In the final weeks of 2021, the option-adjusted spread on the ICE BofA US High Yield Index reached 283 basis points [1], its tightest level since before the global financial crisis, and a figure that, in the context of the preceding two decades, could only be understood as the product of conditions that no prudent investor should have expected to persist. The Federal Reserve had maintained its policy rate at the zero lower bound since March 2020. The European Central Bank had done the same, its deposit rate locked at negative forty basis points, its asset purchase programme absorbing sovereign and corporate paper at a pace that would have been inconceivable a generation earlier. In this environment, the spread between what a highly leveraged American corporation paid to borrow and what the United States government paid for the same privilege had compressed, over the preceding five years, to a level that implied either a dramatic improvement in corporate credit quality or a dramatic mispricing of credit risk. The evidence, on reflection, pointed rather more firmly to the latter.
The temptation to treat conditions of this kind as permanent is not new, nor is it irrational. Investors respond to the incentive structures before them, and the incentive structure of the zero-rate era rewarded the assumption of duration and credit risk with a consistency that, sustained over five years, began to resemble a structural feature of the market rather than a policy choice subject to revision. What the present moment requires is not a retrospective indictment of those who positioned accordingly, but an honest assessment of the three structural changes that have now combined to make the pre-2022 credit paradigm unrepeatable, and of what their aggregate consequence means for the management of corporate credit risk in the years that follow.
The Lesson of 1994
The most instructive historical parallel is not the global financial crisis of 2008, which was a solvency event of a particular and unusual kind, but the bond market episode of 1994, which was a duration and liquidity event of the sort that the corporate credit market is now, more slowly, revisiting. In the early months of that year, American credit markets were characterised by a comparable complacency. The Federal Reserve had maintained its funds rate at three per cent since September 1992, and a substantial proportion of institutional investors had positioned accordingly, financing long-dated securities with short-term borrowings in a carry trade that was profitable precisely because the yield curve had been stable for long enough to seem immovable. When the Federal Open Market Committee raised its target by twenty-five basis points on the fourth of February 1994, the reaction was violent and disproportionate to the size of the initial move. By November, the funds rate stood at five and a half per cent, the ten-year Treasury yield had climbed from five point three to eight per cent, and the aggregate loss to holders of American bonds was estimated at five hundred billion dollars, with a further five hundred billion attributable to markets elsewhere in the developed world [2]. The episode has been described, with some justice, as the Great Bond Massacre.
What that episode teaches is not simply that rates can move unexpectedly, a proposition requiring no special instruction, but that the velocity of rate change, acting upon a market structured for stability, produces non-linear consequences that no individual participant's duration model incorporates. The losses were not, in the main, the product of outright default. They were the product of repricing: the mechanical consequence of carrying positions financed at short-term rates that rose faster than the income generated by the long-dated assets they supported. The corporate credit market of 2025 confronts a version of this problem that is structurally different in its mechanism but recognisably similar in its underlying character: a large stock of debt, issued at the interest rates of the zero-rate era, that must now be serviced, and eventually refinanced, at rates that are materially and, on present evidence, durably higher.
A Market Built on Three Assumptions
The corporate credit market of 2015 to 2022 was not merely exposed to the risk of rising rates. It was constructed upon three assumptions that, taken together, created a fragility whose full dimensions are now becoming visible. The first was the permanence of central bank support: the conviction, validated by the interventions of 2008, of 2011 to 2012, and of 2020, that any serious deterioration in credit conditions would be met by monetary accommodation sufficient to prevent a cascade of defaults. This conviction was not irrational; it was grounded in observable central bank behaviour across three successive crises. It was, nonetheless, the kind of conviction that Hyman Minsky identified as the engine of financial instability: the belief, generated by a sustained period of stability, that the stability itself is a feature of the architecture rather than a policy choice that can be, and eventually is, withdrawn [3]. Minsky's taxonomy of financial positions, which distinguishes between hedge finance, in which income covers both interest and principal; speculative finance, in which income covers only interest; and Ponzi finance, in which it covers neither, describes not merely individual borrowers but the aggregate condition of a credit market that has been organised, for a decade, around the assumption that cheap refinancing will always be available at the moment it is needed.
The second assumption was that debt issued at the interest rates prevailing between 2015 and 2021 would not require refinancing at materially higher rates within the relevant investment horizon. This assumption has encountered reality in the form of what market participants have taken to describing as the maturity wall: the concentration of corporate and leveraged debt maturities between 2026 and 2029. According to data compiled from leveraged loan and high-yield bond indices, approximately $1.2 trillion in levered credit matures during this window [4], comprising some $580 billion in leveraged loans and $625 billion in high-yield bonds, with overall corporate debt maturities, across all rating categories, expected to rise from approximately $2 trillion in 2024 to approximately $3 trillion by 2026 [4]. The borrowers who issued this paper at coupons of four to six per cent are now confronting a rate environment that, following the 525 basis point Federal Reserve tightening cycle of 2022 to 2023, imposes coupons in the range of eight to ten per cent on comparable risk. The aggregate incremental interest burden across the universe of affected issuers is not modest. It falls disproportionately on the borrowers least equipped to absorb it: those whose revenue generation is already insufficient to service existing obligations at existing coupons.
