HY OAS: why this credit spread is a leading signal of recession and equity reversal

The HY OAS — published by FRED under code BAMLH0A0HYM2 — bundles three risk premia (perceived default, liquidity, aggregate risk aversion) that have historically reacted ahead of equities. Its sub-300-basis-point compression in late 2025 puts that canonical reading under interrogation.

A documented leading indicator across six stress episodes since 1997, the HY OAS now sits outside the historical correlation envelope. Three competing readings circulate in market literature; none of them clinches. This indicator belongs to the broader family of systemic-risk indicators and market-stress signals, a sub-pillar of the Eco3min market regimes, liquidity, real rates and financial dynamics pillar.

1. The HY OAS as a Composite Signal Built on Three Premia

The High Yield Option-Adjusted Spread is a daily measure of the yield premium investors demand to hold US corporate bonds rated BB+ and below, over matched-maturity Treasuries. The series is built by ICE Bond Indices (a subsidiary of Intercontinental Exchange since the 2017 acquisition of Bank of America Merrill Lynch Indices) and republished by the Federal Reserve Bank of St. Louis under the FRED code BAMLH0A0HYM2, adjusted for the embedded optionality of the underlying bonds (call, put, prepayment). The FRED ticker BAMLH0A0HYM2 definition and the methodological details are unpacked separately in the foundation satellite of this cluster ; the underlying BAMLH0A0HYM2 dataset provides the live series and a downloadable historical extract. What deserves analytical attention here is the composition of the spread as an object — because that composition determines what one can read from it.

1.1 Three premia bundled into a single number

When an investor demands 400 basis points of additional yield to hold a high-yield bond rather than a matched-maturity Treasury, that premium bundles three distinct elements responding to different economic drivers.

The perceived default premium compensates the investor for the market-implied probability, at a given point in time, that the issuer defaults on coupon or principal. This premium is not a measure of the realized default rate — it measures market expectations about that default rate over the life of the bond. It moves with corporate earnings expectations, activity indicators, reference single-name credit spreads (the Markit CDX HY index, in particular), and the perception of where in the economic cycle the economy stands. Eco3min develops this reasoning in the reading of credit spreads by regime. Over 1997-2025, this component accounts for somewhere between 60% and 80% of the variance of HY OAS, depending on which decomposition methodology is used by the Federal Reserve Bank of New York’s structural models (see Gilchrist-Zakrajšek 2012 on the EBP — Excess Bond Premium — and its annual updates).

The liquidity premium compensates for the implicit cost of difficulty selling the bond quickly without a price concession. It is generally modest in normal regimes — a few tens of basis points — but widens sharply during stress episodes, when market-makers reduce inventory risk. This component was particularly visible in March 2020 during the COVID shock: according to Bank for International Settlements estimates (Quarterly Review, June 2020), the liquidity premium accounted for as much as 40% of the initial HY OAS widening during the first two weeks of March, before Fed facilities (PMCCF, SMCCF) pulled this component back toward normal levels.

The aggregate risk aversion premium is the least directly observable but perhaps the most important for the macro reading of the spread. It captures the price markets put on aggregate risk, independent of default fundamentals or liquidity constraints. It is what makes HY OAS widen at times without any identifiable imminent default, during sentiment reversals. This component correlates historically with the VIX and realized equity volatility, which is precisely why the divergence with the VIX in 2024-2026 is treated in a dedicated satellite.

This three-premium decomposition is not a theoretical construct: it corresponds to estimation frameworks used by central banks to evaluate the informational content of credit spreads. The Federal Reserve has employed since 2012 the Gilchrist-Zakrajšek (GZ) decomposition, which separates the observed credit spread into a “fundamental” component (justified by issuer-level default risk characteristics) and a residual — the Excess Bond Premium (EBP) — interpreted as the aggregate risk aversion premium. The decomposition is updated monthly in Federal Reserve Notes and was revised after the COVID episode to account for the new liquidity dimension activated by the PMCCF/SMCCF programs. The fact that official central-bank frameworks perform this decomposition confirms that the three premia are not a stylized abstraction: they are distinct analytical objects whose separation requires specific modeling effort.

