomes negative 6.

We can calculate the amount of debt consistent with a specified coverage requirement:

Here, C is cash available before interest, k is required cash-flow interest coverage and i is the all-in cash interest rate.

Use an illustrative 1.25x coverage requirement throughout. At an 8.5% rate, 60 of cash supports approximately 565 of debt. At 11%, it supports approximately 436.

The existing 600 debt balance therefore exceeds the amount supported at 1.25x coverage in both cases. At 8.5%, the borrower still generates positive post-interest cash flow. At 11%, it cannot meet the interest bill from the assumed cash generation.

The continuous sensitivity makes the thresholds clearer. At an 8% rate, existing debt has 1.25x coverage. At 10%, interest consumes all 60 of available cash. At 12%, coverage falls to 0.83x, and restoring 1.25x coverage requires debt to decline to 400—a one-third reduction.

Debt capacity and cash-flow interest coverage both decline inversely with the financing rate.

Operating cash generation need not deteriorate for the capital structure to require deleveraging.

The 1.25x requirement is an illustrative assumption, not a refinancing threshold that applies across borrowers. A cyclical business with weakening earnings can be difficult to finance even at higher coverage. A business temporarily at 1.0x may retain access to capital if lenders expect a credible recovery and sufficient liquidity is available until then. The cash-flow outlook and its volatility determine how much weight to put on current coverage.

Interest deductibility can reduce cash taxes and soften the financing-cost increase. If the cash-tax forecast already incorporates the deduction, an additional shield would double-count it. Our illustration holds cash taxes constant and excludes that incremental benefit.

Dude, where’s my equity? — higher discount rates and asset coverage

The second channel is enterprise valuation.

The risk-free rate is a component of both the cost of debt and the required return on equity. Holding credit spreads, equity risk premiums and financing weights constant, a higher risk-free rate raises the business’s cost of capital. That higher required return reduces the present value of its operating cash flows.¹⁵

For a simple growing perpetuity:

Here, F₁ is next-period unlevered free cash flow, g is its long-term growth rate and r is the discount rate—the required return on the capital invested in the business. In a conventional enterprise DCF, this is the weighted average cost of capital, or WACC. It is not the risk-free rate alone.¹⁵

Holding projected unlevered cash flows and growth constant, a higher cost of capital mechanically reduces enterprise value. Debt principal remains unchanged, so debt-to-enterprise-value increases.

Assume next-period unlevered free cash flow of 56, growth of 2% and EBITDA of 100. This is a separate valuation sensitivity; the 56 represents next-period unlevered free cash flow, while the preceding interest-coverage example uses 60 of current cash available before interest.

At a 9% cost of capital, enterprise value is 800. Debt of 600 leaves an equity cushion of 200.

Increasing the cost of capital to 10.5% reduces enterprise value to approximately 659 and removes about 71% of the equity cushion.

At approximately 11.3%, enterprise value equals debt. At 12%, it falls to 560, leaving a 40 shortfall against the debt balance.

EBITDA and projected operating cash flows remain unchanged. The higher cost of capital has reduced asset coverage below 100%.

The credit has already weakened before enterprise value falls below debt. At 10.5%, debt remains fully covered, but the equity cushion has fallen from 200 to approximately 59. Creditors have less protection against the next valuation or earnings decline.

The example changes the cost of capital, not the risk-free rate directly. The actual transmission depends on changes in credit spreads, equity risk premiums, financing weights and tax treatment; a 100bp rise in government yields need not produce a 100bp increase in WACC.

The good, the bad and the subordinated - where creditor losses arise

The valuation loss is allocated according to creditor ranking. A subordinated CCC security in a larger capital structure can suffer a complete loss while senior debt remains fully covered.

Consider 5.0x EBITDA of senior debt and 1.0x of subordinated debt. At 6.0x enterprise value, both claims are fully covered in a restructuring. At 5.5x, the subordinated claim receives a recovery of 50% of its claim. At 5.0x, the recovery is 0%.

If we assume, the business is worth 6.0x EBITDA initially, a 16.7% decline in enterprise value eliminates subordinated recovery while leaving senior debt fully covered. The decline in asset value could either come from a reduction in profit growth prospects (lower multiple) or lower profits (lower EBITDA) or both.

A higher issuer rating does not protect the junior claim. Recovery depends on the residual value after senior claims have been satisfied and the value of any post reorg debt or equity received in a restructuring.

A series of unfortunate credit events -default risk and recovery

The credit triangle links the credit spread to two inputs: default risk and recovery.

Here, λ is the default intensity implied by market prices, and R is the assumed recovery if default occurs. The spread compensates investors for default risk and the loss they would suffer. If default becomes more likely (λ increases), the spread should widen. Likewise, if expected recovery falls (R decreases), the loss given default increases and the spread should widen. This is a simplified CDS-pricing relationship, not a direct forecast of realised defaults.¹¹

Higher interest rates can affect both inputs. A higher refinancing rate reduces the amount of debt the business can sustainably service, increasing the risk of restructuring or default. A higher discount rate reduces enterprise value, holding projected cash flows constant, and can lower recovery.

For floating-rate borrowers, cash interest costs change as the reference rate resets, subject to contractual floors, reset periods and hedging.

An illustrative 800bp spread over a 0% underlying rate implies an 8% financing cost. The same spread over a 5% rate implies 13%. With unchanged cash flow and the same coverage requirement, that increase reduces debt capacity by 38.5%.

An unchanged spread can coexist with a substantial deterioration in financing capacity.

The example holds the spread constant to isolate the underlying-rate effect; it does not predict that the spread would remain unchanged after the borrower’s credit profile deteriorates.

Par fiction — valuing repayment, restructuring and default

Contractual yield to maturity assumes receipt of every scheduled coupon and repayment of principal at par.

When restructuring or default is a material possibility, yield is an imprecise yardstick for an investor’s expected return.

Take three illustrative outcomes:

• Bankruptcy: 40% probability and terminal payment of 20.

• Out-of-court restructuring: 40% probability and terminal value of 70.

• Repayment at par: 20% probability and terminal payment of 100.

The probability-weighted terminal value is:

The probability-weighted terminal value is 56 per 100 of face value. But 56 is not today’s fair price: it is an undiscounted expected terminal value. To value the debt today, we also need to account for any coupons received under each outcome, the timing of payments, restructuring costs and an appropriate discount rate. If a restructuring delivers replacement debt or equity, the relevant recovery is the economic value of that package—not its nominal amount.¹⁰

Higher rates can reduce both the probability of repayment at par and the value received in restructuring. The scenario analysis should model those changes explicitly.

The put who loved me — the option embedded in corporate debt

I’m not a particularly quanty guy, but I find it useful to consider aspects such as asset volatility when thinking about the right value for different parts of the capital structure.

In Merton’s simplified model, equity is a call option on firm value. Debt is a risk-free bond minus a put option on firm value.¹²

Simplified, if the asset value rise, the option of the call increases and the put option becomes worth less (which is positive for equity and credit). If the volatility of the asset increases, as the range of outcomes increases, put and call gain in value, which is positive for equity but negative for credit.

Let V denote the value of the firm’s assets, D the promised debt payment at maturity, and B the value of the debt. The subscripts indicate the date: t is today and T is maturity.

At maturity, creditors receive the asset value if it is below the promised payment, or the promised payment in full if the assets are worth more:

Before maturity, debt is worth the present value of the promised payment minus the value of the put:

Here, Pₜ is the put’s value today, r_f is the risk-free rate, and T − t is the remaining time to maturity. This representation assumes a single payment at maturity (zero-coupon bond).

Holding the other inputs constant, a higher asset value increases the value of equity’s call and reduces the value of the put embedded in debt. Both shareholders and creditors gain. Higher asset volatility has a different effect: it increases the value of both options. Equity gains because shareholders own the call. Debt loses because creditors are short the put. At unchanged asset value, higher volatility therefore transfers value from creditors to shareholders.

The put has positive gamma. Debt, which is short that put, has negative gamma with respect to firm value.

As firm value falls towards the default boundary, debt becomes increasingly sensitive to further valuation losses. At high firm values, incremental value predominantly belongs to equity. Far below the boundary, debt increasingly behaves as a claim on the remaining assets.

The transition between debt that is likely to be repaid in full and debt that is likely to suffer a loss is not a sharp cutoff. Higher asset volatility or more time until maturity makes that transition more gradual: there is a wider range of firm values over which repayment remains uncertain. Within this range, the debt’s sensitivity to changes in firm value changes most rapidly—this is where its negative gamma is greatest in magnitude. That point is generally near the default boundary, but it does not necessarily occur exactly where the firm’s asset value equals the debt’s face value.

Higher required returns reduce enterprise value, all else equal. The resulting valuation decline has a larger credit effect when the borrower starts with a small equity cushion. The Merton framework explains that increasing sensitivity.

Higher refinancing costs add a separate cash-flow constraint: the borrower may be unable to refinance the existing debt even before a valuation model puts enterprise value below debt.

If you want more insight into how I think about option pricing in the context of credit—specifically LMEs—check out...

Show me the money — underwriting the refinancing

The CCC-minus-B differential is not necessarily an early-warning signal for the rest of high yield. It can also reflect a more concentrated problem: capital structures that cannot support their existing debt at today’s financing costs.

Higher refinancing rates reduce debt capacity, while higher discount rates erode the equity cushion protecting creditors. Borrowers with less leverage and stronger cash generation can absorb changes that force more fragile structures into restructuring. Stress at the bottom therefore need not produce comparable widening further up the ratings spectrum.

Lower financing costs would help—but the reason they fall matters. Disinflation with resilient cash flows can restore debt capacity. Rate cuts accompanied by deteriorating earnings may not.

CCCs are warning that some capital structures no longer work. Whether that problem remains concentrated—or extends further into leveraged finance—is the question for Part II.

Notes

1. BofA Global Research and ICE Indices LLC, US high-yield senior non-financial bonds, option-adjusted spreads by broad rating; observations through 30 September 2026. Market levels and rating weights are taken from the dataset underlying the accompanying graphics. Rating weights are measured by market value.

2. XAIA Investment, Quarterly Letter 2024|01, “Higher rates bring credit raiders to Europe.” The letter discusses the valuation and financing effects of higher interest rates on leveraged capital structures.

3. Author’s calculations reported in the accompanying graphics: regressions of monthly changes in the CCC-minus-B spread differential on monthly changes in the 10-year Treasury yield, using FRED series DGS10. Samples: 2014–2021 and 2022–September 2026. Coefficients are scaled to a 100bp yield change. These are uncontrolled associations, not identified causal effects. The supporting rolling analysis uses 12-month windows and excludes March–April 2020.

4. S&P Global Ratings, “Altice USA Inc. Downgraded To ‘B-’ From ‘B’ On Lagging Credit Metric Turnaround; Outlook Stable,” 21 February 2024.

5. CreditSights, “Altice USA Debt Tumbles as Management Hints at Potential LME, Creditor Group Swells,” 3 May 2024.

6. S&P Global Ratings, “Altice USA Inc. Downgraded To ‘CCC+’ On Capital Structure Sustainability Concerns,” 15 May 2024. The approximately $500 million free-operating-cash-flow comparison assumes capex reduced to approximately $1.2 billion; it is not the forecast at planned capex. Secondary-market yields are not executable refinancing terms.

7. S&P Global Ratings, “Optimum Communications Inc. Downgraded To ‘CCC’,” 8 June 2026. The downgrade cited increased refinancing risk around approximately $6.2 billion of 2027 maturities.

8. Reshmi Basu, “Creditors Sue Optimum, Patrick Drahi Alleging ‘Fraudulent’ Deal-Making,” Bloomberg, 28 September 2026. The reporting describes creditor allegations, not established judicial findings.

9. John R. Mills and Jeanne H. Yamamura, “The Power of Cash Flow Ratios,” Journal of Accountancy, October 1998. Cash interest-coverage definitions differ. This article defines its numerator as cash remaining after capex, working-capital requirements and cash taxes, before cash interest. The 1.25x requirement is illustrative, not a universal lending standard.

10. Moody’s, Distressed Exchanges in Europe. The value received in a distressed exchange depends on the allocation and economic value of the replacement securities, including the effect of seniority.

11. The credit triangle is a simplified market-pricing relationship between CDS spread, risk-neutral default intensity and assumed recovery. Market-implied default intensity is not a forecast of realised defaults; the relationship should not be treated as an exact decomposition of a distressed cash-bond spread.

12. R. C. Merton, “On the Pricing of Corporate Debt: The Risk Structure of Interest Rates,” Journal of Finance 29 (1974), 449–470. The simplified framework uses firm asset value, one debt class and repayment at maturity. To relate operating enterprise value to the model’s asset value, assume no surplus cash, non-operating assets or additional claims outside the modelled debt.

13. All debt-capacity, valuation, subordination and scenario examples are illustrative. The financing calculation holds cash generation and cash taxes constant and excludes amortisation, fees and the incremental interest tax shield. The valuation calculation holds next-period unlevered free cash flow and growth constant. Neither is a complete refinancing or restructuring valuation.

14. Altice USA, “Altice USA Reports First Quarter 2024 Results,” 2 May 2024, section “Debt Maturities Addressed Near Term.” Documents the $2.05 billion 11.75% guaranteed-note issuance and the $750 million revolver draw.

15. Aswath Damodaran, “Discounted Cashflow Models: What They Are and How to Choose,” NYU Stern, section on firm valuation using free cash flow to the firm and weighted average cost of capital.

16. Octus, “Optimum Likely Dropped Down Conn., NJ, Most of Its NYC Area Assets in Unrestricted Subsidiary Transaction,” 5 December 2025. Describes the November refinancing sequence, removal of B-6-specific protections and JPMorgan Chase Funding’s financing role.

17. Optimum Communications, Form 8-K, filed 29 May 2026, describing the internal reorganisation and unrestricted-subsidiary structure.

18. Optimum Communications, “Optimum Launches Effort to Drive a Consensual Repositioning of its Capital Structure,” 1 June 2026.optimum

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe to Fixed + Floating on YouTube · Spotify · Apple Podcasts

Follow Fixed + Floating on X · Fixed + Floating on LinkedIn

Follow Josef on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

Borrowers did two things to CLO equity. LBOs financed in 2021, when SOFR was zero, defaulted at higher rates. Recoveries fell to around 50 cents, against the 70 that CLO equity buyers typically budget. Performing borrowers, meanwhile, repriced their loans tighter.

CLO liability spreads are contractual. Spread compression on the loans therefore comes out of the equity tranche one-to-one.

Loans in a CLO prepay at par all the time. In a stressed market the CLO manager reinvests that cash in loans trading at a discount, and the gains offset the defaults. CLO investors call this the self-healing mechanism. For two years managers have rarely been able to use it, because performing loans are bid above par.

Shiloh Bates, Chief Investment Officer of Flat Rock Global, has invested in CLO equity and BB notes since 2012. We start with how a CLO earns its money and why self-healing stopped working. We then cover recoveries and liability management exercises (LMEs), and private credit CLOs. We close on captive CLO equity funds and the outlook.

“There’s been no self-healing in the last two years… the loans that are not defaulting and are doing well are all bid above par.”
Shiloh Bates, 24:20

What is in it

• The two legs of the −8%. Defaults at low recoveries were loan losses, while spread compression cut income against contractual liability spreads.

• The loss budget, re-run at 2025 defaults. Underwriting 2% defaults at 70 cents implies about 60bp a year, against about 140bp on Josef’s 2.8% and 50 cents.

• Broad discounts, not defaults alone. In 2015–16 energy loans fell while the rest stayed at par; today 2021 LBOs default while performing loans trade above par.

• Recovery rates moving lower. Underwriters targeted 80 cents, then 70, and recent recoveries sit closer to 50, which Shiloh ties to documentation, asset-light borrowers and LMEs.

• CLOs in LMEs. Equity and excess CCCs give no credit in the CLO’s tests, so CLOs now sweep equity cash to fund restructuring equity instead.

• Private credit CLO economics. Roughly 175bp more loan spread and about 40bp more AAA cost leave the equity about 500bp ahead, in Shiloh’s estimate.

• Captive CLO equity funds. Funds obliged to buy one manager’s next four or five deals keep CLOs forming at projected equity returns of about 10%.

Chapters

00:00 Cold open
00:39 Introduction: CLO returns by tranche and Shiloh Bates (Flat Rock Global)
03:47 How CLO equity makes money
06:05 CLO equity −8% in 2025: 2021-vintage LBO defaults
08:23 Loan repricing and spread compression
10:13 CLO equity and BB notes: default history and spread
12:54 CLO structure: closing date and non-call period
16:18 Reinvestment period, amortisation and calls
21:00 The CLO self-healing mechanism
24:20 Why self-healing stopped working: loans bid above par
26:03 Loan dispersion, default rates and the 2021 vintage
29:50 Loan recovery rates: a structural shift lower
31:02 Loan documentation and lender protections
34:35 Liability management exercises and recoveries
36:43 CLOs in LMEs: CCC limits and equity restrictions
38:21 Distressed funds, debt-for-equity swaps and CLO tests
41:49 Private credit vs broadly syndicated loans in restructurings
46:44 Private credit CLO equity: the extra 175bp of spread
49:46 Self-healing and diversification in private credit CLOs
53:50 Middle-market vs large borrowers: credit quality and spread
57:06 Competition for private credit loans and LBO activity
1:01:46 Captive CLO equity funds and market discipline
1:06:46 Secondary CLO equity returns and NAIC capital charges
1:11:10 Outlook for CLO equity and CLO BB notes
1:15:34 The CLO Investor podcast and the CLO Investing book

About the guest

Shiloh Bates is Partner and Chief Investment Officer of Flat Rock Global, which he joined in 2018. He was previously a Managing Director at Benefit Street Partners. After two years of investment banking he joined CLO managers as a credit analyst, and has invested in CLO equity and BB notes since 2012. He wrote CLO Investing, with a focus on CLO equity and BB notes, and hosts The CLO Investor podcast.

LinkedIn · Flat Rock Global · CLO Investing · The CLO Investor podcast

Support material

E22 Shiloh Bates Transcript

115KB ∙ PDF file

Download

Download

Clo 101 05 12 2026 No Animations

1.74MB ∙ PDF file

Download

Download

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe to Fixed + Floating on YouTube · Spotify · Apple Podcasts

Follow Fixed + Floating on X · Fixed + Floating on LinkedIn

Follow Josef on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

That sounds like 50% loan-to-value. If the borrower defaults, the lender appears to have a substantial cushion: the business could lose half its value before the loan is impaired.

But that analysis is only valid if the lender actually has a claim on the full $1 billion.

Suppose the credit agreement allows the borrower to transfer $300 million of assets or enterprise value to an unrestricted subsidiary. Once those assets leave the restricted group, they may no longer secure the original loan. The lender still has a $500 million claim, but now against only $700 million of value.

The real loan-to-value is therefore 71%, not 50%.

That is not a technical detail discovered during the workout. It is a recovery risk created when the documents were negotiated and the loan was underwritten. The lender’s eventual loss may be determined less by the headline leverage at closing than by what the borrower was contractually permitted to do afterwards.

This is the distinction Michael Gatto wants credit investors to understand.

Gatto is a partner at Silver Point Capital and heads the firm’s private-side businesses. He learned credit on Goldman Sachs’s special situations desk, led Silver Point’s restructuring group and helped build its private credit business. Few investors have assessed the same type of borrower from all three perspectives: as a trading opportunity, as a restructuring situation and as a private loan.

His central argument is straightforward: the industry spends too much time modelling the probability that a borrower will default and too little time modelling what the lender will recover if it does.

A fund does not lose money merely because a borrower defaults. It loses money because the default produces a loss—and because the loss is larger than the underwriting assumed.

That asymmetry makes lending different from venture or equity investing. An equity investor may write off several investments and still be rescued by one ten-bagger. A lender normally lends close to par. There is no ten-times winner elsewhere in the portfolio to offset a loss. Each loss is largely taken in full, which makes the left tail—the amount lost after default—the defining risk.

In this conversation, we examine where that left tail is created: in EBITDA definitions that weaken leverage tests, in drop-downs that move collateral outside the lender’s security package, in uptier transactions that subordinate one group of lenders to another, and in the assumptions managers make about recovery values.

We also discuss the three legal categories into which liability-management exercises generally fall, what the Serta and Incora cases mean for creditors, why private credit is particularly exposed to drop-downs, and the questions allocators should ask when evaluating a direct-lending manager.

“Everyone talks about the probability of default, but as a credit investor, there’s two things. What’s the probability of me making this loan, the company getting into problems and defaulting? But then the second question — if that happens and I get into a workout, what is my severity of loss?”

— Michael Gatto, 25:44

Chapters

00:00 Cold open
01:11 Introduction
02:04 From Goldman’s prop desk to Silver Point
02:58 Trading desk, restructuring, private credit: three roles compared
07:22 Does a trader underwrite a credit differently than a direct lender
10:43 Who thrives on a trading desk, and who in direct lending
12:23 The Quinn restructuring: a violent workout in Ireland
17:20 First Brands, Tricolor and the loss of underwriting discipline
23:00 How a credit fund stays disciplined while it grows
24:04 Why living through cycles makes a better credit investor
25:44 Probability of default versus severity of loss
31:35 The seven-step credit analysis process
33:26 Why the current ratio is analytically worthless
36:11 Where credit analysts go wrong: reading the documents
38:11 How a 50% loan-to-value is really 71%
41:35 What is a liability management exercise
43:18 Drop-downs and uptiering explained
46:32 Serta, Incora and the three buckets of LMEs
52:49 Do LMEs just delay bankruptcy
58:07 Why director duties differ in the US, UK and Germany
59:33 Will private credit see its first LME
1:03:16 Why software is 25-30% of private credit portfolios
1:09:55 Software, AI and permanent capital impairment
1:15:53 Red flags in a private credit manager’s portfolio
1:19:54 The Credit Investor’s Handbook and the second edition

What is in it

• Loan-to-value depends on what the lender can actually claim. A $500 million loan against a business valued at $1 billion appears to represent 50% loan-to-value. But if the credit agreement allows the borrower to transfer $300 million of assets outside the original credit group, the lender may be left with a $500 million claim against only $700 million of remaining value. The effective loan-to-value is then 71%, not 50%.

• A leverage covenant is only as reliable as the EBITDA definition behind it. If the borrower can add back costs, projected savings or other adjustments at its own discretion, reported EBITDA can increase even while the underlying business deteriorates. The leverage ratio may therefore remain within its limit, and the lender may lose the right to intervene when it expected the covenant to provide an early warning.

• A drop-down moves assets outside the original lenders’ collateral package; the debt remains where it is. The borrower transfers assets to an unrestricted subsidiary, which is outside the original covenant and security group. That subsidiary can then borrow against the transferred assets. The original lenders still hold their original loan, but their security may no longer cover the assets they underwrote as part of the recovery value.

• An uptier changes the ranking of creditors without necessarily moving any assets. A majority group of lenders agrees with the borrower to create new debt that ranks ahead of the existing debt. The participating lenders provide new money and often exchange some or all of their existing loans into the new senior facility. The lenders that do not participate keep their original loans, but those loans now rank behind the new debt.

• The legal analysis of an LME starts with the credit documents, not with whether the transaction seems fair. The documents may clearly permit the transaction, clearly prohibit it, or leave the position genuinely ambiguous. That classification determines the strength and likely value of litigation. The Serta and Incora cases illustrate how courts have examined whether the relevant documents permitted the transactions and whether the necessary lender majority was validly established.mayerbrown+1

• An LME can change who gets paid first, but it does not necessarily repair the company. Bankruptcy can provide a stay, allow the rejection of certain contracts and result in the discharge or restructuring of debt. An LME generally does not provide those tools. It may raise new money, extend maturities or shift value towards participating creditors, but it may also postpone rather than solve the company’s underlying solvency problem.

• Private credit is generally less exposed to a classic uptier, but it is not protected from liability-management risk. A private-credit fund often holds the entire loan or controls the relevant lender group, making it harder for another lender group to prime it through a creditor-on-creditor exchange. A drop-down can be different: if the documents permit the transfer, the borrower may be able to move assets outside the lender’s collateral package without creating a divided lender group.

• A manager’s underwriting discipline can be tested by examining what happens after a credit weakens. Ask the manager to identify investments where covenants were amended, cash interest was converted into PIK interest or the maturity was extended. Then ask how those positions are valued and what evidence supports the marks. A position that has experienced all three events and is still marked above 95 may be defensible, but it requires a clear explanation rather than automatic acceptance as a normally performing loan.

Chapters

00:00 Cold open
01:11 Introduction
02:04 From Goldman’s prop desk to Silver Point
02:58 Trading desk, restructuring, private credit: three roles compared
07:22 Does a trader underwrite a credit differently than a direct lender
10:43 Who thrives on a trading desk, and who in direct lending
12:23 The Quinn restructuring: a violent workout in Ireland
17:20 First Brands, Tricolor and the loss of underwriting discipline
23:00 How a credit fund stays disciplined while it grows
24:04 Why living through cycles makes a better credit investor
25:44 Probability of default versus severity of loss
31:35 The seven-step credit analysis process
33:26 Why the current ratio is analytically worthless
36:11 Where credit analysts go wrong: reading the documents
38:11 How a 50% loan-to-value is really 71%
41:35 What is a liability management exercise
43:18 Drop-downs and uptiering explained
46:32 Serta, Incora and the three buckets of LMEs
52:49 Do LMEs just delay bankruptcy
58:07 Why director duties differ in the US, UK and Germany
59:33 Will private credit see its first LME
1:03:16 Why software is 25-30% of private credit portfolios
1:09:55 Software, AI and permanent capital impairment
1:15:53 Red flags in a private credit manager’s portfolio
1:19:54 The Credit Investor’s Handbook and the second edition

About the guest

Michael Gatto is a Partner at Silver Point Capital and Head of the firm’s Private Side Businesses. He learned credit on Goldman Sachs’s special situations proprietary desk and joined Silver Point in 2002, shortly after two of the partners he worked for at Goldman founded it. Across 24 years there he has run positions on the trading desk, built and led the restructuring group, the team that takes over an investment once a complex negotiation begins, and helped build the private credit business, which makes him one of the few people who has underwritten the same borrower from three different seats. He is an adjunct professor at Columbia Business School and Fordham Gabelli, and wrote The Credit Investor’s Handbook, now the standard text for analysts entering leveraged and distressed credit. A second edition is roughly 18 months away and will add international bankruptcy and a chapter on using AI in credit analysis.

Michael on LinkedIn · Silver Point Capital · The Credit Investor’s Handbook

Support material

E21 Michael Gatto Transcript

125KB ∙ PDF file

Download

Download

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe on YouTube · Spotify · Apple Podcasts

Follow on LinkedIn · X · Fixed + Floating on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

In June 2020 Hertz was in Chapter 11. Its unsecured bonds were trading around 40 cents. Its shares still had a positive price.

The company asked the bankruptcy court for permission to sell new stock into that bid. It got permission on 12 June, placed roughly $29m, and pulled the offering within the week after the SEC said it would review the filing.

People bought that stock knowing the company was insolvent, and they were not necessarily wrong to. A Chapter 11 filing accelerates the debt, but it does not settle the claim on day one. The case runs for months. During those months the fleet gets remarketed, buyers appear, and the estate finds out what the assets are actually worth. Buying the equity in June 2020 was a bet that the business would recover enough inside the process to clear the debt. That is what happened. Used-vehicle prices rose through 2020 and 2021, the fleet turned out to be worth far more than its marks, and competing sponsor bids emerged for the estate. Hertz’s plan of reorganization, confirmed in June 2021, paid every creditor in full and delivered more than $1bn of value to the old shareholders.¹

The equity of a company worth less than its debt is a call option on the firm, struck at the face amount of the debt. An amend-and-extend lengthens that call without a filing. The lender is short the put at the same strike, so the put gets more expensive. The question is whether the extra tenor is in the price.

The Model: Equity Is a Call Option

Robert Merton applied to corporate balance sheets the option pricing work Black and Scholes had published the year before.² The payoffs of options look very similar to the payout profiles of equity and debt.

Equity is long a call on the assets struck at face. Debt is a risk-free claim minus a put on the same assets.

Write the market value of the firm’s assets as V and the face amount of the debt as D. At maturity the shareholders choose: repay D and keep the business, or hand the keys. A right to buy the assets for a fixed amount on a fixed date is a call struck at D.

Equity = C(V, D, T, σ)

That call is worth more the longer the time to expiry T, the higher the asset volatility σ, the lower the strike D, and the higher the risk-free rate. Tenor and strike are in the docs. Time and vol move in the extension.

The lender is on the other side of that trade. He owns a risk-free claim on D and he has sold a put on the assets, struck at D. If the assets are worth less than the debt at maturity, the shareholders hand the keys. Lenders take the assets at V and eat the loss of D minus V. That loss is the put paying off against them.

Debt = D·e^−rT − P(V, D, T, σ)

Debt and equity are the only two claims on the firm, so their market values have to add to V. That is put-call parity rearranged. An amendment does not change enterprise value. It reallocates it.

Two inputs get used wrong.

V is the market value of the assets, cash included. Corporate-finance EV does the opposite: it nets cash out. Be consistent on the liability side — count the unfunded pension and the leases too.

σ is the volatility of the assets, not of the equity. Equity volatility rises with asset volatility and with leverage. Delta itself moves with leverage: as the equity goes further out of the money delta falls, but equity volatility still explodes because the V/E term dominates.

σ_E ≈ σ_A · (V/E) · N(d₁)

What Does Two More Years Cost the Lender?

