When Liquidity Becomes Structure

Abstract

In March 2020, the market for the world’s benchmark safe asset stopped behaving like one. U.S. Treasuries did not become credit-impaired. They became difficult to intermediate at precisely the moment investors wanted cash. Dealers absorbed what they could; the Federal Reserve ultimately supplied the balance sheet the market could not. The episode exposed a distinction that conventional measures of liquidity often miss: an asset can be liquid in ordinary trading and still become destabilising when the structure holding it is forced to move faster than the market can absorb.

This paper begins from that distinction. Liquidity is not only a property of an asset. It is also a property of time: the interval between a demand for cash and the ability of a balance sheet to meet it without forced transactions. Redemptions, margin calls, deposit withdrawals and funding maturities shorten that interval. Patient capital, usable collateral, deep intermediation and official backstops lengthen it. The relevant question is therefore not simply can the asset be sold? It is can it be sold, financed or held long enough for price discovery to remain orderly?

The argument is tested against five distinct settings: the March 2020 dislocation in U.S. Treasury markets; the UK liability-driven investment episode of 2022; the 2022-23 Credit Suisse run; the rapid expansion and 2026 redemption pressure in parts of private credit; and the concentrated fund stress observed in Türkiye in September 2026. The cases are intentionally different. Their value lies not in claiming equivalence, but in asking whether common structural variables recur: a loss or confidence shock creates cash demand; that demand forces portfolio or funding adjustment; and the outcome depends on leverage, liability speed, market depth, intermediation capacity and the availability of patient capital or official backstops. The literature provides strong support for these mechanisms, but the evidence is not one-directional. Traditional private-credit vehicles are still largely funded by long-duration capital; post-crisis bank liquidity buffers matter; and localised stress can remain contained. The thesis is therefore conditional rather than alarmist.

Independent research. Views and data are stated as of the research cut-off except where a later revision is explicitly marked. The paper is analytical and informational; it is not investment, legal or tax advice.

Citation convention

The paper uses author-date citations in the text. Current observations are attributed to the original public agency or regulator wherever practicable. Chapra & Co. calculations and conceptual exhibits are labelled as such. Historical comparisons are used to isolate mechanisms rather than to inherit chronology.

01.

The moment choice becomes compulsion

Markets do not usually become disorderly merely because somebody has lost money. They become disorderly when somebody who has lost money also loses the ability to wait. An unlevered holder with patient capital can sit through a mark-to-market decline. A leveraged fund facing variation margin cannot. A bank losing deposits cannot. An open-ended vehicle meeting redemptions cannot. The common transition is not from profit to loss. It is from choice to compulsion.

Financial economics has described the pieces of this mechanism for decades. Brunnermeier and Pedersen (2009) connect funding liquidity to market liquidity. Shleifer and Vishny (2011) show why forced sellers can push prices away from fundamentals when natural buyers are constrained. He and Krishnamurthy (2013) put intermediary capital at the centre of crisis pricing; Duffie (2010) explains why apparently attractive capital cannot always arrive quickly enough. Read together, the literature points to a simple proposition: the timing of the cash obligation matters as much as the eventual value of the asset.

That does not make leverage inherently unstable, nor every redemption a run. It establishes a boundary condition. A structure becomes fragile when the clock governing cash demand runs faster than the clock governing asset sales, refinancing or external support. The same economic loss can therefore remain a valuation event in one portfolio and become a liquidity event in another.

02.

Six propositions the evidence must survive

The propositions are deliberately demanding. The paper does not claim that a Treasury dislocation, a pension collateral spiral, a bank run and a gated private-credit vehicle are the same event. It asks whether the same structural variables — liability speed, leverage, forced cash demand and external balance-sheet capacity — help explain why very different institutions can move from loss to disorder.

Evidence status. Throughout the paper, an established mechanism is one supported by peer-reviewed theory or empirical work; case-consistent evidence means an episode fits that mechanism without independently identifying causality; live or unresolved evidence is contemporaneous and incomplete; counterevidence records observations that limit the thesis. The selected cases are mechanism tests, not a ranking of crisis severity and not an assertion that institutionally different structures are equivalent.

