Basel Output Floor Could Shift UK Bank Lending Toward Corporate Loans and Riskier Mortgages

The Basel III output floor could change which borrowers UK banks find less costly to serve. By limiting how far internal risk models can reduce capital requirements, the floor may encourage banks to expand corporate lending and some riskier mortgages while pulling back from the safest mortgages. The effect could begin before the measure reaches its full level in 2030, as banks have previously adjusted to anticipated capital rules well ahead of implementation.

Why the floor changes incentives

Capital requirements are tied to risk weights: the higher the assigned risk, the more capital a bank must hold against a loan. That makes risk weights relevant to lending decisions, since holding capital carries a cost. Across the banking system, changes in those costs can influence the mix and price of credit available to businesses and households.

Banks may calculate risk weights through approved internal models, using borrower information such as the likelihood of default and the potential loss. The rationale is that banks with suitable expertise can assess their own portfolios in detail. But studies have found that internal models can produce different risk weights for portfolios with similar underlying risks. This variation has raised questions about whether model-based capital requirements are consistent across banks.

The 2017 Basel III finalisation package introduced the output floor as a common benchmark. A bank’s risk-weighted assets (RWAs) cannot fall below 72.5% of the amount implied by standardised approaches. If internal models produce a lower figure, the floor becomes binding and raises the RWAs used to calculate the bank’s capital requirement.

Standardised approaches use fixed risk weights. Those weights have less sensitivity to differences in risk than internal models, but applying the same benchmarks across banks can limit unwarranted differences in capital requirements. The floor therefore puts a boundary around the reductions that internal models can produce.

In the UK, the floor is scheduled to phase in from 1 January 2027, beginning at 60% and rising to 72.5% on 1 January 2030. As it rises, it is expected to bind for more banks, particularly those with large exposures to assets where internal and standardised calculations differ substantially. The analysis focuses on banks whose choices are not already constrained by the leverage ratio, a separate backstop based on capital and total exposures rather than risk weights.

Different loan types face different capital costs

A Bank of England staff working paper by Acosta-Smith et al (2026) examines how banks might respond over the economic cycle. The authors extend a macroeconomic model developed by Angelini et al (2014), with banks lending to households through mortgages and to nonfinancial companies. They compare a system based entirely on internal risk weights with one that also includes the output floor.

In a simulated expansion driven by a positive technology shock, the floor moderates the rise in the credit-to-GDP ratio. It prevents RWAs from falling as far as they would under internal models alone, keeping capital requirements tighter during the expansion. Lending still increases, but the mix changes: with the floor, mortgage growth is more modest and corporate lending grows more strongly than in the model without it.

By the end of the simulation period, mortgages are 0.4 percentage point lower with the floor, while corporate loans are around 0.2 percentage point higher. The differences persist across the simulated period. These are results from a model, rather than observed changes in UK lending, but they show how a capital constraint can affect both the total supply of credit and its allocation.

The mechanism depends on the gap between internally modelled and standardised risk weights, known as the IM–SA gap. Standardised weights are generally higher, but the size of the gap varies by asset class. Where the gap is large, the floor is more likely to raise a bank’s capital requirement. Where it is small, the floor may have little effect and can even sit at or below the internally modelled weight.

For UK banks, mortgages generally have lower RWAs than corporate loans under internal models, but a larger IM–SA gap. That means the floor tends to be more binding for mortgages. In the simulation, the floor-implied weights—72.5% of standardised risk weights—are above internally modelled weights for mortgages but below them for corporate loans. Once binding, the floor therefore makes mortgage lending more capital-intensive relative to corporate lending, giving banks an incentive to shift their portfolios towards company loans.

The model groups mortgages together, though actual home loans span a wide range of borrower and property risks. The researchers’ interpretation is that safer mortgages tend to have larger IM–SA gaps, while riskier mortgages tend to have smaller ones. If that relationship holds and other factors stay constant, the floor could steer banks away from the safest mortgages and towards some riskier ones, alongside corporate lending. The model’s simplified balance sheets make this a directional implication, not a precise forecast of how every bank will respond.

Banks may adjust before 2030

The floor reaches its full level in 2030, but banks need not wait for that date to change their decisions. Their response will depend in part on how they anticipate the future requirements. Research on earlier changes suggests that banks often adapt when proposed rules become clear, before the new arrangements take effect.

Fritsch and Siedlarek (2022), studying the Basel III reforms in the United States, found that banks began adjusting their regulatory capital positions shortly after proposed rules were announced in 2012—years before the new regime took effect. The response was especially clear among regional banks, which were typically more sensitive to supervisory scrutiny and had stronger incentives to remain comfortably above capital thresholds.

Anticipatory changes can extend beyond capital ratios. Hendricks et al (2023) documented changes in reporting, lobbying and business models that reduced banks’ exposure to proposed Basel III rules before they were finalised. Looking at a different change, Guillaume et al (2020) found that UK banks eligible for Pillar 2A capital relief began adjusting their asset mix after the relevant statement was published, rather than waiting until the relief formally applied.

That history makes early portfolio adjustments a plausible part of the output floor’s effects. Banks expecting the floor to bind may start changing lending strategies during the phase-in, although evidence on whether and when that happens will be needed to assess the scale of any shift.

What remains uncertain

The model points to a reallocation towards corporate loans and some riskier mortgages, with less lending to the safest mortgage borrowers, all else equal. The effect could support credit for productive investment, but it also changes which households and businesses face the strongest lending incentives. The direction and size of the response will depend on how banks’ portfolios compare with the model’s assumptions and which capital constraints bind for them.

A key open question is how the leverage ratio interacts with the floor. The analysis focuses on banks for which that separate constraint does not determine lending choices. Further empirical work can test whether a binding leverage ratio dampens the portfolio shifts suggested by the model, and whether banks begin adjusting before the floor reaches its 2030 level.