1.2 Why a multi-premium signal is more robust than a single-factor one

The analytical value of the HY OAS rests precisely on this multi-premium composition. A single-factor indicator — published default rate, VIX, 2s10s Treasury spread — captures one risk dimension but misses others. The Moody’s default rate is ex post: it documents what has already materialized, with a lag of several months. The VIX is forward-looking but purely on equity volatility, with no information on corporate credit deterioration. The Treasury spread embeds monetary-policy and growth expectations, but little information on corporate risk.

The HY OAS, by contrast, simultaneously combines a forward-looking dimension on defaults, a measure of liquidity stress, and a thermometer of aggregate risk aversion. That combination means it rarely moves for a single reason — and when it does move, the move already integrates several channels of information. That is what makes it a robust composite signal: a false positive on one of the three premia is generally contradicted by the other two. Conversely, when all three premia widen simultaneously, the signal is powerful.

1.3 Four historical anchor points

Over 1997-2025, the HY OAS has traded across an exceptionally wide range, with four reference points that structure any level-based reading. The 1997-2025 historical mean sits near 510 basis points (Eco3min calculations on the daily FRED series BAMLH0A0HYM2). The absolute peak recorded during the Global Financial Crisis of 2008-2009 touched roughly 2,020 basis points in November 2008, the most extreme level ever observed since the series began in 1997. The March 2020 COVID-shock peak reached roughly 1,100 basis points before the Fed’s massive intervention. Finally, the 2022 inflation-cycle peak stayed contained around 600 basis points, despite the violence of monetary tightening — a point that itself interrogates the spread’s reaction function to monetary shocks.

The current late-2025 level, around 300 basis points, places the HY OAS in the historical lower range: approximately 40% below the 1997-2025 mean, in the immediate neighborhood of levels observed in 2006-2007 (just before the GFC) and in 1997-1998 (just before the Asian crisis). That proximity is not an operational signal — it is an observable fact that the rest of this framework attempts to interpret.

2. The Credit-to-Equities Transmission: Empirical Observation and Mechanism

The editorial feature that distinguishes the HY OAS from other stress indicators is its documented lead time relative to equities. The empirical observation is solid; the mechanism less so, because it aggregates several distinct channels.

2.1 Six documented episodes since 1997

Empirical analysis of the major US equity reversals since 1997 shows that the HY OAS begins widening before the S&P 500 drawdown in nearly every case. The 1997-2025 chronology of HY OAS widening episodes details event by event: the Asian crisis of 1997-1998, the tech bubble of 2000-2002, the financial crisis of 2007-2009, the 2015-2016 commodity rout, the 2020 COVID shock, the 2022 Fed tightening cycle. Across these six episodes, the lag between the beginning of spread widening and the start of the equity drawdown ranges from several weeks to several months, with a historical median falling in a 4-to-8-week range depending on dating conventions. The case is built out in Eco3min’s panorama of historical market crises.

This empirical regularity across six heterogeneous episodes — external shock (Asia), valuation shock (dotcom), balance-sheet shock (GFC), commodity shock, health shock, monetary shock — is what makes the HY OAS robust as a leading indicator. It is not a correlation contingent on one type of crisis; it is behavior reproduced across very different economic configurations. This suggests that the underlying mechanism is not trigger-dependent but structural to credit-market microstructure.