03.

Core markets and the changing intermediary balance sheet

The argument matters more because the ownership of core market risk is changing. Between 2021 and 2025, the BIS estimates that non-bank financial institutions increased their share of advanced-economy sovereign debt from 44 to 53 percent. Domestic central-bank holdings moved the other way, from 27 to 17 percent; the foreign official share also declined. Nothing in those numbers proves fragility. They do, however, change who must provide the next unit of balance sheet when volatility rises.

The U.S. Treasury market provides a concrete illustration. Federal Reserve Form PF analysis published in June 2026 estimates that large hedge funds' gross U.S. Treasury exposure reached about $4.0 trillion by September 2025, with $2.4 trillion of long exposure and $1.6 trillion of short exposure. Repo cash borrowing reached about $3.0 trillion. The same analysis estimates the cash-futures basis trade at approximately $830 billion, about double its early-2020 peak. Treasury activity is also concentrated: the 50 largest funds accounted for approximately 90 percent of gross Treasury exposures, up from 84 percent at the beginning of 2023 (Monin, 2026). These strategies can improve price alignment and liquidity in normal conditions. Their systemic relevance comes from scale, concentration and funding dependence rather than from arbitrage itself.

The point also extends the mechanism developed in Chapra Research No. 02. That paper argued that sovereign financing reaches private discount rates through the balance sheets that absorb and finance duration. The current paper narrows the lens: if those balance sheets are more leveraged, more funding-dependent or more concentrated, the liquidity characteristics of the market can change even when the underlying security remains a sovereign obligation.

04.

March 2020: when the safe asset needed liquidity

March 2020 is the cleanest rebuttal to the idea that liquidity travels permanently with the asset. U.S. Treasuries are among the world's deepest markets, yet the Federal Reserve documented severe dislocations as a broad range of investors sought cash. Bid-ask spreads widened, market depth fell and some dealers reached or approached their willingness or capacity to absorb further sales (Federal Reserve Board, 2020).

The episode was not a single-cause event. Foreign investors sold more than $400 billion of Treasuries during March; some relative-value hedge funds reduced positions; mortgage REITs faced margin calls; mutual funds and other investors also sought liquidity. The Fed's own retrospective cautions that the evidence for hedge-fund deleveraging as the primary driver is weak. Peer-reviewed evidence from the broader debt-market disruption points in the same direction: Haddad, Moreira and Muir (2021) identify acute liquidity needs among bond investors as a central mechanism, while Kargar et al. (2021) document the collapse in corporate-bond market liquidity. The case therefore supports a broad cash-demand and intermediation mechanism, not a monocausal hedge-fund story.

The decisive evidence lies in what happened next. Between 9 and 17 March the Federal Reserve expanded repo operations; on 15 March it authorised at least $500 billion of Treasury purchases and $200 billion of agency MBS purchases; on 23 March it removed quantity limits for purchases needed to restore market functioning; and it subsequently expanded dealer and foreign-official liquidity facilities. The interventions absorbed securities that would otherwise have remained on dealer balance sheets and coincided with improved market functioning (Federal Reserve Board, 2020).

The striking feature of March 2020 is that credit quality was almost beside the point. What became scarce was the balance sheet willing to warehouse what investors wanted to turn into cash. Leverage and basis trades mattered, but the dash for cash was broader than either. The episode is therefore better read as a failure of intermediation under simultaneous demand for liquidity than as a hedge-fund story.

05.

UK LDI 2022: solvent in years, illiquid in days

The UK LDI episode provides a cleaner illustration of leverage converting a long-horizon economic position into an immediate cash requirement. Defined-benefit pension schemes use liability-driven investment strategies to hedge long-duration interest-rate and inflation exposures. Rising gilt yields can improve the economic position of a pension scheme by reducing the present value of its liabilities. Yet leveraged LDI funds simultaneously face falling asset values and higher collateral or margin requirements.