Start with a company carrying $300m of asset value, $200m of debt and $100m of equity.³ The business deteriorates and asset value falls to $150m. The debt is now the fulcrum security. On the maturity date the lenders should be taking the business in a debt-for-equity swap, with the whole $150m going to them. That is a 75-cent recovery. The shareholders hold a call struck at $200m that is $50m out of the money.

The debt matures in a year. The equity holder wants more runway, so the equity option has a chance of finishing in the money. An amend-and-extend is a relatively mild LME: it terms out the debt by at least two years. The official story is recovery. The company heals, creditors get repaid, and equity keeps both control and the optionality of its call.

So the maturity moves out by two years and the face amount stays at $200m. The company continues to operate the same way, so assume asset volatility is unchanged.

On asset volatility, 45% is a cyclical single-asset industrial: an auto supplier, a building-products name, a shipowner in a weak market. A large diversified investment-grade issuer runs in the high teens to low twenties.

Instead of taking 75 cents in a debt-for-equity swap, the lender is left holding a claim worth 56.5. No cash left the estate, ops did not change, vol held flat — and 11.6 points still moved.

The two claims add to 75 cents at both maturities, and they would add to 75 cents at any point in time, as long as asset value is unchanged. Whatever one side gains, the other loses.

Where the 11.6 Points Go

The 11.6 points is two things.

Time value of a two-year term-out: 7.4 points. Expected payback moves from year one to year three. At 4%, $200m due in three years is worth $177.4m against $192.2m due in one, so the lender gives up $14.8m of present value, or 7.4 points on $200m of face. I discount at the risk-free rate throughout, which is the model’s convention. A higher rate on the same duration, to capture the risk premium, widens the gap.

Extension of the put: 4.2 points. The put you are short just got two years longer. More time means more paths on which the assets finish below $200m, and the strike you are short is discounted for longer. You are short it, so its gain is your loss. It goes from $55.9m to $64.4m, which is $8.5m more of put written against you, or 4.2 points on $200m of face.

Here is the claim built from its two parts, so the put values above tie back.

So two thirds of the transfer is duration and the rest is the longer put.

Can a Higher Coupon Compensate the Lender?

The instinct is to take it back in coupon. Unfortunately, that rarely works.

A cash-pay coupon comes out of the firm’s assets. Every dollar of interest that leaves the company reduces V. What the lender holds at the new maturity is a claim on what is left. He is paying himself out of his own collateral.

Interest is tax-deductible, so a dollar of coupon costs the firm less than a dollar. Depressed companies are usually cash-constrained. Cash paid out as interest is cash not available to right-size the operations.

So the calculation is: pay the coupons out of V, then value the residual claim on the asset value that remains at the new maturity, and add back the coupons the lender has actually received.

The lender needs about 24 points of present value. Charged over the two years you are granting, that is 13.5% a year, or $27m of cash interest. In this example the company needs to pay out 18% of its assets in cash, every year, out of a business that has just been restructured.

This is a large part of why names that exit with a heavy cash coupon are back in a restructuring usually inside 18 months. Lenders who insist on a high cash coupon destroy asset value, and the asset value is their own recovery.

Coupons also only get paid if the credit survives. The table treats them as certain. Value them only on the paths where the company stays a going concern, and the required rate goes higher again.

No cash-pay coupon a company in this condition can afford compensates its lender for a two-year extension. It is paid out of recovery.

What If You Bought the Loan at a Discount?

The claim is worth about 68 cents with a year left and about 56.5 with three. Seeing 56.5 next to 68 does not mean you bought cheap.

Absent further deterioration, buying at 56.5 is about 11 points cheap if the loan is not extended. If it is extended two years, 56.5 is fair value.

If you already expected the extension, buying at 56.5 does nothing for you. That is what the three-year claim is worth. You paid fair for the outcome you underwrote, with no risk premium left. Distressed funds underwrite to IRRs in the high teens and up. To clear that on this claim you need a discount to the post-extension value, not a discount to par.

The model ignores the tax shield, assumes one class of debt where real structures have several, holds asset volatility constant, and settles everything on one date. Merton settles the option once, at T. In practice each maturity is another date at which equity can pay to extend. Price that option as well as this one.

Once you own a distressed credit like this, it behaves less like a credit and more like a partial equity stake. At 56.5, the claim has a delta of 43 to the assets. If the assets move by a dollar, the claim moves by 43 cents.

What Happens During the Extra Runway

A shareholder who has just bought two more years on an option that is $50m out of the money has every reason to make that option more valuable. With two more years, the cheap way to lift the call is to raise asset vol. The docs usually do not stop that.

Mike Harmon put the mechanism plainly in the ABI Journal in January: “The Black-Scholes theory tells us that an out-of-the-money call option increases in value with more time and volatility. Thus, the equityholder has an incentive to use its governance power to increase both elements within a company’s business strategy, which might work against maximizing the value of the company as a whole — or even destroy value.”³

The shareholders’ loss is bounded at zero. Their gain above the face amount is not bounded at all. Widen the distribution without changing the expected asset value and you raise the equity and lower the debt by the same amount.

Take the equity side of the same numbers. Moving from one year to three raised the shareholders’ call from $13.8m to $37.0m. Take asset volatility from 45% to 60% on top and it goes to $52.0m. That is $38.2m created for the equity, or 19 points of a $200m claim, on a business that has not improved. Vol on a three-year option is worth more than the same vol on a one-year. The extra tenor is what makes the vol grab so attractive.

To raise asset vol you raise operating leverage — how much EBIT moves for a given move in revenue. Deferring maintenance capex buys cash now and widens the range of operating outcomes as reliability degrades. Selling the stable division and keeping the cyclical one raises the cyclicality of what is left, and it happens almost by default, because the stable division is the one a buyer pays a full price for. A sale-leaseback transaction gets sold as deleveraging, cash in and debt down. It also takes cash out of operating cash flow every year, and the rent ranks ahead of the debt.

It is very hard to tell a deliberate increase in operating leverage from one the company had no choice about. Suppliers pull credit from a distressed name. The disposal programme sells what a buyer will pay for. Maintenance slips because there is no cash. For investors, the only thing that matters is whether vol is going up or down, whatever the intention behind it.

Doesn’t the Failure Rate Undercut All This?

There is evidence on what companies do with the extra runway. Roe and Rotaru hand-collected 89 coercive, non-pro-rata liability management exercises. Of the companies with at least two years of post-transaction history, 56% had filed for bankruptcy. Within two years, only 22% had avoided both a filing and a further default.⁴ That is the aggressive tail of the market and not a sample of plain amend-and-extends, and the authors say so themselves. Even in that tail, extra runway is often not enough.

So the obvious objection. If the option expires worthless four times in five, how much can the shareholders really have gained?

Remember, the shareholders are already out of the money. They have nothing to lose. The gain is on the amendment date, in the mark. The option does not have to finish in the money. From that day the debt claim is worth 56.5c instead of 68c, and the shareholders’ option is worth $23m more. In a cyclical industry the runway can be worth far more than that. Hertz is the case from the top of this letter: the equity was economically worthless in June 2020 and paid out more than $1bn a year later, on a move in used-car prices.

Distressed names are rarely good companies with a bad balance sheet. There is usually something wrong with the business too. Over-levered, the business starves. Capex gets cut to make interest. Working capital tightens. People leave. Customers dual-source.

So the lender gives up 11.6 points on amendment day. Equity takes them. Then V keeps falling through the aftermath, and that later drop is deadweight: neither side collects it. The first loss can be negotiated. The second cannot.

Could an Exchange Be Better Than a Bankruptcy?

Most of the LME criticism is fair. The counter is narrower.

Chapter 11 is not the catastrophe a European reader assumes. In the United States it is a normal restructuring tool. Companies file, reorganise and trade through it, and Hertz is the example from the top of this letter: it filed and its creditors were paid in full.

The better argument is mandate: who can take the equity.

Most lenders do not want the equity and many cannot hold it. A CLO is the clearest case. The indenture puts hard limits on equity, reorganisation equity it does receive generally has to be sold quickly, and CLO recoveries have run 20 to 30 points below other par holders in situations like Deluxe Entertainment, Acosta and JCPenney.⁵ A separately managed account with a credit mandate is in the same position. Plenty of funds could legally take the equity but have no operating people to run what they take. The loan buyer base has changed. A large part of it cannot own equity.

So the choice is not always between a clean debt-for-equity swap and a coercive exchange. For a lender who cannot hold the equity, the swap is not available at any price, and the only executable trade is the extension.

That is a real argument and it is why these deals get done.

Three Things I Take Away

I found it hard to accept that a simple two-year extension can cost the lender 11.6 points with no cash moving and no change in the business. Put-call parity sends every point of it to equity.

Extending the runway can be the right trade if the business can actually earn it. But it also gives management the scope and the incentive to raise operating leverage, and raising the volatility of the business transfers more value from creditors to equity. Both observable actions, like selling stable business lines, and less observable ones, like deferring maintenance capex, raise operating leverage. Which effect dominates depends on what the amendment lets them do to vol.

Face value and the yield or spread that goes with it tell you little about expected return. Build the scenarios and price them, and you get a realistic return expectation. Either way, extensions are less attractive than they look.

Notes
1. Hertz Global Holdings plan of reorganization, confirmed June 2021; recovery to existing shareholders per the company’s announcements of May and June 2021.
2. R. C. Merton, “On the Pricing of Corporate Debt: The Risk Structure of Interest Rates”, Journal of Finance 29 (1974), 449–470.
3. Mike Harmon, “Do Coercive Liability-Management Exercises Destroy Firm Value?”, ABI Journal, Vol. XLV, No. 1, January 2026. The hypothetical structure is his; all option values in this piece are my own calculation.
4. Mark J. Roe and Vasile Rotaru, “Liability Management’s Limited Runway: Corporate Restructuring Today”, 136 Yale Law Journal (forthcoming 2026). Sample of 89 coercive, non-pro-rata LMEs.
5. O’Melveny & Myers, “CLO Issues in Workouts and Debt Restructurings”.

Fixed + Floating is for informational purposes only. Not investment, legal, or tax advice.

That is residual-value risk. Ankur Patel, a partner in the Ares Credit Group, is not trying to underwrite it as an asset call.

The loan tenor is three to six years and sits inside the offtake, not alongside the economic life of the chip. Amortisation is set so the remaining principal at maturity is meant to meet whatever the GPU is still worth. If a gap remains, a creditworthy name writes a residual-value guarantee over it.

Tenor, amortisation and the RVG are there so you are not taking leftover chip risk. Ankur’s line on the recording is the underwriting conclusion: he is not relying on the assets to get out. The credit is the customer contracted to pay for the compute. The hardware is secondary collateral.

“One of the main goals from a lender perspective is to make sure you have as little residual value at the end. Residual value is the value of the GPU at the end of three years, four years, five years.”
— Ankur Patel, 44:07

If a GPU financing lands on the desk, these are the credit questions. How much residual is left at year three, four, five? Is the offtake longer than the loan, and what are the termination rights? Who writes the residual-value guarantee, and does that name still pay if the offtaker is gone? If the sector overbuilds, do you still have contracted cash flow, or a depreciating box with no bid?

We spent the hour on how that is documented: what asset-based finance actually funds in a data centre, why power is the binding constraint, and how ABF, term loans and high yield take different slices of GPU risk. Ankur sits in Ares’ alternative credit business — just under $60 billion across 40 to 50 asset classes — with a concentration in heavy assets, including digital and AI.

What is in it

• What ABF actually funds in a data centre. Powered shells and turnkey construction, then substations, turbines, engines and battery systems, then the contractor fleet of excavators, cranes and dump trucks on site — and increasingly the GPUs themselves.

• Tenor sits inside the offtake, not alongside it. Three to six years, deliberately shorter than the contract paying for the compute, because any portion of the loan that outlives that contract is exposure to a resale market that may not exist.

• Amortisation is set against residual value. The balance is walked down to meet whatever the chip is still worth at maturity, and where a gap remains a creditworthy name writes a residual-value guarantee over it.

• The underwriting order puts the chips last. Offtake rating first, then the operator and its uptime SLAs, because a terminated contract leaves you with full racks and no payer. Then the guarantee and who stands behind it. Chip generation — B300 and Rubin over H100 — is a preference, never the exit.

• Power is the binding constraint. Data centres take five to six percent of US grid power today and go north of ten percent within three to five years, against a 30 to 35 gigawatt shortfall — which is why developers are buying options on land near gas and building generation on site while they wait for an interconnect.

• Pre-lease behind-the-meter power is the newest ABF trade. That spend used to be equity funded. ABF now bridges the twelve to twenty-four months to a signed lease with an SPV, a pledge of the power collateral and a dynamic borrowing base, before a bank takeout recycles the capital into the next project.

• The overbuild answer is supply-side. Ankur is not especially worried, and his three reasons are power availability, equipment supply chains and zoning approvals — brakes on the pace of building rather than a view on whether the compute gets used.

Chapters

00:00 Financing the AI data centre build-out

02:34 What ABF actually finances: shells, power, GPUs, site equipment

06:26 How ABF differs from bank lending and working capital

08:22 Asset-based finance in plain English: the three buckets

11:24 Lending against assets versus owning them

12:24 Where ABF sits: investment grade to opportunistic

17:04 Insurance balance sheet versus Pathfinder funds

19:00 Sourcing edge and a 100-person investment team

20:30 The asset classes inside infrastructure finance

23:17 Collateral and covenants versus contracted cash flows

24:38 Why ABF underwrites portfolios, not single assets

27:29 Digital versus non-digital infrastructure

31:54 Power as the binding constraint on data centre build

35:42 Behind-the-meter power as an ABF opportunity

39:17 Where ABF capital enters the build phase

40:39 The GPU financing market and who borrows

42:29 How GPU financings are structured: SPV, tenor, offtake

46:07 Segments of GPU financing: high yield, term loans, ABF

47:39 Underwriting GPU risk and lender protections

50:22 The data centre overbuild bear case

52:01 Structural protections against overbuild

53:07 Red flags and lessons

54:57 Three-year outlook: how big ABF gets and who funds it

About the guest

Ankur Patel is a Partner in the Ares Credit Group, working in the firm’s alternative credit business — the asset-based finance platform, just under $60 billion of AUM across 40 to 50 asset classes, with over half of that in the Pathfinder family of funds. He is a generalist on a roughly 100-person investment team, with a concentration in heavy assets: real estate, infrastructure, digital and AI. His career runs close to thirty years — a neuroscience degree in the 1990s, then strategy consulting, then investment banking, then more than fifteen years in asset-based investing, first at Fortress and for the past eight years at Ares.

Ankur on LinkedIn · Ares Management · In the Gaps

Support material

F+f 19 Ankur Patel Transcript

116KB ∙ PDF file

Download

Download

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe on YouTube · Spotify · Apple Podcasts

Follow on LinkedIn · X · Fixed + Floating on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

Processes are getting narrower and faster. What they are not doing is reducing debt.

Mike Harmon spent twenty-one years at Oaktree Capital in the Special Situations Group, investing in distressed companies for control or significant influence, and now teaches financial restructuring at Stanford Graduate School of Business. He wrote The Financial Restructuring Tool Set because the practical guide he wanted did not exist. We spent an hour on why balance sheets stop getting fixed, and what would have to change for that to reverse.

What is in it

• Equity as an out-of-the-money call option. Once a company is worth less than its debt, the option gains value from time and from volatility rather than from enterprise value — so shareholders extend maturities instead of repairing the balance sheet.

• Covenant-lite removed the timetable. With no maintenance test there is no early event of default, so a lender watching a borrower deteriorate has no right to enforce anything until there is an actual default.

• By then, priority is all that is left to negotiate. And priority is zero-sum: one group moves up only because another moves down. The negotiation stops being about the company.

• Cutting debt means somebody has to take the equity. Most holders cannot. A CLO has an equity bucket with a hard limit; a mutual fund has a mandate written for debt. Only genuine distressed funds want the position.

• Which is why the swaps that still get done are in private credit. One or two lenders instead of forty, able to coordinate, and able to own the business afterwards.

• The missing rung. The US has liability management at one end and Chapter 11 at the other with nothing in between, while the UK, Japan, France and China all run a lighter court-supervised path. Mike makes the case for one.

  ---

Chapters

00:00 Zombie Companies in Credit Markets
02:23 Financial Distress vs Restructuring Outcomes
04:35 Equity as an Out-of-the-Money Call Option
07:24 Covenant-Lite Loans and the Gheewalla Precedent
08:46 Leverage at 10x EBITDA: Anatomy of a Zombie Company
10:57 Chapter 11 Valuation and DIP Financing Control
15:44 Value-Creating Interventions in Distressed Companies
19:13 Maintenance and Growth CapEx Under Debt Overhang
21:31 Scenario Analysis vs Cash Flow Forecasting
23:19 Economic Distress vs Financial Distress
24:54 Covenant-Lite Documents and Lender Size Advantage
26:20 Deal Documents in Distressed Investing
28:54 The Majority Lender Playbook in Liability Management
32:05 Early Creditor Organisation vs Waiting for Default
34:26 Liability Management Exercise Success Rates
35:59 Debt Maturities and Forced Deleveraging
36:48 The Holdout Problem in Out-of-Court Debt-for-Equity Swaps
41:31 Private Credit and Creditor Coordination
44:14 Absolute Priority Rule and Cram-Down Settlements
46:58 StaRUG, WHOA and UK Part 26A Restructuring Plans
48:46 Fossil and New Fortress Energy: Restructuring in London
50:30 The Case for a US Chapter 16 Process
53:58 In-Court Restructuring Tools and Bankruptcy Costs
57:57 Where Distressed Investors Add Value
1:02:05 Assessing Whether a Distressed Business Is Fixable

About the guest

Mike Harmon is a Lecturer in Finance at Stanford Graduate School of Business, where he has taught financial restructuring for six years. Before that he spent twenty-one years at Oaktree Capital Management in the group now called the Special Situations Group, investing in distressed and dislocated companies for equity control or significant influence. He is the author of The Financial Restructuring Tool Set: How to Fix Your Broken Balance Sheet (Columbia University Press).

Mike on LinkedIn · Stanford Graduate School of Business · the book on Columbia University Press and Amazon

Support material

Slides For Fixed And Floating

260KB ∙ PDF file

Download

Download

F+f 20 Mike Harmon Transcript

309KB ∙ PDF file

Download

Download

Fixed + Floating

Fixed + Floating is a credit podcast hosted by Portfolio Manager Josef Pschorn. Long-form conversations with credit market practitioners - portfolio managers, analysts, restructuring advisers and academics - on private credit, high yield, distressed debt, CLOs, liability management and credit policy.

Subscribe on YouTube · Spotify · Apple Podcasts
Follow on LinkedIn · X · Fixed + Floating on LinkedIn

If there is a guest or a topic you want covered, reply to this post or message me on LinkedIn.

Fixed + Floating is for informational purposes only. Not investment, legal or tax advice. Guest views are their own and do not represent their employer.

References

The Financial Restructuring Tool Set: How to Fix Your Broken Balance Sheet, Mike Harmon, Columbia University Press — Columbia University Press · Amazon

Liability Management’s Limited Runway: Corporate Restructuring Today, Mark J. Roe, Vasile Rotaru - Oxford Business Law Blog

This episode is a full walk-through of how a long/short credit book is actually run: which credits are eligible in the first place, how positions get sized from the downside before conviction enters, when a short is worth the borrow and the carry, and where the tail hedge sits.

🎧 Also on Spotify · Apple Podcasts · YouTube

📊 KEY INSIGHTS FOR CREDIT INVESTORS

• Median return to an LME is now one year: Co-op exercises rewrite the documents and leave the operating problem intact, so the issuer is back at the table fast.

• A co-op seat is not a steerco seat: Without the committee you cannot see who holds senior, who holds junior, or who is running CDS alongside a cash position.

• Size decides eligibility before credit quality does: A 5% holding in a 22bn structure is 1.1bn of bonds, so a few-hundred-million fund is never on the steerco.

• Orphan bonds are where the edge survives: A 400m single-bond structure is too small to matter for a $13bn manager, and sell-side coverage has thinned over the last decade.

• Position size comes from the downside: 30 to 40 points of loss against a 50bp per-position budget caps the trade at roughly 1.4% of the fund.

• A short needs a catalyst, not a disappointment: Earnings 5 to 10% below expectations does not pay for the borrow and the carry.

• Bid-offer widens as the bond falls: At 30 cents, three or four points of spread is more than 10% of the price of the bond.

• Equity index put spreads hedge the credit tail: Hedging high yield with high yield removes the exposure he is paid to hold.

⏱️ TIMESTAMPS

00:00 Long/Short Credit and the Wine Analogy 00:37 Frits Lieuw-Kie-Song and Ironshield Capital 04:29 Orphan Bonds: Why Large Managers Skip 400 Million Issues 06:53 Defining Value in Credit: The Sail-Around-the-World Test 08:56 Credit Research Process: Website Red Flags Before the Model 12:13 Scenario Modelling and Probability-Adjusted Returns 15:09 Virgin Media Case Study: Two Products, Four Competitors 18:26 Liberty’s Four-to-Five Turn Leverage Playbook 21:50 Fibre Migration, Altnet Consolidation and BT Pricing Risk 24:43 LMEs vs Pre-COVID Restructuring: A Plaster on the Wound 26:28 One-Year Median Return to LME 28:53 Position Sizing From the Downside 31:06 The 50 Basis Point Loss Budget and the Risk Committee 34:55 WorldCom: Margins Too Good to Be True 37:37 Shorting Credit Without a Catalyst 41:22 Global High Yield Since 1989 and the Equity Put Tail Hedge 45:18 Market Neutral Mandate: 6% Net, Flat in a 10% Equity Drawdown 46:54 Bank QIS Tail Overlays and Structuring Fees 48:43 Evoke Trade: UK Gaming Tax, 70 LTV and a Takeover Catalyst 53:31 Volta Grid Trade: Factory-Built Generators and Delay Risk 56:21 CoreWeave Issuance and the AI Credit Technical 57:56 Tech Winner-Take-All: The Intel Lesson From 2000 1:01:45 2008 Counterparty Risk and the Morgan Stanley Trade 1:05:54 TARP, the Fed and the Price-Insensitive Buyer 1:08:45 Packaging Credits: Pass-Through Pricing and Excess Capacity

🎙️ ABOUT THE GUEST

Frits Lieuw-Kie-Song, Portfolio Manager, Ironshield Capital, London. 25+ years in global high yield, event-driven and distressed credit, including eleven years at Allianz Global Investors. Joined Ironshield in 2025 to run the HY Alpha / Market-Neutral (UCITS) strategy. Ironshield is a European leveraged finance boutique founded by David Nazar with restructuring experience across European jurisdictions, and publishes long-form research on software credit and private credit marks.

Ironshield Capital: [INSERT IRONSHIELD URL] · LinkedIn: [INSERT FRITS LINKEDIN URL]

🔥 SUBSCRIBE TO FIXED + FLOATING: https://youtube.com/@FixedFloating

📌 CONNECT WITH JOSEF PSCHORN Substack: https://fixedfloating.substack.com LinkedIn: https://linkedin.com/in/josef-pschorn Twitter/X: https://x.com/FixedFloating

⚠️ DISCLAIMER: Fixed + Floating is for informational purposes only. Not investment, legal, or tax advice. Guest views are their own and do not represent their employer. Consult qualified professionals before making investment decisions.

Now take the same market from a fund of a few hundred million. A EUR 400m single-bond capital structure is the core hunting ground, and sell-side coverage of those names has thinned over the past decade, so fewer people are doing the work than were doing it ten years ago.

That is the standard case for a boutique. What I had not worked through before this conversation is that the same size advantage turns into a disadvantage once a credit gets into trouble.

A small fund can buy issues a large fund will not bother with. It cannot get onto a creditor committee, because a seat requires holding a meaningful share of the debt, and in a large capital structure that is more money than the fund has. So the small fund has an edge while the credit performs and no influence once it restructures.

That would matter less if restructurings were rare. They are not. Companies now return to the table faster than they used to, which means more of a credit’s life is spent in the phase where a small fund has no say. The set of names such a fund can responsibly own has shrunk, and none of those credits had to deteriorate for that to happen.

The rest of this piece is a structured walk-through of the sequence a name has to survive before it becomes a position, the arithmetic that caps how large that position can be, and why the set of eligible names keeps getting smaller.

Two Managers, One Credit — Why the Mandate Precedes the Analysis

Virgin Media O2 carries roughly 22bn of debt. Holding 5% of that structure costs 1.1bn. A fund of a few hundred million cannot get there, so it cannot sit on a steer holder committee in that name under any circumstances.

That is the constraint Frits Lieuw-Kie-Song of Ironshield Capital described, and it is not an information problem. The advisors know who the large holders are, and so does everyone else in the market; the identity of the drivers is not the secret. The point is that he cannot be one of them. In a 22bn structure his fund is a price-taker in a negotiation over its own recovery, so his rule is to be involved only where he can hold enough to drive the process, and to stay out otherwise.

The same constraint runs the other way lower down the size range. A boutique can enter EUR 400m names the large fund will never open a file on, and has to exit names the large fund can sit inside comfortably. Neither is running the better strategy. They are running different perimeters, and the perimeter fixes the opportunity set before a single number is checked.

Inside that perimeter a name still has to survive an ordered sequence, and the order matters more than any individual step. Building the model early is the most common way to lose two days, because it commits the research time before the eligibility questions have been answered.

The Sequence — Seven Questions Before a Name Becomes a Position

1. Does the name fit the fund?

Can it be owned or shorted in size that matters? Is liquidity adequate at that size? Which stage of the lifecycle is it in — performing, stressed, or close to an exercise where a committee seat becomes necessary? Frits’s stated zone is 300-500 over, performing credit, deliberately upstream of the workout desk that Ironshield’s flagship special situations fund covers. If the name fails here, stop.

2. What does the business do, and what are the sector KPIs?

Understand the business before touching a spreadsheet: what it sells, how customers and competitors see it, whether the product is growing, whether regulation or technology is moving against it. Then reduce the sector to the variables that drive margin.

Read the customer-facing part of the company’s website, then read the investor relations section. If the two describe a different business, that gap is the first red flag.

3. What is changing?

Value comes from a change in the business, not from a bond looking cheap against comps. The change can be external — regulation, technology — or internal: new management, shareholder pressure, a badly run company being fixed.

The best version is a change in what the shareholders want. Shareholders normally want dividends and acquisitions, and bondholders want neither. Occasionally the equity holders conclude their shares are worth more with two or three turns less leverage and the M&A stopped. For as long as that holds, both sides want the same outcome: leverage falls, spreads tighten, and the tightening has a cause you can name. Identify it before the market does and the position runs for several quarters.

4. What does the bear case say?

Read both sides before entering, not after. If the short case against a long you like is weak, that is the setup — particularly when the bond has been sold hard, because the selling is then either technical or a misunderstanding, and you can be paid for both.

The negative version of this came out of WorldCom, one of the largest losses of Frits’s career. The numbers looked excellent, EBITDA margins ran well above AT&T’s, and it was a fraud with a recovery close to zero. The rule that survived it: when something looks too good to be true, name the structural reason for it — a monopoly, a patent, a real cost advantage. If you cannot name it, you do not own it.

5. What is the scenario-weighted return?

Only now does the model appear. Cash flow projections, three or four scenarios covering improvement, stasis and deterioration, probabilities against each, and an expected return. The sense-check Frits applies alongside it has nothing to do with multiples:

If you say you will go and sail around the world for a year, which of the investments are you not worried about? Which do you think will survive, whether the market is up or down ten or fifteen percent, just because the business model is robust and they’ve got a strong balance sheet?

6. What size does the downside support?

Apply the arithmetic in the next section. If the answer is too small to matter to the fund, go back to step 1 and mark the name ineligible rather than taking a token position.

7. Long, short, or neither?

The same credit can be a long at one point in its life and untouchable at another. Virgin Media O2 is the example: a business Frits currently reads as a melting ice cube, a name he could build a long case for in the front-end 2029s if sentiment turns, and one he expects to be out of entirely if an exercise arrives. One body of work, different answers depending on where the credit sits.

Exit discipline

Three situations, three different responses:

• The thesis is wrong — a fact misread, a business misunderstood. Cut immediately, regardless of price.

• The thesis is intact and the position is moving against you. Scale down as it approaches the loss budget instead of making one all-or-nothing decision at the end.

• The position has breached the budget and you like it more than when you put it on, because it is ten points cheaper. At Ironshield this goes to a committee of the portfolio manager, the CFO and the CIO, and the decision has to be unanimous.

If you and I talk and we have the same investment philosophy, and I really like a position, I should be able to persuade you. If I can’t persuade you that it’s a good position, then maybe it’s not such a good position after all.

Conviction is lowest exactly when you are asked to defend it, because the market has spent a month telling you that you are wrong.

The Arithmetic That Caps a Position

The conventional order is to find the name, do the work, form a view, then size. Running step 6 as a gate instead of a conclusion changes which names reach the model at all.

Start with what the position loses in the bad case. A front-end bond bought in the nineties that ends in a restructuring costs 30-40 points. Set against that a loss budget — the maximum any single position may cost the fund. Frits works to roughly 50bp, and it flexes by situation; one position last year was set at 30bp.

Maximum position size (% of NAV) = loss budget ÷ downside in points

A 35-point downside against a 50bp budget caps the position at 1.4% of the fund, before anything has been said about how much you like the trade.

