The Three Cases Problem
[I] THE SHARED FACT, THREE READINGS
Between early April and late August 2026, the baseline benchmark for American borrowing costs executed a move that macroeconomists have studied for a century: the yield curve un-inverted.
In early April, the gap between the yield on a 10-year U.S. Treasury bond and the yield on a 3-month Treasury bill stood at approximately 0.61 percentage points. By late August, that spread had widened to 0.83 percentage points. Crucially, this un-inversion was not driven by the Federal Reserve rapidly slashing short-term interest rates to rescue a failing economy. Instead, it was driven by the long end of the curve moving higher: the 10-year Treasury yield held persistently above 4.60%, touching 4.89% in the spring before settling near 4.68% by late August.
In market terminology, this is a "bear steepener" — a regime where long-term borrowing costs rise faster than short-term rates.
When a yield curve un-inverts after a prolonged period of inversion, historical textbooks offer a straightforward interpretation: a recession is imminent. Yet today, financial markets, central bankers, and corporate treasurers are looking at this exact same dataset and drawing radically different conclusions.
To understand why, it helps to borrow a concept from physics: the three-body problem. In classical mechanics, predicting the movement of two gravitational bodies orbiting each other is mathematically clean and predictable. But introduce a third massive body, and the interaction becomes chaotic. The system is not unpredictable because any individual gravitational force is miscalculated; it becomes unpredictable because three independent, competing forces act on the same objects simultaneously, producing an emergent outcome that no single closed-form equation can resolve.
Macroeconomics today is confronting its own three-body problem. The yield curve's movement is being acted upon by three separate, internally rigorous, and evidence-based economic mechanisms. None of these three frameworks is wrong in isolation. The current uncertainty stems from the fact that all three forces are acting on the market at the exact same moment.
[II] CASE A — WHEN CALM MEANS DANGER (THE MINSKY & CREDIT QUALITY ENGINE)
The first framework is the Cyclical or Minskyan view. Its core claim is simple: financial markets that remain calm for a long time do not eliminate risk — they quietly accumulate it. As Hyman Minsky posited, stability is inherently destabilising. During extended periods of economic calm, investors become accustomed to low default rates and gradually stop demanding adequate compensation for the risks they take. Underwriting standards deteriorate, corporate leverage expands, and the economy shifts through Minsky’s classic taxonomy — moving from Hedge finance (where cash flows cover principal and interest) to Speculative finance (where cash flows cover only interest, requiring principal debt rollover).
Proponents of Case A look at current corporate credit markets and see classic late-cycle euphoria masking underlying balance-sheet fragility.
Consider the headline numbers. The option-adjusted spread on U.S. high-yield corporate bonds — which measures the extra interest rate speculative-grade companies must pay above risk-free Government Treasuries to compensate investors for default risk — sits at approximately 2.63 percentage points (263 basis points). By historical standards, this is exceptionally compressed. Similarly, the Excess Bond Premium — a specialized measure developed by Federal Reserve economists to isolate pure investor sentiment by stripping out individual firm default probabilities from corporate bond spreads — currently sits at negative 0.32 percentage points (-31.9 basis points). This places investor risk appetite in the 16th percentile of its multi-decade history, indicating that lenders are actively discounting risk and providing cheap capital to corporate borrowers.
On the surface, this looks like a picture of pristine corporate health. But Case A advocates point to a glaring internal fracture beneath the headline numbers: structural credit tiering.
While upper-tier speculative borrowers rated BB enjoy comfortable spreads of just 1.53 percentage points above Treasuries, distressed borrowers at the bottom of the credit stack — those rated CCC or lower — face borrowing spreads of 10.31 percentage points (1,031 basis points). The ratio between CCC and BB spreads currently stands at 6.74-to-1. Under Minskyan analysis, these CCC borrowers have entered the dangerous realm of Ponzi finance — units whose cash flows cannot cover even current interest commitments, making them entirely dependent on continuous refinancing or asset sales.