In late September 2022, the Bank of England recorded a four-day increase of 140 basis points in 30-year gilt yields — more than twice the largest comparable move since 2000. LDI funds and pension schemes faced more than £70 billion of margin and collateral calls between 23 September and 14 October. Over the same period, LDI funds sold around £23 billion of gilts and DB pension schemes around £14 billion, using other liquid assets and existing cash buffers to meet the balance of the requirement (Bank of England, 2022).

The mechanism is close to Brunnermeier and Pedersen's liquidity spiral. Falling gilt prices raised leverage and collateral needs. If cash could not be raised from investors or other assets quickly enough, funds had to sell gilts into an already illiquid market, putting further downward pressure on prices and creating still larger cash requirements. The Bank intervened with temporary long-dated gilt purchases and expanded collateral facilities to give funds time to rebuild resilience.

The paradox is sharper than the usual language of leverage suggests. A pension scheme could become better funded in actuarial terms while becoming less liquid by the hour. Economic solvency and liquidity solvency were moving on different clocks. The danger arose because collateral had to move immediately while the institution’s economic horizon extended for decades.

06.

Credit Suisse 2022-23: buffers are stocks, withdrawals are flows

Credit Suisse is institutionally different from LDI, but it isolates another part of the same problem: liability speed. FINMA concluded that repeated strategic failures, risk-management deficiencies and scandals eroded confidence. The bank still complied with regulatory capital and liquidity requirements, and it held comfortable liquidity buffers in summer 2022. Those buffers did not prevent a rapid run once confidence deteriorated (FINMA, 2023).

Credit Suisse reported Group net asset outflows of CHF 123.2 billion in 2022, of which CHF 110.5 billion occurred in the fourth quarter. That is a measure of assets under management, not deposits. Separately, the Swiss National Bank estimates that outflows between October 2022 and March 2023 were approximately CHF 200 billion, more than half of the bank's customer deposits or roughly 30 percent of its September 2022 balance sheet; around CHF 90 billion of deposit outflows occurred in October alone. These measures are not additive, but together they show the speed and breadth of the confidence shock (Credit Suisse Group, 2023; SNB, 2024).

Credit Suisse isolates the same timing problem in a very different institution, but the distinction matters. Bank deposits are not mutual-fund shares. The common mechanism is a runnable claim: the liability owner can demand cash quickly relative to the maturity, saleability or encumbrance of the asset base. Diamond and Dybvig (1983) establish the coordination problem theoretically; modern fund evidence extends related incentives beyond banks. FINMA's finding that digital communication and transfer channels accelerated the Credit Suisse run illustrates how technology can compress the time available for balance-sheet adjustment without making banks and investment funds economically equivalent.

The case also demonstrates the limits of static liquidity ratios. Buffers are stocks; withdrawals are flows. FINMA found that Credit Suisse met regulatory liquidity requirements and held comfortable buffers in summer 2022, yet rapid outflows later brought it to the brink of insolvency. The lesson is not that liquidity regulation is futile, but that stress design must capture the speed, concentration and legal-entity usability of liquidity as well as its nominal amount.

07.

Private credit: patient liabilities can make illiquidity resilient

Private credit offers the opposite problem. Public-market stress becomes visible immediately through prices. Private-market stress can remain less visible because loans trade infrequently and are often valued through models or periodic appraisals. That can reduce observed volatility without eliminating economic risk. The IMF and FSB both emphasise that private credit's traditional closed-end structure is an important stabiliser: long-duration capital is usually well matched to long-duration assets, and the sector's current systemic risks appear more contained than the rhetoric around "shadow banking" often implies (IMF, 2024; FSB, 2026).

The reason to include private credit is both structural growth and the changing terms on which investors can access it. The FSB estimates total private credit at $1.5-2.0 trillion at end-2024 and highlights leverage, valuation opacity, interconnections, liquidity mismatch and concentration as vulnerabilities, while also noting that the market has not been tested at its current scale through a severe downturn. Traditional closed-end structures remain an important absorber because investor capital is comparatively patient (FSB, 2026).