The short side behaves differently, and this is where the calculation is most often done badly. Frits frames the downside on a short of a par bond with two years to run as the present value of two years of coupons. That is the correct standalone accounting, but it overstates the constraint inside a long-biased book, because the long positions generate carry that covers most of the coupon paid away on the short. What is left is the capital loss — how far the bond can appreciate against you — and on a two-year bond already at par that is three or four points.

The last line is the useful one. At three or four points of capital risk the 50bp budget permits a position larger than any book would run, so the loss budget no longer sets the size. How many bonds you can borrow does.

The same arithmetic screens the research list. If the loss budget only supports 20bp of the fund, the position cannot add performance, so the work is not worth doing.

Why the Short Side Clears a Higher Bar

A long/short credit book is not built to a target ratio. It gets built from what clears the perimeter, in three layers: core longs where value is stable or improving and the return is asymmetric; selective shorts where the business is structurally weaker than the market believes; and tail hedges against dislocations that would be expensive to cover through single-name shorts.

The second layer stays small because a short has to clear three hurdles.

There is a behavioral version of the same problem. A bond running from 100 to 20 over two years is a wonderful trade, and almost nobody is still short at 20. Most of the theoretical return in a credit short goes to a holder who does not exist.

Shorts therefore fail the eligibility test more often than longs, the book ends up net long, and the work a balanced pair book would have done has to happen somewhere else.

That somewhere else is the overlay. Frits hedges the tail with out-of-the-money put spreads on equity indices, 5-15% out, six months, rolled to stay at that part of the curve. Hedging high yield with high yield removes the exposure he is paid to hold. His supporting number: since 1989, global high yield hedged into euros has been down in five calendar years against six for BBB- European credit, and high yield outperformed over the period. The hedge is there for the tail, not for the beta.

Reducing a Sector to Three Numbers

The most useful analytical habit in credit is reducing a sector to the variables that decide the outcome, then monitoring only those. Telecom and cable make the point cleanly, because the business is subscribers multiplied by ARPU, adjusted for churn. Those three drive the margin and, as Frits put it, really nothing else does.

Each has a driver underneath it, and that is where the analysis happens:

• Subscribers is where market structure shows up. A water utility has one seller. Four companies will sell a UK household home internet, and the multiple the market once assigned on utility logic does not survive that.

• ARPU faces constant downward pressure, because the price per unit of data keeps falling and the operator is relying on consumption rising fast enough to hold the monthly bill flat.

• Churn shows whether customers are leaving over price, service quality or a competitor, and it moves first.

Leverage is not a fourth KPI. Subscribers, ARPU and churn tell you whether EBITDA is durable. Leverage tells you what happens to the credit when it is not.

UK telecom shows both halves at once. BT, listed, trades around 4x EV/EBITDA, or 5-5.5x adjusting for pension liabilities. A competitor carrying 6x against that comp is priced for a stability the market is not delivering. Liberty Media ran Charter and its European holdings at 4-5x for years and made a great deal of money doing it, which is much of why the same playbook was held onto here for too long.

The Perimeter Has Narrowed — What Changed in Liability Management

A pre-COVID restructuring did operational work alongside balance sheet work. It was the moment a company handed back leases, closed loss-making segments and reset the cost base, because the process carried the authority to force those things through. The co-op-driven exercises that replaced it are built for speed and paid on execution. Documents get rewritten, the company returns to the market, and the operating problem sits where it was. Level 3 and Carvana worked, and their success is much of why the template spread.

I put a figure to Frits that I had from Ed Altman on this show: roughly 50% of companies completing an exercise are back in one within three years. His number is a median of about one year, and shortening.

That changes what a name is worth to a small fund. If the first exercise is the opening move rather than the resolution, holding the credit through it requires influence over the next one. Influence requires a seat, a seat requires size, and size is capped by the arithmetic in Figure 1. For a book of a few hundred million, a credit with real exercise risk is often ineligible in both directions: too small to be in the room, too exposed to sit outside it.

The pie is the pie. The company is worth what it is worth, and in the LME it gets carved differently — and it doesn’t get carved according to the original docs, or the spirit of the docs.

Remember, none of these credits had to get worse for the opportunity set to shrink. Names simply spend more of their life in the phase that requires scale. I do not think that is priced anywhere. Spreads compensate a holder for default risk and for illiquidity. I have yet to see one that compensates a holder for having no vote in the negotiation that sets the recovery.

What I Take Away From This — And What I Will Be Watching

1. The reachable market has narrowed for a small book without a single credit deteriorating, and I do not think that is in the price. More of a credit’s life is now spent in the phase where a few-hundred-million fund has no vote, so more names are ineligible even when the analysis is good. Frits draws that boundary explicitly and declines the situations that fall outside it. I recognize the first half from my own book at XAIA. The second half I have been slower to accept, because turning down a name I have already analyzed and like still feels like giving something up rather than staying inside a boundary.

2. Reduce every sector to the three or four numbers that move the margin, and keep leverage out of that list. In telecom it is subscribers, ARPU and churn, and a monitoring routine built on those three catches deterioration earlier than a quarterly model rebuild does. Leverage sits one level below: it does not tell you which way EBITDA is going, it tells you what the credit suffers when EBITDA goes the wrong way. Running the two together is how a 6x structure gets defended as cheap against a 4x comp while subscriber numbers are already falling.

3. Holding through the whole lifecycle and relying on being large enough when it counts is what most large managers are doing by default rather than by decision, and it has not been tested. The exercises of recent years have mostly happened in markets with a bid underneath them. A median return to the table of one year has not had to survive a period when nobody wants to fund the next one. That is what I will be watching: not default rates, but whether the co-op template still works when the incremental money is unwilling. If it stops working, a committee seat becomes worth considerably more, and the perimeter narrows again.

This article is based on Episode 18 of Fixed + Floating, featuring Frits Lieuw-Kie-Song of Ironshield Capital. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit ironshieldcapital.com.

Fixed + Floating is the premier podcast for institutional investors and finance professionals exploring the forces shaping global credit markets. Hosted by Portfolio Manager Josef Pschorn, the show features conversations with leading voices from investing, research, and academia. We analyze the technical mechanics of High Yield, Private Debt, and Distressed Situations — from covenant evolution and liability management to macro policy impacts on credit cycles — providing forensic depth for the global fixed-income community.

Support Material:

E18 Transcript

246KB ∙ PDF file

Download

Download

Everything that follows concerns the US market. The vehicles are American constructs — BDCs, interval funds, evergreen wrappers — and so is the transfer machinery: NY-law credit agreements, LSTA forms, ClearPar. European private credit shares the direction of travel but runs on different documentation and different regulation; it deserves its own note.

The vehicles outgrew the assets

Private credit grew out of the post-2008 retreat of bank lending. Non-bank lenders filled the gap banks left behind, helped by the first fintech wave, which changed how consumer and small-business assets were originated. The capital came from ten-year, closed-end funds, and the LPs were pension funds and endowments with long horizons and stable liquidity needs.

That architecture contained no duration mismatch. A ten-year fund holds a five- or six-year loan comfortably inside its own life; fund lifetimes and asset lifetimes were aligned by design, and the asset class’s buy-and-hold culture grew out of that alignment. Nobody built transfer infrastructure because nobody needed to transfer anything.

The 2020s changed the liability side. BDCs, interval funds and evergreen structures brought private credit to high-net-worth and retail investors — the funds are now available through Schwab and Fidelity — and even long-dated institutional vehicles have drifted toward evergreen formats. A five-year, originate-to-hold asset now sits inside a wrapper promising quarterly liquidity, held by an investor whose liquidity profile resembles nothing the original model was designed for. The end point of the wrapper logic arrived in 2025 with the State Street–Apollo PRIV ETF, which places up to 35% private credit inside a daily-liquidity ETF, supported by Apollo’s commitment to bid on the fund’s private positions — up to a daily limit the filings do not define.

Under stress the sequence is fixed. Redemptions rise; the liquid sleeve is sold first; what remains is private credit with no ready bid. The vehicle then gates withdrawals, sells at a fire-sale discount, borrows against the book, or marks down NAV and dilutes the holders who stay. Forced sellers are created on a schedule that has nothing to do with borrower performance — vehicle-design risk sits in front of asset risk.

The first half of 2026 supplied the numbers. Blackstone’s $82.5 billion BCRED received $3.8 billion of redemption requests in the first quarter — 7.9% of assets. Blue Owl’s OCIC faced requests on 21.9% of its shares in the same period; non-traded BDCs as a group ran at roughly 12%. (Partners Group gated an evergreen vehicle in June — a European manager, and the one non-US data point here; the pattern travels.) The gates are doing what they were built to do. What their use signals is the harder point: vehicles whose stated liquidity terms exceed what the assets can fund.

No infrastructure below the syndicated market

Corporate bonds settle through CUSIPs, broker-dealers and clearing houses. Broadly syndicated loans run through agent banks with established transfer mechanics. CLOs and securitisations carry their own plumbing. Below that, in true private credit — the bilateral loan between a private lender and a private borrower — the infrastructure stops.

There is no clearing house for these loans and no CUSIP equivalent. No central record of who holds what exists; the documents sit with the parties. Selling a position means partners working their personal networks, deal by deal:

“There is no formalized capital market. You’ve got your partners and your other investment professionals essentially doing roadshows every time they want to sell a deal — going to folks that they know and essentially saying, ‘Hey, do you wanna buy some of this private loan?’” — Alex Cordover, CEO, Tradable

Twenty years of non-bank lending built origination capability without transactional muscle memory. Positions trade anyway — club deals, quiet syndications, transfers between friendly firms — but unadvertised, because a market raised on buy-and-hold treats visible selling as a confession. The volume that exists is the market everybody uses and nobody discusses.

Anatomy of a $25 million transfer

The first decision is structural, and it is binary. An assignment is a true sale: the schedule of lenders changes, borrower consent is usually required, the credit agreement may be amended, and the buyer becomes lender of record. A participation moves the economics only: the seller remains lender of record, keeps the borrower relationship, and typically charges a fee for the arrangement. A sub-participation — a participation written on a participation — adds a further layer between the holder and the claim.

The second decision is the counterparty universe. Sellers exclude direct competitors — the firms they meet in fundraising — before showing anything. Distribution then runs through the seller’s own network, through a bank or broker at significant fee cost, or through a formalised venue.

The process itself: NDA and anonymised teaser; data room containing the loan tape, company financials and, in most secondary sales, the seller’s original investment committee memo; full re-underwriting by the buyer (a stage AI is compressing from weeks to days, with buy-side teams loading data rooms directly into language models); non-binding indication of interest; consents; documentation; settlement. Non-circumvent clauses are fought over in every NDA — sellers fear a buyer walking around them to refinance the borrower directly — but in practice buyers meet the borrower during diligence, and outright circumvention is rare because it is reputationally terminal in a market this small.

The four to six weeks decompose as follows: roughly a week to get under NDA, two weeks of building competitive tension, two to three weeks of diligence, then negotiation through two investment committees. AML, KYC and closing paper move quickly when the credit agreement is untouched. A secondary clears faster than a primary for one reason — the covenants exist, the covenant testing runs, and the borrower already produces the reporting package. Nothing is invented, only transferred.

The paper itself

An assignment does not create a new credit agreement. The buyer accedes to the existing documents through an Assignment and Assumption — a short-form exhibit, standardised by the LSTA, attached to most NY-law credit agreements — and the agent updates the register, the definitive record of who is lender of record. The register, not the credit agreement, answers the question of who holds the loan; the underlying documents themselves sit with the parties. Broadly syndicated loans settle this electronically through ClearPar. Private credit has no equivalent — settlement is email, signature pages and wires.

The consent architecture is where transfers are actually won or lost. In syndicated-style documents, borrower consent to an assignment typically cannot be “unreasonably withheld,” is deemed given if the borrower does not object within five to ten business days, and falls away entirely after an event of default. Private credit documentation is materially tighter. Sponsor-negotiated agreements increasingly carry language making it expressly not unreasonable to refuse distressed-debt or special-situations buyers, and disqualified lender lists — borrower-designated entities barred from ever holding the debt — have expanded in scope: competitor lists updated post-closing, whole categories of institutions excluded, and, in aggressive documents, DQ provisions that survive a default. Whether a DQ list falls away when the borrower stops performing is one of the most consequential and least-read provisions in the document.

This consent asymmetry is the real reason participations exist. A participation requires no borrower consent and does not touch the register — it is the workaround for exactly the loans whose documents make assignment impractical. The price of the workaround is the rights position described below.

Holder structure changes the mechanics again. In a single-lender bilateral loan, the lender is also the agent: selling the position means replacing the counterparty the sponsor originally chose, agency functions must transfer or be re-papered, and consent is a genuine negotiation rather than a procedural step. In a club of two to five lenders, agency and pro-rata mechanics already exist and the consent circle is small. Larger syndicated-style unitranches — with a register, assignment exhibits and minimum transfer amounts — are structurally closest to the BSL market, which is why they trade first. The more the paper already resembles syndicated documentation, the shorter the distance to a functioning secondary.

Structure decides the workout

Structure diligence precedes credit diligence. The transfer mechanism chosen on day one defines where the legal claim sits on the day the credit breaks:

In an assignment, the buyer holds a direct legal claim on the borrower: voting, enforcement, the ability to call events of default. In a participation, the buyer holds a bilateral exposure to the seller. Sacred rights — payment terms, rate, collateral release — can be negotiated into the participation agreement, but they are rights against the counterparty, not the company. A participation holder without explicitly negotiated rights has no seat at the restructuring table.

Enforcement capacity splits by collateral type. Asset-backed lenders operate two decades of machinery — bankruptcy-remote SPVs, borrowing bases, backup servicers — and step into collateral routinely. Direct lending is structurally messier: crowded capital stacks, senior through mezzanine, and collateral that is an operating business rather than a pool of receivables. The cleaner the logical separation of collateral, the more transferable the position.

Where trades die

Trades rarely die on asset quality; a buyer exists for everything at a clearing price. They die in two other places.

The first is the information-asymmetry prior. In a buy-and-hold market, any seller is presumed informed: why is this for sale, what does the seller know. Because selling is rare, selling is a signal; because selling is a signal, selling stays hidden; because selling stays hidden, it stays rare. Concentration limits, deal upsizing and vehicle liquidity needs are routine, legitimate reasons to sell — the market has not yet learned to read them as routine.

The second is agent mechanics. LPs spent decades deferring to a single GP by design; two GPs inside one deal is a newer and unresolved arrangement. A seller staying in a deal resists handing votes on defaults and covenant waivers to additional parties. A buyer resists holding economics while a competitor controls every decision that matters in a workout. No standard framework for shared agent functions exists, and until one does, this remains the largest single killer of otherwise agreed trades.

Marks meet bids

Private credit marks are set by the holder, reviewed by third-party valuation agents working from the holder’s own model inputs, and published on a quarterly cadence. No independent, transactable reference price exists. Every incentive in the chain rewards reported stability: smooth marks support the next fundraise and keep semi-liquid NAVs redeemable at par. The industry’s official term for the result is valuation lag; the blunter term in circulation among practitioners is “valuation laundering.”

Clearing prices are the only test, and the early evidence splits cleanly by segment. Per Tradable’s platform record — concentrated in asset-backed paper, where loan tapes, borrowing bases and collateral performance leave little room for managerial interpretation — every closed transaction has printed at par, usually plus a buyer’s premium. The selection effect cuts both ways: deals whose marks would not survive a bid tend to fail before closing, so the par prints coexist with books that have never been priced at all.

The wider secondary market prices the same question differently. Credit-secondaries volume roughly doubled in 2025 to about $20 billion (Evercore — a global figure, though the market is US-dominated), and pricing is bifurcated. GP-led transactions — continuation vehicles built on selected, performing assets — averaged around 98% of NAV, with TPG Twin Brook’s $3 billion vehicle printing at a slight premium. LP-led portfolio sales cleared near 91%. Par holds where the seller chooses the assets and the buyer re-underwrites them individually; whole portfolios, sold with the tails included, cost high single digits.

A public test also runs daily and is rarely cited in this debate. Listed BDCs trade at median discounts of roughly 15–25% to NAV, with individual names approaching 50%; FS KKR reported non-accruals of 4.2% at fair value in Q1 2026. Equity discounts embed leverage, fees and sentiment, so the read-through to asset marks is inexact. A persistent 20-point gap between exchange prices and reported NAV on the same asset class is, however, the market’s standing estimate of the lag.

Software direct lending is the untested book. Borrowers levered in 2021 saw debt service double as rates rose — $10 million of annual interest becoming $20 million removes $10 million a year from the business — while marks drifted down a few points a quarter. Asset prices do not change only when quarterly filings appear. A mark that cannot be transacted against is an estimate; a functioning secondary turns estimates into prices. Where marks have met real bids, they have held. Where bids are conspicuously absent, the silence is the signal.

Where liquidity forms first

Not all private credit is equally tradable. Liquidity forms first where data is clean and terms repeat at scale: larger club and syndicated-style loans, asset-backed and specialty finance with poolable loan tapes, performing sponsor-backed direct lending, and fund-level LP interests, where a secondary market already operates. It arrives last where every position is a negotiation: bespoke middle-market unitranche, stressed names (a bid implies a forced sale), small one-off positions without a natural buyer, and thinly documented structures where data gaps block underwriting.

The buyer base is broader than the selling stigma suggests: credit funds and BDCs acquiring seasoned, performing paper faster than they could originate it; insurers and pensions deploying large tickets into duration-matched yield; dedicated secondary funds buying at discounts; banks optimising balance sheets; distressed buyers waiting for a clearing price. It is also no longer hypothetical. Dedicated credit-secondaries funds raised a record $16 billion in the first three quarters of 2025 — Ares closed $7.1 billion, Coller $6.8 billion, Pantheon above $5 billion.

The venue model is not the only infrastructure under construction. JPMorgan, Goldman Sachs, Bank of America, Morgan Stanley and Jefferies have begun making markets in private credit loans — circulating runs, quoting bid-ask on names such as Peraton — and Apollo has enlisted five banks to trade its paper. Early reports show more loans offered than bid for at current prices; sellers have arrived before buyers. Whether the market settles on dealer balance sheets or neutral venues remains open. That both are being built simultaneously is the strongest available evidence that the once-every-three-to-four-years cadence is ending.

The standardisation boundary is precise. Process, data and settlement can be standardised; covenants cannot, because the bespoke, bilateral nature of the paper is the product. The realistic end state is not a loan that trades ten times a day — no participant wants that — but a loan that can trade once a month instead of once every three to four years. Tokenization, stripped of its sales pitch, serves that end state in one specific way: co-locating the asset’s own logic — interest entitlement, principal balance, waterfall position, eligible-holder list — with the asset itself, on shared rails that replace bilateral firm-to-firm integrations. It is the DTCC and CUSIP function of the twentieth century, still vacant in this market.

What I take away for credit

1. Forced sellers are made by wrappers, not borrowers. The loans did not change; the vehicles did. Semi-liquid structures holding five-year assets against quarterly liquidity promises will keep producing liability-side stress independent of borrower performance. The screen runs on redemption terms against weighted asset maturity, before any credit work — and in a market without infrastructure, a forced seller often gets no price rather than a bad one.

2. Structure diligence precedes credit diligence. Four provisions dictate a loan’s true transferability: the borrower-consent standard, the scope and durability of the DQ list, the enforceability of sacred rights in a participation, and the mechanics of the agent register. All four are knowable before you do any credit work. Two loans with identical spreads and identical borrowers can have vastly different exit timelines and restructuring rights based entirely on these clauses alone. You must price the exit friction before you price the loan.

3. The absence of a bid is hard information. Asset-backed paper clears at par with buyers paying premiums; software direct lending prints nothing. Where marks can be tested against transactions, they have held; where they cannot, the book is carried at estimates. The first genuine prints in each untested segment will re-anchor entire books, in both directions. I will be watching where bids appear next — and where they conspicuously do not.

This article is based on Episode 17 of Fixed + Floating, featuring Alex Cordover of Tradable. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit tradable.xyz.

Fixed + Floating is the premier podcast for institutional investors and finance professionals exploring the forces shaping global credit markets. Hosted by Portfolio Manager Josef Pschorn, the show features conversations with leading voices from investing, research, and academia. We analyse the technical mechanics of High Yield, Private Debt, and Distressed Situations — from covenant evolution and liability management to macro policy impacts on credit cycles — providing forensic depth for the global fixed-income community.

Support Material:

Tradable F+f V2

366KB ∙ PDF file

Download

Download

F+f 17 Transcript

202KB ∙ PDF file

Download

Download

This piece looks at what actually drove that selloff, how Strategy’s roughly $22bn capital stack is built today, and what the new Digital Credit Capital Framework in the latest 8-K changed for credit investors. The central argument is mechanical rather than conceptual: Strategy has engineered a structure with no covenants and no classical maturity wall through its perpetual preferreds, but remains exposed to one variable — continuous market willingness to fund its securities at a premium to the Bitcoin they represent. As long as Bitcoin retains a bid and Stretch retains liquidity, the true credit risk sits in the capital-raising engine, not in the static balance sheet. The rest of this piece is a structured walk-through of the value drivers, the stack, the selloff mechanics, the new framework, and what I will be watching next.

Three Value Drivers — And One Hard-to-Model Leg

Mark Palmer, the first sell-side analyst to cover Strategy after it adopted its Bitcoin program in August 2020, models the company on a sum-of-the-parts basis, and value accrues in three places. The first is the legacy enterprise-analytics software business, which generates conventional revenue and EBITDA you can attach a standard multiple to; it is small in the overall story, but not zero. The second is the Bitcoin treasury itself, whose projected value is driven by how much additional Bitcoin the company accumulates over time and by the assumed Bitcoin price at the valuation horizon. The third — and the controversial one — is the value created by the treasury activity itself: repeatedly selling securities at a premium to underlying Bitcoin, reinvesting the proceeds into more Bitcoin, and lifting the Bitcoin-per-share metric.

All three legs are ultimately hostage to market confidence. If Strategy’s securities lose their premium and capital-markets access dries up, the engine stalls even though the collateral pool — 847,363 Bitcoin, unchanged through the episode — is exactly where it was.

The treasury-activity leg is the hardest to model cleanly because it presumes the company can, on repeat, raise equity or equity-linked capital at rich valuations, use the proceeds to retire liabilities or buy Bitcoin at distressed levels, and capture a discount on the liability side without permanently raising its future cost of capital. That last step is where most corporates meet reality. Retire converts at 40 cents on the dollar and you book a one-off deleveraging gain, but the next issue prices wider because investors have just watched you equitize or buy back debt in distress. The discount captured on the old paper has to be weighed against the refinancing penalty on the new; the net effect is usually balance-sheet deleveraging, not a clean arbitrage.

Strategy’s bespoke BTC yield, BTC gain and BTC dollar gain metrics were built to quantify this treasury-activity value, but Palmer expects the company to migrate toward a “net BTC yield” framework that better reflects a capital structure now dominated by perpetual preferred rather than converts. For credit investors, that modeling difficulty is precisely why the focus belongs on the robustness of the capital-raising engine, rather than hard-coding treasury activity into a DCF.

How the Stack Actually Works — Debt, Prefs, and Waterfall

My model and Palmer’s account together map a capital stack that is less “debt above, preferred below” boilerplate and more a deliberately tiered structure aligned to Bitcoin volatility. On the asset side, it anchors on 847,363 Bitcoin — about $52.7bn at spot — plus $2.55bn of cash, against $22.2bn of claims at par. Interest coverage is a thin 1.5x, which tells you the model is not built to service debt out of earnings; it is built to be refinanced and compounded.

The hierarchy runs as follows. At the top sit convertible notes of $6.71bn — senior unsecured, pari among themselves, and ahead of all preferred. Six tranches mature between September 2028 and June 2032, with the first investor put in September 2027. The blended coupon is just 0.52%, and two tranches carry a 0% coupon: interest expense is not the risk. The risk is the wall of principal that must be addressed over a roughly four-year window, and the 2029/2030 tranches trading in the high-80s tell you those are the busted, credit-like lines.

Below the converts sit five series of perpetual preferred totalling $15.48bn, and there is a genuine waterfall inside that layer. STRF (Strife), the apex $1.28bn line, is 10% cumulative with real teeth — unpaid dividends compound and step to 18%. STRC (Stretch) is the $10.49bn giant, a variable-rate (~12%) monthly-pay instrument engineered to peg $100, now trading near 88 as a broken-peg, high-supply line. STRE (Stream) is a €904m euro-denominated 10% cumulative, junior to STRF and STRC. STRK (Strike) is an 8% cumulative convertible, effectively a fixed-income claim plus a deep-out-of-the-money perpetual call. STRD (Stride) is the junior-most: 10% non-cumulative, minimal protections, the structural first-loss piece just above common. Below all of it sits the common equity — the residual, absorbing first loss and capturing unlimited Bitcoin upside.

Two design choices stand out. First, the only fixed maturity in the entire stack is the convert wall; everything else is perpetual, which pushes true refinancing risk into 2027–2032 and hands management discretion over when and how to meet it. Second, the preferred layer behaves like quasi-permanent capital: no covenants, no Bitcoin-linked collateral triggers, no forced refinancing, and non-dilutive to common while outstanding.

Ordering the stack this way does more than show quantum — it shows exactly where stress lands. Running the collateral through descending Bitcoin prices, the common equity absorbs first loss all the way down to roughly $26k Bitcoin. Below that, the junior preferreds cascade: STRD is wiped near $24.5k (about 60% below spot), STRK and STRE follow, and Stretch itself only begins to impair below ~$20k. The senior converts and STRF stay whole even at $10k. That is a far cry from a structure on the brink.

At spot (~$62.2k BTC), $55.2bn of assets covers $22.2bn of claims, leaving a ~$33bn common cushion (claims at par).

Recovery by instrument across Bitcoin-price scenarios, seniority top to bottom. Common absorbs first loss to ~$26k; the junior preferreds cascade below that.

What Actually Broke: Retail Leverage on a Non-Money-Market Preferred

PPopular narrative put the 32-Bitcoin sale at the centre of this episode. Palmer’s read is that the sale was about S&P’s issuer-rating model, not about liquidity. S&P’s B-minus write-up had effectively assumed Strategy would never sell Bitcoin under any circumstances — a literal reading of years of “never sell” evangelism — despite disclosures stating that Bitcoin could be sold under certain conditions to protect the structure. The 32-Bitcoin sale was deliberately de minimis, executed to prove the optionality exists. Within days the company resumed buying Bitcoin in far larger size and reiterated that its business remains accumulation. It was a message to a rating-agency model, not a pivot in strategy.

The real stress point sat elsewhere. Structurally, Stretch is a perpetual preferred with a ~12% face coupon after the recent 50bp hike, tax-advantaged distributions, and full exposure to market volatility. Yet a large share of holders had begun treating it as a cash-like instrument. Per the company’s own disclosure, roughly 80% of Stretch is held by retail investors — a holder base that almost never appears in that size in quasi-debt instruments, which tend to be institutionally dominated.

Many of those retail holders had layered margin leverage onto Stretch positions on the assumption they were parking cash in something close to a money-market fund. When Stretch slipped from par, those levered positions hit margin triggers, forcing price-insensitive selling into a falling market and compounding fear-driven redemptions from accounts that simply wanted out before it dropped further. Palmer’s assessment is that the bulk of the drawdown was driven by this forced deleveraging and panic selling, not by any deterioration in the instrument’s fundamentals.

A second misunderstanding sat on top. As Stretch’s price fell from 100 toward the high-80s, its effective yield (a ~12% face rate divided by a lower market price) mechanically rose. Some investors misread the higher yield screen as evidence that Strategy’s cash dividend obligations were climbing and the company was under growing strain. In reality, the cash outlay is set by the face rate and the outstanding notional; it does not re-set when Stretch trades at 88 or 68. What falls is the instrument’s usefulness as a funding tool: issuing preferred far below par is uneconomic and signals weakness. Put bluntly, the selloff was a leverage-and-behaviour event — retail margin on a volatile preferred while Bitcoin itself was falling — not a collapse in collateral coverage. That distinction is the whole game for anyone separating solvency risk from market-structure risk.

Why the 32-Bitcoin Sale Still Matters

If the sale did not cause the doom loop, why care about it? Because rating-agency assumptions feed the path of capital costs. S&P’s original B-minus had implicitly hard-coded Strategy as a buyer-only entity, ascribing no value to the disclosed optionality to sell Bitcoin tactically. By executing a de minimis sale and then immediately resuming large-scale purchases, management signalled two things at once: that it is willing to sell Bitcoin in stress to defend the structure, and that it will do so on its own terms and sizing rather than in a forced, liquidity-driven panic. From a credit lens, that proof point lowers the probability that an agency or lender models absolute rigidity and, from there, a forced-insolvency downside path. Optionality has value even when it is never used at scale.

That optionality is no longer theoretical. The 32-Bitcoin signal was quickly followed by the real thing: a sale of more than $250m, Strategy’s largest on record, to fund preferred distributions and top up the reserve. The escalation from signal to scale is the point: management has now demonstrated, not merely disclosed, that it will sell Bitcoin to defend the capital structure. For a rating model, a proven seller is a different credit from an assumed never-seller.

Perpetual Preferred as (Almost) Permanent Capital

Perpetual preferred sits deliberately between senior debt and common equity, and Strategy’s use of it is close to the cleanest expression of why that middle position exists for a volatile-asset balance sheet. Senior debt — bank loans, secured facilities, even convertible notes — carries covenants, collateral triggers and hard maturities. Strategy has already lived one version of that: a $205m Bitcoin-backed loan with a trigger that forced it to post additional Bitcoin when the price fell below a threshold. The loan was tiny against the overall stack, but it dominated the narrative, because markets fixate on the weakest link rather than the average strength of the structure.