Furthermore, economic research by Greenwood and Hanson (Issuer Quality and Corporate Bond Returns) demonstrates that headline credit spreads are often a lagging indicator. The true barometer of cyclical turning points is the quality mix of debt issuance — the proportion of junk or high-default-risk issuers flooding credit markets during booms. In 2026, primary market underwriting has heavily favored the upper tier, with high-yield issuance accounting for just 12.7% of total corporate bond volume while investment-grade issuance absorbs over 80%. Distressed and lowest-tier issuers are effectively locked out of primary debt markets, concentrating balance-sheet risk at the bottom of the stack even as headline spreads remain superficially calm.
As Reinhart and Rogoff documented in eight centuries of financial crises (This Time Is Different), investors repeatedly misjudge these leverage booms as structural "new paradigms" right before credit markets freeze. Under Case A, the headline calm in corporate spreads is an illusion sustained by mega-cap tech stability. As low-rate legacy debt comes due and must be refinanced at prevailing market yields, stress at the bottom of the credit stack will inevitably migrate upward.
Case A's Core Prediction: The widening gap between upper-tier and lower-tier corporate borrowers will continue to expand, and the calm in headline credit spreads will eventually break as corporate refinancing needs force a repricing of risk.
[III] CASE B — THE PLUMBING HAS CHANGED (AND THE FORTRESS BALANCE SHEET HAS FRACTURED)
The second framework argues that Case A is fighting the last war. Its core claim is that the financial system's back-end infrastructure — the structural mechanisms that connect central bank liquidity, commercial banks, money markets, and corporate balance sheets — has been fundamentally rebuilt since the crises of 2008 and 2019. Consequently, old historical relationships between rising interest rates and corporate distress no longer apply because the plumbing that used to burst under pressure has been reinforced.
Case B relies on four specific operational realities in today's market:
The Corporate Debt Refinancing Delta: Contrary to fears of an impending corporate "debt cliff," the average interest rate gap between maturing corporate bonds and current market refinancing yields sits at just +0.16 percentage points (+15.6 basis points). Because large corporations aggressively extended their debt maturities during the record-low rate environment of 2020–2021, the vast majority of corporate debt is insulated from immediate interest rate shocks.
Bank Lending Standards: In the Federal Reserve's latest Senior Loan Officer Opinion Survey (SLOOS), commercial banks reported a net 0.0% tightening of lending standards for commercial and industrial loans to large and middle-market firms. Banks are not pulling back credit or cutting off access to liquidity as they typically do ahead of a cyclical downturn.
Central Bank Liquidity Backstops: Following short-term funding market volatility in September 2019, the Federal Reserve established the Standing Repo Facility (SRF) to act as a permanent, explicit ceiling on short-term money market rates. Today, usage of this facility is near zero, and overnight repo rates remain strictly aligned with the Fed's target rate. The liquidity circuit breakers are fully operational, but funding markets are so stable that no institution currently needs to draw on them.
Superstar Firm Capex and the Shift to Debt Financing: Grounded in research on market concentration (Autor et al., The Fall of the Labor Share and the Rise of Superstar Firms), dominant tech hyperscalers maintain high profit markups and massive market share. The major technology infrastructure companies are deploying roughly $720 billion to $745 billion in capital expenditures this year alone for data center and AI buildouts.
However, Case B's internal narrative has undergone a critical shift regarding where the money comes from.
The initial market assumption — that this unprecedented buildout would be funded purely out of internal cash flows and cash reserves, insulating tech giants from debt markets entirely — has not held up. Annual capex intensity has officially surpassed aggregate operating cash flows, pushing Free Cash Flow into negative territory for Alphabet (in Q2 2026) and Amazon (in Q1 2026) for the first time in years.
In response, tech giants have turned aggressively to credit markets. Global AI-related public debt issuance is projected to approach $570 billion in 2026 (highlighted by mega-bond deals like Oracle's $18B and Alphabet's $20B sales). Behind public markets sits an additional $3.1 trillion in off-balance-sheet shadow borrowing — including $1.2 trillion in uncommenced leases and $1.9 trillion in Special Purpose Vehicle (SPV) and private credit purchase commitments.