The risk is no longer purely hypothetical. The ECB's May 2026 Financial Stability Review documents sizeable redemption requests at U.S. semi-liquid private-credit vehicles and business development companies; some vehicles met requests in full while others limited redemptions in line with contractual gates. The ECB simultaneously judges direct euro-area exposure to private credit to be limited enough that private credit alone is unlikely to be a source of systemic instability at present. That combination is analytically useful: redemption pressure has materialised, but systemic propagation has not been established (ECB, 2026).

The liability side is where the private-credit question becomes observable rather than merely prospective. Illiquid loans held in locked-up funds may be structurally resilient precisely because investors cannot demand frequent liquidity. Semi-liquid vehicles change that liability structure. Chen et al. (2010), Goldstein et al. (2017) and Schmidt et al. (2016) provide the academic basis for first-mover and strategic-withdrawal mechanisms in funds; the 2026 U.S. private-credit redemptions provide a contemporary observation of the liquidity mismatch, while the limited direct euro-area exposure is important counterevidence against treating the episode as systemic.

08.

Türkiye 2026: a live test of containment

The Turkish episode of September 2026 is useful only if it is kept in proportion. On 17 September the Capital Markets Board (SPK) closed TEFAS transactions for funds founded by seven portfolio-management companies and ordered selected funds to be liquidated. Bulletin 2026/61 subsequently updated the liquidation list and set out a bank-led process for converting portfolio assets to cash with explicit regard to investor interests, market depth and liquidity conditions. The Financial Stability Committee described the disturbance as credit and liquidity problems concentrated in a limited segment of the fund market and said it did not identify a fundamental or structural risk to the broader operation of Borsa Istanbul or Turkish capital markets.

This is also a mixed case rather than a clean redemption experiment. On 18 September SPK said the regulatory process had begun after observations from late 2025 that certain funds, particularly some free funds and money-market funds associated with specific managers, were contributing to price movements in low-free-float shares that the regulator judged difficult to explain by fundamentals. The current episode therefore contains market-conduct, concentration and liquidation dimensions alongside liquidity stress. This paper uses the liquidation and containment mechanics; it does not infer a generic fund run where public flow data have not yet established one.

That official assessment is not proof that contagion has been eliminated; it is the proposition currently being tested. What is observable is the containment architecture: transaction closures, a defined liquidation process, custodian-bank responsibility, and broader liquidity measures. Actual fund-level redemption and financing data remain incomplete. The episode therefore belongs in this paper as a live H5 case, but causal language should remain provisional until the transmission channels are observable in data.

It would be a stretch to claim that the event validates the fiscal-supply regime as a whole. The available evidence does not show that persistent sovereign borrowing or a rising duration premium caused the Turkish fund stress. What the episode does provide is subsequent evidence consistent with a narrower proposition from Research No. 02: contemporary leverage and liquidity stress can be concentrated in particular financial structures without immediately becoming a system-wide deleveraging cycle.

09.

Different institutions, the same question: which clock runs out first?

Across the cases, the initiating event is less informative than the contractual or institutional reason somebody must obtain cash. The common transition is from voluntary to constrained behaviour. That transition is what gives balance-sheet structure explanatory power.

The comparison also rejects a simplistic "interconnectedness is bad" view. Acemoglu, Ozdaglar and Tahbaz-Salehi (2015) show that denser networks can diversify relatively small shocks while propagating sufficiently large ones; Elliott, Golub and Jackson (2014) similarly find non-monotonic effects of integration and diversification on cascades. Greenwood, Landier and Thesmar (2015) show how common asset holdings can transmit fire-sale losses across balance sheets. The empirical cases are consistent with that conditional view: networks can absorb stress while capital is available and become transmission channels when balance-sheet constraints bind simultaneously.

10.