Perpetual preferred avoids all three problems. It carries no maturity wall — some of these instruments could, in Saylor’s words, remain outstanding a century from now. It carries no covenants or Bitcoin-linked margin calls, sidestepping the “tiny loan, big headline” problem entirely. And it is non-dilutive to common while outstanding; dilution only appears if the preferreds are eventually equitized or taken out via share issuance.

Layer on the tax treatment and it edges closer still to permanent capital. As long as Strategy reports no positive net income — and management does not expect to for at least ten years — preferred distributions are treated as return of capital rather than ordinary income. For retail, high-net-worth and family-office holders, that is a meaningful after-tax uplift on top of a ~12% face rate. Capital-gains treatment still applies if an investor buys the preferred at a discount and sells higher, but the dividend income — the bulk of the total-return story — escapes ordinary income tax under current conditions. That combination is why Palmer describes Strategy and Strive (via SATA) as the two true pioneers of perpetual-preferred financing in the digital-asset-treasury space, and why treating Stretch as cash was always dangerous: you cannot simultaneously demand a 12% tax-advantaged yield, Bitcoin exposure, and money-market price stability.

Two-Way Capital Management: Four Levers, One Ring-Fence

Before the latest 8-K, Strategy’s playbook was strictly one-way: issue securities, take the proceeds, buy Bitcoin. That regime works while markets pay a premium for new paper and Bitcoin is rising, but it leaves the company passive when demand dries up and its own securities trade at a discount. The Digital Credit Capital Framework formalizes a shift to two-way capital management, giving the board four explicit levers.

The Digital Credit Capital Framework turns a one-way engine into four board levers around the Bitcoin treasury.

First, it can issue perpetual preferred to buy Bitcoin — the core flywheel, run when pricing is attractive and demand is deep. Second, it can sell Bitcoin tactically, up to roughly $1.25bn, to defend the structure, fund buybacks or manage stress, ending the de facto “never sell” era without turning Strategy into a net seller. Third, and for the first time, it can buy back its own preferred, a direct lever to support Stretch when it trades below par beyond simply hiking the dividend. Fourth, it can buy back common — mechanically raising Bitcoin-per-share just as surely as buying more Bitcoin, but without depending on external capital.

Running across those levers is a new constraint: the cash reserve backing the perpetual preferreds is now ring-fenced. After management used $1.5bn of that reserve to retire converts — shrinking what retail holders had viewed as a multi-year dividend runway down to months — the framework now prohibits using the reserve for debt retirement or any other purpose. Retail holders, who own about 80% of Stretch, forced that correction in real time. This is the real governance upgrade: it codifies that the preferred reserve is not a silent piggy-bank for balance-sheet clean-up, and it leaves the convert wall to be addressed by other means. The framework did not stay on paper for long: within days of the 8-K, Strategy sold more than $250m of Bitcoin specifically to fund preferred distributions and refill this reserve — the two-way playbook used exactly as designed, and a signal that protecting the preferred dividend now sits ahead of dogmatic accumulation.

ATMs: Who Funds the Story, and How Big Is It?

Who stands on the other side of the at-the-market equity offerings that can raise over $1bn in a single week? Palmer’s answer is that the mix has evolved. It used to be a handful of anchor “true believers” plus a long tail of retail; today the ATM program is more institutionalised, run through a syndicate that resembles IPO book-building, with broker-dealers allocated blocks they place across client bases. Retail still dominates the total shareholder base and drives much of the day-to-day volatility, but institutions increasingly take up the offerings.

For credit investors, the critical point is that those ATMs only work while investors — retail and institutional — keep believing the story enough to double down at scale. The moment the narrative shifts from “proven Bitcoin-treasury compounder” to “over-engineered capital-structure trade,” the ability to sell $1bn in a week at attractive levels disappears. Stretch’s trading level and the common’s are therefore not just screen prices; they are continuous referenda on whether the market will keep funding the strategy. That is exactly why Palmer spends so much time on BTC yield: it is an attempt to prove that ATM capital is being converted into Bitcoin-per-share efficiently enough to justify the next round.

Different From 2022 — But Not Risk-Free

It is important not to blur this episode with the 2022 bust. In 2022 the equity traded below convert strike ranges, the converts collapsed into deeply distressed territory in the 30s–40s, and the market explicitly priced a non-trivial probability of coercive equitization and an equity wipe-out. A Bitcoin-backed bank loan with collateral triggers was still in the structure, and the risk conversation centred on senior instruments — covenants, margin calls and lender negotiations. That episode was about potential default on the debt.

This cycle is structurally different. The stack is anchored by perpetual preferreds and deeply in-the-money converts at a 0.52% blended coupon, and the fragility sat in retail margin accounts on Stretch rather than in bank covenants or standstill talks. The convert book still carries a 2027–2028 wall, but trades as short-dated credit, not near-default paper. The preferreds have become the true hinge: Stretch trades at a discount to par, but with large, liquid two-way flows and a ~12% coupon intact.

The existential caveats have not gone away. Everything here hinges on Bitcoin remaining relevant, non-corrupted and technologically robust — surviving quantum, protocol attacks, regulatory overreach, or a structural shift in digital-asset preferences. None of those risks is diversifiable inside Strategy’s model; they hit both the asset side and the narrative that justifies the capital-raising engine. Palmer’s long-term Bitcoin framework rests on fixed supply (21m coins by 2140), ongoing monetary debasement, and incremental institutional adoption via stablecoins, tokenisation and regulatory normalisation such as the CLARITY Act. If those drivers fail, no amount of capital-structure engineering ultimately saves the equity story.

What I Take Away From This — And What I Will Be Watching

1. The real credit risk is the capital-raising engine, not today’s balance sheet. With $55bn of assets over $22bn of par claims and a first-loss cushion that survives to ~$26k Bitcoin, solvency is not the near-term question. The near-term question is whether Strategy’s securities keep the premium needed to fund further treasury activity.

2. Perpetual preferred is powerful but fragile. Structurally, Stretch is close to optimal for a volatile-asset balance sheet: no maturities, no covenants, tax-advantaged, non-dilutive. But its “permanence” is a market-price fiction that holds only while it trades near par and can be issued at attractive levels. With ~80% retail ownership meeting margin leverage, that fiction broke in weeks.

3. The convert wall is a calendar problem, not a coupon problem. A $6.71bn stack at 0.52% blended is not an interest-burden story; it is a 2027–2032 refinancing-and-equitization story. How management meets it — Bitcoin sales, ATMs, preferreds, or a mix, now with the reserve ring-fenced off — defines the next chapter.

4. The single number worth watching is Stretch’s level relative to par. Bitcoin’s price matters, but the immediate tell for whether the flywheel can restart is whether Stretch normalises back toward 100 and stays there. A stable, liquid preferred market at par makes new perpetual issuance viable; a persistent deep discount turns Stretch from “permanent capital” into stranded capital.

The episode did not reveal a broken balance sheet. It revealed how sensitive Strategy’s model is to retail leverage, instrument misunderstanding, and the continuous vote of confidence expressed through the Stretch and common screens. For a credit investor, that is where the underwriting work now sits.

Listen to the full episode for Mark’s takes on Strategy.

This article is based on Episode 16 of Fixed + Floating, featuring Mark Palmer (StoneX). The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.benchmarkcompany.com/leaders/1601/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

F+f 16 Mark Palmer Transcript

207KB ∙ PDF file

Download

Download

I. The arithmetic ceiling and who is long it

Global listed companies generated roughly 142 trillion of revenues last year. Against that, total employee expenses ran between 20 and 25 trillion. Those two numbers put a hard ceiling on any “AI replaces labour” story: even in the extreme case where every worker on earth is automated away, the prize for AI is bounded by the wage bill, not by whatever number appears in a pitch deck.

When you see AI TAM slides pointing to 20–30 trillion of annual revenue, you are not looking at an aggressive scenario; you are looking at something that is arithmetically impossible in aggregate. Damodaran’s point in his “big market delusion” work is precisely this aggregation problem: each AI company can be priced off an internally consistent story, but the sum of those stories does not fit into any plausible market.

Equity can live with that for a while. If you own a basket of AI names, you only need a handful of big winners to offset the write‑offs. Credit cannot. Credit gets paid a capped coupon on every name and takes a full loss when the story breaks. Once this delusion is being financed with leverage rather than retained earnings, it stops being a harmless equity bubble and becomes a credit problem.

II. How the delusion moves into credit

The funding mix of the AI build‑out has changed. On estimates cited in the pre‑recording brief, 500–600 billion of spend since 2023 has been funded largely from hyperscaler retained earnings. The marginal dollar is now debt, ABS and off‑balance‑sheet JVs. Morgan Stanley expects 250–300 billion of hyperscaler‑related issuance in 2026 alone.

CoreWeave is the cleanest live example of how equity narratives are now setting debt terms. Its unsecured risk trades around 500 bps in CDS, roughly where you would expect a speculative‑grade name with volatile cash flows and heavy capex. At the same time, its asset‑backed infrastructure stack includes an 8.5 billion delayed‑draw term loan, rated A3 / A‑low, and a 3.1 billion syndicated, GPU‑ and contract‑backed facility, both on investment‑grade terms. The whole capital structure tightens as the equity trades up and the backlog narrative improves.

From a lender’s perspective, that is the wrong way round. The unsecured and structurally subordinated pieces are being paid for single‑B risk. The “infrastructure” debt is being paid for as if it were a mature, regulated asset with stable tariffs and demand, not a highly levered bet on early‑stage AI workloads and a fragile ecosystem of counterparties. You are getting IG spreads on assets whose economics are still in price‑discovery mode.

Damodaran’s comment is blunt: if your spread is being set by the market cap and the story, you have lost the script as a lender. You lend against current cash generation and tangible assets, not what equity investors hope the business might be worth in five years.

The real issue is structural. Within the same issuer, the best assets are pledged to the most senior tranches. Unsecured and holding‑company creditors are effectively lending against residual value in a technology stack with short half‑lives and no guaranteed secondary market bid. If the delusion breaks, the losses do not sit in the IG‑labelled paper that funded the GPUs; they sit in the parts of the structure that were priced as if they were one notch above “core infra” but are in fact equity‑dependent residuals.

III. Financing should act its age

Damodaran’s corporate life‑cycle framework is useful for credit because it forces an honest conversation about when debt makes sense at all. The rule of thumb is simple: the balance sheet should act its age.

• Start‑ups and young growth companies: value is almost entirely in future growth. They should be financed with equity or, at most, convertible instruments. Straight debt introduces a hard default condition into what should be an option on growth.

• Mature companies: stable cash flows and taxable income justify using debt to exploit the tax shield.

• Declining companies: the asset base is shrinking; the debt should shrink with it. If dollar debt stays flat, default becomes a timing problem, not a tail risk.

The AI build‑out is violating this in both directions. Cash‑rich, mature hyperscalers are levering up for a young‑growth story, including off‑balance‑sheet JV exposure that is structurally more fragile than their core businesses. At the same time, young AI companies whose value is pure optionality are taking straight debt — venture debt — to avoid dilution. Damodaran’s view on that is clear: venture debt “violates every rule in corporate finance”, because it puts the entire growth option at risk to save a handful of percentage points on the funding line.

From a credit standpoint, the diagnosis is straightforward. Lending into a business whose value sits mostly in the right tail of a wide distribution is a short‑vol trade. If you are not being paid like a volatility seller — via equity participation, tight covenants, security, or control — you are mis‑pricing your own risk. In that environment, a convertible is the honest instrument. It keeps the cash‑flow drag low for the issuer and gives you upside in the same dimension you are short via the debt.

IV. Overcapacity, leverage, and why some sectors never recover

We also need to be honest about what happens when you build too much capacity on too much leverage.

Damodaran’s distinction between “good” and “bad” overcapacity is helpful. In a young market, creating capacity ahead of demand can deter new entrants and improve long‑run economics. In a shrinking or structurally challenged market, it is the opposite: excess capacity becomes a trap, because nobody wants to buy what you need to sell. His tobacco example — a plant built for five percent growth in a market shrinking five percent per year — is the pure case.

The AI data‑centre build‑out sits between these two poles. Today’s narrative treats overcapacity as strategic. Hyperscalers and specialist operators build ahead of workloads and assume utilisation will follow. The problem is that the funding does not look like growth equity; it looks like term loans, ABS and long‑dated leases with hard payment profiles.

My own working thesis — that overcapacity plus leverage is close to unfixable — matters here. If an industry is oversupplied and mostly equity‑funded, it can consolidate. Equity takes the hit, assets are written down or scrapped, and the survivors earn better returns. Once you add heavy leverage, nobody can afford to be the first to shut capacity and crystallise losses while coupons still need to be paid. Covenants and security packages are structured to protect existing debt, not to allow the economically rational outcome.

For credit, the key is to recognise when a sector has moved into this regime. In that state, “mean reversion” is the wrong default; the right assumption is persistent low returns, forced capex just to stay in the game, and a much higher probability that, when something finally gives, it happens through default rather than through an orderly M&A cleanup.

V. Distress, gap risk, and the pricing vacuum in the middle

On distress, Damodaran’s framework has one point credit should steal and one point it should push back on.

The point to steal is his insistence that default risk belongs in the cash flows, not the discount rate. In his distressed‑valuation work, he values the firm as a going concern with a cost of capital that reflects business risk, not default, and separately estimates a distress or liquidation value and a probability of ending up there. Equity value is survival‑weighted going‑concern value plus distress value times the probability of distress. For equity, the distress leg is usually zero because of limited liability.

Translated into credit language, that is nothing more than PD×LGD done properly: an explicit survival path and an explicit failure path, instead of widening the WACC until the DCF spits out a number you are comfortable with. The discipline is to model what actually happens in default — which assets are saleable, what the forced‑sale haircut looks like, how much value goes to employees, governments and senior secured creditors — rather than hiding those questions inside an arbitrary premium on the discount rate.

The point to push back on is how far to take the option analogy. In deeply levered firms, equity does behave like a call option on firm value with a strike equal to the face value of debt, and debt behaves like risk‑free debt minus a put on the assets. But the standard implementation of that insight assumes a diffusion: continuous asset‑value paths and well‑behaved volatility. Real defaults do not look like that. They gap. Covenant breaches, ratings actions, liquidity freezes and counterparty reactions create multi‑standard‑deviation moves in a single step.

Credit’s “no man’s land” is exactly where this becomes relevant. Early on, a performing credit prices roughly linearly. Expected loss is PD×LGD, and spread is a reasonable proxy for compensation. As the credit deteriorates, the payoff profile becomes convex and option‑like. Small changes in enterprise value have outsized effects on recoveries, and the instrument’s behaviour depends heavily on path and liquidity. Yet these names still sit in “performing” portfolios, and desks still try to quote them as “+450 over” with a point price.

In that zone, it is more honest to admit that you cannot summarise the risk with a single spread. The right tool is scenario analysis: a distribution of outcomes with explicit recovery and timing assumptions, and an IRR for each. If the range of plausible outcomes spans equity‑like returns on the upside and full capital loss on the downside, you should stop thinking of yourself as a lender and start thinking of yourself as a provider of rescue equity in disguise. If you are not being paid for that, step away.

VI. Pricing vs valuing in AI hardware

The GPU depreciation debate is a good example of where Damodaran’s pricing vs valuing distinction bites credit.

On his numbers, a realistic two‑to‑three‑year useful life for GPUs would increase depreciation charges by roughly 176 billion over 2026–2028 compared with current schedules. That is 176 billion of cost being pushed out of reported earnings, even though the cash has already left. At the equity level, that matters because P/E and EV/EBITDA are still the language most investors use. At the credit level, it matters because leverage, coverage and covenant tests are often expressed off EBITDA, not off cash flow.

If you apply standard leverage metrics (net debt / EBITDA, interest coverage) to a set of issuers whose EBITDA is overstated by aggressive depreciation, you will conclude that balance sheets are stronger than they are. The structure is the same as any late‑cycle capex boom in a commodity: EBITDA looks fine until the cycle turns, because the denominator has not yet recognised the asset decay.

The other issue with AI hardware is cyclicality. In sectors like memory or GPUs, the real drivers are supply, demand and capacity decisions. The P&L confirms the story late. Damodaran’s point is that a numbers‑first investor in cyclicals is structurally lagging the cycle. For credit, the implication is simple: do not wait for the earnings miss. Underwrite against capex plans, utilisation, pricing trends and the technology roadmap. If those are deteriorating while EBITDA is still growing, your metrics are lying.

What i take away from the conversation

For a credit investor, the Damodaran conversation is only useful if it shows up in how you actually run money. For me, the concrete changes are:

• Stop letting equity do my PD work. Stock prices, IPO marks and private rounds are telling me almost nothing about default probability or recovery. I will treat them as sentiment indicators, not inputs to the default probability (PB) or the recovery rate.

• Be explicit about truncation risk. If I am going to adjust WACC for distress, I will do it as a translation of explicit survival probabilities and failure scenarios, not as a fudge factor. Alternatively, I can keep WACC clean and handle default entirely through PD×LGD and scenario trees. Mixing the two is not acceptable.

• Recognise when I’m short volatility. In highly levered, option‑like names — late‑stage venture credits, AI infra SPVs, structurally oversupplied sectors — unsecured and weakly covenanted exposure is economically a short put on the asset base. If I am not receiving equity‑like compensation, I should not pretend it is a carry trade.

• Drop “spread talk” in the convex zone. Once a name moves into the performing‑to‑distressed “no man’s land,” I will stop thinking in terms of “+X over” and start thinking in terms of scenario IRRs and recovery distributions. If I cannot produce a credible set of downside scenarios that still justify my entry price, the right position size is zero.

The AI build‑out will not make every credit investor look foolish. But it will be a clean A/B test. Lenders who insisted on acting their age — and made their structures act their age — will lose less. Those who let equity write their underwriting for them will find out what it feels like to be on the wrong side of a big‑market delusion.

Listen to the full episode for Aswaths’s takes on the interconnection of equity and credit markets.

This article is based on Episode 15 of Fixed + Floating, featuring Aswath Damodaran (NYU Stern). The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://pages.stern.nyu.edu/~adamodar/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

F+f 15 Valuation Aswath Damodaran Clean Transcript

327KB ∙ PDF file

Download

Download

This part looks at what happens when operations alone cannot carry the maturity wall — which instruments are actually being used, how they work in practice in auto suppliers, and where their limits sit in light of German law and jurisdiction choices.

The menu of instruments

The toolkit Hauke works with runs along a spectrum:

• Purely consensual, out‑of‑court deals — exchange offers, waivers, covenant amendments agreed on a contractual basis.

• UK schemes and Part 26A restructuring plans — used where English law governs a meaningful portion of the debt.

• German StaRUG procedure — the preventive framework that allows a court‑supervised restructuring of financial liabilities before insolvency.

• Full insolvency under InsO — when liquidity is exhausted or a going‑concern plan cannot be agreed.

Which tool fits is a function of the capital structure, not legal taste: where the liabilities sit, which law governs them, where COMI lies, how fragmented the creditor base is, what shareholders will accept and, as Hauke keeps emphasising, how much cash runway remains. A solid commercial plan can often be implemented consensually; the statutory tools become decisive when holdouts, shareholder resistance or timeline make that impossible.

LEONI: StaRUG at scale in an auto supplier

LEONI is the first large, listed auto supplier to run a full StaRUG process, and it is the template most people now have in mind. The financial restructuring covered roughly EUR 1.6bn of debt and combined three key elements:

• A capital reduction to zero and subsequent capital increase of EUR 150m, subscribed solely by a vehicle of Stefan Pierer, followed by a delisting.

• The transfer of around EUR 708m of bank and Schuldschein claims to that vehicle and their contribution into LEONI’s capital reserves, effectively removing those liabilities from LEONI’s balance sheet.

• A recovery instrument giving former financial creditors a 45% economic participation in any value rebuild via a virtual profit‑sharing mechanism, without formal shareholding.

Existing shareholders were fully wiped; the free float of roughly 25,000 retail and institutional investors exited at zero; Pierer’s vehicle became sole shareholder. The debt stack was reduced by just over EUR 700m and the company received EUR 150m in fresh liquidity. All of this was implemented via a StaRUG plan using cross‑class cram‑down against dissenting shareholders and some creditors; appeals were dismissed and the plan is now binding.

For auto‑supplier credit, the relevance is clear. StaRUG can now be used to:

• Take a listed tier‑1 private.

• Wipe out equity completely.

• Impose a deep haircut and equitisation structure on distributed bank and Schuldschein lenders, provided the 75% class majorities are reached.

It is a demonstrated route to a genuine deleveraging of a tier‑1 auto supplier — not just an amend‑and‑extend — but it addresses financial claims only.

What StaRUG does and does not do

StaRUG reaches financial liabilities, security packages and shareholder positions. It does not reach customer and supplier contracts, employment, leasing or most operational arrangements. It is a financial‑claims tool, not an operational one.

For an auto supplier, that distinction is central. A StaRUG plan can reset the balance sheet, change the shareholder and push out maturities, but it cannot close an unprofitable plant, exit a structurally loss‑making product line or extract the company from a bad long‑term supply contract on its own. Where the core problem is a mismatched capital structure and funding profile, StaRUG can remove the immediate financing constraint. Where the core problem is structural overcapacity and the wrong portfolio, StaRUG can only clear time and capital structure; the heavy lifting has to happen in operations via plant closures, asset sales and footprint change.

In practical credit work the question is therefore: is this a LEONI‑type situation — viable business, wrong balance sheet — or one where the economics of the portfolio and footprint do not work even after deleveraging? StaRUG is well suited to the first category. In the second, it may still be used, but only as the legal wrapper around a much more intrusive operational restructuring.

Trusteeships: holding structures for run‑offs and sales

The other instrument that appears repeatedly in German supplier cases is the double‑sided trusteeship (Treuhandlösung). A trustee holds the shares in the operating company in trust, typically with secured creditors as primary beneficiaries and former shareholders subordinated behind them.

The tool serves two main purposes in auto suppliers:

• In a run‑off, it provides a neutral governance structure while the business is wound down, contracts are performed and value is distributed over time. Equity that is deep out‑of‑the‑money is effectively sidelined; creditors have an economically aligned fiduciary.

• In a distressed M&A, it allows lenders to mandate a controlled sale process without becoming direct owners themselves and without being constrained by shareholder consent problems — whether a difficult controlling shareholder or a fragmented family register.

Trusteeships show up where equity is clearly out of the money, but a full insolvency is seen as value‑destructive or too disruptive to OEM supply chains. They are not a substitute for StaRUG; they are a governance and execution tool that can sit alongside StaRUG or InsO processes.

Cross‑border reach and recovery reality

Auto suppliers are rarely purely German businesses. Plants sit in Central and Eastern Europe, Turkey, North America, sometimes China; JVs and local entities hold working capital and hard assets. StaRUG is a German tool; its effect outside Germany depends on recognition regimes.

• A StaRUG plan compromising German‑law claims will generally be recognised in the UK (under the Rule in Gibbs framework for German‑law claims) and in the US via Chapter 15, as Spark Networks and other cases have shown.

• It does not automatically extend into China; taking and enforcing security there is slow and uncertain; cash and value can be trapped regardless of what a German court orders.

• Value sitting in non‑EU jurisdictions with weak recognition regimes (particular emerging markets) may be only partially recoverable even if headline EV looks acceptable.

For recovery analysis, the relevant question is not only “what is EV?” but “where is EV?”. A structurally important Chinese JV or Turkish plant that cannot be effectively pulled into the perimeter reduces practical recovery, even if consolidated leverage looks manageable on paper.

The creditor mix: who is in the stack and what that enables

The creditor base in German automotive has shifted away from purely house‑bank syndicates towards a mix that includes loans, bonds and Schuldscheine.

That shift is uneven:

• Larger tier‑1 structures with syndicated RCFs, Schuldscheine and listed bonds — attract international restructuring capital and can support StaRUG‑type solutions with deep deleveraging and new‑money packages.

• Smaller Mittelstand suppliers with €100–500m revenue and bilateral lending often sit below the threshold of interest for global funds and remain dominated by local banks, public‑sector lenders and families.

Where specialist capital is present, it changes what is feasible. Funds that bought into the capital structure at a discount can accept equity, PIK instruments and hard haircuts in order to get to a sustainable post‑restructuring profile. Traditional at‑par lenders, constrained by regulation and internal policy, often cannot. The LEONI transaction — roughly EUR 700m of liabilities removed, EUR 150m of new equity, equity wiped, sole new shareholder — is the size of adjustment that typically requires that kind of capital.

Where specialist capital is not present, and bank groups are unwilling or unable to accept that depth of write‑down, the pattern tends to be tighter: smaller amendments, more incremental covenant fixes and a higher probability that the name returns to the workout list when the next shock hits. The legal toolkit is the same; the willingness and ability to use the full extent of it is not.

How I read restructuring choices in auto suppliers (Part 2)

1. Where in the structure the restructuring is run — German OpCo versus holdco / foreign finance entity. A piece i have coauthored lays out the perimeter problem: once a restructuring reaches a German operating entity as issuer or guarantor, director duties, capital maintenance rules, the worse‑off test and IDW S6 bound the economic outcome tightly. At that level, StaRUG and InsO transactions are constrained to a corridor between insolvency‑floor recoveries and what a going‑concern opinion will support; aggressive creditor‑on‑creditor outcomes are the exception, not the rule. By contrast, where debt sits at a Luxembourg or Dutch holdco and German entities are only indirectly involved, US Chapter 11, UK plans or WHOA can still be used to implement materially non‑pro‑rata trades without immediately triggering German filing obligations. In auto suppliers, that split decides whether a case feels like a German corridor‑type process or more like an offshore capital‑structure trade.

2. Whether the problem is “only” financial, or structurally operational — which determines whether StaRUG is enough. StaRUG works best when the business model is still viable and the issue is a balance sheet sized for a different world: LEONI is the clean example. There, the plan combined an equity wipe, c. EUR 700m of debt removal and EUR 150m of new money, followed by a delisting — a genuine balance‑sheet reset for a business that still has a role in future supply chains. Where the core issue is structural overcapacity, a fundamentally wrong portfolio, or a footprint that cannot be made economic, StaRUG can only clear time and capital structure. The decisive question then is not “StaRUG versus insolvency”, but whether the operational changes that are actually needed — plant closures, exits, asset sales — are being executed alongside the legal process.

3. Who sits in the creditor stack and which tools that actually makes usable. For larger tier‑1 structures with bonds, Schuldscheine and international lenders, StaRUG combined with new money from special‑situations funds can deliver LEONI‑style deleveraging — deep write‑downs, equitisation, new control owner and a clean runway. For mid‑market suppliers dominated by local banks and statutory German‑law bonds (SchVG), the tools and economics look different: German bond law hardwires pro‑rata outcomes; bank groups are more constrained on haircuts; restructurings tend to be earlier, more incremental and more consensus‑driven. Overlaying that with the perimeter question from point 1, the pattern is simple: if the capital structure is small, German‑centric and issued at OpCo, expect corridor‑type outcomes. If the structure sits above Germany, has foreign‑law instruments and has specialist capital in the book, more aggressive, documentation‑driven solutions are on the table — but usually only as long as the German entities have not yet tripped into their own filing obligations.

Taken together, Parts 1 and 2 describe one problem: European auto suppliers are running with too much capacity and the wrong portfolios into a world where capital is scarcer and cheaper competitors are setting the price. The operational and legal toolkits can fix that, but only when they are used together and only at the cost of real capacity and equity.

Listen to the full episode for Daniel’s and Hendrik’s takes on auto suppliers.

This article is based on Episode 14 of Fixed + Floating, featuring Daniel Steiner of PwC and Dr. Hendrik Hauke of Willkie Farr & Gallagher. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.pwc.de and https://www.willkie.com.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Autofacts Market Update March 2026

1.37MB ∙ PDF file

Download

Download

Pwc Automotive Update 251014

2.21MB ∙ PDF file

Download

Download

Willkie General Considerations On Recognition Of Restructuring Processes

1.18MB ∙ PDF file

Download

Download

Download

In the podcast, Daniel Steiner and Hendrik Hauke walked through the mechanics from the factory floor to the courtroom. This note is the companion piece: it sets out why suppliers, as a group, no longer earn their cost of capital, why overcapacity makes refinancing the wrong instrument, how China is resetting the profit pool, and how that translates into credit work on names as different as Schaeffler, ZF or a €250m tier‑2.

The plateau, not the cycle

]The starting point is a survey statistic that is easy to quote and hard to ignore. Around 40% of automotive CEOs say their company will not exist in ten years if they do nothing; over 70% of European suppliers signal performance‑improvement needs; about a third are explicitly distressed, meaning they already face difficulty refinancing.

The character of the number is what matters. In previous cycles, a third of the universe in distress marked the bottom. Today that third is not shrinking between surveys. Top‑line revenue has recovered to roughly 2019 levels and then flatlined, while both the cost base and the cost of capital have deteriorated. A sector that has clawed back its pre‑pandemic revenue but runs with a structurally worse balance sheet and cost structure is not mid‑cycle. It is at a new equilibrium, and the equilibrium is uncomfortable.

The Tier‑1 sandwich

Steiner’s P&L picture is a sandwich. From the top you have volume reductions, under‑utilised factories, OEM call‑off fluctuations and regional demand shifts — every one of which compresses revenue with limited warning. From the bottom you have energy, material, transport and personnel costs, rising interest expense and outright financial distress. The supplier sits in the middle with limited ability to pass on pressure in either direction.

