Creditors in the Dark: Yield Speculation, the Fight for Cash, and the Hidden Default Risk Nobody Is Pricing

Credit is not money. That distinction — obvious when stated plainly, systematically ignored in practice — is the foundation of everything that follows. The United States monetary base, M0, the only money the Federal Reserve actually creates and controls, stands at $5.46 trillion. The M2 money supply, which includes bank deposits, savings accounts, and money market funds, is $23.1 trillion. The total US public debt is $39 trillion. And the total credit market — when you layer in corporate bonds, private credit, leveraged loans, CLOs, mortgages, and every other instrument of structured obligation — dwarfs all of it. For every dollar of real base money, the system has manufactured roughly 4.2 dollars of deposit claims through fractional reserve banking. And on top of that foundation, it has stacked a debt edifice that the base money supply could not repay in full even if every dollar of it were devoted entirely to that purpose. The system works — until it doesn't. And what happens when it doesn't is not a slow unwind. It is a scramble.

The credit market functions through a perpetual act of trust. Borrowers issue obligations. Lenders accept them in exchange for yield. Both parties implicitly trust that when the time comes to repay or refinance, the conditions to do so will exist. That trust holds during good times not because the underlying math works, but because enough participants believe it does. What breaks the trust is not a single default. It is the realization, spreading through the market in real time, that the pool of real cash available to absorb distress is a fraction of the pool of credit obligations that have been issued against it. At that moment, the most rational thing any creditor can do is be the first one to get paid — and that collective rationality is exactly what makes it catastrophic.

Over-Leverage Is No Longer a Warning — It's the Baseline

The private credit market has grown to $3.5 trillion in assets under management as of 2025, up from under $1 trillion a decade ago and projected to reach $5 trillion by 2029. This is capital deployed largely outside the regulated banking system, with limited transparency, in illiquid instruments, to borrowers who could not access the public bond market. The structural risk embedded in this growth is not that the loans will default — some will, and that is priced in to a degree. The structural risk is what happens to the system's plumbing when refinancing conditions tighten simultaneously across the market.

Over 90 percent of US leveraged loans issued in recent years are now covenant-lite — a record. Covenant-lite means the lender has surrendered most of the early warning mechanisms that traditionally allowed them to force a renegotiation before a borrower deteriorated beyond recovery. The lender is flying blind, holding an instrument that pays a coupon and promises repayment but offers no effective tripwire before a credit event. Private credit CLOs — pools of these instruments, repackaged and tranched for institutional investors — now account for 20 to 25 percent of total CLO issuance, up from roughly 10 percent a few years ago. And 24 to 29 percent of existing leveraged loans mature in 2025 and 2026, requiring refinancing at interest rates two to three times higher than when the original loans were made. That refinancing wave is not theoretical. It is in the calendar.

The Historical Fight for Cash — and How Fast It Happens

The sequence of a credit liquidity crisis has been observed enough times that it should no longer be surprising — yet each iteration catches the market off guard, because each iteration arrives in a different coat. The mechanism is always the same. Credit spreads widen. Leveraged positions face margin calls. Forced selling produces more price declines. More margin calls follow. The only thing anyone wants in this environment is cash — actual cash, not a credit instrument that promises cash — and the discovery that cash is genuinely scarce relative to the claims against it sends prices through the floor.

In 1998, Long-Term Capital Management collapsed in under four months after Russia defaulted on domestic debt. LTCM had structured convergence trades across dozens of markets, all premised on the same intellectual framework: dislocated spreads would converge. Instead, they diverged everywhere simultaneously. LTCM lost $4.6 billion. The Federal Reserve organized a $3.6 billion bailout from fourteen institutions because the alternative — an uncontrolled unwind of LTCM's positions — threatened to cascade through every credit market in the world. The lesson: in a dash for cash, the correlations you built your model on break down. Everything trades to zero liquidity at once.

In 2008, the high-yield spread index — the ICE BofA HY Index — reached 2,074 basis points by December 8th. A spread that had been trading below 300 basis points in mid-2007 had expanded by more than 1,700 basis points in eighteen months. Investment grade spreads nearly tripled. At those levels, the implied default rates were beyond what had ever been observed in the historical record. The market was not pricing expected defaults. It was pricing the possibility that the system itself could not function — which is a different kind of risk, and one that no yield premium can adequately compensate for.

In March 2020, the dynamic was distinct and in some ways more instructive. The initial shock was not a credit event but a liquidity event. Even US Treasuries sold off simultaneously with risk assets — the canonical safe haven was not safe, because holders needed cash, not yield, and they would sell anything liquid to get it. HY spreads peaked near 1,100 basis points. BBB spreads hit approximately 500 basis points. Corporate bond issuance nearly vanished. The Federal Reserve announced corporate bond purchase programs on March 23, 2020. The market bottomed the same day. Not because fundamentals had changed, but because the cash backstop reappeared. The cash was always the point.

