Stress Testing: Pushing the Boundary

11^th February 2026

Introduction

Stress testing is a key tool for evaluating financial stability. Recent events have demonstrated that capital and liquidity buffers may prove insufficient when feedback loops exacerbate an initial shock and trust erodes. Banks must endure adverse macrofinancial conditions not only on a standalone basis but as wider market liquidity and confidence shocks emerge. Accordingly, stress testing is shifting from standalone solvency and liquidity tests to more comprehensive frameworks that capture a wider array of market risks, scenarios, and interrelationships. Despite notable advancements however, stress testing may still miss critical market dynamics—like the interplay between solvency and liquidity, the impact on profitability beyond the stress test horizon and shifts in market sentiment—potentially leaving systemic vulnerabilities unidentified. This blog discusses how stress tests should systematically integrate solvency and liquidity shocks, embed business model viability, and account for risks from non-bank financial entities. These improvements would help ensure that stress testing remains a critical forward-looking safeguard in today’s interconnected financial landscape.

Why stress testing needs an upgrade

Bank stress testing is a cornerstone of financial stability, designed to measure resilience and systemic risk under adverse conditions. The central question is simple yet profound: how would the banking sector respond to severe shocks? Would institutions absorb the impact, or break under pressure? Stress tests aim to answer these questions by simulating extreme scenarios - typically a deep recession - to assess the impact on bank balance sheets. At its core, stress testing focuses on bank solvency: ensuring banks have enough capital to withstand shocks. History shows however that liquidity shortfalls - not just inadequate capital - often trigger failures. Bear Stearns, AIG, Banco Popular, and Credit Suisse collapsed despite healthy regulatory capital ratios because liquidity evaporated. Recognizing this, authorities have long since undertaken liquidity stress tests. But recent crises have revealed the deep interconnection between solvency and liquidity risk. When asset values fall - a solvency shock - banks can face margin calls, lose access to short-term funding, and see funding costs spike, creating endogenous liquidity shocks. Conversely, liquidity stress can magnify solvency problems as banks resort to costly fire sales or expensive borrowing. Ignoring these feedback loops may lead to a significant underestimation of both solvency and liquidity risk.

Events such as the COVID-19 “dash-for-cash” (2020), Archegos collapse (2021), the UK gilt market turmoil (2022), and the March 2023 banking turmoil underscore the importance of capturing wider market liquidity risks and interconnectedness with non-bank financial institutions (NBFIs). While growing, practical experience of integrating solvency and liquidity risk in microprudential stress testing remains limited, even after a decade of calls for more realistic, macroprudential approaches (BCBS, 2015). The challenge is twofold: first, linking macrofinancial solvency scenarios with systemic and idiosyncratic liquidity shocks; and second, how to calibrate behavioral assumptions that would amplify shocks.

The collapse of Credit Suisse was a stark reminder that strong capital alone does not guarantee survival. Despite meeting enhanced capital and liquidity requirements for systemic banks, Credit Suisse succumbed to a self-fulfilling liquidity crisis triggered by a loss of confidence (SNB, 2023). At the heart of this failure were market concerns about the bank’s ability to generate sustainable profits after years of scandals and losses - a vulnerability that crystallized in the risk-off environment of March 2023. However, stress tests are narratives of “stocks” (capital), rather than “flows” (profits). Persistent shocks to profitability do not matter as long as stressed capital ratios remain above regulatory thresholds within the stress testing horizon. This blind spot should be addressed.

Stress testing is forward-looking but not far enough. Stress tests project capital ratios under calibrated scenarios over a 3–5-year horizon, but they reveal nothing about long-term viability. This contrasts with market valuations, which reflect expectations of sustainable profitability. Under standard valuation models, equity prices represent the discounted value of future dividends over a perpetual horizon. Supervisors are increasingly recognizing the need to assess business model sustainability—ensuring that bank strategies deliver durable returns over time. This is one of the key revisions in the Basel Core Principles for Effective Supervision (BCBS, 2024). Yet, business model sustainability remains largely untested in current stress-testing frameworks.

