Bank of England loosens the reins on bank capital as new risks gather in markets and AI

The Bank of England is preparing to soften parts of the post-crisis rulebook that has governed British lenders since the financial crash, arguing that a system designed for one era is now producing frictions in another. Alongside its latest Financial Stability Report, the Bank set out proposals to ease how capital buffers and leverage constraints apply to the largest UK banks, a move it says will make the regime more “effective” and “better calibrated” to the risks of today’s financial system. The intent is plain: free up capacity in balance sheets so banks can keep credit flowing to households and businesses at a time when the wider economy is uneasy and growth is hard to find.

To understand why this matters, it helps to recall what these buffers were built to do. After 2008, regulators concluded that banks had been allowed to run too thinly capitalised for too long, using borrowed money to inflate profits in good times and leaving taxpayers to shoulder the losses when the cycle turned. Capital requirements and buffers were the hard lessons turned into policy: lenders should hold enough loss-absorbing capital to endure shocks without collapsing, and the financial system should be strong enough to keep functioning during stress rather than amplifying it.

Those reforms succeeded in one important sense. UK banks today are generally better capitalised, better supervised and more cautious than their pre-crisis predecessors. Yet even robust safeguards can create unintended consequences, particularly if rules intended to be used flexibly become, in practice, constraints that bite precisely when the economy is weak. The Bank is now acknowledging that tension. It wants lenders to use their additional buffers in a downturn, rather than retrench to preserve ratios, and it wants to adjust the leverage framework that sits beneath risk-weighted capital rules.

The focus is on domestically oriented lenders such as NatWest, Lloyds Banking Group, Santander UK and Nationwide Building Society. The Bank is proposing changes that would give them greater freedom to draw down parts of their additional capital buffers during periods of stress. That sounds technical, but it speaks to a familiar political and economic anxiety: when households and businesses most need access to borrowing, regulators can inadvertently encourage banks to pull back. If capital rules are perceived by boards and investors as hard floors rather than usable cushions, banks are more likely to protect ratios than extend credit, even if the broader public interest lies in the opposite direction.

At the centre of the plan is a recalibration of the leverage ratio, the blunt measure of how much capital a bank holds against its overall exposure, regardless of how “safe” those exposures are deemed by models. Leverage rules were introduced as a backstop after the crisis because risk weighting can fail when it matters. Assets considered low-risk can suddenly become volatile, and internal models can prove optimistic. A leverage ratio helps prevent banks from expanding their balance sheets too aggressively on the assumption that risks are small or well dispersed.

The Bank’s proposed adjustments would reduce the leverage ratio requirement for large UK banks by around 20 basis points in aggregate. It is not, in the Bank’s telling, a bonfire of regulation. Rather it is presented as a refinement, intended to stop the leverage framework from penalising ordinary domestic lending. The implication is that the current calibration has become overly restrictive relative to the risk of the activity it is capturing, particularly for banks whose business is largely in the UK and whose lending is focused on mortgages and corporate credit rather than global trading books.

That is a judgement call, and the Bank has been keen to frame it as such. Andrew Bailey, the Governor, described the decision as “finely balanced”, language that signals not only prudence but also the awareness that the politics of loosening any rule associated with bank safety are delicate. In the years since the crash, regulators have repeatedly promised not to forget the past. Any hint of complacency invites criticism, especially in a country where public memory still holds the image of emergency rescues and the long, grinding aftermath of austerity.

The Bank is also managing another tension: the industry’s long-standing complaint that the UK’s framework has been more onerous than those of some peers, and that British lenders are disadvantaged internationally. UK Finance, the industry body, welcomed the proposals as “important initial steps”, but its chief executive, David Postings, suggested that the sector is looking for something more substantial. A senior executive at one large bank described the reforms as falling short of the “comprehensive review” needed to keep the UK aligned with evolving international approaches.

This is where the debate becomes as much about national economic strategy as about technical prudential policy. Ministers have been pressing regulators to support growth, and the City has argued that competitiveness has suffered under a regime that sometimes feels designed to prevent the last crisis rather than enable the next decade of investment. Yet the Bank’s job is not to deliver growth at any cost. Its mandate is to protect financial stability, and if that sounds abstract, the report itself provides reminders of what instability can look like in 2026: the sudden repricing of risk, the rapid unwinding of leveraged trades, and the operational vulnerabilities that arise as finance becomes more complex and more dependent on technology.

Indeed, the most striking feature of the Financial Stability Report is the contrast between the Bank’s willingness to loosen certain constraints on traditional lenders and its alarm about the build-up of risk elsewhere. It warned that “risky asset valuations” have become more pronounced since December, with leverage rising in equity markets. This is not a story of high-street banks making reckless mortgage loans, but of market participants using borrowed money to magnify returns in fashionable corners of the market, especially those linked to artificial intelligence.

The Bank pointed to hedge funds increasing leverage as they build positions in companies associated with AI. That matters because leverage does not merely increase gains, it accelerates losses and can force rapid selling when prices move the wrong way. In an era where trading is faster, liquidity can be less reliable than it appears and crowded positions can unwind together, leverage becomes a transmission mechanism for volatility. A sharp correction in a popular trade can ripple into broader markets as funds scramble to meet margin calls and reduce exposures.