A few episode-specific details illustrate this diversity. The 1998 episode saw the HY OAS cross 500 bps in August, about eight weeks before the intra-year S&P 500 trough in mid-October, in the wake of the Russian default and the LTCM collapse. The 2000-2002 episode saw the spread widen gradually from mid-2000, anticipating the March 2000 Nasdaq peak but more so the second leg down in 2001-2002, in a configuration where credit deterioration (notably among telecom and post-Enron energy issuers) preceded the broader earnings revision. The 2007-2009 episode is arguably the cleanest in lead-time terms: the HY OAS had begun widening as early as July 2007 (rising from 280 to 450 bps in a matter of months) while the S&P 500 did not peak until October 2007 — three months of lead time, followed by a massive equity drawdown into March 2009. By contrast, the 2020 COVID episode showed near-perfect synchronization between the two series: HY OAS and S&P 500 turned within days, which partially invalidates the canonical lead-time reading for that episode and has triggered analytical debate about the nature of the shock (exogenous, instantaneous, indiscriminate).

2.2 Lead time is not a timing rule

The analytical temptation is to convert this documented lead time into a timing rule: “when HY OAS widens by X basis points over Y weeks, the S&P 500 corrects within Z weeks.” That conversion does not hold empirically and will not be proposed here. The observed lead time varies from episode to episode, and the magnitude of spread widening needed to herald an equity reversal is not constant. In 2007, a modest widening (from roughly 250 to 350 bps) preceded a major S&P 500 drop; in 2015-2016, a sharper widening (from roughly 350 to 850 bps) was followed by a limited equity drawdown.

This variability does not invalidate the signal — it qualifies its nature. The HY OAS is a directional signal, not a magnitude or precise-timing signal. It indicates that deterioration is underway in the most cycle-sensitive corporate segment, which is macro-relevant information in itself, but cannot be translated into an operational decision rule. The empirical heterogeneity of the lead time also implies that an inflection of the spread should be cross-validated with concurrent indicators (default cycle, equity volatility regime, lending standards) before any analytical inference is drawn from it.

2.3 Three concrete transmission channels

The empirical observation holds because three distinct channels articulate the transmission from credit to equities. These channels are detailed in a dedicated satellite, but their synthesis is worth laying out here to understand why the mechanism exists — and why it could weaken in certain regimes.

The first channel operates through refinancing cost. HY issuers see their marginal financing cost rise as soon as the spread widens, before equity analysts have revised earnings forecasts. The cost increase mechanically pressures margins, and therefore valuations. A 200-bps move in spread translates, for an issuer with an average maturity of 5 years, into a direct 10-bp annualized increase in the interest burden per dollar of refinanced debt — an impact that diffuses to marginal EBIT over 2 to 4 quarters.

The second channel operates through bond covenants. Some HY bonds contain clauses that trigger at defined spread levels, forcing corrective action by the issuer (dividend cuts, asset sales, buyback suspensions). Those actions in turn weigh on equity valuations. The share of HY issuance with strict covenants declined over the 2010-2025 period (the “covenant-lite” segment rose from 25% to roughly 75% of primary issuance), which weakens this transmission channel over time — a point often missed in readings that lean on pre-2010 empirical regularities.

The third channel operates through equity-credit arbitrage run by certain hedge-fund structures, which treat an issuer’s equity and debt as two derivatives of the same underlying asset (the firm) and adjust exposure in mirror. This arbitrage was particularly active in the 2007-2008 and 2015-2016 episodes, where spread deterioration on specific names (Lehman, Bear Stearns in 2008; energy majors in 2015) preceded by several weeks the corresponding equity collapse. The transmission mechanism from credit to equities is unpacked in depth in a dedicated satellite of this cluster, where each channel is analyzed separately and the relative weights of the three channels are discussed regime by regime.

2.4 Why HY reacts before IG

A logical question arises: why HY OAS, and not IG OAS (investment-grade spread)? The answer lies in the differential sensitivity of the two segments to the default cycle. HY issuers have by construction more fragile balance sheets and a higher default probability. A marginal deterioration in the economic backdrop translates immediately into an upward revision of HY default probability, while it has only a second-order effect on the IG segment. This is why HY OAS typically moves 3 to 5 times more in amplitude than IG OAS over a given episode, and with several weeks’ lead time.