Credit rating actions can play a pivotal role in a bank’s downfall. Downgrades increase borrowing costs, restrict access to short-term funding, and limit the ability to renew maturing obligations. More importantly, downgrades trigger outflows from institutional investors - asset managers, hedge funds, and NBFIs - who rely on ratings to manage risk and allocate capital. Sophisticated counterparties, including non-banks, are prone to withdraw funding, cut credit limits, and demand higher collateral when a bank loses its prime-1 credit rating status (highest short-term debt rating). The loss of these clients further weakens the bank’s revenue base and exacerbates liquidity stress. While business risk is not included in Basel capital requirements, rating downgrades have direct revenue implications. Stress tests should quantify systematically the impact of downgrades on funding, liquidity, and profitability.

Market perceptions matter - particularly in risk-off environments. They influence business volumes, liquidity demands, and funding availability. Yet, stress tests tend to not systematically capture these dynamics. The solution? Incorporate market perceptions into stress testing. One option is to consider market-based resilience metrics (e.g., SRISK). However, using market values alone is problematic due to procyclicality. Instead, stress tests should identify “switch” points - moments when resilience is more accurately measured by complementing economic fundamentals with market-based valuations. These switch points likely relate to the size of losses, the profitability outlook, and the relative performance against peers (similar to credit rating actions) - not just to the distance of capital to regulatory minima. Incorporating these elements can advance stress-testing practices by adding idiosyncratic risk to macrofinancial risk.

Large counterparty credit risks (CCR) can amplify a bank’s troubles. Costly missteps led to mounting losses over several quarters, and the massive hit from a single counterparty - Archegos, amounting to USD 5 billion (roughly 10 percent of equity) - signaled structural risk management weaknesses in Credit Suisse. CCR losses can stem not only from outright defaults but also from shifts in exposure and deteriorations in counterparty credit quality, both of which tend to intensify under market stress. NBFI counterparties are particularly vulnerable to such shocks (ECB, 2023). While risk‑mitigation tools such as collateralization and margining practices reduce CCR, they increase counterparties’ liquidity risk. A sharp rise in CCR losses can also erode confidence in a bank’s business model. Uncertainty about bank resilience under different market conditions may trigger large withdrawals. As a result, CCR losses can function as early‑warning signals and generate system‑wide amplification effects (Barbieri, Grodzicki, Halaj, and Pizzeghello, 2025). 

What is being done?

Stress testing continues to evolve in response to emerging vulnerabilities and scenario uncertainty. The Federal Reserve applies a counterparty default scenario component in the CCAR, and in 2025 used two exploratory scenarios to assess systemic risk from private credit and hedge funds. The ECB has complemented system-wide solvency tests with thematic exercises on interest rate risk (2017), liquidity risk (2019), climate risk (2022), cyber resilience (2024), and reverse stress testing (2026). The Bank of England has conducted exploratory exercises on cyber risk (2024) climate risk (2021) and liquidity risk (2019), and more recently launched a system-wide exploratory scenario (SWES) to examine risks arising from NBFI interconnectedness and core funding markets. The IMF has conducted climate risk analysis and systemwide liquidity analysis (Ding, Laliotis, and Toffano) to capture emerging risks and look at the evolution of the wider financial system. These initiatives reflect a growing recognition that stress tests must go beyond recession-related losses to cover counterparty defaults, earnings under stress, and broaden the stress test perimeter to include the behavior of NBFIs.

Incorporating dynamic effects remains challenging. Despite notable strides, most supervisory stress testing frameworks remain static - capturing shocks in isolation rather than tracing intertemporal trade-offs and spillovers. For example, raising liquidity through money or capital markets to meet deposit withdrawals can undermine solvency and deepen the solvency-liquidity spiral, while selling high-risk businesses to boost capital ratios may erode long-term profitability and depress equity valuations (e.g., Credit Suisse’s sale of securitized products to Apollo in 2022). Moreover, actions to strengthen individual resilience can trigger adverse system-wide effects; for instance, reducing liquidity to distressed NBFIs may mitigate counterparty credit risk in the short term but amplify fire-sale dynamics and spill back through valuation losses. While some macroprudential stress tests (measuring the resilience of the system as a whole) tend to capture dynamic effects (Budnik et al, 2024), supervisory stress tests (measuring individual resilience) do not. A key obstacle to incorporating behavioral responses is the resulting complexity, which raises concerns about interpretability and model risk.