This distinction is vital. As macroeconomic research on corporate debt maturity shows (Jungherr et al.), while long-term debt insulates firms against immediate rate hikes, accumulating massive debt without matching near-term earnings creates Fisherian debt overhang. A debt-funded buildout depends entirely on future revenues arriving on schedule to service that debt. If AI monetization and productivity gains lag behind execution schedules over the next few years, debt service burdens could transform hyperscaler balance sheets from stabilizers into amplifiers of credit stress.
Case B's Core Prediction: Corporate default rates and credit spreads outside the AI infrastructure sector will remain muted as long as central bank repo plumbing holds. Within the AI infrastructure sector specifically, resilience depends on whether real revenue growth arrives quickly enough to service the debt being issued today.
[IV] CASE C — IT'S NOT ABOUT CREDIT AT ALL (SOVEREIGN DURATION & PHYSICAL CAPACITY)
The third framework steps outside the private credit channel entirely. Its core claim is that private lending markets and corporate risk appetite are secondary actors in today's economy. The dominant driver of interest rates, inflation, and economic activity is the public balance sheet: persistent government budget deficits, massive Treasury bond supply, and real physical supply-side capacity constraints in energy, labor, and global trade.
From the perspective of Case C, analyzing corporate credit spreads to forecast the economy is a fundamental category error.
Consider the sovereign financing picture. U.S. federal outlays reached $7.01 trillion over the past fiscal year, with annual primary deficits remaining historically large during an economic expansion. To fund this expenditure, the U.S. Treasury must continuously issue massive volumes of government debt into global capital markets.
This unprecedented supply of sovereign debt has directly impacted the pricing of long-term risk. The 10-year U.S. Treasury term premium — the extra yield investors demand to hold a long-term government bond rather than rolling over short-term bills — has expanded to +0.87 percentage points (+86.8 basis points). Econometric research by Thomas Laubach (New Evidence on the Interest Rate Effects of Budget Deficits) proves that long-end yields reprice directly to fiscal projections: a 1 percentage point increase in projected deficit-to-GDP raises 10-year forward Treasury rates by 20 to 29 basis points. Investors are demanding higher yields not because they fear private corporate defaults, but because they require higher compensation to absorb the sheer volume of public debt issuance.
Furthermore, inflation mechanics under Case C are anchored in the Fiscal Theory of the Price Level (Bianchi, Faccini, and Melosi, A Fiscal Theory of Persistent Inflation). When government spending expansions are perceived as unfunded debt shocks — expenditures not backed by expected future tax surpluses — they generate structural inflation drifts that central banks cannot easily extinguish through short-rate hikes alone.
This is compounded by real physical supply constraints. Over the trailing eight quarters, the correlation between real GDP growth and GDP deflator inflation has been negative (-0.36). In plain language, when economic growth decelerates, price pressures accelerate. As research on state-dependent multipliers demonstrates (Michaillat and Saez), aggregate demand stimulus in a supply-constrained economy yields low or negative multipliers, crowding out private investment while aggravating inflation.
Under Case C, the yield curve is un-inverting because the market is pricing in the long-term reality of persistent public borrowing and physical capacity limits.
Case C's Core Prediction: Long-term Treasury yields and the shape of the yield curve will continue to track government borrowing trajectories and real-world supply bottlenecks, remaining largely detached from traditional private corporate credit indicators.
[V] WHAT THE FED CHAIRMAN ACTUALLY SAID
On August 28, 2026, Federal Reserve Chairman Kevin Warsh delivered his address at the Kansas City Fed's annual economic symposium in Jackson Hole, Wyoming. Examining his remarks provides a live test of how policy makers are weighing these three competing frameworks.
Warsh's speech distributed evidence across all three cases without committing the central bank to any single reading:
Adjacent to Case A: Warsh emphasized that inflation is "not self-executing, nor... necessarily mean-reverting," reminding the audience that the Federal Reserve remains accountable for 65 months of sustained, elevated price pressures. His refusal to assume inflation will automatically return to target aligns with Case A's warning against market complacency.