The distinction from the investment-grade spread is treated as a standalone object in a satellite of the cluster, because it constitutes a useful intermediate reading to qualify the nature of a spread move: an isolated HY widening points to segment-specific stress, while a combined HY + IG widening points to broader macro stress.

3. Reading the HY OAS by Regime: Compression, Widening, Peak

The absolute level of HY OAS does not say much taken in isolation. What matters analytically is to situate the current level within a grid of historical regimes, each with its own behavioral signature.

3.1 Compression regime (sub-400 basis points)

A HY OAS below 400 basis points reflects a regime of complacency or favorable macro absorption. Over the 1997-2025 series, this regime has been observed across three main windows: 1997-1998 before the Asian crisis, 2004-2007 before the GFC, and 2017-2018 and 2024-2026 more recently. The common feature of these windows is not a single economic state — the macro contexts were very different — but a simultaneous compression of all three premia: low default expectations, liquid bond market, low aggregate risk aversion.

This regime is analytically tricky because it is asymmetric in signal terms. When HY OAS is in durable compression, the downside is mathematically bounded (a spread cannot go negative without exceptional structural distortions), while the upside remains very large. This distributional asymmetry implies that the risk-reward ratio of “holding credit risk” deteriorates as compression persists — an empirical observation, not a positioning prescription.

3.2 Moderate widening regime (500 to 800 basis points)

A HY OAS between 500 and 800 basis points corresponds to the “average” regime, close to the 510-bp historical mean. This regime is compatible with moderate economic growth, an HY default cycle running at 3-5% annualized (Moody’s Trailing 12-month Speculative Grade Default Rate), and normal equity volatility. In this regime, the HY OAS is at its least informative in directional-signal terms: it says nothing abnormal is unfolding, but it does not point to a direction.

Exit from this regime to the upside — widening beyond 800 bps — historically constitutes the most actionable surveillance signal, without being a binary operational trigger. Over the six episodes 1997-2025, a sustained crossing of 800 bps coincided in five out of six cases with an NBER-dated recession ex post (the 2015-2016 case is the exception, the episode being qualified ex post as a slowdown without formal recession). The structural reasons explaining why credit spreads widen in recessions are unpacked separately in a top-of-funnel piece that complements this framework.

The 2015-2016 episode is worth detailing because it is analytically instructive about the limits of the 800-bp threshold. HY OAS then peaked around 887 bps in February 2016, driven principally by the energy segment (the ICE BofA US High Yield Energy index alone exceeded 1,700 bps at the peak) in the context of Brent crude collapsing below $30. That widening reflected concentrated sectoral deterioration, not systemic stress. The NBER never dated a recession around that episode, and the S&P 500 corrected temporarily before resuming its uptrend. The lesson is general: a crossing of 800 bps driven by a single sector does not carry the same meaning as a widening diffused across the HY universe, and any threshold-based reading of the aggregate spread should be cross-checked against the dispersion across sub-segments. Analysis of intra-segment dispersion (BB vs B vs CCC, and by sector) is a second critical reading of the aggregate signal, often more diagnostic than the headline level.

3.3 Peak regime (above 1,000 basis points)

A HY OAS above 1,000 basis points is a severe stress regime, observed across only two episodes since 1997: the GFC (peak ~2,020 bps in November 2008) and the COVID shock (peak ~1,100 bps in March 2020). This regime reflects systemic stress where all three premia of HY OAS widen simultaneously to historical extremes. The signature of the GFC is analyzed as a case study in a dedicated satellite: the anatomy of the 2008-2009 widening documents the sequence of phases (August 2007, March 2008 Bear Stearns, September 2008 Lehman, November 2008 peak) and allows separation between “normal” default-cycle widening and systemic-stress widening.

The analytical challenge of the peak regime is not so much to anticipate it — by definition it is almost always posterior to an identifiable shock — but to understand, once it has set in, which premia are moving, in what order, under what impulse. That fine decomposition is not directly accessible from the HY OAS level alone; it requires cross-referencing the spread with other signals (CDS, MOVE Index, FRA-OIS, offshore funding spreads) to identify the nature of the unfolding stress.

