This Time Is Different? Speech By Sarah Breeden, Bank Of England, Deputy Governor, Financial Stability, Given At The HLS-PIFS Symposium On “Building The Financial System Of The 21st Century: An Agenda For Europe And The United States”

Speech

It is no understatement to say that we have been living through extraordinary times.  And it’s not obvious there is a calmer outlook ahead of us. What might that mean for financial stability?

History has taught us that crises do not arrive without warning. In the run up, optimism overrides caution, resilience is reduced, leverage and complexity build up, risk migrates to opaque and hidden corners. When a shock hits, we see rapid deleveraging, contagion and crisis. As the Bank of England’s Deputy Governor for Financial Stability, my focus is on making sure we learn the lessons from the past. So will this time be different?

Let me start with some good news.

The global financial system has proved resilient in the face of extraordinary shocks.

In the space of the last six years we have experienced six big shocks – a pandemic, Russia’s invasion of Ukraine, the failure of a number of overseas banks including a global systemically important one, severe stress in the UK gilt market following the mini-budget in 2022, repeated disruptions to world trade, and the worst energy shock in my living memory. Conflict in the Middle East continues to add to uncertainty and unpredictability, with sharp movements in energy prices and bond yields.

Yet, through all of this, the system has largely absorbed shocks rather than amplified them. It has continued to provide essential services to households and businesses.

That resilience matters – and it is not an accident. It reflects the hard‑won lessons of the global financial crisis. Banks now hold significantly more, and higher‑quality, capital and liquidity, and we are much better prepared for a failure (Ramsden, 2026). Macroprudential tools – from countercyclical capital buffers to housing measures – lean against the build‑up of risk, so that debt is less likely to magnify downturns.

So it’s natural to ask: is this time different? Or is there more we need to do?

Today I will look at the lessons from history, assess where vulnerabilities now sit, and explain how recent events affect the odds that those vulnerabilities will crystallise at scale. And I will finish by setting out what we are doing – at home and internationally – to reduce the likelihood of that happening.

Lessons from history

History rarely repeats itself precisely – but it does tend to rhyme.

Across financial crises, a common pattern emerges. Optimism slowly outruns realism. Confidence replaces caution. And risk migrates, via complex routes, to the least visible, least regulated, and least well‑understood corners of the system.

Galbraith’s The Great Crash 1929^footnote[1] describes a period convinced that innovation had rewritten the rules. Equities were treated as a one‑way bet. Investment trusts promised diversification while quietly amplifying leverage^footnote[2]. Warnings were not absent – but they were discounted.

Accounts of the global financial crisis, like Andrew Sorkin’s Too Big To Fail^footnote[3], tell a strikingly similar story. Complexity provided reassurance. Models offered comfort. Risk was judged to be small and dispersed – and therefore contained.

The point is not that equities in the 1920s were the same as sub‑prime mortgages in the 2000s. Each era invents its own instruments. Innovation changes the packaging; human behaviour supplies the content. The question to ask is not whether history will repeat itself—but whether we can hear its echoes and whether we act in time.

There are real reasons to think the system is considerably stronger than in the past.

First, the banking system. Banks are far better capitalised and more liquid than before the global financial crisis. UK bank stress tests show that the banking system can continue to lend through severe but plausible scenarios, including sharp increases in commodity and energy prices and major supply‑chain disruptions – a scenario that feels somewhat familiar^footnote[4]. A credible resolution regime further reduces the risk that an individual bank failure becomes systemic.

Second, households and businesses. Corporate balance sheets today appear more resilient than in the mid‑2000s and vulnerability indicators are well below past peaks. Household indebtedness is close to its lowest level since the early 2000s, arrears are low, and debt‑servicing burdens remain manageable.

Risks have not disappeared – they have shifted

But resilience in parts of the system does not guarantee the resilience of the whole.

Activity has increasingly migrated from banks to market‑based finance, which now accounts for around half of UK and global financial‑sector assets. That shift brings benefits. It diversifies funding and moves risk from entities with fixed nominal liabilities – such as bank deposits – to investors who should expect and are better placed to absorb losses. But losses, and the behaviours they trigger, can still amplify shocks – particularly when leverage, liquidity mismatch and complexity are involved.

At the same time, risk has shifted towards sovereign balance sheets. Before the global financial crisis, vulnerabilities were concentrated in households and firms. Since then, governments have borrowed extensively – first to stabilise the system after the GFC, and later to support businesses and households through the economic effects of the pandemic and war in Ukraine.

Sovereigns have long horizons and the greatest capacity to manage risk. But public debt in advanced economies is now close to post‑war highs. That reduces fiscal space, limiting governments’ ability to respond to future shocks, and can increase sensitivity to shifts in market confidence.