Research No. 02 revisited

The distinction is deliberate. Research No. 02 stated that subsequent evidence should be recorded as strengthening, weakening or retiring the framework rather than being treated as automatic confirmation. Research No. 03 follows that rule. The Turkish episode is subsequent to the 1 September cut-off of Research No. 02, but it tests only the narrow concentration-versus-systemic distinction. It does not demonstrate that the broader fiscal-supply mechanism caused the event.

The continuity between the two papers is therefore not that one “confirms” the other. It is that both place the marginal balance sheet at the centre of transmission. Research No. 02 asked who absorbs a growing stock of sovereign duration and at what discount rate. Research No. 03 asks what happens when that same class of private balance sheets is required to adjust faster than its funding, collateral or investor base will comfortably allow. Supply determines what must be held. Structure determines how violently holdings may have to change.

11.

The architecture of resilience

A paper about liquidity should not be written only from the failure side. The same structural framework identifies absorbers. Long-duration capital is a stabiliser when it is genuinely locked up. Liquid buffers matter when they are usable at the legal entity where cash is needed. Dealer and bank balance sheets absorb flows when capital and risk constraints leave room. Central-bank facilities can provide time when private balance sheets cannot. Governance and operational speed matter because delays can convert a manageable margin call into forced selling.

This produces a practical definition of structural liquidity: the capacity of a balance sheet to meet cash demands without requiring transactions that materially worsen its own solvency or the functioning of the market in which it trades. That definition is intentionally broader than bid-ask spreads. It includes time, funding, liabilities and the availability of other balance sheets.

12.

What would prove this paper wrong?

The framework should be revised rather than defended if the following observations persist:

Asset-level explanation dominates. Market depth and trading-cost measures explain stress episodes without material incremental information from funding, liability speed or investor structure.

Leverage is not associated with faster adjustment. Comparable shocks produce no meaningful difference in forced selling, margin pressure or funding reduction between leveraged and unleveraged structures.

Redemption incentives are weak. Open or semiliquid vehicles holding illiquid assets do not exhibit stronger outflow sensitivity, first-mover behaviour or spillovers in periods of poor performance.

Intermediary constraints do not matter. Large forced-sale episodes are absorbed without meaningful relationship to dealer capital, repo conditions, natural-buyer capacity or official balance sheets.

Local stress propagates without channels. Systemic events occur repeatedly without common holdings, funding links, collateral chains, confidence effects or other identifiable transmission mechanisms.

Private credit remains resilient as liquidity expands. Semiliquid and retail-access private-credit structures continue to grow through a full downturn without redemption stress, stale-valuation problems or broader spillovers.

A loss becomes dangerous when it creates cash demand faster than the surrounding system can supply time. The relevant object of analysis is therefore not simply the asset that falls in price, but the capital structure that determines who must act next.

13.