What Yields Are Telling You — and What They Are Not

The 10-year Treasury yield sits at 4.49 percent as of mid-May 2026. The 5-year TIPS auction in April cleared at a real yield of 1.367 percent, implying a 5-year inflation breakeven of 2.58 percent — the highest at auction since October 2022, and well above the 15-year average breakeven of 1.98 percent. CPI for the twelve months ending March 2026 was 3.3 percent, still a full 130 basis points above the Federal Reserve's 2 percent target. The Fed funds rate is at 4.25 to 4.50 percent, on hold since late 2025. The market expects one to two cuts in the second half of 2026.

Against this backdrop, high-yield credit spreads are at approximately 279 basis points over Treasuries. Investment grade spreads are around 80 basis points. Both are described, accurately, as historically tight — offering limited compensation for the risks embedded in the instruments they price. This is the essence of yield speculation. Investors are being paid a yield that reflects the expected value of credit losses under a relatively benign scenario — and they are systematically not being paid for the tail risk: the scenario where liquidity evaporates, refinancing conditions tighten simultaneously, and mark-to-market losses force further selling from leveraged holders.

The hidden default risk has several components that do not show up in reported spreads. The first is refinancing risk — $39 trillion in public debt plus the private credit refinancing wave hitting in 2025 and 2026, all at rates that would have been considered catastrophically high as recently as 2021. A borrower who issued five-year debt at 2 percent in 2021 is now rolling into a market offering 6 to 7 percent. The debt service increase is not a credit event. It is a slow margin squeeze that shows up in earnings before it shows up in defaults — and by the time it shows up in defaults, the spread is already blowing out. The second component is illiquidity risk. Private credit instruments — $3.5 trillion of them — do not have a liquid secondary market. In a stress scenario, there is no price discovery. There is only a last-done trade that nobody wants to do. The third is correlation risk: the same shock that triggers credit stress also tends to trigger dollar strength, rate increases, and equity market declines simultaneously, eliminating the diversification that most portfolios rely on.

Japan: The $1.14 Trillion Pillar Starting to Wobble

Japan holds $1.14 trillion in US Treasury securities — the largest single foreign creditor position on earth. For decades, this was an almost mechanical arrangement. Japan's interest rates were zero or negative. US rates were higher. Japanese institutional investors — life insurers, pension funds, regional banks — borrowed in yen at near-zero cost and deployed the proceeds into US Treasuries, pocketing the rate differential. This carry trade, estimated at over $4 trillion globally, was one of the structural pillars of US debt financing. It kept US yields suppressed. It kept the dollar bid. And it kept the entire system functioning at a debt load that would otherwise have required significantly higher rates to clear the market.

The Bank of Japan began exiting this arrangement in March 2024 when it abandoned Yield Curve Control — its explicit cap on 10-year Japanese government bond yields. By late 2025, the BOJ had raised its policy rate to 0.75 percent, the highest in decades. Japan's 10-year JGB yield hit a 25-year high in December 2025. The 40-year JGB yield hit a record in January 2026. The carry trade that sustained foreign demand for US Treasuries is narrowing in profitability. Japanese net purchases of US Treasuries decelerated from $180 billion in full-year 2024 to $45 billion in 2025 — a collapse of 75 percent in the annual buying pace. The question of when this becomes net selling is no longer hypothetical. It is a matter of timing.

The paradox here is subtle but important. As Japanese rates rise, the cost of hedging yen exposure on US Treasury purchases falls — which could, near-term, keep some Japanese demand intact as hedge-adjusted yields remain attractive. But the marginal new demand that carried US Treasury auctions for years is structurally diminished. And in a risk-off environment — exactly the environment where Treasury yields should fall as a safe-haven bid — Japan could be a net seller, inverting the historical relationship between fear and Treasury prices. The market has not fully priced this inversion. It is priced for Japan to remain a passive, permanent buyer. That assumption deserves to be stress-tested.

Hormuz, Inflation, and the Rate Path That Cannot Cut Its Way Out

The Strait of Hormuz crisis has introduced a complication that the Federal Reserve cannot address with the tools it has. Brent crude surged approximately 65 percent — $46 per barrel — in March 2026 alone, the single largest monthly gain ever recorded. The IEA declared the current disruption the worst energy crisis in history, worse than the 1973 and 1979 oil shocks combined, with over 12 million barrels per day removed from accessible global supply. Energy prices are forecast to surge 24 percent in 2026. And CPI, already running at 3.3 percent before the full effects of the oil shock worked through the supply chain, is now facing a second inflationary impulse from the energy side precisely as the market was expecting the Fed to begin cutting.