For most of the last two years, the bottom did the damage: German material and labour costs rose while productivity fell; energy prices sat at multiples of Asian levels. More recently, the top has reasserted itself because OEMs are under margin pressure and are pushing it back down the chain. For credit, the relevant point is that the squeeze is now bilateral. This is not a 2009‑style volume collapse with a clean rebound; both sides of the P&L are slow‑moving headwinds.

The profit pool inverted — and China is setting the benchmark

The single chart that frames the episode is the one that shows supplier and OEM profit pools crossing. In 2017–2018 the average supplier out‑earned the average carmaker; by 2024 that relationship had inverted and then sagged for both. Through the pandemic, supplier EBIT margins fell from roughly 7% to about 4%, while OEMs — helped by tight supply and mix — improved their profitability. The part of the chain that historically captured the better margin became the part that captured the worse one.

Since 2023 both supplier and OEM margins have been falling together. That removes the usual safety valve. In a normal cycle, a stressed supplier leans on a healthier OEM for volume, price relief or bridge support. When the OEM’s own margin is contracting, that option disappears. The OEM that might once have absorbed a supplier’s cost inflation is now exporting its own. Distress does not dissipate in that configuration; it propagates.

Underneath the margin story sits a content story. A 2026 BEV simply contains less of what the German supplier base was historically best at. The exhaust system, classic transmission and fuel system are gone. The value migrates towards the battery and the software‑defined EE stack — battery management, ADAS, power electronics, thermal — where Chinese suppliers now offer solutions at a fraction of Western prices with acceptable performance. For a supplier whose problem is portfolio relevance rather than cost control, the first question is no longer “how do I cut costs” but “do I still make anything the vehicle needs.”

Steiner’s China trip made that tangible. In Beijing, drivers sit through red lights that last several minutes, shopping or doing karaoke on large in‑car screens; the car is competing with the smartphone for attention in a traffic jam, not with an autobahn for handling at 200 km/h. The local OEMs design around that use case: rear‑seat comfort, entertainment, and software first; horsepower much further down the list. A “global car” designed for every market at once is under pressure when the Chinese customer is optimising for time in traffic and the German one for time on the motorway.

The OEM profit hole: 30–40% of earnings tied to China

For German OEMs, China is not a side market. At peak, roughly 30–40% of their group profit came from China. Since then their China market share has fallen from about 25% of the market to closer to 15%, taking a disproportionate share of the profit pool with it. The local competition now includes well over 100–150 Chinese OEMs, many focused on EVs with shorter development cycles, localised software and integrated domestic supply chains.

That shift is why OEMs have much less room to “carry” weak European suppliers than they did in the 2010s. A P&L that used to be supported by China is now squeezed there, and the natural response is to push price and volume pressure back into the domestic supply base. For suppliers, the China problem does not show up as a line item in their own accounts; it shows up as fewer levers left at their customers.

China’s supplier margins — and the neck‑breaker scenario

The contrast with China is uncomfortable for suppliers. In the domestic NEV ecosystem, supplier margins still sit above OEM margins; suppliers in China currently earn materially better returns than their European or US peers on comparable content, helped by scale, vertical integration, cheaper labour and subsidised inputs. When Steiner compares Western ADAS stacks with Chinese ones, he finds Chinese offerings priced at roughly a tenth of the Western equivalents, with technology that “works pretty well” — something he tested himself on the streets.

That cannot last indefinitely. With around 50% overcapacity and more than 100–150 OEMs in the field, Steiner expects China to consolidate down to closer to ten domestic OEMs by 2030. In that process, supplier margins will be squeezed as well, and domestic Chinese competitors will replace some of today’s suppliers. The neck‑breaker for European suppliers is what happens if that consolidation leaves a handful of Chinese OEMs and suppliers who can still earn decent economics at a price level that sits structurally below the European cost curve. In that world, the global price anchor for key components is set in a geography where the European cost base simply cannot follow.

The constraint is capital, not margin

This is the point where the operating narrative and the credit narrative diverge. If you only watched margins, you would expect distress to ease wherever a supplier stabilises its operations. That is not what the balance sheets are telling you.

On Steiner’s numbers, German suppliers run the highest interest‑expense‑to‑EBIT ratio of any region. Part of that is price; part of it is that capital structures were built for pre‑2020 volumes and pre‑2022 funding markets. The legacy coupons on bank debt and Schuldscheine do not look extreme in isolation. The problem is that the refinancing that is available today clears at much higher yields, if at all, into a system with less appetite for automotive risk.

A good supplier can still produce an 8–10% EBITDA margin on paper. If the all‑in cost of funding after refi sits above that level, the structure contains a break point that no amount of factory‑floor work removes. Hauke’s shorthand — if your financing cost is in the low‑ to mid‑teens and your EBITDA margin is eight or nine, you have a breaking point — captures the mechanism. The same operating result that stabilises a US supplier can leave a German one stuck, because the German supplier is paying more to carry the same nominal leverage and, crucially, may not be able to refinance at all on like‑for‑like terms.

From a credit perspective, the conclusion is blunt. In aggregate, German suppliers do not earn their cost of capital at current leverage and spread levels. Most are overlevered for the cash flows their portfolios can realistically produce. Debt reduction is not a nice‑to‑have; it is the core of the investment case.

Overcapacity makes consolidation the cure

If capital is the constraint, overcapacity is the disease underneath it. Europe is running with roughly 25–30% excess capacity; China with closer to 50%; North America, having already consolidated, sits in better shape. German plant utilisation has fallen by about 25 percentage points and now sits below the level at which plants are economically viable.

You cannot refinance your way out of structural oversupply. The cash needed to service new debt is exactly what the surplus capacity destroys. North America makes the point in reverse: capacity came out first, and supplier margins only normalised after that process. For Europe, the relevant lever is physical: which plants shut, which lines are sold, which portfolios are broken up. Refinancing, on its own, extends the period over which the excess capacity does damage.

For divestitures, this shifts the arithmetic. A business line that clears at 3–5x EBITDA, below the 4–5x that used to be the norm, can still be the right sale if it removes capacity and debt from a system that has too much of both. Steiner’s “Tetris” picture is accurate: you move blocks to build a more efficient stack, retire the pieces that have no future, and keep some legacy capacity alive longer than you would like because ICE content still needs to be supported through a slower‑than‑planned phase‑out. None of that is free, which is precisely why it cannot be funded with can‑kick refinancings.

The obvious objection is competition law. Hauke’s view is that the fear is overstated. The market is still highly fragmented; OEMs have no interest in creating a single 60–80% supplier in any given component; and “last‑man‑standing” situations, where one supplier continues to produce a critical component, are easier to defend than textbook monopolies in growth markets. The real constraint is not doctrine but coordination: aligning OEMs, regulators and lenders around a consolidation plan that takes capacity out rather than pushing maturities out.

Sorting the survivors

Not every distressed supplier is a wind‑down. The discipline that separates a turnaround from a managed exit is directly usable as a screening tool.

Steiner works with four questions:

1. Portfolio relevance: Does this product still have a role in the vehicle, and on what time horizon? With ICE living longer than expected, some legacy lines stay profitable for longer than consensus assumed; others have no role in a 2030‑plus fleet.

2. Portfolio reshaping: Are there credible M&A or partnership options — joint ventures, regional deals, carve‑outs — that can move a sub‑scale or non‑core unit into a better owner’s hands?

3. Capabilities: Does the company have the operational and engineering depth to execute in the segments that matter — or is it structurally behind in battery, ADAS, power electronics, software and the relevant EE architectures?

4. Capital access: Even with good answers to (1)–(3), is there a shareholder and funding base willing to support a three‑ to five‑year restructuring process?

Two further conditions gate any real turnaround. There has to be a stable, committed owner prepared to fund a multi‑year process, and there has to be basic execution capacity — the right footprint, transparency on unit economics and management depth. Where those are missing, the honest label is “managed wind‑down”, not “turnaround”.

Forvia: the framework in practice

Forvia is the example that shows the framework being applied. Under Martin Fischer, the company suspended its dividend, ran cost programmes, and sold its interiors business to Apollo at a sub‑group margin. The sale did not look attractive if you only looked at the multiple. It made sense because it freed capital, took a lower‑relevance unit off the books and signalled that the group was prepared to reshuffle the portfolio rather than carry everything into the new regime.

That is what a working consolidation trade looks like in a large tier‑1: not a single “heroic” turnaround, but a sequence of disposals and restructurings that move capacity and capital into the parts of the chain that can still earn their cost of capital, and away from those that cannot.

What I take away for credit (Part 1)

1. German suppliers, in aggregate, do not earn their cost of capital — debt reduction is the thesis. On Steiner’s numbers, German suppliers run the highest interest‑expense‑to‑EBIT ratios globally, off balance sheets built for a different volume and spread regime. Even where reported EBITDA margins sit in the high single digits, the all‑in cost of funding after a 2027–29 refi is above that level for a large part of the universe. For IG names with real scale (ZF, Continental), that shows up as a tighter headroom to fund capex and M&A.

2. Consolidation is the cure, but the path is issuer‑specific. Europe’s 25–30% overcapacity is a sector fact; no amount of refinancing changes it. At the top end (ZF, Forvia, large private groups) the consolidation trade runs through portfolio reshuffles, asset sales and selective exits — Tetris played with multi‑billion‑euro blocks. In the Mittelstand, consolidation is more brutal: unprofitable or stranded plants shut; portfolios get carved up between stronger peers; small suppliers disappear without a headline because nobody has the balance sheet or the incentive to carry them. ed is a warning sign, not a comfort.

3. The neck‑breaker is a China‑driven reset of price and margin expectations into a European cost base that cannot follow. Chinese suppliers currently earn better margins than European peers on the same electronic and battery content, helped by scale, subsidised inputs and integrated ecosystems. If Chinese OEM and supplier margins compress further, the global price anchor for key components will be set even further below the European cost curve. For premium German OEMs already losing share and profit in China, the natural response is to push that pressure back into their own supplier base. In that world, the European supplier problem is not just leverage and cost of capital; it is that the margin pool used to service that capital is being repriced by a competitor with structurally lower unit economics.

Part 2 looks at what happens when these operational fixes are not enough and the maturity wall is in sight — and how the current restructuring toolkit gets used in auto suppliers.

Listen to the full episode for Daniel’s and Hendrik’s takes on auto suppliers.

This article is based on Episode 14 of Fixed + Floating, featuring Daniel Steiner of PwC and Dr. Hendrik Hauke of Willkie Farr & Gallagher. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.pwc.de and https://www.willkie.com.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Autofacts Market Update March 2026

1.37MB ∙ PDF file

Download

Download

Pwc Automotive Update 251014

2.21MB ∙ PDF file

Download

Download

Ff 14 Auto Supplier Clean Transcript

382KB ∙ PDF file

Download

Download

Markets have explained this as a macro story: weak housing starts, elevated mortgage rates, the post‑pandemic volume hangover, the natural run‑off of IRA and CHIPS Act manufacturing capex. All of those are real. But they are accelerants. The fuel was already there in capital structures sized at peak 2021–22 EBITDA for a volume trajectory that never persisted, and in business models that have very different floors when volumes move against them.

The rest of this piece is a structured walk‑through of why that matters for credit — beginning with the structural asymmetry between heavyside and lightside as credit propositions, then the demand and operating‑leverage mechanics that are now exposing the difference, then the anatomy of distress in JELD‑WEN and Pfleiderer, and closing with the three things I will actually be watching in HY building products over the next twelve to twenty‑four months.

The Structural Asymmetry: Why the Floor Is Different

The heavyside/lightside distinction in building materials is usually presented as a product taxonomy — cement and aggregates on one side, windows and cabinets on the other. In credit terms, it is something more basic: a difference in how much protection you have when volumes move against you.

Heavyside: how the floor is built

What these businesses actually do.

• Cement, aggregates, ready‑mix concrete, asphalt, and, in some definitions, gypsum wallboard and concrete products.

• High‑bulk, low‑value‑per‑ton products that are expensive to transport; competition is local and logistics‑driven, not brand‑driven.

Why the market structure is so strong.

• Each cement plant typically serves a limited regional market with only a handful of viable competitors.

• Replicating a modern plant requires hundreds of millions to low billions in capex, plus quarry access, environmental approvals, zoning, logistics and energy infrastructure.

• Permits in many US metro areas take ten to fifteen years — and in some markets are effectively unavailable.

How that translates into the credit floor.

• Local oligopolies sustain rational pricing and steady market shares in all but the worst demand shocks.

• Over the last twenty years, cement prices have compounded at roughly four to five percent annually; lumber, as a light‑side proxy, has managed one to three percent.

• In distress, quarries and kilns trade as strategic infrastructure — scarce, difficult‑to‑replace assets with replacement value far above book — rather than as generic manufacturing capacity.

In Belton’s shorthand, heavyside assets are “local infrastructure monopolies with permitting moats, scarce resources, and sustainable pricing power.” The credit market has not always priced that combination as such.

Lightside: where the floor gives way

Business model in one line.

• Windows, doors, cabinets, flooring, siding, insulation, roofing systems, interior finishes — diverse products, but a remarkably similar credit profile.

Three structural weaknesses.

• Low barriers to entry: a windows or cabinets factory can be set up quickly and cheaply relative to a cement plant; capacity can be added in small increments, onshore or offshore, with much lighter permitting.

• Weak bargaining position: large retailers and builder groups — Home Depot, Lowe’s and their European equivalents — control shelf space, specifications, and ordering terms. The manufacturer is usually a price taker into a concentrated channel, not a price maker into a fragmented contractor base.

• Little hard‑asset protection: machinery is generic, many sites are leased, and there is no quarry‑style scarcity in distress. Once you move down the cap stack, there is very little that obviously forces lenders or buyers back to the table.

How that shows up in numbers.

• Product lines run to five‑figure SKU counts; every SKU is open to substitution and downgrade when affordability bites.

• Even before you overlay ownership, CreditSights’ work shows heavy‑side producers earning higher margins, converting more of EBITDA to free cash flow, and running with lower leverage than lightside peers over one to two full cycles — roughly 26.6% vs. 20.9% EBITDA margins, 44% vs. 32% FCF/EBITDA, and 1.4x vs. 1.8x leverage in IG indices.

What this means for HY building products

At investment‑grade level, the pattern is clear: heavyside beats lightside on margins, cash generation, and leverage, in both US and European indices. At high‑yield level, it becomes a solvency question.

• The US HY building products sleeve is “much more lightside‑focussed, and mostly lower‑margin,” dominated by businesses with low barriers to entry, weak bargaining power with retailers, and limited hard‑asset cover.

• Layer private‑equity capital structures — nine‑times‑plus leverage sized at peak 2021–22 EBITDA — on top of that base, and you get exactly the pattern we see now: heavyside HY can “support HY cap structures”, light‑side HY produces stressed and distressed clusters when the cycle turns.

The McKinsey counter‑argument — that intra‑segment ROIC dispersion exceeds inter‑segment dispersion, so execution beats exposure — is not wrong for equity. But for credit, the question is not the ceiling; it is the floor. And the floor in lightside building materials, in a sector downturn, is structurally lower than the floor in heavyside — not because lightside companies are badly run, but because the economics and the asset base allow it.

The Demand Picture: Volume, Pricing, and Where It Hurts

Between 2019 and 2022, the nominal revenue picture in building materials looked deceptively strong. COVID‑era household formation, stimulus, and migration drove a genuine demand surge in housing and R&R. Supply chains were tight. Pricing power was real. Top lines grew quickly, and the sector looked healthy.

Underneath, volume was already rolling over by 2022. Single‑family starts peaked in 2021 at around 1.13 million, then fell by 11% in 2022 and another six percent in 2023, before a brief seven percent rebound in 2024 that never fully rebuilt the pipeline. Total home starts have “ground steadily lower,” down three percent in 2022, eight percent in 2023, four percent in 2024, and one percent in 2025. Multi‑family followed a different, but equally volatile, profile. Upstream suppliers felt the boom and the bust more acutely than the builders, because their utilisation and fixed‑cost absorption moved one step earlier in the chain.

The ACA’s spring 2026 forecast has US real construction spending down three percent in 2026 after a four percent decline in 2025 — with non‑residential the weakest at minus eight percent in 2026, residential down just under two percent, and public/infrastructure down three percent. Residential is the number that matters most for lightside HY credits. A windows or cabinets plant is built around units moving through a system. When starts fall from 1.5 million to 1.1 million, factories lose throughput, fixed costs stop being theoretical, distributors destock, and EBITDA can unravel quickly.

Content per start helps at the margin — bigger houses, more premium finishes, higher‑spec glazing. But in a hard downturn, the premium specification thesis runs in reverse. Builders shrink kitchen footprints, remove islands, simplify bathroom layouts, and downgrade from semi‑custom to stock cabinets. A cabinet package that ran to 20–25 thousand dollars can be cut to 12–15 thousand. Window openings shrink, architectural shapes disappear, glazing specifications are simplified, and frame quality is reduced. Doors can be standardised and cheapened. Sidings see less mix pressure, because the envelope still has to exist, but even there builders can move from fibre cement to vinyl, remove stone veneer, and simplify trim. Insulation is the relative outlier: building codes and energy standards make deep downgrades difficult, so it behaves more like a quasi‑heavyside category.

The R&R market is the other potential release valve. On paper, the fundamentals are excellent: a median US home age of around forty years, roughly 300 thousand dollars of homeowner equity, and mortgage lock‑in keeping people in existing houses longer. In practice, R&R is highly sensitive to psychology and financing costs. Homeowners may have equity and a kitchen in need of replacement, but high contractor quotes, floating rates on HELOCs, and a background of economic and geopolitical uncertainty are enough to stall projects. Fixed‑rate second‑lien mortgages — which let homeowners preserve ultra‑low first‑lien mortgages while financing renovation with payment certainty — could unlock the next leg if rates drift structurally lower. Until then, R&R looks more like a sideways market than a cycle‑saver.

On the heavyside, the same macro environment looks less dramatic. Aggregates and cement pricing have remained strong despite negative real growth in construction, with aggregates up four to seven percent, cement flat to modestly up, and ready‑mix and asphalt broadly positive. Vulcan reported a 3.5% rise in volumes in Q1 2026; Martin Marietta reported thirteen percent, both guiding for low‑single‑digit volume growth and three to four percent price inflation for the year. Public construction’s apparent three percent real decline is as much a denominator story — inflation in construction spending includes materials costs — as it is a demand story. For the heavyside IG names selling into infrastructure, that inflation is a revenue benefit, not a problem.

One Trade in Disguise: The HY Index

At index level, the US HY building products segment is not ten different trades; it is one housing bet dressed up as diversification.

The anatomy is straightforward. On the lightside, you have windows, cabinets, roofing, flooring, siding, and building products distributors — nominally distinct businesses with different SKUs, customer relationships, and operating histories. On the macro side, you have a small set of drivers: housing starts, mortgage rates, affordability, R&R spending, and dealer inventory cycles. In good markets, investors can focus on company‑specific execution, operational improvement, and idiosyncratic positioning. In a downturn, the correlations spike and the sector trades as leveraged housing beta.

“I am increasingly viewing the high‑yield building products market in the States as essentially a leverage play on the US housing cycle, disguised as effectively a diversified quasi‑industrial index.”

Private equity ownership is not an incidental detail here; it is a structural feature. Lightside businesses are easier to buy, easier to consolidate, easier to “improve” on visible metrics, and less capital‑intensive than heavyside quarries and kilns. The US HY building products snapshot in Belton’s deck carries two flags in bold on the lightside side of the ledger — “Heavy LevLoan Element to Space” and “High Degree of PE Ownership.” In practice, that means light‑side HY investors are not just underwriting operating risk; they are underwriting sponsor behaviour, covenant flexibility, and the willingness to use liability‑management tools when the numbers do not add up.

The floating‑rate loan market has been the first place this showed up. As the Fed’s hiking cycle from 2022 onward repriced interest expense almost immediately, coverage ratios in cov‑lite structures compressed long before maturities became an issue. The bond market, with fixed coupons, enjoyed a temporary insulation; that insulation is now disappearing as the 2027–2029 maturity wall moves into view and operating performance remains under pressure. CLO tolerance for amend‑and‑extend, incremental facilities, and other forms of maturity‑wall engineering has kept some of the weakest names alive longer than a traditional cycle would have, but it has not solved the underlying problem.

The Anatomy of Distress: JELD‑WEN and the Lightside Model

JELD‑WEN is a useful specimen of what is going wrong on the lightside, precisely because it is not a cartoon villain. It is a broadly representative business, with real scale, and its problems line up closely with the structural weaknesses of the model.

The portfolio is about two‑thirds doors, around nineteen percent windows, and the rest other building products, with a meaningful European footprint accounting for roughly a third of revenues. End‑market exposure is around 87% residential, split between roughly half new construction and just over forty percent R&R, with thirteen percent non‑residential. The revenue base is concentrated: the top ten customers account for about half of sales; Home Depot and Lowe’s together represent around thirty percent, and Home Depot alone roughly seventeen.

In credit terms, that is a highly levered price taker into a concentrated channel, not a price maker into a fragmented contractor base. When volumes fall and the company attempts to push price to offset cost inflation, it is pushing into two retailers who can — and do — use their bargaining power on pricing, service standards, shelf space, and ordering terms. When operations are under strain, service slips; when service slips, volumes can be lost quickly to alternative suppliers or to destocking decisions made at the retailer.

The current numbers capture that spiral. American core revenue was down roughly fourteen percent year‑on‑year in the first quarter, driven by a thirteen percent volume and mix decline and a one percent price decline. The business is tracking both new construction weakness and R&R softness at the same time. EBITDA margins have compressed towards four percent, a long way below Belton’s “steady‑state” high‑single‑digit view for a business with this positioning and customer concentration, and further still from the mid‑teens margins better‑placed building products names can sustain. Leverage sits around nine times.

The capital‑structure pressure point is the December 2027 unsecured maturity, which goes current in December 2026. The term loan is higher in the stack and has collateral; it trades in the low‑to‑mid‑90s. The unsecureds, reflecting the maturity wall and the low probability of a vanilla refi at nine times leverage on four‑percent margins, sit in the 70s. In practice, they are already being treated as the shock absorber for whatever liability‑management exercise comes next.

There is optionality in the European business: a disposal at a decent multiple could provide some deleveraging and runway, even if it does not fully solve the operating problem. But assets sold under stress rarely clear at the multiples needed to transform the balance sheet, and every quarter of weak performance erodes bargaining power. Belton’s base case is not an immediate US‑style court process, but a bond‑focused liability management exercise built around the unsecured notes. If North America does not stabilise and the European option is not realised, a broader restructuring remains possible, but the window narrows as the “current” date approaches.

The European Parallel: Pfleiderer and Silekol

The story that makes JELD‑WEN feel uniquely American — cov‑lite loans, LMEs, and sponsor engineering — looks less exceptional once you look at Pfleiderer.

The company — a German wood‑panels manufacturer that had already restructured once — executed a drop‑down in December 2025 that moved its Silekol adhesives business outside the restricted group, selling ninety percent of the equity into newly created unrestricted subsidiaries. Around 110 million euro of new debt was raised by Monarch and 25 million by SVP, secured on the Silekol shares. The new money sits structurally senior to the existing bonds; the guarantees and collateral that Silekol had provided to the notes fall away. What is left in the restricted group is a weaker asset base and a subordinated bond stack — and a schedule of maturities and covenants that now have to be negotiated from a position of reduced collateral coverage.

The twist is that this was done in an ostensibly “cleaned‑up” post‑restructuring document. Covenant tightening in the prior process — priming caps, investment limits — was supposed to prevent exactly this kind of structure. Instead, the company appears to have used an asset‑sale path to bypass those tighter limits. The lesson for European HY is simple: if you still believe you are in a fundamentally more consensual, less aggressive covenant environment than US HY, you are underwriting a regime that is fading quickly.

The most exposed cluster is not the CCC tail where equity is already gone, but the stressed single‑B names trading in the 60–85 range, with some equity optionality still worth preserving, material refinancing uncertainty, and a 2027–2029 maturity wall ahead. In that zone, sponsors have every incentive to use drop‑downs, portability, and other structural tools to preserve optionality rather than write cheques. The lightside distributors — leveraged spread businesses dependent on inventory turnover, supplier terms, and housing volumes, with very limited hard‑asset protection — look like the natural place for this to surface first.

Operating Leverage: Where Utilisation Actually Bites

The part of the discussion that probably carries the highest practical value for HY investors is operating leverage, and it sits later in the conversation than you might expect.

On the heavyside, operating leverage is a utilisation story with hard thresholds. Cement plants and aggregates quarries are “hugely asset‑intensive” — you still have to pay maintenance, labour, environmental compliance, and depreciation, whether you are at full tilt or not. The sweet spot for cement utilisation is roughly 80–90%. Once plants drop below around 75%, fixed‑cost absorption deteriorates quickly; below about 70%, margin compression accelerates, because kilns must stay hot and energy efficiency drops sharply as loads fall. Aggregates quarries run slightly lower, with a healthy band around 70–85%, but the principle is the same: blasting, hauling, and reserve‑development costs are largely fixed, so unit economics worsen quickly if tons fall too far.

Today, US heavyside utilisation still sits on the right side of those thresholds. Residential is softer, but infrastructure and data‑centre work have kept plants in “supportive rather than stretched” territory. A cement producer running at 78% utilisation can still generate good cash flow because pricing holds and the local oligopoly remains rational.

On the lightside, the cliff is steeper and much less forgiving. Plants and branch networks were built for a 2021–22 world of 1.5 million starts and strong R&R, not for a 1.2–1.3 million environment with flat renovations. Many factories are already running below their optimal throughput; the fixed‑cost base — labour, SG&A, logistics, footprint — was sized for a bigger system. From that starting point, Belton’s estimate is that an incremental five percent volume decline can translate into a ten to twenty percent EBITDA drop and 100–300 basis points of margin compression, because you are spreading fixed costs over fewer units and have limited ability to move price.

The core difference is that heavyside has three things lightside does not: high barriers to entry, a stronger negotiating position, and hard assets that still matter at the end of the capital structure. Lightside HY has low barriers to entry, a weaker negotiating position against retailers, and very little in the way of unambiguously valuable hard assets once you are in distress. Operating leverage in that context is not an abstract concept; it is the path from “expensive but refinanceable” to “fulcrum security” over two or three reporting periods.

The QXO/TopBuild Antidote

The one apparent counterexample in this cycle is QXO/TopBuild: sophisticated strategic capital paying around 17 billion dollars for a lightside installation and distribution platform at the same time as multiple other lightside HY names are in active distress.

The way Belton framed it — “the exception that proves the rule” — is useful. TopBuild sits at the part of the light side where scale can materially change the economics: installation and distribution, rather than manufacturing. At that end, economies of scale in pricing, procurement, and logistics are real; route‑to‑market control matters; and the capital structure is not obviously mis‑sized for the underlying business. The deal says less about the average lightside HY issuer and more about what happens when you combine a decent business with genuine scale and a clean balance sheet.

For HY, the take‑away is to refine the framework, not abandon it. Within lightside, there is a difference between manufacturers who sell into concentrated retailers with low asset protection and distributors/installers who can use scale and local density to earn better economics. The former dominate the US HY building products index. The latter are rare — and, when they exist, get bought by people like Brad Jacobs.

The Infrastructure Moat and the M&A Shift

The heavyside credit story is not the one implied by top‑down construction spending headlines. Infrastructure and public construction in the US is forecast to dip in real terms in 2026, in part because the natural run‑off of IIJA and other programmes coincides with higher measured inflation in construction spending. But inflation in US construction put‑in‑place includes materials costs, so higher inflation is actually positive for those heavyside credits that sell into infrastructure projects: nominal spending can hold up even as real volumes are flat.

From the perspective of aggregate producers like Vulcan and Martin Marietta, the volume story remains quietly robust. Both are guiding to modest volume growth alongside mid‑single‑digit price increases, and both have successfully pushed through price in recent years even as parts of the construction complex softened. The more important shift is in M&A. The big transactions in the sector — Vulcan/CalPortland divestitures, CRH’s purchase of Weaver in Texas, Martin Marietta’s asset swaps with Quikrete — are not primarily about current EBITDA; they are about acquiring permitted reserves in markets where entitlements are scarce and community opposition is organised.

In high‑growth metro areas like Dallas, Phoenix, or Atlanta, securing a new quarry can take a decade or more of permitting, or simply may not be feasible. Once a company holds a permitted quarry in one of those basins, it is effectively holding a regulated‑like infrastructure asset with a multi‑decade life and a high barrier to entry, not a generic commodity plant that can be undermined by incremental capacity. That is what heavyside valuations are paying for when they trade on 15–17x EV/EBITDA: not next year’s earnings, but fifty years of irreplaceable reserves and embedded local pricing power.

The cross‑Atlantic piece sits behind this. For the first time in years, Europe’s civil‑engineering spending is forecast to outpace the US: Euroconstruct has European construction growth at 3.3% in 2026, with civil engineering the main driver. European IG building materials credit currently trades roughly ten basis points inside the broader IG index; US IG building materials sits about one basis point outside. The market is still pricing the old narrative — weak Europe, strong US — just as the data begin to suggest the opposite for infrastructure‑exposed heavyside credits.

Names like CRH, which combines US aggregates with European cement and infrastructure exposure and has relisted on the NYSE, sit squarely on that pivot. Belton’s house view has been CRH as the flight‑to‑quality trade in both EUR and USD IG. The long‑CRH / short‑Mohawk framing — Europe‑as‑call‑option plus US infrastructure on the long side, European consumer exposure and energy intensity on the short — is one expression of the idea that the market has not fully re‑rated heavyside as a structurally different asset class, especially in Europe.