4. What the Spread Says, and What It Does Not

A meaningful share of market literature treats the HY OAS as a binary signal — widening = sell equities, compression = buy risk. That reading is analytically thin and empirically contradicted. Four partial readings deserve to be named in order to be set aside.

4.1 Confusing level with direction

A HY OAS at 300 basis points does not mean the same thing if it is coming from 250 (widening underway, weak but directional signal) or from 500 (compression underway, easing signal). The level taken in isolation ignores the dynamic information, which is precisely what makes HY OAS a leading signal: it is spread inflections that anticipate equity inflections, not absolute levels.

4.2 Confusing signal with threshold

There is no universal threshold above which HY OAS “signals” a recession or an equity reversal. The six episodes 1997-2025 show highly variable trigger thresholds depending on the starting regime and the nature of the shock. The 800-bp figure cited above is a statistical regularity, not a decision rule; it was crossed without a recession in 2015-2016, and the 2020 recession was declared as HY OAS crossed 800 bps within days, with no transition phase.

4.3 Ignoring the spread composition

The same HY OAS level can result from very different combinations of the three underlying premia. A compression to 300 bps driven by a simultaneous drop across all three premia does not say the same thing as a 300-bp compression driven by a collapse of the liquidity premium (rare but possible under massive QE) while the default premium holds. The signal is only readable rigorously if one attempts to decompose the spread, which requires structural models (NY Fed EBP, GZ models) that are not accessible in real time to most readers — hence the importance of triangulating with other series (HY CDS, Moody’s default rate, liquidity indices).

4.4 Ignoring the cyclical context

The predictive power of HY OAS on equities is conditional on the macro regime in which it sits. In active QE phases (post-2009, post-COVID), the credit-to-equities transmission was partially short-circuited by direct Fed absorption of risk assets. In Fed tightening phases (2022), transmission was amplified by simultaneous exit from both markets. Reading HY OAS without accounting for the monetary context and the cycle phase amounts to applying a signal grid to a signal whose transfer function has changed.

5. The Current Regime: Record Post-2022 Compression Under Two Paradoxes

In late 2025, the HY OAS published by ICE/BofA prints around 300 basis points — a level only observed across three short windows since 1997 (1997, 2006-2007, 2017-2018) and never sustained for long. This atypical compression deserves close examination because it coexists with dynamics that, taken individually, would argue for a wider spread.

5.1 Four rarely-compatible dynamics

First dynamic: the Federal Reserve has entered an easing cycle starting in late 2024, after the 2022-2023 tightening sequence. Historically, easing-cycle entries have coincided with contrasted HY OAS behavior — compression in some cases (the 2019 mini-cycle), widening in others (2007, where initial easing coincided with the start of credit deterioration). A clean compression of the spread during the initial phase of an easing cycle is not a standard case.

Second dynamic: the effective HY default rate, measured by the Moody’s Trailing 12-month Speculative Grade Default Rate, stands at around 4% in early 2026 — above the post-2000 historical mean of 3.5%, without being extreme. That default level indicates that credit risk has not vanished, contrary to what a 300-bp spread compression might suggest. This dynamic is mapped in our guide to frequent errors about credit and spreads.

Third dynamic: the 2026-2027 maturity wall in the US HY segment is documented by Bank of America Research (Maturity Wall Update, first quarter 2026) and JP Morgan Credit Strategy. Roughly $320 billion of HY debt matures in 2026-2027 and will need to be refinanced at market conditions that, depending on segment and signature quality, may differ substantially from the conditions of the original 2020-2021 issuance. This maturity wall is the classic Damocles sword of a deferred default cycle: if a meaningful share of refinancings fails or is done at elevated spreads, the realized default rate would mechanically accelerate.