When government bond yields rise sharply, financing conditions tighten across the economy. Sovereign bond markets anchor pricing for mortgages and corporate credit and provide core collateral for repo and derivatives markets – the plumbing of the financial system. When that plumbing fails, stress can propagate quickly, as the UK experienced in autumn 2022.

Against that backdrop, I wanted to highlight three vulnerabilities that are salient today.

Risk one: the growing importance of private markets.

Private markets play an important role in financing productive investment. They broaden the financing landscape and can allocate risk to investors best placed to bear it.

Growth has been rapid. Global private‑market assets under management are estimated to have risen around six‑fold since the low‑rate era began in 2008/9, to roughly $18 trillion by 2025. They have not yet been tested, at that scale and complexity, by a broad‑based macroeconomic shock in a higher‑rate environment^footnote[5].

The Financial Policy Committee has long highlighted vulnerabilities in risky credit markets, notably private credit (Benjamin, 2024). Transparency is more limited. Valuations may lag reality. Underwriting standards have weakened. Leverage is a layer cake – at the borrower, fund and sponsor level – making it hard to measure. Loans are sometimes sliced and diced into CLOs and often held or distributed on the basis of ratings agency assessments. And complex connections to the rest of the financial system, including banks, insurers and reinsurers, create an ecosystem where losses that are already hard to size become even harder to trace.

Recent events have sharpened attention on this sector. Investor sentiment towards riskier credit has weakened, reflecting concerns about asset quality, valuation discipline and liquidity. A series of defaults – MFS, Tricolor, First Brands Group – has reinforced concerns. Redemption pressures have also risen in a number of international retail private‑credit funds, largely distributed through wealth management channels. Retail involvement remains small in aggregate. But as history shows, stress often emerges first at the margins – and then spreads.

A broad-based credit crunch in private markets could tighten financing conditions for the UK real economy. If valuations or rating assessments are challenged – through defaults, markdowns or discovery that risk models have been mis-calibrated – lenders may pull back. Akerlof’s “lemons” problem may then apply. When investors struggle to distinguish between good and bad risks – as well they might given opacity and complexity – they price the worst case. In that environment, sound borrowers can end up paying “bad‑firm” prices simply because, under stress, it’s hard to differentiate. And limited incentives to do that differentiation can further tighten lending conditions.

Risk two: leverage in government bond markets.

Hedge funds have become increasingly important players in government bond markets. And their use of leveraged strategies has been a key driver in the recent growth of leverage in UK cash gilt markets. In early 2026, net gilt repo positioning reached its highest level since data collection began in 2017, with borrowing concentrated among a relatively small number of firms pursuing similar strategies. These strategies support liquidity in normal times. But as price‑sensitive, leveraged investors, hedge funds can amplify shocks and cause jumps to illiquidity^footnote[6].

This risk is not unique to the UK, with similar developments notable across many advanced economy bond markets. And many hedge funds operate across jurisdictions and multiple asset classes, increasing the risk that stress in one market spills into others through shared exposures or correlated deleveraging.

Our first system-wide stress exercise (SWES) showed how hedge fund activity can amplify stress in gilt markets if the costs of repo financing increase sharply or banks limit funding. The extent of this pull back from banks was often greater than expected by their counterparts. The discretionary deleveraging of concentrated positions can also amplify market moves, as we saw in UK short rates when hedge funds unwound commonly held positions after the onset of the Iran conflict. Markets effectively absorbed the high volatility and volumes. But if market interest rates move by more than they would have done otherwise, we might see an additional tightening in financing conditions.

Risk three: stretched asset valuations.

A third vulnerability lies in asset valuations. Despite a materially uncertain macroeconomic outlook, global risk premia remain compressed, close to levels last seen before the global financial crisis. Valuations in parts of the US technology sector – particularly firms closely linked to artificial intelligence – appear especially stretched even as the path to monetisation of this new general purpose technology remains unclear.

At the same time, investment by AI firms is increasing materially – estimated at over $5trn over the next five years. To date, much of this investment has been financed by equity and cash. But debt financing – across public markets, leveraged finance and private credit – has risen rapidly and is expected to increase further^footnote[7].

While we have already this year seen some adjustment in market valuations in AI‑related sectors, they remain elevated. A reassessment of the future earnings potential could lead to abrupt price declines and raise questions over how debt will be repaid. Any broader stress across credit markets could affect UK borrowers and investors.

Echoes of the past

Across all three of these risks you can hear echoes of the past. The combination of leverage, complexity, concentrations and opacity rhyme with the vulnerabilities brought about by the rise of CDOs in 2007 and, more distantly, the development of investment trusts in 1920s. All at a time when the disconnect between high risky asset prices and real economy uncertainty seems marked.