AI, in the Bank’s account, is not only a source of speculative fervour in markets but also a driver of a more traditional vulnerability: debt. The report highlighted the rapid acceleration of debt financing for AI-related companies, particularly to fund the construction of data centres and other infrastructure. The pace of this investment, it suggested, is historically unusual. Behind the excitement about productivity gains lies an older financial question: what happens if revenue projections fail to match the scale of capital expenditure and borrowing? If the assumptions underpinning the boom prove optimistic, the holders of debt issued by these businesses could face losses, and defaults could rise in ways that spread through credit markets.

There is a further complication with AI infrastructure: technological obsolescence. Data centres and specialised hardware are expensive, and their economic life can be shorter than investors anticipate if new architectures emerge or energy costs rise. Debt, unlike equity, does not absorb disappointment easily. It requires payments regardless of whether the original narrative remains compelling. The Bank’s concern is not an argument against AI investment, but a reminder that credit booms are often built on extrapolation, and extrapolation is fragile when the underlying technology evolves rapidly.

The report also ventured into a territory that, until recently, would have seemed peripheral to core financial stability: frontier AI and cybersecurity. Bailey described the risks posed by advanced models as “a big issue”, reflecting the growing worry that powerful systems can be used to scale attacks, automate vulnerability discovery or destabilise operations. For financial firms, which rely on continuous availability and trust, cyber resilience is not a matter of convenience. It is systemic. The more finance depends on complex software, the more the sector must update, patch and integrate, and the greater the chance of glitches or operational failures that can interrupt services or expose critical infrastructure.

This is one reason the Bank’s move on capital buffers cannot be interpreted simply as a concession to industry lobbying. The central bank is not suggesting that the world is safer. It is suggesting that risks have changed shape. Some are moving away from bank balance sheets and into markets, non-bank financial institutions and technology-driven vulnerabilities. That shift raises a difficult question for regulators: how tightly should banks be constrained if the next shock is more likely to come from outside the banking system?

The report offered one clue by highlighting increased leveraged trading by hedge funds in UK government bond markets. Gilt markets are the foundation of the financial system, a pricing reference for mortgages, corporate borrowing and pension liabilities. Episodes in recent years have shown how quickly dysfunction can appear when leverage and forced selling collide, and how hard it is to restore calm once confidence slips. If hedge funds are employing strategies that depend on stable liquidity and low volatility, a sudden shock can become self-reinforcing, prompting central bank intervention even if banks themselves are not the epicentre.

Private credit was mentioned as another area of opacity and potential fragility. The attraction of private credit has been its promise of steady returns and bespoke financing outside traditional bank lending. Yet opacity can conceal concentrations of risk, and the interconnectedness between private funds, insurers, pension schemes and banks can be hard to map in real time. If stress emerges, the absence of transparent pricing can delay recognition of losses and complicate the process of stabilisation.

It is in that context that the Financial Policy Committee warned of its “particular concern” that several vulnerabilities could crystallise simultaneously. The lesson of modern finance is that crises rarely arrive with a single cause. They emerge when valuation, leverage, liquidity and confidence interact. A sell-off in stretched equity markets could coincide with losses in AI-linked credit, operational disruption from a cyber incident, and strained liquidity in gilt markets. None of those elements alone guarantees a systemic crisis, but together they can produce one, particularly if participants do not fully understand the “size, concentration, and interconnections” of their exposures.

The Bank’s recalibration of capital rules, then, can be read as an attempt to preserve room for manoeuvre. If banks are not the primary source of risk, they still need to be resilient enough to absorb shocks transmitted from markets. But they also need to be capable of supporting the economy through downturns rather than amplifying them. Allowing buffers to be used as intended may improve that capacity, assuming investors and rating agencies accept the logic and do not punish banks for acting counter-cyclically.

Critics will argue that loosening leverage requirements, even by 20 basis points, risks eroding a backstop that exists precisely because models can be wrong. Supporters will counter that a backstop that bites too early can distort behaviour and reduce lending unnecessarily, and that the Bank is not abandoning prudence but adjusting calibration after years of experience. Both positions contain truth. The more important test is whether the reforms are accompanied by clarity on when buffers should be used, how supervisors will respond, and what metrics will signal that resilience is being maintained.

For the banking industry, the proposals are a signal that the central bank is listening, but not capitulating. The consultation period will give lenders and other market participants the chance to push for broader changes, and it will also provide an opportunity for sceptics to probe the assumptions behind the Bank’s confidence. The word “proportionate” does a lot of work in regulatory debates. It is often invoked to justify easing, but it can also justify tightening if risks grow. The Bank is effectively asserting that it can be proportionate and still tough, more flexible and still credible.

For the wider economy, the immediate question is whether any freed capacity translates into meaningful additional lending, or whether it is absorbed by other priorities such as shareholder distributions and balance sheet optimisation. The Bank will insist that a resilient system is a prerequisite for sustainable credit, and that the point of flexibility is to sustain lending during stress, not to encourage an indiscriminate expansion. Yet the political narrative will be shaped by outcomes, not intentions. If households remain squeezed, businesses struggle to invest and credit conditions stay tight, pressure will build for further adjustments.

In the end, the Bank is attempting to walk a narrow path between two memories: the memory of 2008, when undercapitalised banks nearly brought down the system, and the memory of more recent market disruptions, when leverage and liquidity outside banks threatened stability. The first created the regime now being adjusted. The second is a reminder that stability cannot be secured simply by demanding ever more capital from lenders while risks migrate elsewhere. The challenge for policymakers is to keep banks safe without making them so constrained that they cease to perform their essential economic function, and to confront the uncomfortable reality that the next crisis may be driven less by mortgages and more by markets, technology and the concentrated bets of institutions operating beyond the traditional perimeter.