Despite progress, a critical gap persists. Liquidity shocks are often modeled independently of solvency positions, creating fragmented risk assessments that overlook their interplay. Funding liquidity can quickly morph into solvency risk, especially when market perceptions amplify stress. The next step is to tackle this issue systematically by quantifying the impact of confidence shocks for weak banks.

The solvency-liquidity nexus: the ultimate trigger

To address this gap, some exploratory frameworks integrate solvency and liquidity dynamics. For instance, Cont, Kotlicki and Valderrama (2020) capture key transmission channels such as credit downgrades, business risk, and the interplay between solvency and liquidity. The 2025 IMF EA FSAP bank stress test enhances this framework to include CCR arising from endogenous defaults of NBFIs, to reflect real-world contagion mechanisms. Here is how the framework works:

• Credit Downgrades: When leverage ratios approach regulatory minimums, credit downgrades restrict access to unsecured funding and credit-sensitive markets.
• Business Risk: Credit downgrades trigger reputational damage and reduce client activity, eroding bank profitability.
• CCR: NBFIs’ liquidity stress - driven by redemptions and margin calls - lead to missed variation margin payments, prompting banks to liquidate derivative positions and incur CCR losses.
• Endogenous Liquidity Risk: A perceived weak bank faces increasing liquidity demands, escalating costs and collateral constraints, deepening the “death spiral” between solvency and liquidity risk.

The system-wide effects of these interactions can be significant. Covi and Škrinjarić (2025) extend this framework to the UK and find that solvency–liquidity feedback loops account for one-third of default risk on average. This implies that running separate solvency and liquidity stress tests may significantly underestimate systemic risk. 

How should stress testing evolve?

1. Integrate Solvency and Liquidity Stress Tests

Solvency shocks - such as CCR losses - can trigger endogenous liquidity shocks, amplifying losses through increased funding costs and fire sales. “Liquidity at Risk,” can be a useful forward-looking measure in these cases, as it quantifies the liquidity resources a bank needs under adverse scenarios, conditional on its unique balance sheet structure and risk sensitivities. By integrating solvency and liquidity stress testing, supervisors and banks can better capture idiosyncratic risks and anticipate how market shocks propagate through both channels.

2. Stress Business Models

Regulatory capital is a “stock,” while the capacity to generate organic capital is a “flow.” A bank’s ability to generate future profits determines its resilience, confidence restoration, and survival under shocks. Stress tests should enhance projections of future profitability beyond the testing horizon. They could incorporate potential credit rating downgrades driven by profitability concerns. Performance could be assessed not only against regulatory hurdle rates but also relative to peer outcomes - a proxy for real-world downgrade risk. Applying discount rates to future profits based on bank-specific uncertainty under stress could help align regulatory and market metrics under adverse conditions.

3. Capture Dynamic Effects

Most supervisory stress tests assume static balance sheets for conservatism, as higher risk-weighted assets typically outweigh incremental net interest income. Yet, this assumption is not truly conservative for complex, systemic banks, which often experience revenue declines from reduced client activity in investment banking and wealth management. Incorporating market perceptions of resilience - especially when valuations “shift” from fundamentals to market sentiment - can better capture the dynamics of business risk and funding needs for banks perceived as “weak” by the market.

4. Perform Reverse Stress Tests

Broadening the scope of stress testing by adding multiple scenarios (Federal Reserve) and incorporating reverse stress testing to pinpoint critical shock amplitudes that could lead to defaults (ECB Reverse Stress Test). Reverse stress tests can also help address scenario uncertainty, geopolitical risks, and non-linear effects from sudden breaks in historical correlations. In addition, combining banks’ granular bottom‑up models with top‑down dynamic models can help mitigate model‑risk uncertainty.