I am not predicting the next crisis. But history suggests that when these conditions coincide, the system becomes more fragile.

The conflict in the Middle East raises the odds that one or more of these vulnerabilities could crystallise at the same time. Shocks to growth, inflation and interest rates hit all borrowers simultaneously. Because the system is interconnected, stress in one area can amplify behaviour in others.

The global financial crisis taught us that crises often involve multiple vulnerabilities crystallising together. Almost all of the vulnerabilities highlighted in our pre-GFC Financial Stability Reports crystalised at the same time^footnote[8]. In such a scenario, correlations and losses might shift outside of historic norms. So, firms and regulators alike need to be responsive and flexible in our risk management and stress testing now, with a focus on overlapping shocks and the likelihood of many shoes dropping at the same time.

What we are doing differently

Let me finish with what we are doing about current risks – and why the institutional framework today gives us greater confidence than in the past. A standing macro-prudential framework that focuses on identifying systemic risks has been the key driving force behind the increased resilience we have seen. I want to highlight three priorities as we look ahead.

First, a shift towards genuinely system‑wide surveillance.

As the system becomes more complex, interconnected and opaque, surveillance must follow interconnections, not institutional boundaries (Breeden, 2024). Our first SWES – an international first – examined how stress could propagate across core UK markets (Bank of England, 2024).

This year we will conduct our second SWES focused on private markets, bringing together alternative asset managers, large banks and institutional investors such as insurers and pension funds (Bank of England, 2025). The objective is simple: to understand how market participants behave in stress; what that means for markets as a whole; and how shocks might transmit to the real economy.

And since capital markets are global, our system-wide analysis must not stop at national borders. Already, other jurisdictions are running SWES-type exercises. Improved transparency and international coordination, including through the work of the FSB’s Non‑bank Data Task Force, are essential.

Second, building greater ex‑ante resilience in market‑based finance.

Surveillance alone is not enough. Where effective and practical, vulnerabilities should be addressed directly to build ex ante resilience – just as bank capital requirements have done for the banking system.

A priority area is gilt‑repo resilience. Our reforms following the 2022 stress have strengthened LDI funds (Breeden, 2023), which have remained resilient through recent volatility. But results from our first SWES suggests more needs to be done. Last year we published a discussion paper on a range of proposals, including greater central clearing for repo and minimum haircuts for non‑centrally cleared repo trades (Bank of England, 2025). We will work with firms and authorities on next steps and provide a comprehensive update in early 2027.

Again since these markets are global, progress also depends on international collaboration. The Financial Stability Board’s work on non‑bank leverage is an important foundation, and our analysis continues to build on it (FSB, 2025Opens in a new window ).

Third, the development of new ex‑post tools.

We cannot anticipate every crisis. And it would be neither feasible nor desirable to build ex‑ante resilience for every possible state of the world.

We have a number of market-wide and contingent facilities that enable us to deliver liquidity when needed. While, central banks have long acted as lender of last resort to banks, friction in the system means liquidity does not always reach the areas where it is most needed in stress. And so most recently, we have developed the Contingent NBFI Repo Facility—a tool that can be activated to address severe gilt‑market disruption, lending cash against gilt collateral to participating insurance companies, pension funds and liability-driven investment funds where financial stability is threatened (Ramsden, 2024).

It is precisely because of the ex-ante resilience that these firms have that we are able to provide them with such support ex post. And the same principle applies as we design our regulatory regime for systemic stablecoins (Bank of England, 2025). The ex-ante resilience delivered by our regime supports our ability to provide a standing liquidity facility that backstops their ability to monetise their backing assets in a stress.

Conclusion

Recent events are a reminder that large, correlated shocks can arrive with little warning. The global financial system has been resilient so far, reflecting deep structural improvements since the global financial crisis – especially in our banking system. So is this time different?

The vulnerabilities that have preceded past crises have not disappeared. They have re-emerged elsewhere. Across private markets, government bond markets, and in stretched valuations, you can hear the familiar echoes of leverage, complexity, concentration and opacity. If some of these crystallise simultaneously, we may be in for a rocky ride.

But what is different is our more resilient banking system and our focus. We look at the system as a whole. We invest in surveillance, build resilience where possible, and have targeted tools in case stress emerges.

History suggests that the most dangerous moments are not simply those when risks are high – but those when they are too easily dismissed. Our task is to ensure that they are not.

I’d like to thank Nickie Shadbolt for her assistance in drafting these remarks. I would also like to thank Martin Arrowsmith, Andrew Bailey, Lee Foulger, Grellan McGrath, Liz Oakes, Andrea Rosen, Tom Smith and Michael Yoganayagam for their helpful input and comments. The views expressed here are not necessarily those of the Monetary Policy Committee (MPC) or Financial Policy Committee (FPC).