Fourth dynamic: active geopolitical fragmentation, with effects on supply chains, corporate margins, inflation expectations, and capital flows. These effects are diffuse and do not feed mechanically into HY OAS, but they constitute a latent risk factor that, in theory, should justify a higher risk-aversion premium than the one implicit at 300 bps. The reshoring of supply chains imposes additional capital expenditure on a significant share of HY issuers in the consumer-discretionary and industrial sectors, while the bifurcation of trade and technology flows between major blocs introduces a structural uncertainty that traditional spread models do not capture cleanly.

The coexistence of these four dynamics with a record post-2022 spread compression is not contradictory — it is statistically atypical. The 1997-2025 series contains no prior episode in which a durable sub-300-bp compression coexisted with a documented maturity wall, an above-average default rate, and an initial easing cycle. The closest historical analogue — 2006-2007 — featured a similar compression but in a fundamentally different monetary configuration (terminal hikes, not initial easing) and without an identified upcoming maturity wall.

5.2 Three competing readings circulate in the literature

Three non-exclusive interpretations are advanced in the credit-strategy notes published since 2024.

The first is structural compression. The HY investor base has been profoundly reshaped over the past fifteen years, with the rise of two structurally asymmetric demand pockets: CLOs (Collateralized Loan Obligations), whose purchases are driven by portfolio-construction rules more than by cyclical reading, and retail HY ETFs, whose flows largely track equity-market momentum and retail yield-seeking. This structural demand, independent of fundamental credit risk reading, exerts permanent downward pressure on spreads and could explain why HY OAS remains compressed despite fundamentals that would justify a higher level.

A few orders of magnitude anchor this reading. According to SIFMA and JP Morgan CLO Research data, the US CLO market reaches roughly $1.1 trillion in outstanding volume in 2025, up from $350 billion in 2014 — a tripling in ten years that makes CLOs the largest buying pocket of US leveraged loans. While CLOs invest primarily in loans rather than HY bonds directly, arbitrage between the two markets transmits buying pressure to HY spreads (issuers who can choose between funding vehicles tilt toward the cheaper one, aligning spreads by arbitrage). On the ETF side, combined assets in HYG (iShares iBoxx High Yield Corporate Bond) and JNK (SPDR Bloomberg High Yield Bond) exceed $35 billion in 2025, with a growing share held by retail investors via retail platforms. These ETF flows are procyclical by construction and amplify compression in upward regimes — a mechanism documented in academic work (Ben-David, Franzoni, Moussawi 2017 on ETF liquidity in illiquid underlyings).

The second reading is macro absorption. Under this interpretation, US corporate balance sheets have improved meaningfully since 2020 — lower leverage ratios, longer maturities, higher cash piles. That structural fundamental improvement would empirically justify a lower default premium, and therefore a more compressed HY OAS. In this reading, the current compression is not an anomaly; it reflects a modified fundamental reality.

That reading is supported by several fundamental series. Median net leverage (Net Debt / EBITDA) for issuers in the ICE BofA US High Yield index has dropped from around 4.5x in 2019 to about 3.8x in 2025, according to Bank of America Credit Research — an improvement of about 15% that is consistent with a lower forward default premium. The weighted-average debt maturity has also lengthened: roughly 5.4 years in 2025 versus 4.8 years in 2018, which reduces short-term refinancing risk and therefore the residual liquidity premium. Finally, interest-coverage ratios (EBITDA / Interest Expense) remain above their cycle average at 4.2x median (Moody’s Corporate Debt Update, Q4 2025), despite the rise in nominal rates.

The analytical limit of this reading is that it looks at the aggregate and hides dispersion. The median can improve while the tail of the distribution (CCC issuers, recent LBOs) deteriorates — a pattern consistent with a default cycle concentrated on a smaller fraction of the index, compatible with an aggregate spread compression. Distinguishing a broad improvement from a median improvement masking a deteriorated tail requires granular data that public sources provide only partially.