Adjacent to Case B: Warsh explicitly highlighted that corporate credit spreads sit "near the low ends of their historical ranges" and that bank lending standards remain "on the easier end of their historical range." Rather than viewing this calm with suspicion, Warsh interpreted these metrics as evidence of "remarkable resilience" on Main Street and Wall Street. He also characterized AI investment as a genuine "hinge point in history," noting that capital expenditure in equipment and software is growing at 9% annually — with more than half driven by AI buildouts.
Adjacent to Case C: Warsh presented disaggregated inflation telemetry, noting that 54% of consumer price categories experienced annual price increases above 3% over the past year. While down from pandemic peaks, this is well above the 32% baseline that prevailed prior to 2020. This concentration of price pressures in specific categories supports Case C's diagnosis of localized supply bottlenecks rather than broad, generalized demand excess.
Perhaps the most revealing detail in Warsh's speech was an omission: across his entire address, the Federal Reserve Chairman made zero mention of the yield curve, the 10-year yield, or the Treasury term spread. The single market signal that anchors Wall Street's recession debate was completely absent from the Fed chief's explicit policy framework.
Warsh concluded with a direct warning regarding the reflexive feedback loop between central banks and financial markets — a dynamic he described as a structural hall of mirrors: market participants look to the central bank for signals about economic health, while policy makers look to asset prices for signals about financial conditions. The danger is an insular loop where both institutions end up reacting to reflections of each other rather than pricing the real-world physical and fiscal constraints shifting underneath them.
[VI] THE CROWDING-OUT COLLISION (CASE B VS. CASE C)
The transition of Case B from a cash-funded capex cycle to a debt-funded infrastructure boom introduces a critical point of friction: a direct collision between corporate tech borrowing (Case B) and sovereign Treasury borrowing (Case C).
When tech hyperscalers fund their investment out of internal cash flows, they operate independently of fixed-income supply dynamics. But when hyperscalers hit bond markets with $570 billion in annual public issuances — supplemented by $3.1 trillion in off-balance-sheet SPV commitments and leases — they are competing directly for the exact same marginal dollar of long-duration capital as the U.S. Treasury's multi-trillion-dollar deficit financing.
This dual-supply surge creates a compounding effect on long-term interest rates. As corporate AI debt and Treasury issuance flood bond markets simultaneously, long-end yields face structural upward pressure (the 10-year yield holding at 4.68%).
In turn, this crowding-out effect feeds directly back into Case A's vulnerability: while mega-cap tech and the sovereign can comfortably absorb long-term borrowing costs near the 4.68% baseline, lower-tier corporate borrowers (facing 10.31% CCC spreads) are effectively priced out of debt markets. The collision between Case B debt supply and Case C fiscal supply actively accelerates the balance-sheet fragility warned of by Case A.
[VII] CONCLUSION — THE DISAGREEMENT ITSELF IS THE SIGNAL
There is a version of this moment that has happened before. In the 1970s, a single dominant economic framework — the Keynesian consensus that had governed policy for three decades — broke apart under stagflation, a condition it had no coherent way to explain. What followed was not a quick handoff to a new consensus. It was a decade where monetarists, supply-siders, and Keynesian holdouts each had genuine partial evidence, each pointed to real data the others couldn't fully account for, and none of them were simply wrong. That period of open disagreement among serious frameworks wasn't a failure of economics. It was what a genuine regime transition looks like while it's still in progress, before enough evidence has accumulated for one explanation to visibly win.
The three-body problem playing out in the current yield curve may be the same kind of moment. That three rigorous, evidence-based camps can look at identical data and reach incompatible conclusions is not a sign that macroeconomics is broken. It's a sign that the underlying regime itself may be shifting, and the frameworks built for the old regime haven't yet been replaced by one built for the new one.
That reframes the task. The useful question isn't "which case is correct" — it's learning to notice, in real time, which piece of evidence actually discriminates between them. Most data releases are compatible with all three stories at once, which is exactly why the disagreement has persisted this long. The releases that matter are the rare ones that only one case can explain, and that the other two would have to actively contradict. A reader who trains themselves to ask that question of every new headline — does this fit one story and break the other two, or does it comfortably fit all three — will see the resolution coming well before it becomes obvious in hindsight, without needing to wait for anyone's dashboard to declare a winner.