References

1. Acemoglu, D., Ozdaglar, A., & Tahbaz-Salehi, A. (2015). Systemic Risk and Stability in Financial Networks. American Economic Review, 105(2), 564-608.
2. Adrian, T., & Shin, H. S. (2010). Liquidity and leverage. Journal of Financial Intermediation, 19(3), 418-437.
3. Bank for International Settlements. (2026). Annual Economic Report 2026. Chapter II: High public debt and shifting financial markets: challenges for central banks.
4. Bank of England. (2022). Financial Stability Report, December 2022. Section 5: The resilience of liability-driven investment funds.
5. Brunnermeier, M. K., & Pedersen, L. H. (2009). Market Liquidity and Funding Liquidity. Review of Financial Studies, 22(6), 2201-2238.
6. Chapra & Co. (2026). The Fiscal-Supply Regime. Chapra Research No. 02. Chapra & Co.
7. Chen, Q., Goldstein, I., & Jiang, W. (2010). Payoff complementarities and financial fragility: Evidence from mutual fund outflows. Journal of Financial Economics, 97(2), 239-262.
8. Coval, J., & Stafford, E. (2007). Asset fire sales (and purchases) in equity markets. Journal of Financial Economics, 86(2), 479-512.
9. Credit Suisse Group AG. (2023). 4Q22 Results and Annual Report 2022.
10. Diamond, D. W., & Dybvig, P. H. (1983). Bank Runs, Deposit Insurance, and Liquidity. Journal of Political Economy, 91(3), 401-419.
11. Duffie, D. (2010). Asset Price Dynamics with Slow-Moving Capital. Journal of Finance, 65(4), 1237-1267.
12. Elliott, M., Golub, B., & Jackson, M. O. (2014). Financial Networks and Contagion. American Economic Review, 104(10), 3115-3153.
13. European Central Bank. (2026). Stress in global private credit markets and its implications for euro area financial stability. Financial Stability Review, May 2026.
14. Federal Reserve Board. (2020). Financial Stability Report, November 2020. A Retrospective on the March 2020 Turmoil in Treasury and Mortgage-Backed Securities Markets.
15. Financial Stability Board. (2026). Report on Vulnerabilities in Private Credit. 6 May 2026.
16. Financial Stability Committee of Türkiye. (2026). Statement on capital-market developments, 17 September 2026.
17. FINMA. (2023). Lessons Learned from the Credit Suisse Crisis. 19 December 2023.
18. Goldstein, I., Jiang, H., & Ng, D. T. (2017). Investor flows and fragility in corporate bond funds. Journal of Financial Economics, 126(3), 592-613.
19. Gorton, G., & Metrick, A. (2012). Securitized banking and the run on repo. Journal of Financial Economics, 104(3), 425-451.
20. Greenwood, R., Landier, A., & Thesmar, D. (2015). Vulnerable banks. Journal of Financial Economics, 115(3), 471-485.
21. Haddad, V., Moreira, A., & Muir, T. (2021). When Selling Becomes Viral: Disruptions in Debt Markets in the COVID-19 Crisis and the Fed's Response. Review of Financial Studies, 34(11), 5309-5351.
22. He, Z., & Krishnamurthy, A. (2013). Intermediary Asset Pricing. American Economic Review, 103(2), 732-770.
23. International Monetary Fund. (2024). Global Financial Stability Report, April 2024, Chapter 2: The Rise and Risks of Private Credit.
24. International Monetary Fund. (2026). Global Financial Stability Report, April 2026.
25. Kargar, M., Lester, B., Lindsay, D., Liu, S., Weill, P.-O., & Zuniga, D. (2021). Corporate Bond Liquidity during the COVID-19 Crisis. Review of Financial Studies, 34(11), 5352-5401.
26. Monin, P. J. (2026). Decomposing Hedge Funds’ U.S. Treasury Exposures. FEDS Notes, Board of Governors of the Federal Reserve System, 22 June 2026.
27. Schmidt, L., Timmermann, A., & Wermers, R. (2016). Runs on Money Market Mutual Funds. American Economic Review, 106(9), 2625-2657.
28. Sermaye Piyasası Kurulu. (2026). SPK Bulletin 2026/61, 17 September; guidance-process statement, 18 September; liquidation-period statement, 21 September; investor-count statement, 23 September; liquidation-timing clarification, 28 September.
29. Shleifer, A., & Vishny, R. (2011). Fire Sales in Finance and Macroeconomics. Journal of Economic Perspectives, 25(1), 29-48.
30. Swiss National Bank. (2023). Financial Stability Report 2023.
31. Swiss National Bank. (2024). Financial Stability Report 2024. Retrospective analysis of Credit Suisse liquidity outflows.

RESEARCH DISCIPLINE — Figures are claim-led analytical exhibits; source definitions are not combined where they are not comparable. Post-cut-off evidence is labelled explicitly.

IMPORTANT INFORMATION — This publication is independent research prepared by Chapra & Co. for informational purposes. It does not constitute investment, legal or tax advice, nor an offer, solicitation or recommendation in relation to any security, financial instrument or transaction. Views and data are stated as of the research cut-off except where later evidence is explicitly labelled.