The third reading is cyclical complacency. This is the 2024-2026 analogue of the diagnosis applied ex post to 2006-2007: the market systematically underprices a risk it cannot yet identify, through excess confidence in the Fed reaction function, in debt absorption, or simply through behavioral inertia. This reading is by construction uncomfortable because it is never verifiable ex ante — only ex post resolves it, and ex post has not arrived.

The 2006-2007 parallel deserves careful articulation, because it is analytically fragile if mishandled. On a pure statistical plane, the HY OAS in October 2006 (~265 bps) and in late 2025 (~300 bps) sit in the same low historical range, and several sentiment indicators show similarities (realized-volatility compression, narrow intra-segment dispersion, compressed equity risk premia). But the macro context differs significantly: in 2006, the Fed hiking cycle was near its end and the market was beginning to price easing; in 2025-2026, the easing cycle is underway but incomplete. In 2006, corporate balance sheets were on average less robust than today in net-leverage terms. The “cyclical complacency” reading is therefore not a copy-paste of 2006-2007 — it points toward a potentially comparable behavioral dynamic, in a partially different fundamental frame. A closely related framework: systemic Risk Indicators and Financial Market Stress Signals. A closely related framework: systemic Risk Indicators and Financial Market Stress Signals. A closely related framework: systemic Risk Indicators and Financial Market Stress Signals.

5.3 The three readings are not mutually exclusive

An analytical particularity is worth noting: these three readings can partially coexist. Part of the compression may be structural (CLO + ETF), part macro (improved balance sheets), and part cyclical (complacency). Identifying the respective share of each driver in the observed 300-bp compression is not accessible from publicly available data at the present moment; it would require structural models integrating flows by investor category, which neither FRED nor public sources provide.

The honest analytical reading consists in recognizing that the current compression is over-determined — several credible explanations coexist — and in monitoring the variables that would allow regime distinction over time: refinancing absorption rate for the 2026-2027 maturity wall, evolution of the Moody’s default rate, intra-HY-segment dispersion (a widening gap between BB and CCC would suggest aggregate compression masks deterioration of lower-quality debt). The record compression of the spread since 2023 is treated in depth in the dedicated satellite on the current regime.

Three granular observables deserve particular attention in the coming quarters. The first is dispersion by rating: the spread differential between BB-rated and CCC-rated bonds has oscillated between 400 and 1,200 basis points over 2010-2025, with sharp widenings at every entry into a stress cycle. A BB-CCC gap that remains contained (300-400 bps) despite observable deterioration of CCC-issuer fundamentals would be consistent with the “structural compression” reading; a gap that widens rapidly at stable aggregate HY OAS would suggest the tail of the distribution is deteriorating and that the median masks the deterioration. The second observable is the 2026-2027 maturity-wall absorption rate: if issuers refinance without difficulty at spreads close to original issuance level, the macro-absorption reading is supported; if a meaningful share of refinancings is deferred, extended, or done at substantially higher spreads, the cyclical reading gains consistency. The third observable is the spread’s reaction to minor shocks: if HY OAS continues to absorb shocks without widening beyond a few tens of basis points, the dominant-structural-demand hypothesis remains valid; if a compression break occurs on a shock of moderate magnitude, that is the most actionable signal of regime transition.

6. What This Configuration Challenges in the Canonical Reading

If the current compression is partly structural, then the HY OAS’s signal function as a leading indicator is in the process of being modified. That does not mean the signal disappears, but that its translation into a macro reading requires a calibration different from the one inherited from the six episodes 1997-2025.

6.1 The gap between historical and contemporary regime

The canonical reading of the HY OAS rested on an implicit assumption: that the investor base and the market structure are stationary over the observation period, or at least that structural changes do not affect the spread’s sensitivity to fundamentals. That assumption is more fragile in 2024-2026 than it was in 2007 or in 2015 — not because the HY segment composition has changed radically, but because the investor base has polarized between structural pockets (CLO, ETF) and active pockets, which modifies the price function.