The obvious downside risk is energy. If the Iran conflict escalates into a sustained oil and gas shock, with Brent holding above the 120–130 dollar range and European gas and power prices responding accordingly, the economics for glass and cement producers could deteriorate rapidly, particularly where furnaces and kilns cannot be idled. In that scenario, the heavyside premium would need to be re‑examined; until and unless that shock materialises, the infrastructure moat looks under‑priced.

What I Take Away From This — And What I Will Be Watching

1. Lightside HY is low barriers, weak bargaining power, and thin hard‑asset cover — and that combination is what is breaking, not “building materials” in general.

The stressed names — Cornerstone, JELD‑WEN, Oldcastle BuildingEnvelope, USLBM, Flytec, Pfleiderer — sit in the part of the sector with the least protection: low barriers to entry, structurally weak negotiating positions against retailers and builders, and very little in the way of hard assets that carry obvious value in distress. On top of that, they have PE‑style capital structures sized at peak 2021–22 EBITDA. When volumes slipped back toward pre‑pandemic levels and pricing normalised, leverage did not come down. We now see exactly the combination you would expect: margin compression from operating leverage, limited ability to push price, and sponsors using structuring tools rather than cheques to buy time. I will be watching the intersection of those three features in any new HY lightside underwriting: barriers to entry, bargaining power in the channel, and the quality of hard assets beneath the stack.

2. Operating leverage on the light side bites earlier and harder than most models admit.

A lot of investment‑grade work treats operating leverage as a gentle slope; the recent lightside HY experience suggests a cliff. Plants are already running below optimal utilisation, and the fixed‑cost base was built for a larger system. From there, a further five percent volume decline can take ten to twenty percent off EBITDA and 100–300 basis points off margins, with very little room to manoeuvre if you are at eight, nine, or ten times leverage. At the same time, the heavyside names most investors worry about have utilisation bands and pricing power that keep them away from the dangerous part of the curve for longer. I will be watching reported utilisation, not just revenues, in light‑side names, and asking the simple question: where is the cliff, and how far away are we in units, not in narrative?

3. The HY building products sleeve should be underwritten as one housing trade with structuring optionality attached, not as a diversified industrial basket.

Belton’s “one housing trade dressed as ten tickers” line is not a metaphor; it is how the index actually behaves in stress. The macro drivers are shared. The ownership patterns are shared. The capital‑structure design patterns are shared. In practice, that means sizing and hedging the exposure as a housing‑cycle bet, distinguishing carefully between loans and bonds, and assuming that every cov‑lite name has LME and drop‑down optionality unless you have read the documents closely enough to conclude otherwise. It also means recognising that the tools we saw first in US retail and US HY — J.Crew‑style transfers, Pfleiderer‑style Silekol structures — are now firmly part of the European playbook.

Listen to the full episode for Andy’s takes on building materials.

This article is based on Episode 13 of Fixed + Floating, featuring Andy Belton of Creditsights. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://know.creditsights.com/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Ep13 Transcript Andy Belton

381KB ∙ PDF file

Download

Download

Building Products Podcast Heavyside V Lightside Revised Deck

2.94MB ∙ PDF file

Download

Download

This growth has attracted not just asset managers and pensions, but also the attention of supervisors and the Basel Committee, which now treat synthetic risk transfers as a systemically relevant market that needs close monitoring for rollover risk, bank–NBFI interconnectedness, and the quality of “significant risk transfer” itself.

The rest of this piece is a structured walk‑through of the asset class, anchored on the questions an institutional analyst would actually ask: what SRT is and is not, how the structures work, why the market has grown, where the real risks and incentives sit, and what supervisors are worried about as the market scales.

What an SRT Actually Is — And What It Is Not

At its core, an SRT or CRT trade is a way for a bank to hedge credit risk on a portfolio of on‑balance‑sheet loans and reduce the risk‑weighted assets (RWAs) the regulator charges against that portfolio. The bank selects a pool of loans — typically corporate, SME, mortgage, infrastructure, project finance, subscription line, or trade‑finance exposures — and transfers the first‑loss or mezzanine slice of portfolio credit risk to an external investor, while retaining the assets, client relationships, and senior risk.

In the canonical example, the bank holds a 1 billion dollar corporate loan portfolio and buys protection on, say, the first 80 million dollars of losses, often corresponding to the 0–7% or 0–12.5% tranche. If the transaction meets the definitional and operational criteria for a synthetic securitisation under Basel and local rules (segmentation into tranches, performance linked to underlying exposures, eligible credit protection, etc.), the bank can recognise a substantial RWA reduction on the protected portfolio.

The economic question for the bank is straightforward: is the coupon paid on the protected tranche lower than the cost of the capital that is freed up? If yes, the trade improves return on tangible equity without issuing new equity or shrinking the loan book; if not, the bank either keeps the loans unhedged or does not originate them in the first place.

It is worth being precise about what SRT is not. It is not a vanilla CDS: there is no requirement for daily margining in fully funded structures, the bank must own the underlying exposures, credit events are narrowly defined (typically failure to pay, bankruptcy, and tightly drafted restructuring), and there is usually no scope to buy “naked” protection on assets the bank does not hold. It is also not a traditional CLO: SRT deals reference assets already originated and held by the bank, with the loans staying on the bank’s balance sheet and the purpose being capital relief and risk management rather than funding or broad loan distribution.

Market Size, Growth, and Who Is Involved

Because SRT is largely a private market, numbers are approximate, but three data points line up reasonably well: manager surveys, law‑firm primers, and the Basel Committee’s recent analysis all suggest annual issuance of roughly 30–35 billion dollars of tranches, 350–400 billion dollars of underlying loans hedged each year, and a cumulative stock of hedged assets above 1 trillion dollars. SRT volumes have more than tripled since the mid‑2010s, and the number of issuing banks has risen into the dozens in Europe and is climbing in North America.

Geographically, the market began as a European phenomenon but is now broadening. European banks — particularly those with relatively tight capital positions — remain the most active, joined by Canadian, US, and a growing group of Central and Eastern European institutions as local rules clarify the treatment of synthetic securitisations. Recent US Federal Reserve guidance and reservation‑of‑authority letters have explicitly acknowledged that well‑structured bank‑issued CLNs can qualify for capital relief, subject to structural caps and strict operational criteria, which is accelerating US issuance from a low base.

On the investor side, the dominant players are specialist credit funds, hedge funds, insurers, and, increasingly, pension plans and sovereign wealth entities, all attracted by double‑digit gross spreads, relatively low expected loss, and low mark‑to‑market volatility. Regulators, including the Basel Committee, have noted that this shift moves risk from regulated banks to non‑bank financial intermediaries (NBFIs), and have flagged the need to monitor bank–NBFI interconnections and leverage in the investor base.

A Sample Trade — What the Capital Relief Actually Looks Like

A concrete example from Clifford Chance illustrates how SRT economics work for a bank under a standardised capital regime.

• The bank holds a 1 billion dollar portfolio of corporate or second‑lien mortgage exposures, risk‑weighted at 100% under general credit rules, so RWAs are 1 billion and minimum total capital requirement at an 8% ratio is 80 million.

• The bank executes a fully funded synthetic securitisation, buying protection on a 12.5% first‑loss tranche (125 million notional) from investors, collateralised with cash on deposit.

• Post‑trade, the bank’s exposure is split into:

  • a senior 875 million retained tranche risk‑weighted at 20%, producing 175 million of RWAs and 14 million of required capital; and

  • a protected 125 million first‑loss tranche, which receives a 0% risk weight because it is fully hedged with eligible collateral.

Total RWAs fall from 1 billion to 175 million, and minimum capital requirement falls from 80 million to 14 million, implying 66 million of capital “freed up”, against which the bank now pays an annual premium to investors. The trade makes sense so long as the cost of that premium is below the bank’s internal cost of the 66 million of capital it no longer has to hold.

This simplified example captures the basic mechanics behind the “75% RWA reduction on a typical 0–7% or 0–12.5% tranche” that practitioners quote: the hedged first‑loss piece takes the expected credit hits, and the retained senior piece becomes materially cheaper in capital terms.

The Mechanics: Three Main Structures

Market practice and law‑firm guides converge on three principal SRT structures.

1. SPV with fully funded guarantee / CDS.
   The bank enters into a credit‑protection contract with an orphan SPV, which in turn issues credit‑linked notes (CLNs) to investors and places the proceeds in a collateral account pledged to the bank. Losses on the reference portfolio reduce the tranche notional and are paid out of the collateral, while premiums from the bank plus collateral yield fund investor coupons. This is structurally close to a synthetic securitisation and, when drafted to meet all criteria in Regulation Q (US) or the EU CRR, can deliver capital relief without specific ex‑ante supervisory permission.

   

2. Direct CLN issued by the bank.
   Here, the bank itself issues CLNs directly to investors referencing the portfolio, and receives the note proceeds as cash on its balance sheet, with contractual mechanisms to ensure those proceeds function as collateral in an identifiable account. Losses on the reference pool reduce the CLN principal and future coupons proportionally. Because the risk‑transfer instrument can be “hypothetical” and collateralisation is more complex, these deals often require explicit reservation‑of‑authority relief or detailed dialogue with supervisors to confirm capital treatment, and the Fed has imposed caps on the total reference portfolio size for some bank CLN programmes.

   

3. Bilateral funded or unfunded protection.
   In bilateral trades, the bank signs a credit derivative or financial guarantee directly with an investor or insurer, which may post full collateral into a pledged account (funded) or rely on its own balance sheet as an eligible guarantor (unfunded). Fully funded bilateral deals can be capital‑efficient but require careful structuring of collateral arrangements and margining to meet operational criteria; unfunded deals can be simpler from a funding perspective but expose the bank to protection‑seller credit and downgrade risk, which UK and EU supervisors explicitly highlight as a prudential concern.

   

Across all structures, three hard constraints apply: the bank must own the reference assets; the credit protection must meet detailed eligibility tests (e.g., defined credit events, enforceable protection, permissible call features); and the loss measurement process must be robust enough for supervisors to accept that risk has genuinely been transferred.

True‑Up, True‑Down, and Why Losses Are “Real”

One of the distinct features of SRT relative to index CDS is the true‑up / true‑down mechanism. When a loan in the reference pool experiences a credit event — typically failure to pay, bankruptcy, or contractual restructuring — the bank records an initial expected loss and adjusts the protected tranche notional accordingly.

As workout proceeds and actual recoveries are realised, the bank and investor reconcile to a final loss number: if ultimate recoveries are higher than first assumed, the tranche is “trued up” and past coupon calculations are adjusted so that investors are compensated for having had more notional at risk than initially reflected; if recoveries are lower, the tranche is written down further. Law‑firm primers emphasise that this two‑step approach is now standard in capital‑relief trades, precisely to align the protection with the bank’s realised credit experience rather than modelled or index‑based proxies.

For investors, the implication is that performance is tied to realised default and recovery on the underlying loans, not to mark‑to‑market spread moves. That largely explains the low NAV volatility and focus on realised loss rather than valuation noise that consultants and managers highlight as a key attraction of the strategy. The trade‑off is that investors must rely heavily on the bank’s workout and provisioning processes; disputes over loss calculations are rare but can be complex, and documentation has become increasingly granular to minimise room for disagreement.

Portfolio Construction and What Gets Hedged

The SRT market is heavily skewed toward granular, performing corporate and SME loan portfolios, with law‑firm and consultant surveys pointing to large‑cap corporates as roughly half the market by volume, SMEs as another mid‑teens share, and the remainder spread across commercial real estate, mortgages, infrastructure, project finance, subscription lines, and specialised exposures like data centres or trade finance.

Banks choose portfolios based on three lenses:

• Size and scalability: reference portfolios are typically at least 1 billion dollars in notional to justify fixed transaction costs and support a programme of repeat deals.

• Capital efficiency: banks favour asset classes where the regulatory risk weight is relatively high compared with the bank’s internal view of credit risk, because hedging those assets delivers the most RWA relief per unit of expected loss.

• Data and systems: SRT requires deep historical performance data and robust ongoing reporting; some loan books that look attractive on a risk‑return basis simply lack the data infrastructure to support a capital‑relief trade.

Portfolios are intentionally granular: 200–500 names in corporate pools, and sometimes thousands of obligors in SME or trade‑finance deals, so that investors are not overexposed to idiosyncratic single‑name risk within a thin first‑loss tranche. Eligibility criteria, concentration caps, and replenishment rules (where portfolios are revolving) are central to managing adverse selection and ensuring that reference pools reflect the bank’s core lending rather than a dump of troubled names.

Lifecycle, Amortisation, and What Happens When Loans Repay

Most SRT trades run for three to five years in legal maturity, with a replenishment period in the first two or three years during which the bank can replace repaid loans with new eligible assets, followed by an amortisation phase in which the portfolio and tranche notionals shrink in line with loan repayments. BIS and Basel Committee work suggest that, in European data, original maturities tend to cluster around two to four years, with effective investor exposure roughly three to five years once replenishment and calls are taken into account.

During replenishment, portfolio composition is governed by eligibility criteria and reporting: new loans must meet pre‑agreed rating, sector, geography, and concentration limits, and breaches typically lead to exclusion or substitution. After replenishment ends, amortisation may be pro rata across tranches or sequential (reducing senior tranches first), with many transactions including switch mechanisms tied to performance triggers.

This medium‑term profile is exactly why rollover risk matters. SRTs are not permanent capital solutions: when protection matures, the capital benefit runs off unless the bank executes new trades. If a bank is relying heavily on SRT for CET1 optimisation, and the issuance window is closed or prohibitively expensive when existing deals roll off, RWAs will creep back up while the bank may still be holding or originating similar loan books. In normal markets, this is manageable; in a stress scenario, it can become a non‑trivial constraint.

Defaults, Bank Credit Risk, and the Credit Suisse Example

Default mechanics, as discussed, run through the credit‑event definitions of failure to pay, bankruptcy, and qualifying restructuring (forgiveness of principal or interest). The two‑step true‑up / true‑down process governs how losses are allocated and reconciled, with investors ultimately on the hook for realised losses on the reference portfolio up to the size of their tranche.

It is important to distinguish portfolio credit risk from bank credit risk. The investor is primarily taking credit risk on the reference portfolio, but in some structures it is also exposed to the bank itself. In a direct CLN or certain unfunded formats, the bank’s own credit quality and place in the capital structure matter because coupons and return of principal depend on the bank’s ability to pay; in a fully collateralised SPV structure with segregated collateral, that originator exposure is more limited and investors are largely looking to the collateral account rather than the bank for repayment.

The Credit Suisse failure is a useful stress‑test. When Credit Suisse AT1 instruments were written down to zero in the UBS rescue, SRT positions linked to the bank’s loan portfolio continued to perform in line with documentation. In those trades, SRT investors sat structurally senior to AT1 and, in some cases, close to insured depositor level in the creditor hierarchy, with risk exposure driven mainly by the underlying portfolio and collateral, not by the bank’s subordinated capital stack. That episode illustrates what well‑structured, fully funded SRTs can look like in a genuine bank failure, but it is not a universal rule: the degree to which losses on unsecured debt or deposits spill into an SRT depends on where the SRT sits in the capital structure and how it is collateralised.

Pricing, Returns, and the Comparison That Matters

On pricing, banks and investors approach the same trade from different directions. Banks price off capital relief: for a given portfolio, they compute how much RWA is saved by transferring a tranche and ask whether the premium implies a cost‑of‑capital that is below their internal hurdle. Investors price off expected loss and risk‑adjusted return, using portfolio data and stress scenarios to estimate defaults and recoveries and then comparing net spread to other illiquid credit opportunities such as CLO equity, private credit, or mezzanine lending.

Across sources, a broad range emerges: bank‑issued or SPV CLNs on large‑corporate portfolios often pay cash plus 6–11%, with consultant data suggesting long‑run spreads around the high single to low double digits, and expected losses for investment‑grade corporate pools in the 0.1–0.9% per annum range and often lower for balance‑sheet securitisations relative to true‑sale ABS. EBA analyses and law‑firm summaries emphasise that lifetime default rates on balance‑sheet securitisations have historically been materially lower than on comparable public corporate credit indices, although the observed window has been benign.

Where SRT is most clearly differentiated is not headline yield but volatility and loss profile. Typical SRT corporate portfolios sit around BBB on average rating, with high granularity and limited exposure to higher‑beta sectors such as leveraged software, whereas CLO equity and some private credit strategies deliberately concentrate in single‑B and BB borrowers with more cyclical cash flows. The result is a return stream that, historically, has exhibited low NAV volatility and limited drawdowns, but also relatively capped upside compared with highly levered or opportunistic credit strategies.

How SRT Behaves Through the Cycle — and What Regulators Worry About

The real question for any capital‑relief product is how it behaves in stress. The COVID episode offered a partial test: spreads on new SRT issuance widened, but banks also reconstituted portfolios, removing the most COVID‑sensitive sectors (retail, hospitality, travel) from new pools and accepting slightly wider coupons on structurally lower‑risk portfolios. That pattern — both pricing and portfolio composition adjusting — is now a standard part of how sponsors describe the product’s cyclicality.

Supervisors, however, are focused on a different risk: rollover risk and procyclicality. If banks come to rely heavily on SRT for capital relief, and the investor bid for new protection evaporates or becomes prohibitively expensive in a downturn, then existing trades will amortise, RWAs will creep back up, and CET1 ratios will fall exactly when credit losses are rising. BIS work and recent Basel Committee commentary suggest that, while average capital relief across issuing banks is currently modest (on the order of tens of basis points of CET1, with only a handful of banks above 1 percentage point), the potential for more material dependence exists as issuance grows.

So far, three factors have prevented rollover risk from becoming acute: strong credit performance in underlying portfolios, the absence of a “maturity wall” where many SRTs expire at once, and still‑limited aggregate dependence on SRT for capital ratios. But both the Basel Committee and the ECB are explicit that rollover and back‑leverage risks are areas for ongoing supervisory attention, particularly as more banks adopt programme‑style issuance and NBFIs fund SRT investments with bank credit lines or repo.

Adverse Selection, Moral Hazard, and Residual Risks

The textbook critique of SRT is intuitive: if a bank can offload first‑loss risk, it may have weaker incentives to monitor credits, and if it has private information about deteriorating loans, it might place those into SRT pools and let investors take the hit. Law‑firm papers and consultant notes acknowledge this concern but emphasise three structural mitigants:

• Regulatory risk retention. Banks must retain a material net economic interest in any securitisation or SRT, and cannot reference 100% of any exposure; they take losses alongside investors whenever the portfolio underperforms.

• Programme economics and reputation. Most issuers are repeat users who rely on the SRT market multiple times per year. Any evidence of cherry‑picking or under‑disclosure would damage pricing, investor appetite, and, potentially, capital relief; issuance costs and reputational risk make “one and done” behaviour uneconomic.

• Documentation and eligibility criteria. Detailed eligibility tests, replenishment rules, and ongoing reporting, combined with due diligence on underwriting standards and historical loss data, give investors tools to push back on portfolio construction and monitor alignment.

That said, none of these mitigants eliminate adverse‑selection risk entirely. Data for SRT portfolios are often less transparent and more heterogeneous than for public securitisations, and regulators point out that imperfect information and complex structures can make it harder for investors and supervisors to detect weak alignment ex ante. The more concentrated the investor base and the more programmatic the issuance, the more important governance, internal controls, and supervisory review become.

Regulatory Scrutiny — What Law Firms and Supervisors Actually Focus On

Regulatory frameworks in Europe, the UK, and the US all revolve around two core tests: definitional requirements (does the trade qualify as a synthetic securitisation or SRT under the rules?) and operational requirements (are there any features that undermine genuine risk transfer?). Law‑firm primers distil the supervisory focus into several recurring themes:

• Genuine risk transfer and “no implicit support”. Supervisors want comfort that risk has truly moved from bank to investor for the life of the trade. Features that could re‑route losses back to the bank — implicit guarantees, over‑generous calls, side‑letters, or high‑cost protection that effectively pre‑pays losses — are viewed sceptically and can lead to capital relief being denied or later withdrawn.

• Pricing and “high‑cost” protection. Rules and guidance constrain structures where the total premium over the life of the trade can exceed the maximum protection amount, or where coupons are clearly misaligned with expected loss; such trades may be treated as funding or profit‑smoothing mechanisms rather than genuine risk transfer.

• Portfolio management, calls, and early termination. Clean‑up, time‑call, and regulatory call options are permitted but tightly circumscribed; investor‑friendly termination rights that might allow protection to disappear when risk increases are generally incompatible with capital relief.

• Choice of protection instrument and counterparty. In the US, only certain guarantors qualify as “eligible”, and unfunded guarantees by insurers or corporates must meet detailed tests. UK and EU supervisors explicitly flag additional prudential risks from unfunded protection, including late payment, non‑payment, and downgrade risk, relative to fully funded structures.

In practice, this means that SRT transactions are heavily lawyered, require legal opinions on enforceability, and often involve pre‑issuance dialogue with supervisors, especially for novel structures or direct bank‑issued CLNs. The regulatory mood is not hostile — the Basel Committee explicitly notes that SRTs appear more prudently structured than pre‑GFC synthetic deals — but it is watchful, with an emphasis on substance over form.

Where SRT Sits in the Bank–Private Credit Architecture

At the deal level, SRT does not dictate individual loan terms or create an originate‑to‑distribute model in the way leveraged finance syndication or CLO arbitrage can; origination teams generally structure loans based on client and risk considerations, and SRT desks later aggregate portfolios of already‑originated assets. At that micro level, the incentives remain those of a traditional relationship‑lending bank.

At the macro level, however, SRT has become an important capacity tool between bank capital and the broader non‑bank credit system. By reducing RWAs on existing exposures, it frees regulatory capital that can be redeployed into new lending, which means banks can originate more risk‑weighted assets than they otherwise could under the same capital base. In practice, that can mean maintaining or growing balance sheets under tighter capital rules, or continuing to support core clients without issuing new common equity.

From a system‑wide perspective, what matters less is who originates a given loan and more where the risk ends up, how levered that holder is, and how the structure behaves when stress hits both banks and NBFIs at the same time. The Basel Committee’s recent work makes exactly this point, calling for continued monitoring of bank–NBFI channels, incorporation of SRTs into stress tests, and attention to rollover and back‑leverage dynamics.

What I Take Away From This — And What I Will Be Watching

Three conclusions shape how I think about SRT after layering the practitioner view with law‑firm and regulatory material.

First, SRT is now an institutional asset class. A market that references over 1 trillion dollars of bank assets, has overtaken European CLO and ABS issuance in some segments, and is explicitly analysed by the Basel Committee and BIS is part of the core capital plumbing of the banking system. Documentation standards, regulatory frameworks, and the investor base have all matured to the point where SRT is likely to remain a permanent tool for bank capital management rather than a transient trade idea.

Second, the cycle test is still in front of the market. Historical performance has been excellent: low realised losses on corporate loan pools, stable spreads, and no documented systemic failures of SRT structures through the post‑GFC period, including COVID. But this period has also been unusually benign for credit, with heavy central‑bank support and no full‑scale banking crisis in the core jurisdictions where SRT is largest. We do not yet have a live data point for how SRT issuance, spreads, bank reliance on capital relief, and realised losses behave when both bank capital and NBFI risk appetite are simultaneously under genuine pressure.

Third, rollover risk and NBFI interconnectedness are the variables to watch. The structural elegance of SRT — converting capital pressure into an annual premium — becomes a vulnerability if a bank’s ability to refinance protection becomes correlated with macro stress. At moderate levels of reliance, the arithmetic of tens of basis points of CET1 relief is benign; at higher levels, and with a concentrated or leveraged investor base, the feedback loop could matter for both bank capital and credit supply.

For credit investors analysing banks, that means SRT pipelines, investor concentration, and the quality of risk‑transfer structures now belong alongside traditional CET1 and RWA disclosures in the analytical toolkit. For investors in SRT itself, it means underwriting not just the reference portfolio but also the legal structure, regulatory context, and the potential for regulatory or market regime shifts over the life of the trade.

Listen to the full episode for Frank’s takes on SRTs.

This article is based on Episode 12 of Fixed + Floating, featuring Frank Benhamou of Cheyne Capital. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.cheynecapital.com/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Ep12 Transcript

368KB ∙ PDF file

Download

Download

Introduction To Srt

119KB ∙ PDF file

Download

Download

The article closes with a five-line checklist you can run on any software deal in ten minutes to decide whether you’re being paid for the documentation risk.

This is the third part of a software arc on Fixed + Floating. For context on software business models and leverage, see:

Covenant Erosion and the Normalisation of LMEs

Octus data shows the average LBO and refinancing covenant quality score deteriorated from 3.33 in 2023 to 3.53 in Q1 2026, on a scale where higher means worse. That number understates the problem. The starting point was already weak — covenant quality had been eroding systematically through the 2010s, meaning the recent decline compounds on a base that would have alarmed most credit committees a decade ago.

In 2024, 34 liability management exercises were executed across leveraged credit markets, up from 23 in 2023, before moderating to 28 in 2025. LMEs have become a recurring feature of credit markets, not an anomaly. And software credits — the sector that now dominates leveraged finance deal flow — carry the same covenant packages as any other leveraged buyout, despite having an asset profile that makes them uniquely vulnerable to the mechanisms LMEs exploit.

First Principles: What Covenants Are Actually Protecting

Before examining specific LME structures, the analytical framework requires grounding in what covenants are designed to do. The most important promise in a credit instrument is repayment of principal at maturity. Every other covenant exists to protect that promise — either by protecting the credit quality of the borrower or by protecting the lender’s position in the capital structure.

• Credit quality covenantskeep value inside the restricted group.

• Restricted payment covenants prevent the borrower from distributing cash to equity holders.

• Asset sale covenants ensure that when assets are disposed, proceeds either stay in the business or are used to repay debt.

• Affiliate transaction covenants prevent value transfers to related parties at non-arm’s-length terms.

• Change of control provisions give lenders an exit if ownership shifts in ways that could alter the credit profile.

• Capital structure covenants protect the lender’s relative priority.

• Debt and lien covenants control the borrower’s ability to raise additional secured or structurally senior debt, preventing dilution of the existing lenders’ recovery position. These are the covenants that, when loosely drafted, enable up-tier and double-dip transactions.

Imagine the borrower exercises every flexibility the documentation permits — every permitted payment, every permitted investment, every incremental facility, every asset sale basket. Would you still lend at this coupon? If the answer is no, the covenant package is mispriced.

“If you invest based only on your analysis of the credit, you’re only investing based on half the story because the contract tells the other half.”

That framing is deceptively simple. In practice, most credit analysts spend their time on enterprise value, cash flow, and leverage — the credit fundamentals. The contract analysis is often delegated or treated as a checklist exercise. This approach systematically underweights the optionality embedded in the documentation — optionality that becomes decisive precisely when credit quality deteriorates and LME teams begin looking for openings.

Pricing Covenant Quality: Why a Single Number Falls Short

One question I raised with Sabrina was whether covenant quality can be reduced to a basis-point adjustment — some quantitative measure that maps weaker covenants to wider spreads. Her answer was unambiguous: it cannot.

The difficulty is that covenant provisions are not independent variables. A generous permitted investment basket is benign for a borrower with no transferable assets but lethal for a software company whose entire enterprise value resides in intellectual property that can be moved between entities overnight. A loose FMV determination clause — one that allows fair market value to be assessed by the board without independent third-party valuation — is irrelevant when the borrower holds physical plant and equipment but becomes a mechanism for value extraction when the primary asset is IP that can be valued flexibly.

In the Akzo Nobel deal, lenders were able to force a materially higher coupon in exchange for keeping specific asset sale and distribution protections. That transaction shows that the market will sometimes price documentation — but the spread premium was highly specific to that issuer, capital structure, and set of clauses. Generalising from a handful of such cases into a “covenant basis-point premium” is analytically unsound.

In simpler terms, covenant quality is a facts-and-circumstances analysis. It depends on who the borrower is, what assets they own, what they are likely to do under stress, and which specific provisions interact to create or foreclose flexibility. The Octus score is a useful directional indicator. It is not a substitute for reading the documents.

The LME Taxonomy: Four Mechanisms, One Objective

LMEs are not a single transaction type. They are a family of related strategies, all aimed at improving the position of the borrower (or a subset of lenders) at the expense of remaining creditors. The taxonomy has expanded rapidly since 2020, and each new structure tends to prompt a corresponding blocker clause — which in turn prompts the next innovation.

Drop-Down Transactions

The drop-down is the oldest and most intuitive LME structure, and the J.Crew transaction of 2017 remains its archetype. The mechanics are straightforward: the borrower identifies valuable assets — typically intellectual property — and transfers them to an unrestricted subsidiary. An unrestricted subsidiary sits outside the restricted group, meaning it is not bound by the covenants in the credit agreement and its assets are not encumbered by the existing security package. Once the IP resides in the unrestricted subsidiary, the borrower raises new debt against it, effectively priming existing lenders by removing their best collateral.

The process in a software credit follows a predictable sequence. The borrower designates or creates an unrestricted subsidiary. It transfers IP to that entity, relying on permitted investment baskets that allow investments in unrestricted subsidiaries up to a defined capacity — often linked to a percentage of total assets or an EBITDA-based formula. The board determines the fair market value of the transferred IP, typically without any requirement for an independent third-party valuation. The IP is then licensed back to the operating company, ensuring no operational disruption. The unrestricted subsidiary raises secured debt against the IP.