The three forces are still pulling. But disagreement of this depth, among frameworks this rigorous, has historically been the leading edge of change, not a permanent stalemate. It doesn't last. Something will move first.
References & Telemetry Guide
Ambrosino, Ludovica, Jenny Chan, and Silvana Tenreyro - 2026 - Trade Fragmentation, Inflationary Pressures and Monetary Policy
Autor, David, David Dorn, Lawrence F. Katz, Christina Patterson, and John Van Reenen - 2020 - The Fall of the Labor Share and the Rise of Superstar Firms
Bank for International Settlements (BIS) - 2023 - International Banking and Financial Market Developments
Bank for International Settlements (BIS) - 2024 - Challenges to the Resilience of US Corporate Balance Sheets
Baqaee, David and Emmanuel Farhi - 2022 - Supply and Demand in Disaggregated Keynesian Economies
Bianchi, Francesco, Renato Faccini, and Leonardo Melosi - 2022 - A Fiscal Theory of Persistent Inflation
Bretscher, Lorenzo, Alice Hsu, and Andrea Tamoni - 2020 - Fiscal Policy Driven Bond Risk Premia
Briggs, Joseph and Devesh Kodnani (Goldman Sachs) - 2023 - Progress and Peril: The Economic Potential of AI
Bruno, Valentina and Hyun Song Shin - 2015 - Capital Regulation, Risk-Taking, and Monetary Policy Transmission
Congressional Budget Office (CBO) - 2026 - The Budget and Economic Outlook: 2026 to 2036 and Monthly Budget Reviews
Context Terminal - 2026 - Comparative Analysis Telemetry (April - August)
Copeland, Adam, Darrell Duffie, and Yilin Yang - 2021 - Reserves Were Not So Ample After All
D’Amico, Stefania and Thomas B. King - 2013 - Flow and Stock Effects of Large-Scale Treasury Purchases
Engstrom, Eric C. and Steven A. Sharpe - 2019 - Financial Market Perceptions of Recession Risk
Favara, Giovanni, Simon Gilchrist, Kurt F. Lewis, and Egon Zakrajšek - 2016 - Recession Risk and the Excess Bond Premium Calculation
Federal Reserve Bank of Chicago - 2026 - National Financial Conditions Index Time Series (NFCI)
Federal Reserve Bank of New York - 2022 - Understanding the Repo Market and Federal Reserve Tools
Federal Reserve Bank of New York - 2026 - Adrian-Crump-Moench (ACM) 10-Year Treasury Term Premium Model (THREEFYTP10)
Federal Reserve Bank of New York - 2026 - Money Market Reference Rates and Policy Facility Disclosures (SOFR, IORB, TGCR, Standing Repo Facility)
Federal Reserve Bank of New York - 2026 - Primary Dealer Statistics and Treasury Coupon Holdings Disclosures
Federal Reserve Bank of New York - 2026 - Weekly Foreign Official Overnight Reverse Repurchase Agreements Time Series (WLRRAFOIAL)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA AAA US Corporate Index Effective Yield (BAMLC0A0CMEY)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA CCC & Lower US High Yield Index Option-Adjusted Spread Daily and Weekly Series (BAMLH0A3HYC)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA Corporate & High Yield Effective Yield Series (BAMLC0A0CMEY, BAMLH0A0HYM2EY, BAMLH0A1HYBBEY, BAMLH0A3HYCEY)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA Single-B / BB US High Yield Index Option-Adjusted Spread Time Series (BAMLH0A1HYBB)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA US Corporate Index Investment Grade Option-Adjusted Spread Time Series (BAMLC0A0CM)
Federal Reserve Bank of St. Louis (FRED) - 2026 - ICE BofA US High Yield Index Option-Adjusted Spread Time Series (BAMLH0A0HYM2)