5. Conduct System-Wide Stress Tests

To enhance realism and strengthen conservatism, stress tests should extend coverage beyond banks to include NBFI behavior under stress. This would enable a more comprehensive quantification of CCR losses and support the assessment of resilience of funding markets under coherent scenarios. Stressed CCR losses should capture the cost of unwinding portfolios and potential market losses from the unmatched hedging positions. Recent initiatives using granular transaction level data - such as the 2024 Bank of England's SWES exercise, and the 2025 IMF EA FSAP system-wide stress test - provide valuable blueprints for advancing this approach.

Conclusion

To conclude, financial crises have shown that liquidity and solvency risks are deeply interconnected. NBFIs can amplify stress through CCR and disruptions in funding markets - while market perceptions on business model viability can magnify vulnerabilities - especially when a bank’s strength or profitability comes into question relative to its peers. Advancing stress testing to capture these dimensions would enable supervisors and macroprudential authorities to strengthen their capacity for proactive risk mitigation, helping to foster a safer and more resilient financial system.

Note

The views expressed herein are those of the author and should not be attributed to the IMF, its Executive Board, or its management.

References

Adrian, T., Grippa, P., Gross, M., Haksar, V., Krznar, I., Lepore, C., Lipinsky, F. Oura, H., Lamichhane, S. and Panagiotopoulos, A., (2022), “Approaches to Climate Risk Analysis in FSAPs”, IMF Staff Climate Notes.

Arnould, G., Leika, M., Panagiotopoulos, A., Spector, M., and Valderrama, L. (2025), “Stress Testing the Banking Sector”, Euro Area FSAP, IMF Country Report No. 25/211 available at EA FSAP TN Stress Testing the Banking Sector

Bank of England (2024), “The Bank of England’s system-wide exploratory scenario final report”, November, available at SWES

Barbieri, C., Grodzicki, M., Halaj, G., and Pizzeghello, R. (2025). System-wide implications of counterparty credit risk

Basel Committee on Banking Supervision (2015), “Making supervisory stress tests more macroprudential: Considering liquidity and solvency interactions and systemic risk”, BIS Working Paper No. 29

Brownlees, C., and Engle, R., (2017), “SRISK: a conditional capital shortfall measure of systemic risk”, ESRB Working Paper No. 37.

Budnik, K., Ponte Marques, A., Ben Hadj S., Georgescu, O., Giglio, C., Grassi, A., Durrani, A., Figueres, J.M., Konietschke, P., Le Grand, K., Metzler, J., Ortl, A., Población García, F.J., Shaw, F., Trachana, Z., Chalf, Y., Gross, J., Sydow, M., and Franch F., (2024), Advancements in stress-testing methodologies for financial stability applications, ECB Occasional Paper Series, No. 348.

Cont, R., Kotlicki, A., and Valderrama, L., (2020), “Liquidity at risk: Joint stress testing of solvency and liquidity”, Journal of Banking and Finance 118.

Covi, G., and Škrinjarić, T., (2025), Measuring the stability of the banking system: capital and liquidity at risk with solvency-liquidity interactions | Bank of England, Staff Working Paper 1,140.

Ding, X. Laliotis, D., and Toffano, P., (2024), A Framework for Systemwide Liquidity Analysis, IMF Working Paper 24/104.

European Central Bank (2025a) Stress testing at the ECB: mastering risk analyses and impact assessments with top-down models

European Central Bank (2025b) ECB to assess banks' stress testing capabilities to capture geopolitical risk

European Central Bank (2025c), Stress tests in uncertain times: assessing banks’ resilience to external shocks, The Supervisory Blog, Chair of the Supervisory Board, Claudia Buch.

Federal Reserve (2023), Multiple scenarios in stress testing, Speech by Vice Chair for Supervision Michael S. Barr.

Letizia, E., and Wintenberg, H., (2025), “Systemic Risk Analysis – NBFI”, Euro Area FSAP, IMF Country Report No. 25/211 at EA FSAP TN Systemic Risk Analysis - NBFI 

Swiss National Bank (2023), Financial Stability Report. Financial Stability Report 2023