Software IP is uniquely suited to this mechanism. It is the most valuable asset in the enterprise. It is the most portable — unlike a factory or a fleet, it can be transferred between entities with a board resolution and a licensing agreement. It can be valued flexibly because there is no liquid market establishing a reference price. And it can be licensed back without any observable change to the business’s operations. A manufacturing company that transferred its plant to an unrestricted subsidiary would need to shut down production. A software company that transfers its source code licenses it back the same day.

“Software deals have the same covenant packages as other leveraged finance deals — but the flexibility applies like a ticking time bomb because of IP portability.”

This is the central paradox. The documentation was designed for industrial borrowers with fixed assets. Applied to software companies, the same provisions create vastly more flexibility because the asset characteristics are fundamentally different.

Up-Tier Transactions

The up-tier — with the Serta transaction as its defining case — operates differently. Rather than moving collateral out, the borrower provides preferential security or priority to a select group of existing lenders willing to provide new money or consent to amended terms. The remaining lenders — those who did not participate — find themselves subordinated within the capital structure without having agreed to anything.

The mechanics rely on open market purchase provisions or voting thresholds that allow a simple or qualified majority of lenders to amend key terms, combined with covenant flexibility that permits the issuance of new super-priority debt. A subset of lenders negotiates directly with the borrower, often under NDA. Non-participating lenders may be “finding out in the paper the next day.”

The provisions that enable up-tier transactions are primarily found in the debt and lien covenants — specifically, in the definitions of permitted debt, permitted liens, and the conditions under which new debt can be issued with priority over existing obligations. When these provisions allow secured debt to be raised at entities above or alongside the existing borrower group, existing lenders’ recovery priority is diluted.

Double-Dip Transactions

The double-dip — with the At Home transaction as the reference case — is structurally more complex. It uses existing contractual provisions in combination to provide certain lenders with multiple avenues of recovery against the same asset base. The mechanism typically involves transaction security provisions that allow specific intercompany arrangements to create additional claims — effectively enabling a single dollar of new money to have two separate paths to recovery.

In practice, a double-dip might involve a lender providing financing at one entity and simultaneously receiving an intercompany loan claim at another entity within the group, such that both claims have security over group assets. The result is that the participating lender’s effective recovery rate is significantly enhanced relative to non-participating lenders, without additional capital being injected into the business.

Extend and Exchange

The most recent evolution in the LME toolkit is the extend and exchange — a category that encompasses consent solicitations to extend maturities, create new loan classes, and restructure existing instruments. The transactions at Better Health, Oregon Tool, and ION Platform illustrate this approach.

These transactions have become more prominent in part because post-Serta contractual innovations have made traditional up-tier transactions harder to execute. Borrowers and their advisors have adapted by using exchange offers and consent solicitations that achieve similar economic objectives — improving the position of participating lenders and extending the borrower’s liquidity runway — while working within or around the contractual restrictions specifically designed to prevent Serta-style up-tiers.

Why Software Is the Sector That Should Concern You

The intersection of deteriorating covenant quality and software-specific asset characteristics creates a compounding vulnerability.

Software credits have no maintenance covenants in their loan agreements — a trend that has become near-universal across leveraged lending. Maintenance covenants — tested quarterly, requiring the borrower to maintain leverage or coverage ratios — provided an early warning mechanism. When a borrower breached a maintenance covenant, lenders had the right to negotiate, amend, and potentially accelerate. Without them, the first sign of credit stress may be a missed coupon payment, by which time the borrower’s restructuring advisors have already been working on an LME for months.

The IP transfer mechanism described above applies with particular force to software. The most valuable asset is also the most portable. Board discretion over fair market value, the absence of third-party valuation requirements, and the ability to license IP back without operational disruption — all standard features of leveraged finance documentation — combine to make the drop-down a low-friction exercise for software borrowers.

The covenant packages themselves are undifferentiated. Software deals receive the same incurrence-based covenant architecture as any leveraged buyout — chemicals, healthcare, industrials. The permitted investment baskets, restricted payment capacity, and unrestricted subsidiary designation mechanics are functionally identical. But the consequences of exercising that flexibility are radically different when the borrower’s enterprise value is concentrated in transferable intangible assets rather than dispersed across physical plant, inventory, and customer contracts.

Case Study: Xerox and the Joint Venture Innovation

The Xerox transaction illustrates how LME innovation outpaces the contractual blockers designed to prevent it.

Xerox is a legacy hardware issuer in the process of pivoting to software, carrying a heavy maturity profile. The company needed to raise capital and restructure its liabilities. The standard drop-down playbook — transfer IP to an unrestricted subsidiary, raise debt against it — was foreclosed by a J.Crew blocker in the existing documentation. J.Crew blockers, which became standard in post-2017 documentation, specifically prohibit the transfer of material intellectual property to unrestricted subsidiaries.

Xerox’s advisors found an alternative. Instead of transferring assets to a subsidiary, they moved assets into a joint venture. The JV was structured so that Xerox retained majority economic ownership but held less than 50% of voting power. The critical insight is definitional: under standard leveraged finance documentation, a subsidiary is defined by reference to voting power — specifically, ownership of more than 50% of the voting stock. A joint venture in which the borrower holds less than 50% of the votes is not a subsidiary under the documents, even if the borrower receives the majority of the economic returns.

Because the JV was not a subsidiary, it was not an unrestricted subsidiary. Because it was not an unrestricted subsidiary, the J.Crew blocker — which applies specifically to transfers to unrestricted subsidiaries — did not apply. The IP was transferred to the JV, which then co-invested it to produce proceeds that flowed back to the Xerox structure.

This is what we can call the Easter egg approach to documentation. Restructuring teams execute a novel transaction, and then advise the capital markets teams at their firms to embed the enabling provisions — or the absence of blockers — in the next round of new issuance documentation. The market adapts, but always with a lag. The Pfleiderer transaction demonstrated a related innovation: rather than transferring assets to an unrestricted subsidiary, the borrower simply sold them outright, generating proceeds outside the restricted group.

The broader takeaway from Xerox is that capital structures are becoming more complicated without solving the underlying credit problem. No deleveraging occurred. Maturity was extended. Liquidity was created. But the fundamental question — whether the enterprise can generate sufficient cash flow to service its obligations — was deferred, not resolved.

Case Study: ION Platform and the Limits of Blocker Clauses

ION Platform — a software conglomerate with a complex capital structure spanning bonds and loans — provides a real-time case study in how blocker clauses interact with LME incentives.

In September and October 2025, ION Platformexecuted a dual transaction: issuing new notes and exchange notes simultaneously, creating two distinct classes of creditors with different levels of protection. The exchange note holders had leverage — the borrower needed their consent to extend maturities — and negotiated materially tighter covenant protections. The new note holders, facing time pressure and competitive dynamics in the primary market, received a looser package.

The resulting capital structure contains a J.Crew blocker, but its scope is limited. The blocker applies only to material intellectual property, defined with reference to conducting business “in all material respects” and to the concept of material adverse effect. ION operates as a conglomerate of distinct software businesses. The question is whether transferring the IP of one business unit would constitute a transfer of “material” IP when the remaining businesses continue to operate. A reasonable argument exists that dropping a single platform would not prevent ION from conducting its business in all material respects — precisely because the remaining platforms would still function independently.

ION’s documents also predate the Xerox transaction, meaning there is no Xerox blocker — no provision preventing the JV structure that Xerox pioneered. ION could, in principle, replicate the Xerox approach. Paradoxically, the market has already begun responding: Recent new‑issue documentation is already starting to include Xerox‑style blockers, demonstrating how quickly the innovation–response cycle turns.

There is a parallel to Altice, a French TelCo, which restructured recently. Altice had one tightly covenanted tranche within its capital structure. When that tranche was refinanced in 2015–2016, the tight protections disappeared, and the borrower gained substantial additional flexibility. The lesson for ION is directional: the exchange notes — the tighter tranche — represent a constraint on the borrower’s flexibility. ION has a clear incentive to refinance or retire those notes, unlocking the full optionality embedded in the looser new note documentation.

A double-dip transaction at ION is less probable, because the intercreditor agreement contains restrictions on the subordination of intercompany debt that would complicate the mechanics. But the other avenues — drop-down via the JV route, or refinancing the tight tranche to unlock flexibility — remain available.

The Five-Line Software Credit Checklist

Sabrina provided a framework for rapid initial assessment of LME vulnerability in any software credit — a ten-minute exercise that identifies whether deeper covenant analysis is warranted.

• EBITDA definition: Check the breadth of permitted addbacks and whether synergy addback caps exist. An EBITDA definition with uncapped pro-forma adjustments and wide management addback discretion inflates the denominator of every leverage-based test in the documentation, expanding capacity across all incurrence-based baskets simultaneously.

• Leakage possibilities: Identify whether there are non-guarantor entities within the restricted group and whether unencumbered collateral exists that could be used as the basis for an up-tier transaction. The presence of significant assets outside the guarantor and security package is a precondition for several LME structures.

• Restricted payments and permitted investments capacity: Assess the ability to move value out of the restricted group — through dividends to equity, investments in unrestricted subsidiaries, or other permitted transfers. This is the capacity that enables drop-down transactions. The relevant metrics are the builder basket, the general permitted investment basket, and any grower-based components that expand with EBITDA.

• Blockers: Determine whether J.Crew blockers, Xerox blockers, or other specific anti-LME provisions exist. But blockers are not binary. A J.Crew blocker limited to “material intellectual property” provides less protection than one covering all IP. A blocker that applies only to unrestricted subsidiaries does not address the JV route. The quality and scope of the blocker matter as much as its presence.

• Co-ops: Assess whether lenders are organising. Are creditor groups forming? What is the broader lender dynamic? Co-ops — formal or informal groups of lenders coordinating to protect their position — have become an essential part of the LME landscape. Their existence signals that sophisticated holders have identified risk; their absence may signal complacency.

The Blocker-LME-Blocker Cycle and What Could Break It

The evolution of LME practice follows a predictable pattern: a novel transaction exploits a gap in existing documentation, the market responds with a new blocker clause, restructuring teams find the next gap, and the cycle repeats. J.Crew prompted J.Crew blockers. Serta prompted Serta restrictions. Xerox prompted Xerox blockers. Each iteration adds complexity to documentation without eliminating the underlying incentive — which is that borrowers under stress will always seek to extend their liquidity runway, and advisors will always find mechanisms within the documentation to enable it.

The most significant potential development for 2026 are the pending court decisions on creditor co-ops. Two cases are currently testing the legality and scope of co-op activity. If a court ruling makes lenders nervous about joining co-ops — by finding, for example, that co-op participation creates liability or violates credit agreement provisions — the consequences would be asymmetric. Borrowers would retain all their existing flexibility. Lenders would lose their primary mechanism for collective defence.

A negative ruling on co-ops “would be open season for borrowers to do LMEs.” The coordinated lender response that has, in some cases, deterred or moderated LME activity would be undermined. Individual lenders acting alone have neither the information nor the leverage to resist a well-advised borrower executing a complex restructuring transaction.

The Structural View

The analytical conclusion from this conversation is uncomfortable but clear. Software credits carry the same covenant architecture as industrial leveraged buyouts, but the asset characteristics of software businesses — concentrated IP value, portability, licensing flexibility, board-determined FMV — interact with that architecture to produce vastly more LME optionality than the documentation was designed to accommodate. The blocker clause cycle will continue to iterate, but each new blocker addresses only the last transaction, not the next one.

For software credit investors, the decisive variable in the near term is not spread or leverage — it is documentation. The five-line checklist is a starting point. The Xerox and ION case studies demonstrate that even credits with blocker clauses carry residual vulnerability. And the pending co-op litigation could remove the last meaningful check on borrower behaviour in distressed situations.

I do not think the LME wave is cresting. I think it is becoming more sophisticated. The practical question for credit investors is whether their covenant analysis has kept pace with the restructuring advisors on the other side of the table — especially in software. For most, the honest answer is no.

Listen to the full episode for Sabrina’s takes on covenants and LMEs.

This article is based on Episode 11 of Fixed + Floating, featuring Sabrina Fox of Fox Legal Training. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.foxlegaltraining.com/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Ep11 Transcript

172KB ∙ PDF file

Download

Download

Flt Lme Podcast Deck

213KB ∙ PDF file

Download

Download

This is not inherently sinister. Insurance balance sheets are a natural home for illiquid assets. But beneath this demographic shift lies a more complex engine: regulatory arbitrage. The true story of the private credit and life insurance boom is how capital efficiency is being optimized to the absolute limit. The way these assets are structured, financed, rated, and aggressively arbitraged across state lines and into offshore captive reinsurers creates a systemic mechanical tension—one that warrants close examination by any credit professional.

(Note: For a foundational primer on the mechanics driving this shift, check out our previous deep dive with Jakub Lichwa on Episode 3)

Understanding US Life Insurance Products

To understand why private credit has become central to insurance asset allocation, one must first understand the structure of the liabilities being funded. Most US life insurance products can be classified into two broad categories (variable annuities fall somewhere in between):

• Protection-oriented products: Term life, whole life, and guaranteed universal life (GUL) earn revenue primarily from mortality spread—the difference between premiums charged and claims paid. The insurer bets that claims will be lower than expected. These products have no stated maturity, and claims can accelerate in unpredictable fashion (e.g. pandemic, natural disaster), and as a result the business has natural need for more liquid assets that can be sold quickly.

• Spread-based products: Fixed annuities, Multi-Year Guaranteed Annuities (MYGAs), and fixed index annuities (FIAs) earn profit from the investment spread—the difference between yield earned on invested assets and the crediting rate paid to the annuitant. An insurer might invest at 5.5% and credit at 4.0%, keeping 150 basis points. These products are acutely interest-rate and duration-sensitive. They also have clear maturity profiles which can be aligned with those of private credit assets.

The mix of these categories, and the rate environment in which they are sold, determines the insurer’s duration profile, asset allocation, and leverage. Pre-1980s, the US population was young (median age 30) and demand was for protection products. The 1980s brought rising rates, and annuities exploded as consumers wanted to lock in yields. Then 2008 and the GFC crushed annuity-heavy books through spread compression between 2010 and 2015.

This created conditions for private equity buyouts of insurance firms. Sponsors brought capital, operational expertise, and critically, access to origination channels in private credit markets. Post-2022, rising rates and an aging population have triggered an annuity renaissance.

The Asset-Liability Match: Why This Model Works

The match between annuity liabilities and private credit assets is structural, not accidental. Annuity liabilities carry durations of 3 to 10 years. Private credit placements across various underlyings—such as real estate lending, infrastructure financing, and asset-based finance—can be structured to match precisely.

Liability stickiness and predictability amplify the value. Annuity policyholders face surrender charges of 6% to 8% in year one, stepping down to zero after 10 years. An insurer placing capital into a 7-year loan knows the annuity funding it will remain in place for at least 5 to 7 years.

Pricing mechanics create a pull toward higher-yielding assets. Annuities are typically priced 1% to 2% below asset yield. Private placements deliver roughly 80 basis points of pickup versus public corporates (approximately 60bps from the general liquidity premium, 78bps from financial placements, and 156bps from ABS segments). Without this pickup, the spread narrows to levels insufficient to cover administrative costs and desired profitability, whilst retaining competitive pricing of annuities.

PE-backed insurers solve this through their virtuous flywheel: the sponsor controls origination, structures assets for desired ratings and duration, deploys them into the insurer, then uses captive reinsurance to release capital for the next round. Banks retreating post-GFC left a lending gap that this model efficiently fills.

Asset Quality: The Ratings Puzzle

When one examines the credit profile of insurance portfolios, a striking pattern emerges: externally assessed ratings tell a significantly better story than regulatory ratings. NRSRO ratings (assigned by the nationally recognized agencies) show approximately 90% of insurance private credit holdings are investment grade. Analysis of public filings from major players shows the proportion of IG-rated holdings has either held steady or improved over time.

But NAIC ratings—assigned by the Securities Valuation Office (SVO), which has authority to challenge and override external ratings—tell a different story. The divergence is meaningful and growing for some insurers. This matters because capital requirements are tied directly to ratings:

• An A- rated asset carries a capital charge of approximately 1%.

• A BBB- rated asset carries approximately 2.2% (more than double).

• A BB- rated asset carries approximately 6%.

A portfolio does not need to suffer actual defaults to experience stress. A wave of downgrades alone can increase capital requirements six-fold, instantly pressuring the insurer’s balance sheet.

The credibility of these ratings is under active scrutiny. The IMF has flagged the explosive growth in specialized rating agencies since 2019, specifically agencies focused on private credit where large incumbents have stepped back. In November 2025, the SEC launched an investigation into Egan-Jones Ratings Company over its private credit rating practices, noting they assigned significantly more ratings per analyst than peers like Moody’s or S&P. If originators gravitate toward the most generous rating agency, portfolio distributions will skew optimistic.

Opaque Disclosures and Cross-Sector Comparisons

Anecdotal evidence reinforces the opacity concern. Between Q2 and Q3, Security Benefit reclassified its ABS holdings (representing 40% of total assets) from corporate exposure to a more granular breakdown. Nearly half of the ABS portfolio, or 20% of total asset allocation, was disclosed as being private equity. While the reclassification does not prove deterioration, it reveals an opacity that had previously concealed the true composition of the portfolio.

It is critical to distinguish insurance private credit from BDC private credit. BDCs hold largely single-B rated corporate loans with heavy software exposure. Insurers hold primarily IG-rated, diversified portfolios of private credit across various underlyings (real estate, infrastructure, ABS). The key risk for BDCs is actual defaults; the key risk for insurers is downgrades and regulatory recalibration.

The Funding Side: Stability and the Liquidity Myth

Negative commentary often mischaracterizes the liquidity risk of life insurers. Recently, it was suggested on a major podcast that institutional investors holding funding agreements could theoretically call up Apollo and ask for $3 billion back from Athene Global, triggering a run.

This is structurally false. Funding Agreement-Backed Notes (FABNs) function like bonds, not demand deposits. They are pari passu with general policyholders, have termed-out maturities, and contain zero acceleration clauses triggered by asset deterioration, ratings downgrades, or market dislocation.

The real risk is a spread-widening and refinancing-cost vulnerability. If an insurer is downgraded or credit conditions tighten, the cost of refinancing matures liabilities spikes.

Beyond funding agreements, Apollo’s liability structure provides meaningful stability.

Looking at representative structures, approximately 50% of annuities carry withdrawal fees, and roughly 33% cannot be surrendered at all. Policyholders can withdraw up to 10% of principal penalty-free per year, and state-level guarantees provide coverage of $250,000 to $500,000 against an average annuity size of approximately $150,000. However, unlike FDIC bank deposit insurance, state guaranty funds are post-funded—capitalized through industry assessments only after a failure occurs, which can represent an element of concern for annuity holders (especially in a situation of several insurers failing in close succession).

Regulatory Arbitrage: The Structural Risk

Insurance regulation in the US is primarily a state-level function, not federal, enabling arbitrage both across states and offshore. The Bank for International Settlements reported that approximately 40% of reserves have been ceded to offshore reinsurers.

The mechanism that has attracted the most scrutiny is Asset-Intensive Reinsurance (AIR). In AIR, the primary insurer transfers both assets and liabilities to a reinsurer—often an affiliated captive in Bermuda or the Cayman Islands. The return on transferred assets pays for the liability reinsurance, releasing significant capital.

A concrete example highlights the leverage:

• Start with a $1 billion annuity block ($950 million in reserves, $50 million in capital).

• Cede 90% to the reinsurer. Reserves drop to $95 million.

• Add a counterparty risk charge of approximately 5% at a BBB rating ($45 million).

• New required assets equal $140 million. Capital tied up is just $7 million.

• The transaction releases $43 million in capital, enabling the acquisition of $860 million in new assets.

The tail risk is deeply concentrated. If the offshore reinsurer fails, the primary insurer must recapture both assets and liabilities, and that $43 million in capital relief evaporates instantly. In a revealing data point, Apollo flagged regulatory arbitrage in their own investor update as “the most serious threat to the financial system”.

2008 Echoes and Stress Scenarios

If the system faces stress, a clear hierarchy emerges:

1. Rating downgrades triggering capital calls: A move from A- to BBB- doubles the requirement. Evidence suggests ratings may warrant scrutiny, and the capital arithmetic is unforgiving.

2. Reinsurance counterparty stress: Concentrated offshore ceding means a single reinsurer failure forces recapture across multiple primary insurers simultaneously.

3. Default clusters: Widespread defaults would require a severe macro shock given the high IG concentration.

4. Policy surrender spikes: A tail risk, heavily mitigated by withdrawal penalties and liquidity buffers.

The structural parallels to the pre-GFC system are meaningful. Non-bank lenders are again filling a lending gap vacated by banks. Funding mismatches persist—surrenderable annuities funding illiquid private credit. Ratings-driven capital opacity remains, and regulatory arbitrage through offshore captive reinsurance mimics the off-balance-sheet SPVs of the past. The key difference today is the existence of state guaranty funds and NAIC regulatory oversight.

Conclusion

Insurance is a great home for illiquid assets, but only at the right price and when properly regulated. The key risk is mainly that the asset quality on private credit across various underlyings is lower than expected, which impairs the asset quality of the broader portfolio and drives up the capital requirements, eroding regulatory metrics.

Massive amounts of these risks have been shed to offshore captive reinsurers. If underlying assets underperform, those risks could come crashing back onto the primary balance sheets. Consequently, a vast amount of risk has been transferred into the insurance sector. Because this is a highly regulated sector, it could find itself undercapitalized—not purely from a sudden wave of borrower bankruptcies, but from the acute risk of regulatory recalibration and rating downgrades.

Listen to the full episode for Jakub’s takes on Insurance and private credit.

This article is based on Episode 10 of Fixed + Floating, featuring Jakub Lichwa of TwentyFour Asset Management. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://www.twentyfouram.com/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Podcast March 26 Final

883KB ∙ PDF file

Download

Download

Ep10 Transcript Clean

126KB ∙ PDF file

Download

Download

The premise of much technology writing — that innovation is inherently democratizing, that smaller firms can out-maneuver giants with better products and lower costs — misses something essential about how competitive advantage actually works at scale. Bessen has spent two decades watching this play out across industries. His conclusion is counterintuitive and material for credit investors: technology is making it harder, not easier, for dominant firms to lose their position.

From Software to Economics: How a CEO Learned What Textbooks Got Wrong

When Bessen ran a software company in the 1980s and 1990s, he noticed something odd. Economics textbooks insisted that patents and proprietary technology were essential drivers of innovation. Yet the software industry — which had far fewer enforceable patents than, say, pharmaceuticals or manufacturing -- was innovating rapidly. Competitors were thriving. Knowledge was leaking everywhere. And the industry was not collapsing.

This observation became the foundation for Learning by Doing, published in 2015. Bessen’s core finding was simple but at odds with standard economic theory: the source of competitive advantage in knowledge-intensive industries is not legal protection. It’s accumulated know-how. The firm that learns fastest, and learns from doing, wins. The firm that can absorb lessons from each iteration, build organizational memory, and operationalize that learning into faster product cycles and better decision-making — that firm outpaces competitors. Patents matter far less than economists assumed.

The insight applies directly to how we should think about technology moats today. A moat built on learning — on the organizational ability to extract value from data, experience, and operational depth -- is far stickier than a moat built on a clever algorithm or a proprietary piece of code. This distinction matters enormously for credit investors trying to assess whether a tech company’s competitive position is durable or fragile.

The Walmart Model: Why Technology Alone Doesn’t Explain Market Dominance

Bessen often uses Walmart to illustrate this principle. In the 1970s and 1980s, Walmart built a proprietary IT system that tracked every cash register, every store, every product, and every inventory movement across its entire supply chain. The system gave Walmart a decisive advantage: it could see demand patterns in real time and respond faster than competitors. If a product was selling well in one region, Walmart could replicate that signal across stores. If demand was dropping, it could adjust orders and floor space allocation faster than Kmart or other regional chains.

But the moat was not simply “Walmart had better software.” The moat was that Walmart’s entire organization had been built from the ground up to use that software effectively. Store managers understood how to interpret the data. Procurement teams had restructured supplier relationships around just-in-time ordering. The logistics network had been designed to support rapid inventory turns.

When Kmart attempted to replicate Walmart’s system in the 1990s, it failed. This wasn't because Kmart lacked the capital or the scale—at the time, Kmart was a retail giant. It failed because it lacked the organizational agility to deploy that software with the same effect. Kmart’s stores, suppliers, and management culture were entrenched in an older operating model. The software, in isolation, was useless without the institutional know-how to extract value from it. This shows why these moats are so durable: even well-funded peer competitors struggle to replicate decades of organizational learning.

But while organizational complexity protects dominant firms from their giant peers, a different dynamic protects them from smaller upstarts: the brutal economics of proprietary IT. This is a crucial threshold Bessen identifies. A firm needs massive scale to amortize the fixed costs of building and maintaining a custom, proprietary software system.

Below a certain size, it is simply not economically rational for a smaller firm to build a system like Walmart's. Smaller rivals can buy off-the-shelf SaaS or build specialized niche tools. But they cannot match the scope, depth, or integrative power of a multi-billion-dollar proprietary platform. Software's high fixed costs and low marginal costs naturally favor the biggest player, allowing the Goliaths to continuously widen the productivity gap over everyone else.

Banks, Data, and the Fintech Challenge

The same dynamic appears in banking and finance. Large banks like JP Morgan have invested heavily in proprietary credit assessment systems. These systems don’t just automate what an older credit officer would have done manually. They ingest vast amounts of customer transaction data, behavioral data, and historical loan performance data. The bank’s data scientists continuously retrain the models. The models become more precise over time. And because JP Morgan has millions of customers, the data set compounds — each new loan, each default, each repayment adds signal to the next iteration.

Fintech firms have challenged traditional banks in certain segments — payment processing, for instance, or targeted lending to underserved customers. But in the U.S., fintech firms face a structural barrier that Bessen emphasizes: they do not have direct access to the customer transaction and deposit data that large incumbent banks possess. Open banking regulation in Europe has forced banks to share more data; in the U.S., incumbent banks have resisted these requirements. The result is that the largest U.S. banks retain a data advantage that smaller competitors, no matter how clever their algorithms, cannot easily overcome.

For credit investors, this matters. It suggests that disruption risk to incumbent banks — at least from fintech in the U.S. market — is lower than technology commentary often suggests. The moat here is real, and it is being reinforced, not eroded, by technology itself.

Assessing the Durability of Tech Moats: Value, Not Barriers

One common mistake in evaluating technology companies is to assume that a competitive moat must be built on barriers to entry — on something that prevents competitors from even trying. Bessen’s framework is different. The moat comes from delivering real, customer-facing value, and from the difficulty of replicating that value at the same scale.

Walmart’s moat is not that no one is allowed to build a supply chain system. The moat is that customers find what they want more often at Walmart than at competitors. That experience comes from decades of learning. It is not easily replicated. Competitors are not blocked by a legal wall; they face obstacles in building equivalent capability.

This distinction has a practical implication for credit investing. When evaluating a tech company’s competitive position, ask: Does this company deliver something customers value? And is that value difficult to replicate because of accumulated learning, scale economies, or organizational depth? Those are durable moats. The strongest tech moats are often invisible from the outside. They look like operational excellence, not innovation. They look like a company that executes well and learns continuously, not a company with a revolutionary product.

The Counterintuitive Truth: Disruption Risk Has Actually Declined

Here is where Bessen’s research departs sharply from conventional technology discourse. Over the past 20 years, the top four firms in most industries have become less likely to drop out of the top four than they were 20 years ago. This is true even as technology is changing rapidly and new entrants are appearing constantly.

The explanation lies in a paradox: technology increases competition in the aggregate. It lowers barriers to entry in many segments. It enables new business models. But at the very top of the market — among the largest, most established firms — technology is actually reinforcing incumbency.

Why? Because dominance at scale now rests on accumulated data, operational complexity, and organizational learning rather than on a single proprietary secret or a technological breakthrough. These advantages are compounding and harder to displace. A dominant platform that has learned how to integrate AI into its operations does not become more vulnerable to disruption because of AI — it becomes more powerful.

Consider IBM. In the popular narrative, IBM is a disruption cautionary tale. The company dominated mainframes in the 1960s, faced pressure from minicomputers in the 1970s, nearly collapsed when personal computers arrived in the 1980s. This story is correct in outline but misleading in implication. IBM did not disappear. It remained profitable and large for decades after its “disruption moment.” Revenue only began to decline around 2010, a full 25 years after the PC supposedly rendered the mainframe obsolete.

Nokia presents a similar case. The company dominated mobile phones. The smartphone was indeed disruptive. Nokia lost the phone business. But Nokia did not vanish. It shifted into telecommunications infrastructure and became, in Bessen’s blunt characterization, “more of a patent troll” — extracting rents from historical patents rather than competing on products.

The broader point is that disruption, when it happens, is slow. The dramatic stories in technology journalism — the moment the giant falls — are the tail end of a long process. For credit investors, this means that disruption risk to truly dominant firms is lower than headline attention suggests. The firm you think is vulnerable today may remain profitable and large for another decade, even as its market position erodes in real time.

Automation, Employment, and the Historical Pattern That Still Holds

One area where technology commentary becomes most overheated is employment. The fear that machines will displace human workers has been constant since the Industrial Revolution. Bessen’s historical analysis suggests we should be skeptical of that fear — not because disruption is impossible, but because the historical pattern has been remarkably consistent, and we have no convincing evidence that the pattern has broken.

In the 19th century, textile machinery dramatically increased output per worker. Looms became faster, spindles multiplied, automation accelerated. Economic theory suggested employment in textiles should have collapsed. Instead, employment in textiles grew for over a century. Why? Because demand was elastic. As cloth became cheaper, people bought more cloth. Families that previously owned two or three garments could now afford six or seven. Industrial demand for textiles expanded. So even as output per worker rose, the industry needed more workers overall.