Federal Reserve Board - 2026 - Excess Bond Premium (EBP) Historical Data Series
Federal Reserve Board of Governors - 2026 - Financial Accounts of the United States: Nonfinancial Corporate Debt Outstanding (Z.1 Release, L.103 / L.213 / NCBCMDPMV)
Federal Reserve Board of Governors - 2026 - Gilchrist-Zakrajšek Credit Spread and Excess Bond Premium Dataset (FEDS Notes / ebp_data.csv)
Federal Reserve Board of Governors - 2026 - Selected Interest Rates: 5-Year, 5-Year Forward Expected Inflation Rate (H.15 Release / T5YIFR)
Federal Reserve Board of Governors - 2026 - Senior Loan Officer Opinion Survey on Bank Lending Practices (July 2026 SLOOS)
Gilchrist, Simon and Egon Zakrajšek - 2012 - Credit Spreads as Predictors of Real-Time Economic Activity
Greenwood, Robin and Samuel G. Hanson - 2013 - Issuer Quality and Corporate Bond Returns
Guynn, Randall D. and David Gross - 2022 - Bank Resolution and the Fed’s New Standing Repo Facility
Jiang, Zhengyang, Hanno Lustig, Stijn Van Nieuwerburgh, and Shuran Xiaolan - 2021 - The U.S. Public Debt Valuation Puzzle
Jungherr, Gerald, Matthias Meier, Timo Reinelt, and Immo Schott - 2022 - Corporate Debt Maturity Matters for Monetary Policy
Laubach, Thomas - 2009 - New Evidence on the Interest Rate Effects of Budget Deficits and Debt
Martin, Antoine, James McAndrews, and Alexandra Preslopsky - 2021 - Rewiring Repo: Standing Facilities and Market Structure
Michaillat, Pascal and Emmanuel Saez - 2022 - State Dependence of Fiscal Multipliers: The Source of Fluctuations Matters
Morgan Stanley Credit Research - 2026 - Global AI Debt and Shadow Financing Report
Morgan Stanley Research - 2024 - Artificial Intelligence – Bubble or Boom?
Nguyen, Toan - 2022 - Public Debt, Consumption Growth, and the Slope of the Term Structure
Nikolaidi, Maria - 2017 - Minsky's Financial Instability Hypothesis
Office of Financial Research (OFR) - 2026 - U.S. Financial Stress Index Time Series (OFR FSI)
Reinhart, Carmen M. and Kenneth S. Rogoff - 2009 - This Time Is Different: Eight Centuries of Financial Folly
S&P Global Market Intelligence - 2026 - Global Corporate Rated Debt Maturity Schedule (2025–2029 Debt Wall)
S&P Global Market Intelligence - 2026 - Technology Sector Credit Default Swaps (CDS) and Infrastructure Risk Disclosures
Securities Industry and Financial Markets Association (SIFMA) - 2026 - U.S. Corporate Bond Statistics: Monthly Issuance and Trading Volume
U.S. Bureau of Economic Analysis (BEA) - 2026 - NIPA Table 1.1.1: Real Gross Domestic Product and Implicit Price Deflator Quarterly Time Series (GDPC1, GDPDEF)
U.S. Bureau of Labor Statistics (BLS) - 2026 - Producer Price Index Industry and Commodity Databases (NAICS 3-Digit and 4-Digit PPI Flat Files)
U.S. Bureau of Labor Statistics (BLS) - 2026 - U.S. Import Price Index and PCE Services Less Food & Energy Series (IR, JCXFE)
U.S. Census Bureau - 2026 - Concentration Ratios for Manufacturing and Service Sector Industries (NAICS CR4 / CR20 Datasets)
U.S. Department of the Treasury - 2026 - Daily Treasury Statement and Operating Cash Balance Reports
U.S. Department of the Treasury - 2026 - Monthly Treasury Statement of Receipts and Outlays of the United States Government (MTS / FYONET)
U.S. Securities and Exchange Commission (SEC) - 2026 - Form 10-K and Form 10-Q Disclosures for Amazon, Alphabet, Microsoft, and Meta
Warsh, Kevin - 2026 - Remarks at the Federal Reserve Bank of Kansas City Economic Symposium (Jackson Hole, WY)