The consequence of this second failure is visible in the data on zombie firms, defined by the Bank for International Settlements as companies whose interest coverage ratio, the ratio of earnings to interest expense, remains below one for three or more consecutive years. A BIS working paper analysis found that their share of the listed corporate population in advanced economies rose from approximately two per cent in the late 1980s to some twelve per cent by 2016 under a broad measure [5], a rise attributable in part to the sustained decline in interest rates over that period, which mechanically improved coverage ratios and reduced creditor pressure on chronically unprofitable borrowers. Rising rates operate the mirror image of this dynamic: a proportion of currently non-zombie firms will migrate into the category as their fixed-rate debt, issued at the rates of 2018 to 2022, rolls over into the current rate environment. The size of that migration is not yet fully visible in default statistics, which have remained, through 2024, below their long-run averages. It will become more visible in the 2026 to 2028 refinancing cycle.
As the accompanying chart illustrates, the decade since 2015 resolves into two distinct regimes, separated by the rate shock of 2022. In the first, spreads compress with minor interruptions, the brief 2018 widening and the more dramatic but rapidly reversed COVID spike of March 2020, into the extraordinary tightness of 2021. In the second, which begins with the Federal Reserve's first rate increase in March 2022, spreads widen, partially retrace, and then tighten again, but the underlying conditions, an elevated policy rate, a maturing debt stock, a central bank no longer committed to asset purchases, have not reverted. The chart does not show permanent spread widening. It shows permanent uncertainty: a market in which the variance of spread outcomes is structurally higher than it was between 2015 and 2021, and in which the assumption that any episode of widening will be met by policy accommodation is no longer available as an anchor.
The Third Assumption: Passive as a Stabiliser
The third structural change embedded in the pre-2022 credit market is the one most resistant to quantification, and the one whose consequences are most asymmetric in their distribution between benign and stress scenarios. The growth of passive investment vehicles in corporate debt markets has been substantial: fixed-income ETFs in the United States had accumulated approximately $1.7 trillion in assets under management by the middle of 2024 [6], representing a transformation of the institutional structure of corporate bond ownership that has no clear precedent in the history of these markets. The conventional analysis of this development emphasises the benefits: lower transaction costs, democratised access to credit exposure, and the structural demand from ETF inflows that has contributed to the compression of spreads across the quality spectrum. These benefits are real. What the conventional analysis addresses less candidly is the mechanism by which that structural demand operates when its direction reverses.
A bond ETF, unlike an actively managed credit fund, cannot exercise credit judgment. It holds the index. When its holders redeem, it sells; when the market is disorderly, it sells into the disorder. Academic research published in the Financial Review has documented that net outflows from investment-grade bond ETFs increase the volatility of their underlying holdings through the creation and redemption mechanism, as authorised participants arbitraging the gap between ETF prices and net asset values are forced to dispose of securities at prices that the underlying market is not, in conditions of stress, prepared to absorb without significant price impact [6]. This mechanism was observed with unusual clarity during the credit market dislocation of March 2020, when investment-grade bond ETFs traded at discounts to their net asset values implying prices for their underlying holdings that were inconsistent with any observable transaction in those securities. The Federal Reserve's announcement of a corporate bond purchase programme on the twenty-third of March resolved that dislocation within days. The key variable for the next period of stress is whether any equivalent intervention is available, given that the central bank has now established, through the 2022 tightening cycle, that it will not deploy its balance sheet in defence of credit conditions when inflation remains elevated.
The mechanism that contributed to the compression of spreads during the inflow phase is the same mechanism that will contribute to their widening during the outflow phase. A passive vehicle cannot change its mind.
The interaction between passive concentration and the maturity wall is the element of this analysis most difficult to model and most consequential if the modelling proves inadequate. A market in which passive vehicles represent a substantial fraction of marginal demand is a market in which price discovery, in conditions of stress, depends on the behaviour of a relatively small number of authorised participants whose capacity to absorb supply is finite and whose willingness to do so is determined by their own balance sheet constraints. This is not, in itself, a novel observation; it applies, in varying degrees, to any market in which a significant fraction of participants are not engaged in active price formation. What makes the current configuration unusual is the combination of its scale, the $1.7 trillion in fixed-income ETF assets, with the particular timing of the maturity wall, which concentrates the refinancing challenge in the same three-year window, 2026 to 2029, during which the Federal Reserve has signalled only a gradual further reduction in policy rates.
The Counter-Case, Taken Seriously
The objection to this analysis, which deserves to be taken seriously rather than dismissed, is that the credit market has already demonstrated a capacity for orderly adjustment that the pessimists of 2022 did not anticipate. High-yield default rates in 2024 remained below their long-run averages [4]. The maturity wall was visible as early as 2022, and in the intervening period a substantial proportion of the most exposed issuers have refinanced opportunistically, extending maturities and accepting higher coupons rather than awaiting the wall's arrival. Spreads in late 2024 compressed to levels approaching the historic tightness of 2021, suggesting that credit investors remained willing to extend capital on terms that implied confidence in issuers' ability to service higher-coupon debt from revenues that had, in many cases, grown through the post-pandemic recovery. These are not trivial observations. A market with $1.7 trillion in passive structural demand is a market with inflows capable of absorbing supply at tight spreads for longer than valuation-based analysis suggests.