The rate path matters enormously for the credit thesis. Every point of Fed easing that gets delayed by energy-driven inflation is a point of relief that does not arrive for the leveraged loan refinancing wave. Every additional month of 4.25 to 4.50 percent funds rate is another month of debt service at elevated cost for borrowers who structured their obligations on the assumption of a normalized rate environment. The private credit market's $3.5 trillion was largely extended when rates were at zero or close to it. The refinancing of that book at current rates is not merely a headwind. For the weakest borrowers, it is the cliff.

If Treasury Yields Rise, AI Multiples Are the First Casualty

The AI bull run has produced a remarkable concentration of market value in a small number of names trading at multiples that presuppose a long runway of above-average growth. Nvidia trades at a price-to-sales ratio of 26.75 times — 39 percent above its 10-year median of 19.31 times. Its trailing P/E is 46 times. Hyperscalers collectively — Alphabet, Amazon, Meta, Microsoft, Oracle — are expected to spend approximately $520 billion on AI infrastructure in 2026, a 30 percent increase from the already-elevated 2025 pace. The market is pricing not just the current business but the future cash flows that this infrastructure investment is supposed to generate, discounted at a rate that reflects the current yield environment.

This is precisely where yields and equities intersect. Long-duration assets — stocks whose value lies predominantly in future earnings rather than current ones — are mathematically the most sensitive to changes in the discount rate. When the Fed raised rates from near-zero to 4.50 percent in 2022, the Goldman Sachs Non-Profitable Tech Index fell 27 percent in January 2022 alone and roughly 50 percent over the full year. Broader growth sector multiples fell more than 77 percent from their 2021 peaks. The mechanism is mechanical: each 100 basis points rise in the risk-free rate increases the denominator of every discounted cash flow model, compressing the present value of future earnings. A stock trading at 50 times forward earnings is not five times riskier than one at 10 times. In a rising rate environment, it can be fifty times riskier — because the discount rate sensitivity scales with the duration of the earnings stream.

The parallel to today is uncomfortable. The AI bull run is being funded by a belief that interest rates will normalize toward 3 percent, which is the implicit long-run rate embedded in most AI company DCF models. If the Hormuz shock keeps inflation above target, if Japan repatriation pushes 10-year yields above 5 percent, and if the private credit refinancing wave produces visible defaults that widen spreads — the discount rate assumption embedded in every AI multiple is wrong. Not directionally wrong. Wrong by enough to produce meaningful multiple compression across the entire sector, at a time when the sector represents an extraordinary concentration of market capitalization.

Where the Money Goes When the Fear Starts

Capital outflow signals are already visible, even before a formal credit event. The DXY has fallen approximately 10 percent from its late-2024 peak of 109 to around 99 today — the dollar weakening even as the US runs the highest rates in decades, which is an anomaly that typically signals something structural rather than cyclical. In January 2026, net outflows from US Treasuries were $18 billion and from US equities were $22 billion in a single month. Japan's Treasury buying pace has collapsed 75 percent. Gold surged over 70 percent in 2025, the best annual performance in decades, as central banks and institutional investors diversified away from dollar assets — explicitly citing the "weaponization of the US dollar" as the structural driver.

In a full credit scare, the sequence of capital flows tends to follow a pattern. First, risk-off rotation within equities — out of high-multiple growth, into cash-generative value and defensives. Second, a bid for short-duration Treasuries as the safe-haven reflex fires, even if long-end yields simultaneously rise from supply pressure. Third, a bid for physical gold and short CHF positions as dollar confidence wavers. Fourth — and this is where Japan matters specifically — a potential yen strengthening as carry trades unwind and repatriated capital flows home. The yen carry unwind in August 2024 was a preview: a single week of BOJ rate guidance caused a global equity selloff and a swift yen appreciation. That was a warning shot. The underlying position that caused it has not been fully unwound.

For investors thinking about positioning in this environment, the directional signals point toward shorter duration in fixed income, skepticism of credit spread levels that offer little historical margin of safety at 279 basis points for high yield, caution on the highest-multiple names in the AI complex, and some structural allocation to assets that perform in dollar stress scenarios — gold, short-term yen exposure during carry unwind phases, and energy equities that benefit from the very Hormuz shock that is compressing the Fed's room to maneuver. None of these are certainties. They are asymmetric bets in an environment where the risk is asymmetric.