Concretely, this suggests that the widening magnitude required to signal stress equivalent to a historical episode could be lower today than it was in earlier episodes. An acceleration from 300 to 450 bps within a few weeks, which would have been qualified as concerning but not alarming in 2007, could today constitute a more powerful signal — because structural compression normally absorbs minor shocks, so a break in the compression signals a regime change, not background noise.

This implicit recalibration of the reading scale is not a rewriting of historical thresholds; it is a qualification of the signal as a function of the investor regime. A HY OAS moving from 300 to 450 bps today must be read while keeping in mind that structural buyers (CLO, ETF) continue to provide a demand floor, and that the widening therefore signals either a behavioral break in those buyers (cessation of retail ETF flows, for example) or the emergence of a risk factor strong enough to overcome that structural demand. In both cases, the signal carries more informational density than the same magnitude would have in a regime without structural pockets.

A practical consequence follows. The traditional surveillance focus on the 800-bp threshold, inherited from the five-out-of-six historical episodes, may be miscalibrated for the contemporary regime. Earlier inflection points — for example a clean break of 400 bps to the upside, or a meaningful divergence between aggregate HY OAS and CCC-only spreads — could carry more diagnostic value today than the canonical 800-bp watch level. This is not to dismiss the canonical reading, which remains the right framework for severe stress regimes; it is to recognize that the spread’s information content has shifted in compressed regimes, and that surveillance must adapt accordingly.

6.2 The divergence with the VIX in 2024-2026 as signal test

The unusual divergence between HY OAS and VIX since late 2023 provides a partial test of the three competing readings. If HY OAS compression were fully justified by fundamentals (macro reading), one would expect to see the VIX also compressed. If compression were purely structural (CLO + ETF), one would expect to see a divergence with a VIX that continues to capture equity risk without credit relay. If compression were cyclical (complacency), one would expect to see both series compressed without proportional reaction to shocks.

The observed reality — durable HY OAS compression, intermittent VIX around 15-25 — is more consistent with a “structural + cyclical” combination than with a full macro justification. But this inference remains circumstantial; it is not a verdict.

7. What the HY OAS Says in 2026, Without Absolutizing It

At the close of this reading, several observations can be set down without extrapolation. The HY OAS remains a robust composite signal, whose analytical function rests on the combination of three distinct premia. Its empirical documentation as a leading indicator of the S&P 500 across six heterogeneous episodes is solid and not subject to major revision. Its contemporary reading, however, cannot rely on mechanically applying the grid inherited from the 1997-2025 episodes: the investor base has changed, which modifies the spread’s sensitivity to fundamentals and therefore the translation of the signal into a macro reading.

The current compression around 300 basis points is statistically atypical given concurrent fundamentals (initial easing cycle, moderate default rate, 2026-2027 maturity wall, geopolitical fragmentation). Three credible interpretations circulate, none mutually exclusive, and no publicly available data allows ex ante arbitration between them. The variables to monitor for over-time resolution are identifiable: refinancing absorption rate, intra-HY-segment dispersion between BB and CCC, evolution of the Moody’s default rate, comparative behavior of HY OAS and IG OAS. These observables provide a surveillance framework that does not rely on the canonical thresholds, which may be miscalibrated for the current structural regime.

A final point of analytical honesty: while the HY OAS remains the most-watched single indicator of US corporate credit risk, no single series — however composite — should carry the full weight of a macro thesis. The robustness of the signal across 1997-2025 was reinforced by its concordance with parallel indicators (yield curve, lending standards, default cycle). The same triangulation discipline applies today, particularly in a regime where each signal’s fundamental sensitivity has potentially shifted. Reading the HY OAS without absolutizing it means keeping the spread in the conversation as one of several voices, not as the verdict.

The editorial role of HY OAS as a leading indicator is neither invalidated nor confirmed by the 2024-2026 configuration. It is under active interrogation, and that is probably the most honest status one can assign to it at the present moment. The broader question — do credit spreads predict recessions — receives a more general treatment in a top-of-funnel companion piece, which usefully complements the cluster-level reading developed here.

Last updated — 18 June 2026