What this counter-case does not resolve is the asymmetry of the outcome distribution. The base case, in which issuers refinance smoothly, the economy avoids recession, and default rates remain contained within historical ranges, is the most probable outcome. The tail case, in which a deterioration in economic conditions coincides with the peak of the refinancing cycle in 2027 or 2028, is less probable but carries consequences that passive vehicles, by their nature, are not designed to absorb without amplifying. The problem is not that the central scenario is wrong. The problem is that the distribution of possible outcomes around that scenario is substantially wider than the current level of spreads implies, and that the instruments through which most institutional investors now hold credit exposure are precisely those least adapted to navigating a wide outcome distribution.
What the Data Will Tell Us
Three variables will determine, over the next three years, whether the repricing of corporate credit proves to have been a manageable transition or something more consequential. The first is the trajectory of interest coverage ratios in the leveraged loan universe between 2026 and 2028, as fixed-rate bonds mature and variable-rate loans reprice against a base rate that, on current Federal Reserve projections, will decline only gradually. A sustained deterioration in aggregate coverage ratios, measured across the universe of issuers represented in leveraged loan indices rather than merely among current defaulters, would be the clearest early signal that the zombie migration described above is proceeding on a scale that default statistics are not yet capturing. The second variable is the behaviour of passive ETF flows under any period of sustained credit deterioration: whether the redemption dynamics identified in the academic literature manifest at sufficient scale to produce the price amplification that the March 2020 episode foreshadowed, or whether the depth of the buyer base proves sufficient to absorb supply without the discontinuities that would force forced-sellers into an illiquid market. The third, and most consequential, is the degree of central bank flexibility: whether any future deterioration in credit conditions occurs in an environment in which inflation has sufficiently retreated to permit the kind of balance sheet deployment that resolved the 2020 dislocation, or whether, as in the very different circumstances of 1994, the monetary authorities find themselves unable to provide the accommodation that credit markets have, for a decade and a half, expected to be available on demand.
Minsky's central insight was that the financial system does not tend, over time, toward equilibrium. It tends, instead, toward the accumulation of fragility during periods of stability, and toward the revelation of that fragility when stability is disturbed [3]. The corporate credit market of 2026 is not in crisis. Its spreads are tight, its default rates are contained, and its passive vehicles continue to attract inflows. It is also, in the three dimensions examined here, structurally less resilient to disturbance than it was a decade ago, and it is confronting, in the maturity wall of 2027 and 2028, a scheduled disturbance of known magnitude at an uncertain point in the economic cycle. Whether that disturbance proves manageable depends less on the credit quality of any individual issuer than on the behaviour of a system whose architecture has been transformed, in ways that no participant designed and no regulator fully anticipated, by a decade of rates that were, as the events of 2022 confirmed, neither structurally permanent nor, once adjusted, rapidly reversible. The 1994 bond market massacre was forgotten within eighteen months as the carry trades that it had destroyed were rebuilt at the new rate level. The maturity wall does not afford the same luxury of forgetting; it is a scheduled appointment, and it is approaching.
- Federal Reserve Economic Data (FRED), Federal Reserve Bank of St. Louis. "ICE BofA US High Yield Index Option-Adjusted Spread (BAMLH0A0HYM2)." FRED Economic Data. Updated daily. fred.stlouisfed.org; "ICE BofA US Corporate Index Option-Adjusted Spread (BAMLC0A0CM)." fred.stlouisfed.org
- Wikipedia contributors. "1994 bond market crisis." Wikipedia, The Free Encyclopedia. Accessed March 2026. en.wikipedia.org; see also: Tooze, Adam. "Chartbook Newsletter #14: Are bond vigilantes real? The strange case of the 1994 bond market massacre." adamtooze.com. 28 February 2021. adamtooze.com
- Minsky, Hyman P. "The Financial Instability Hypothesis." Levy Economics Institute Working Paper No. 74. May 1992. levyinstitute.org
- S&P Global Ratings. "Global Refinancing: Pressures Linger For The Lowest-Rated Credit." S&P Global. 2024. spglobal.com; PitchBook LCD. "2026 US High-Yield Outlook: Volume to tick higher amid looming maturity wall." PitchBook. 2026. pitchbook.com
- Banerjee, Ryan and Boris Hofmann. "The rise of zombie firms: causes and consequences." BIS Quarterly Review. September 2018. bis.org; Bank for International Settlements. "Corporate zombies: Anatomy and life cycle." BIS Working Papers No. 882. October 2020. bis.org
- Agapova, Anna. "ETFs and the price volatility of underlying bonds." Financial Review. Wiley Online Library. 2025. onlinelibrary.wiley.com; VanEck. "2025 Corporate Bond Market Trends: An Investor's Guide." VanEck. 2025. vaneck.com