The employment decline in textiles did not come until the mid-20th century, when demand finally saturated. Closets were full. The savings from automation no longer translated into increased demand; they translated into lower prices and reduced headcount.

The same pattern is visible in software development today. Programmer productivity has increased roughly 5% per year for the past 20 years, compounded. According to standard economic theory, employment in software development should have fallen. Instead, employment has risen by roughly 3% per year, and the total number of software developers employed is at an all-time high. Demand has grown faster than productivity gains.

Microsoft’s internal study of Copilot found a 26% productivity gain for developers using the tool. The study found no employment effect. The developers worked faster. Projects completed ahead of schedule. And the company assigned those developers to new projects, creating new software rather than eliminating the headcount.

For credit investors, this distinction matters. Companies that are planning workforce reductions tied to AI adoption may be facing a different dynamic than they expect. Demand may simply shift. Productivity gains may be captured through new products and services, not through headcount reduction.

AI Investment and the Winner-Take-All Risk: Why This Bubble Is Different

Bessen’s analysis of AI investment risk departs from cheerful startup narratives but also from doomsday scenarios. The comparison he draws is to the dot-com bubble, and the distinction is clarifying.

During the dot-com boom of the late 1990s, massive capital flowed into fiber optic cable and server infrastructure. When the bubble burst and many companies failed, the recession was mild because the infrastructure did not disappear. Fiber optic cables remained in the ground and eventually came back into service as data demand grew. They lasted 20-30 years. Servers lasted about 10 years. The capital was wasted, in a sense, but it was not destroyed — it was eventually put to productive use.

AI investment is fundamentally different. The majority of current AI capex is not going into durable physical infrastructure. It is going into training frontier models and purchasing GPUs to run them. A frontier AI model becomes outdated on an average timeline of two months. A GPU under continuous training stress lasts a couple of years before degradation makes it inefficient. If there is a recession or a correction in AI valuations, these assets will not come back into service. They will be worthless.

This creates a winner-take-all dynamic that is far more brutal than the dot-com scenario. Dario Amodei, CEO of Anthropic, captured the risk: “It’s not going to support all of them.” If only a small number of AI companies end up viable and profitable, the capital invested in the runners-up will simply evaporate. There will be no recovery when demand returns, because the assets are consumed, not mothballed.

For credit investors, this matters in two ways. First, AI companies that are burning capital in pursuit of frontier model development should be viewed with skepticism about profitability timelines. Second, the broader AI capex boom contains embedded tail risk. Not all of this capital will generate returns. The question is how much will be trapped in failed companies and stranded assets.

Who Actually Profits from AI: Adaptation, Not Disruption

The winners in the AI era will not be the companies building the largest language models or the most sophisticated chat agents. Instead, the winners will be companies that figure out how to adapt AI to real problems in their existing business.

Consider call centers. Many companies have implemented AI chatbots to reduce labor costs. By most accounts, customer experience has degraded. Call quality has fallen. Customers are more frustrated. Yet the capital investment has been made, and sunk costs are driving continued deployment.

Contrast this with organizations that use AI as a tool to augment human workers. A customer service representative using an AI system that suggests relevant policies and account history can handle more complex inquiries. In these cases, AI increases human productivity and value creation. The profit opportunity is real.

Industry concentration has been rising for the past 100 years as economies have matured. This trend existed long before AI. But AI, deployed thoughtfully, can accelerate it. If only the largest firms have the scale to deploy AI in a value-creating way, concentration will continue to rise. Smaller firms will deploy AI in cost-cutting ways and will gradually lose margin. Credit quality will follow.

Antitrust, Unbundling, and the Paradox of Forced Competition

Bessen tells a story about IBM and the U.S. Department of Justice that upends typical antitrust logic. In the 1960s, IBM bundled software with hardware. The DOJ threatened antitrust action. IBM unbundled — it began selling software separately from hardware.

The result surprised everyone. Once the software market was open, third-party developers flooded in. Software development accelerated. New applications emerged. IBM’s software business grew enormously. The unbundling was far more profitable for IBM than the bundled approach had been.

Amazon’s AWS tells a related story. Amazon developed its own internal web services infrastructure to support its e-commerce business. Then it opened this infrastructure to external customers for a fee. AWS became one of the most profitable divisions of Amazon. Again, the demand was hidden until the market was opened.

The implication for antitrust policy is subtle. Regulation intended to increase competition in the product market may, paradoxically, increase technological competition while actually benefiting the dominant firm. Forcing dominant firms to open their systems can benefit the dominant firm more than it benefits rivals, while also creating genuine spillover benefits to the broader economy.

For credit investors, the lesson is that forced unbundling or forced data sharing — potential outcomes of antitrust litigation against large tech firms — may not harm the dominant firm as much as regulation proponents hope. In fact, if recent history is any guide, the dominant firm may eventually profit from the change.

Software Complexity and Regulatory Risk: The Dieselgate Precedent

Dieselgate — Volkswagen’s emissions test cheating software — illustrates how software complexity makes regulation fundamentally harder. A small number of VW programmers wrote code that detected when a car was undergoing an EPA emissions test and reduced emissions during the test, while allowing much higher emissions during normal driving. The software was relatively simple to create. It was extraordinarily difficult to detect. Volkswagen’s cars were emitting pollutants at 10 to 20 times the legal level.

The same dynamic appeared in the 2008 financial crisis. Mortgage instruments became so complex, and their risk characteristics so obscured through layers of software abstraction, that even professional investors could not accurately assess them.

As economies become more software-dependent, regulatory enforcement becomes exponentially harder. This favors incumbent firms with large compliance teams, deep regulatory experience, and the resources to invest in legal infrastructure. For credit investors evaluating firms in highly regulated industries — automotive, banking, pharmaceuticals — the software complexity layer is material.

Evaluating Technology Investment: When Capex Becomes Unpredictable

A traditional capex investment in a steel mill has clear parameters. You invest in the mill, it produces steel at a certain rate, the market for steel determines your profit.

Technology investment is far harder to evaluate. company have profit-making potential? Will its investments deliver capabilities that customers value and will pay for? The answer is almost always opaque. Many AI companies are currently unprofitable. The path to profitability relies on costs coming down — a speculative assumption about a specific path of technological development that may not materialize.

For credit investors, this means traditional capex analysis tools are inadequate when applied to technology companies. The risks are different. The timelines are different. The probability of total capital loss is much higher.

Listen to the full episode for James Bessen’s take on how technology strengthens incumbents, why disruption stories are often overstated, and what credit investors should actually be watching as industries become more software-intensive.

This article is based on Episode 9 of Fixed + Floating, featuring James Bessen of Boston University’s Technology and Policy Research Initiative. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit https://sites.bu.edu/tpri/.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

Bessen Book Talk 30min

650KB ∙ PDF file

Download

Download

09 Transcript

159KB ∙ PDF file

Download

Download

Alec Keblish and Matthew Hughes, both senior analysts at 9fin joined Fixed + Floating to work through the mechanics systematically.

The Foundation: What Made Software the Ideal LBO Candidate

To understand why the current repricing is structurally significant, it helps to recall why software attracted more than $500 billion of leveraged debt in the first place.

Pre-cloud, software was a physical product — a CD, a perpetual licence, a one-time payment. The shift to Software-as-a-Service flipped the cash flow profile entirely. Instead of volatile, lumpy revenue, SaaS companies generate annual recurring revenue (ARR): the annualised sum of all active subscription contracts, typically built annually in advance, often with auto-renewal clauses. The result is a business whose revenue is contracted, predictable, and aligned almost perfectly with debt service requirements.

Net revenue retention (NRR) — the metric that captures what is happening inside the existing customer base — became the secondary heartbeat metric. An NRR above 120% means the company grows organically without acquiring a single new customer (upsells and seat expansions outpace churn). A median SaaS business runs at 105–110%. Below 100% means the base is contracting before new logo growth is even considered. As a reference point from the European high yield market, TeamSystem — the Italian SME software provider — recently reported approximately 85% recurring revenue and an NRR of roughly 108%.

Gross retention of above 90%, combined with ARR predictability, is what lenders have historically underwritten. The stickiness comes from switching costs: data migration, staff retraining, workflow disruption, and contract inertia that make changing platforms a 6–18 month, multi-million-dollar process. For a credit investor, these dynamics translated into an asset-light business with 80%+ gross margins, near-zero marginal cost of serving incremental users, and a free cash flow conversion profile that justified leverage levels of 7–10x EBITDA — well above what most industrial or energy businesses could support.

The SaaS P&L: Where the Margin Structure Was Built

The margin structure that private equity built its software playbook on has a specific architecture. Walking through each line clarifies exactly where AI is now applying pressure.

Revenue is primarily subscription ARR — contracted, recurring, typically billed annually in advance — supplemented in some cases by professional services at implementation, and increasingly by consumption-based elements tied to AI usage.

Cost of goods sold (COGS) has historically been 15–25% of revenue for traditional SaaS, covering cloud hosting (AWS, Azure, GCP), customer support, DevOps, and data centre costs. This is where AI introduces the most significant structural change: inference compute is now a new COGS line item, directly variable with usage. Every AI interaction — every model call — consumes GPU time, memory, and power. The more customers use the AI-enabled features, the higher COGS becomes. For AI-heavy models, COGS can rise to 30–40% of revenue.

The consequence is gross margin compression. Traditional SaaS gross margins run 75–85%. AI-integrated SaaS is compressing that range to 60–75%. For some agentic use cases with uncapped token consumption, gross margins can fall further. This matters because gross profit is the foundation of all operating leverage calculations — it is the engine that funds R&D, sales, and G&A while generating the EBITDA that debt is serviced against.

R&D — engineering salaries, development tools, cloud dev environments — typically runs 15–25% for growth-stage companies and 10–15% for mature or PE-owned names. The temptation post-acquisition is to cut here aggressively. The risk is product stagnation: underinvestment in R&D drives up churn, which is a slow-burn credit problem that shows up in NRR before it shows up in revenue. AI coding assistants are reporting 20–40% productivity gains, and some companies claim multiples of output per engineer — which theoretically allows R&D cost reduction without stagnation. But the same tools are available to every competitor, so productivity gains may simply get competed away through pricing pressure.

Sales and marketing is the largest lever PE firms pull post-acquisition. A growing SaaS company spends 30–50% of revenue here to acquire new logos. A mature or PE-owned company cuts that to 15–20%, relying on renewals and price hikes rather than new customer acquisition — a deliberate shift in strategy that compresses near-term EBITDA while accepting slower growth.

General and administrative (finance, HR, legal, executive comp) runs 8–15% for growth companies and 5–10% for mature names. Offshoring is the primary lever here, but G&A is a smaller portion of the cost base than R&D or S&M, so the impact is correspondingly limited.

The PE playbook is to acquire a company that spends 40–50% on S&M and 20–25% on R&D, cut both aggressively (to 15–20% and 10–15% respectively), push gross margins to 80–85% through hosting optimisation and pricing discipline, and exit with a 35–45% EBITDA margin against which 7–10x debt was underwritten. As Matthew noted: “Achieving 40% EBITDA with 85% gross margins requires a very specific profile. You can’t have 30–50% sales and marketing expense and hit 40% EBITDA. To get there, you must radically alter the operating expense structure.”

The Leverage Math: Why Margin Compression Is Non-Linear

This is the section that matters most for credit investors reviewing software names right now, because the leverage impact of margin compression is not proportional — it accelerates.

Start with a baseline: a 7x levered SaaS company running 40% EBITDA margins. If gross margins compress by 500 basis points — from 85% to 80% — EBITDA declines by approximately 12–15%. The natural assumption is that leverage ticks up by a proportional amount. It does not.

A 10% EBITDA decline does not move leverage from 7x to 7.7x. It moves it to 7.8x. A 20% EBITDA decline — the kind of pressure a company faces when gross margin compression is combined with revenue growth slowing from 15% to 5% — moves leverage to 8.75x. Add a revenue decline scenario and you are looking at 9–9.5x, which makes the capital structure hardly sustainable. A severe scenario — meaningful revenue contraction combined with gross margin compression — produces effective leverage of 10x or above, which is restructuring territory.

The dynamics around LTV compound the problem. Software multiples that once cleared 15–20x EBITDA at entry are compressing toward 10–12x in private markets — and if AI disruption continues, 8–10x exit multiples are not unrealistic. A 7x levered company bought at a 15x multiple has roughly a 5x turn of enterprise value cushion. At an 8x exit multiple, that cushion narrows to 1–3 turns of leverage — a thin margin for error for a business whose cash flow profile is being simultaneously pressured.

Recovery rates follow the same logic. Software credits have historically recovered 40–50 cents on the dollar — above the HY average — because of ARR stickiness and customer retention. But software assets are intangible: take the business and shake it upside down and what falls out is talented engineers, a maintained code base, and a contracted ARR base. Under distress, engineers leave first (they are highly employable), which drives quality slips, which accelerates churn, which drives further revenue decay. The doom loop is self-reinforcing and, once started, difficult to reverse.

AI as a New Cost of Goods Sold

The structural paradox of AI integration is that it is simultaneously mandatory and margin-destructive.

Traditional SaaS has near-zero marginal cost of serving additional users. Once the platform is built, incremental revenue is almost pure gross profit. This is the operating leverage that justifies high multiples. AI flips that dynamic: every AI interaction consumes compute resources, which means gross margins become directly correlated with usage. The more customers use the AI-enabled product, the higher COGS become — an inverse relationship between product performance and profitability.

The strategic dilemma, as Alec framed it, is explicit: “You have to integrate AI because it’s mandatory to be competitive. But doing so does structurally weaken the very margin profile that these companies were underwritten on.” A company that refuses to integrate AI loses customers to AI-native competitors. A company that does integrate AI erodes the gross margins that service its debt. There is no cost-free option.

On the productivity side, AI does offer genuine savings — primarily in engineering and customer support. AI coding assistants reduce developer headcount requirements; AI agents handle tier-one support tickets, reducing customer service cost. The bull case for incumbents is that AI-driven headcount reduction of 20–30% expands EBITDA margins enough to offset inference cost growth. The caveat is that since these productivity tools are universally accessible, any efficiency gains may simply get competed away through pricing pressure rather than dropping to the bottom line.

Fast-Growing vs. PE-Owned SaaS: Two Different Risk Profiles

The AI credit risk is materially different depending on which type of SaaS company you are looking at.

Fast-growing SaaS — 20–40% ARR growth, NRR above 120%, negative or low EBITDA, levered on ARR multiples — faces primarily a growth deceleration risk. AI alternatives make it harder to win new logos. If ARR growth slows, revenue multiples compress, and the equity cushion that justified aggressive leverage disappears. A secondary risk is liquidity: GPU infrastructure spend must be committed before AI features generate revenue, accelerating cash burn for companies that do not yet have the balance sheet to absorb it.

Mature / PE-owned SaaS — sub-10% growth, NRR around 100–110%, 35–45% EBITDA, levered 6–10x on EBITDA — faces primarily a margin erosion risk. These companies must invest in AI to defend their product against disruptors, but every dollar of AI investment directly compresses the EBITDA margins that their debt covenants are tested against. The PE playbook of price hikes and cost cuts runs into a headwind: AI alternatives give enterprise clients new leverage in contract renewals. KPMG recently demonstrated this at an industry level — demanding and receiving a 14% fee reduction from its longtime auditor, citing AI productivity gains. That dynamic scales to enterprise software contracts.

The covenant risk is real: even a 5% EBITDA miss at 7–10x leverage creates default risk. Companies that fail to hit growth targets also fail to hit deleveraging targets, which for the most highly levered names was the central promise underwriting the refinancing.

Lessons from History: Technology Creates Value, Not Necessarily Returns

Alistair Nairn’s Engines That Move Markets — a book I can highly recommend (apart from the 2000s the part about autos is great as well) — studies every major technology transition from railways to the internet and surfaces a lesson that is structurally relevant right now: identifying which industries will be disrupted is substantially easier than identifying which companies will win.

Of hundreds of car manufacturers in 1910, only a handful survived. The internet transformed commerce, media, and communication — and then the dot-com bubble wiped out 80% of equity value in eighteen months. The early leaders in almost every technology cycle failed or were acquired. Competitors disappeared. The market was right about the technology and wrong about the winners.

The implication for credit is uncomfortable: “We can say with high confidence that AI will disrupt software. What we cannot say is which specific SaaS companies will efficiently integrate AI — and we also cannot say which will take the wrong steps and get wiped out.” The current selloff, in this framing, is the market repricing uncertainty rather than expressing a precise view on individual credits. It is that unknown — not a specific set of defaults — that is driving spread widening across the sector.

The data centre parallel is particularly salient. In 2026, hyperscalers are expected to spend approximately $750 billion in annual capex. The most recent HY data centre bond — for Fleet Data Centers, a subsidiary of Trapped Capital — is a 200-megawatt initial build working toward a 25-gigawatt total capacity. Many of these deals are underwritten at 80–100% utilisation. The historical precedents — the railway mania of the 1840s, the $1 trillion fibre buildout of the telecom bubble that peaked at single-digit utilisation — are not reassuring.

Is AI Replacing the Software TAM or Growing It?

The question of whether AI substitutes for existing software spend or expands the addressable market is the right one to ask — and the answer is that three things are happening simultaneously, with the net effect still unresolved.

The first dynamic is substitution: AI agents replace human seats, which represents direct revenue loss for incumbent SaaS vendors. One AI agent replacing multiple human users is straightforward revenue cannibalization for a per-seat pricing model. The second dynamic is expansion into previously unaddressable use cases — autonomous research, 24/7 monitoring, handling unstructured data — which grows the total addressable market. The third dynamic is creation of entirely new infrastructure demand: hyperscaler capex approaching $750 billion annually is additive to existing enterprise software spending.

The credit implication of the first dynamic is the most immediate and most negative: seat-based pricing is structurally challenged. A consumption-based model (cost per token, cost per API call, cost per workflow completed) introduces revenue volatility that credit models have not historically been built to handle. The SaaS ARR metric — the heartbeat number that lenders underwrite against — becomes unreliable as a forward indicator when revenue is variable rather than contracted. Some companies are moving toward hybrid models: a base contract with usage overcharges, capturing both predictability and upside from power users. Twilio is the cited example of a European HY credit already operating on a usage-based model — and while that was flagged as a potential risk 12 months ago, it may now represent a structural advantage in an AI-driven world.

Platform moats beat feature moats in this environment. A feature moat — a smart search, a workflow automation, a nice dashboard — can be replicated by a capable developer in days. A platform moat — owning the system of record, the data flows, the integration layer — cannot be replicated without ripping out the plumbing that a business depends on. Conga, a Thoma Bravo-backed RevOps company built on top of Salesforce, is the cited example: its financing was pulled mid-sell-off because the market could not price the value of a feature moat on top of a platform whose equity had already repriced.

The Disruption Framework: A Credit Matrix for Every SaaS Name

The analytical tool that Alec and Matthew presented is a 2×2 framework mapping two axes: data moat depth (how much proprietary, irreplaceable data does the company control?) and workflow complexity (how structured, regulated, and judgment-intensive are the tasks the software automates?). It was pioneered here:

TSCS

Don't Short SaaS

$1.2 trillion…

Read more

9 months ago · 247 likes · 31 comments · Strategist, Les Barclays, and Architect

The four quadrants define four distinct credit risk profiles.

Kill Zone (shallow data moat, low workflow complexity): Simple, repeatable tasks with no proprietary data advantage. AI directly substitutes. Examples include basic CRM tools, email marketing platforms, simple dashboards, scheduling software. The credit risk is existential. Recovery value is uncertain because the intangible assets that constitute the value of these businesses — a code base actively maintained by engineers who have better options elsewhere — are subject to the doom loop described earlier. Alec’s framing: “Over the next 12–18 months, we may see a lot of these entering distressed territory and finding out how valuable their IP really is.”

Data Fortress (deep data moat, low workflow complexity): Proprietary data with simple delivery. AI improves the product but cannot replicate the data. Examples include Bloomberg Terminal, credit rating agencies, and specialised data vendors. These are defensible credits — but pricing pressure exists if AI can aggregate comparable data more cheaply. If a generic model can produce output that is 80–90% as accurate as the proprietary product, that shifts pricing power toward the client. The real opportunity for Data Fortress companies is to leverage their proprietary data to build AI-powered workflows and extend their moat rather than simply defend it.

Contested Ground (shallow data moat, high workflow complexity): Complex workflows that are not easily replaced because of scale and interconnectivity, but data advantages are limited. Generic ERP, IT service management, and HR platforms sit here. These are not the easiest businesses to displace — ripping out enterprise resource management is a multi-year project with substantial migration cost — but the switching cost is not data-driven, which means motivated enterprises can eventually make the move. The credit opportunity here is that some of these names have been oversold in the current repricing and may offer value relative to the near-term disruption timeline.

AI-Enhanced Monopoly (deep data moat, high workflow complexity): The strongest position in software. Deep proprietary data combined with judgment-intensive, regulated, multi-stakeholder workflows that AI augments but cannot replace. Epic in healthcare, Palantir in government and defence, Veeva in pharmaceuticals. The risk here is low. These businesses are likely net beneficiaries of AI — layering AI onto an existing platform extends the moat rather than eroding it.

Proprietary data, in this framework, is the last line of defence for software valuations. Models are commoditising: open-source alternatives like Llama, Mistral, and DeepSeek are now frontier-capable, and 80% of AI startups reportedly use cost-effective Chinese models. The model layer itself is structurally a low-margin business. What cannot be commoditised is domain-specific training data accumulated over decades — health records, financial transaction histories, supply chain data. As Alec noted: “If you are simply a software company layering a pretty UI over somebody else’s data, somebody else can put a prettier UI over you. The data is the moat.”

Named Credits: Kill Zone and Contested Ground

The conversation covered a range of specific names across both the risk and opportunity ends of the framework.

European names

ION Platform sits in the AI-Enhanced Monopoly quadrant on the framework. It serves large-scale enterprise contracts with asset managers, banks, and trading firms — subscription terms averaging five years and a reported 2.8x weighted-average contract term remaining. Deeply embedded workflows requiring compliance, risk management, and data sovereignty make churn risk low and switching cost high. The credit thesis is that AI gets layered on top of the platform rather than replacing it.

The complication is market mechanics rather than business fundamentals: ION Platform has been one of the worst performers in European HY in 2025, in part due to opacity in its financials and investor concern — perhaps not entirely rational — about the Altice precedent. In a market where creditors are watching for aggressive liability management, one of Europe’s most leveraged software capital structures attracts a risk premium regardless of operational quality.

TeamSystem — Italian SME-focused accounting, payroll, and compliance software — sits in Contested Ground. Deeply embedded in client filings, payroll runs, and regulatory workflows, it benefits from inherent retention. Full displacement by an AI agent in the near term is unlikely. The market has treated it more gently than ION Platform, reflecting the difference between contested ground (showing-me stage, limited AI substitute) and a deeply leveraged monopoly with opaque financials.

Software AG sits in or near the Kill Zone. Its core process management offering — documentation, workflow mapping — is increasingly automatable by cheap generative AI tools, particularly for SME clients looking for cost reduction. Buyers are expecting AI features as table stakes, not as a value-add. The spend required to add those features will pressure margins, and NRR is at risk as customers find cheaper alternatives.

US names

The US market context is important: private credit is substantially larger in the US than Europe, and BDCs like Blue Owl Technology Finance and Sixth Street Specialty Lending have more than 50% exposure to software names. The most recent development — a separate Blue Owl fund permanently restricting withdrawals — underscores that the repricing is not just a public market phenomenon.

Conduent is the clearest Kill Zone name in the US conversation. Business process outsourcing — call centres, repetitive back-office workflows, rules-based tasks — is precisely what AI agents are best suited to automate. Conduent has no proprietary data moat. Its service delivery is commoditised. Despite being lower-levered than many software LBO names, its secured notes fell 20 points in a month — demonstrating that you do not need an over-leveraged capital structure to be in the crosshairs of an existential business model shift.

Quest Software sits close to the Kill Zone boundary — legacy infrastructure middleware and admin tools with limited proprietary data advantage. AI-native tools can increasingly replicate administration, monitoring, and troubleshooting. Quest was already experiencing pricing and mix pressure before the AI acceleration. A refinancing in mid-2025 preserved it through the near term, but if AI erodes the need for standalone admin tooling, margin durability is materially weakened.

Perforce Software — developer tools and enterprise DevOps — sits in the AI blast radius most directly. AI coding assistants reduce reliance on traditional version control, testing, and DevOps orchestration. Perforce is B3 rated, sponsor-backed, and 5x levered. The combination of leverage and frontline AI exposure is the most dangerous combination the framework identifies.

On the other side: Cloud Software Group is cited as a genuine software credit that can adapt — primarily because of scale. As the largest privately owned software company, it has the resources to invest in AI integration that smaller, more concentrated players do not. Qlik Technologies (BI platform) is named as a double-B-rated first-lien credit that has sold off but where lenders remain constructive — a name swept up in headline risk that operates in industries with significant near-term inertia.

Winners: Picks and Shovels, Infrastructure, and Good Credits Sold Off Hard

The real winners of this transition, as Alec framed it, are not software companies. They are the businesses providing the physical infrastructure that AI runs on — regardless of which AI model, which application layer, or which software company ultimately survives.

Vertiv, Generac, and Eaton are the cited picks-and-shovels plays. Vertiv and Eaton supply power distribution, cooling, and UPS systems to data centres. The demand tailwind from AI is structural and does not depend on any individual hyperscaler or AI model winning the market. Historical demand for electrical equipment at data centres is being overwhelmed by AI-driven requirements — companies like Vertiv have already reported significant gains.

The data centre credit story is more complicated. $15 billion of HY data centre debt has been issued since October 2024, almost all of it structured as project finance — pre-revenue builds underwritten at 80–100% utilisation. The construction risk, energy infrastructure constraints, and counterparty concentration (many deals dependent on a small number of hyperscaler tenants) are real risks. The Fleet Data Centers offering is illustrative: a 200-megawatt initial facility working toward 25 gigawatts of total capacity, from a company without a substantial operating track record. The telecom parallel — a $1 trillion fibre buildout at single-digit utilisation — is not alarmist framing; it is the historical base rate for technology infrastructure build-outs funded by capital markets.

Healthcare software companies represent the most interesting near-term opportunity in the “good credit sold off hard” category. Healthcare in the US operates with significant regulatory inertia — slower AI adoption than financial services or technology — which gives companies like R1 RCM, Ensemble Health Partners, and Meddata a meaningful runway before their addressable market shrinks. In the near term, they can slash labour costs as AI tools come online while their hospital and payer customers have not yet adopted AI themselves. The asterisk: on a long enough horizon, as hospitals and payers adopt AI, the TAM for these intermediaries shrinks materially. The credit opportunity is timing — buying at distressed levels during a market sell-off when the headline risk is AI disruption but the near-term cash flow is protected by industry inertia.

Khaki International — defence IT services contractor for the US Department of Defense — is a name mentioned specifically: the DoD’s announcement of a shift toward software-based infrastructure creates a direct revenue tailwind regardless of which AI model ultimately prevails inside that infrastructure.

The Analyst Handbook: How to Screen Any SaaS Credit Now

Matthew offered a practitioner’s checklist at the close of the conversation — four screening questions and five key credit risks for any SaaS name in a post-AI market.

The four screening questions are:

• Revenue durability: What is the NRR? What is gross revenue retention? Are contracts enterprise-level or SME? How will those clients respond to the AI disruption in their own industry?

• Cash timing: Is the company extending payment terms — allowing monthly rather than annual billing — to retain clients who might otherwise churn? This is a leading indicator of ARR stress before it hits revenue.

• Profitability: Rule of 40 (EBITDA margin + revenue growth rate), current EBITDA margins and gross margins, trend direction.

• Leverage tolerance: What happens to covenant headroom if EBITDA margins compress 500bps? If revenue growth drops from 15% to 5%? Build the scenario table — the non-linearity makes it essential.

The five key credit risks to flag for every SaaS name:

• Concentration risk: How many customers account for 20%+ of ARR? Single-customer concentration is catastrophic in a churn scenario.

• Pricing model transition: Is the company on seat-based, usage-based, or hybrid pricing? How dependent is the current ARR on a per-seat model that AI agents directly threaten?

• Cybersecurity exposure: AI-powered attacks on software systems are a growing risk. Third-party breaches that affect a SaaS platform’s data integrity can drive accelerated customer churn.

• Third-party dependency: What happens if a core infrastructure dependency (a hyperscaler, a key API provider) goes down? What is the recovery plan?

• Ownership structure: Founder-led and focused on long-term growth versus PE-owned and optimising for exit. The PE playbook has historically compressed R&D and S&M investment — which is the exact investment the business needs to make right now to defend against AI disruption.

Listen to the full episode for Alec’s and Matthew’s takes on BDCs.

This article is based on Episode 8 of Fixed + Floating, featuring Alec Keblish and Matthew Hughes of 9fin. The views expressed are those of the speakers and do not constitute investment advice. For more information, please visit 9fin.com.

Fixed + Floating is a credit podcast for institutional investors and finance professionals — exploring the forces shaping global credit markets. Hosted by credit PM Josef Pschorn, the show features conversations with leading voices from investing, research, and academia.

Support Material:

F+f Podcast Slides Final

1.48MB ∙ PDF file

Download

Download

08 Software Transcript

115KB ∙ PDF file

Download

Download