What drives differences in commercial banks’ product level returns?

By Colm Manning, Prashant Babu, Ben Bourdais and Robert Włodarski

The UK SME lending market has diversified in recent years, with challenger banks, fintechs, and non‑bank lenders playing a growing role. However, it remains predominantly bank based, particularly for larger and more established businesses, and the non-bank lenders’ share of SME lending remains modest. This article therefore focuses on how incentives vary within large banks across lending segments. It does not assess competitive differences between banks and non‑banks, recognising that banks operate under a distinct prudential regime reflecting their role as deposit‑taking institutions.

The rate of return on bank lending is a key consideration in banks’ choice where to deploy capital.

Like any business, banks aim to generate a RoE in a way that aligns with their strategic objectives. They do this in a way that manages constraints related to capital, liquidity and risk appetite. They adjust their customer lending rates, based on their funding costs and given the competition they face, to compensate them for the risk they take when lending as well as their operating and other costs. The composition of a bank’s product mix is influenced by the structure of its funding base, areas of relative competitive advantage, opportunities for portfolio synergies and cross-selling. Banks also actively manage their portfolios to achieve a target risk-weighted asset profile to ensure that capital is allocated efficiently in line with internal risk-return thresholds and external regulatory requirements.

Our analysis suggests that SME lending tends to result in the lowest RoE for large UK banks.

RoE compares the net income a bank earns from lending (the numerator) to the equity allocated against the associated risk of that lending (the denominator). So different RoEs can reflect differing net incomes or different equity requirements.

Estimating product-level RoE is complex and subjective as we must allocate revenue, costs, and capital across different types of lending. We can measure banks’ group and entity-level RoE from published data, but banks’ typically do not disclose profitability by product or their internal RoE methodologies. We take the average lending rate and deduct our best estimate for average funding, impairment, and operating costs. The result is expressed as a per cent of equity, which we assume to be a fixed share of average risk-weighted assets. We estimate ranges of RoE to reflect monthly variability in customer lending rates, differences across banks, and plausible assumptions regarding cost allocation, particularly fixed and variable costs related to origination, servicing, and funding (Chart 1).

Chart 1: Estimated large bank average product level return on equity is typically lowest for SME lending

Footnotes

• Notes: Data as at 2025 H1. Based on aggregated large bank data except for lending rates, which use all banks. We assume funding costs are same across all products. RoE = ((1-𝜏)×(𝑁𝑒𝑡 𝑖𝑛𝑐𝑜𝑚𝑒))/𝐸𝑞𝑢𝑖𝑡𝑦 where 𝜏 is the tax rate assumed to be 25%, equity is assumed to be 15% of risk-weighed assets, net income = (Customer rate - funding rate) - (impairments + operating costs).
• Sources: Bank of England, Prudential Regulation Authority (PRA) and Bank calculations.

It’s worth noting that our results may look different in both the level and ordering relative to those produced by any specific bank. There are many valid ways to calculate product-level RoE, with assumptions varying across banks due to differing strategic priorities. Some banks may not segment lending this way at all. One challenge not captured lies in credibly quantifying additional benefits associated with lending, such as the potential to deepen client relationships, support affiliated business activities, or the benefits associated with holding customer deposits. Understanding banks’ lending incentives in practice is made more complex as they may take account of marginal earnings and costs when deciding to lend rather than the average figures we’ve used in our analysis. Despite these challenges, we’ve made a best-effort estimate of SME RoE relative to retail and large corporate lending for large banks. While imperfect, this exercise is valuable in prompting further thinking about returns across lending segments and incentives to lend to those segments.

Capital requirements do not appear to be the key driver of different RoE between corporate and SME lending.

Capital requirements, driven by risk weights on lending, are the main determinant of the equity that banks must hold against an exposure. These risk weights vary materially across asset classes based on the inherently different risk profiles of different forms of lending. As a result, large differences in risk weights across asset classes are a significant driver of returns on equity that banks can earn on different types of lending. At a more conceptual level, there is a trade-off. Materially lower risk weights can improve lending returns but may inadequately capture differences in underlying risk that we discuss later.

For the rest of this article, we’ve focused our comparison between SMEs and large corporate lending. As set out below, the differences in risk weights between these forms of lending in practice mean that higher average capital requirements play only a minor role in explaining the RoE gap between SME and corporate lending (Chart 2).

Chart 2: Higher impairments and operating costs are the main driver in lower average SME RoE relative to larger corporates

Footnotes

• Notes: Data as at 2025 H1. The chart shows the main drivers of the estimated RoE gap between SME and large corporate lending, relative to the baseline corporate RoE in Chart 1. Contributions reflect differences in average customer rates, banks’ expected impairment charges (proxied using observed provisioning and impairment data), operating costs, and risk-weighted capital intensity, based on aggregated large bank disclosures and supervisory data. The decomposition is illustrative rather than a statement of banks’ internal profitability frameworks and is intended to highlight the relative economic drivers of the RoE gap under a consistent and transparent set of assumptions.
• Sources: Bank of England, PRA and Bank calculations.

Differences in capital requirements between SMEs and larger corporates are driven by several factors – notably different risk profiles – and are an intended feature of our risk‑sensitive regime.^{footnote [1]} As Table A shows, SMEs have higher idiosyncratic risk than larger corporates reflected in higher average probabilities of default and, for retail SME, higher loss given default. These feed directly into banks’ internal ratings-based (IRB) models and were considered by the PRA when it calibrated standardised approach risk weights under its implementation of Basel 3.1.

All else equal, both unrated corporate SME and retail SME loans attract lower risk weights for a given PD and LGD than larger unrated corporates under the IRB approach.^{footnote [2]} This is because SME outcomes are generally considered to be driven more by firm-specific factors and are diversified across many smaller debtors. They are therefore assigned a lower asset correlation than larger corporates in the regulatory framework under IRB, which pushes down risk weights for SMEs.

In addition, most SME lending actually benefits from a downward adjustment to capital requirements beyond what the purely risk-based assessment implies. Since 2014, this has been delivered through the SME support factor, which reduces risk-weighted assets by between 15% and 24%, which has, all else equal, lowered the equity capital required to support eligible SME lending relative to comparable lending to other corporates. From next year, the PRA will replace the SME support factor with a similar favourable treatment through a firm-specific SME lending adjustment to capital requirements.^{footnote [3]}

Taken together, capital requirements in practice play a small role compared to other drivers of the estimated RoE gap between SME and large corporate lending. This reflects the fact that that some SMEs loans attract lower risk weights for a given PD and LGD than larger unrated corporates, and policies that downwardly adjust SME capital requirements beyond what the purely risk-based assessment implies.

Differences in impairment rates and operating costs are the main drivers of differences in RoE between SME and larger corporate lending.

Expected impairments account for the largest difference in the RoE gap between SMEs and corporates. For lenders, these are expected credit losses under IFRS 9 that are forward‑looking, rather than just realised losses, and incorporate probability‑weighted loss outcomes. For our analysis, we use reported impairment charges and provisions as a proxy for these expected credit losses. They are the largest driver of RoE differences because, as already mentioned, SME lending in practice is riskier than lending to larger corporates. There are several reasons for this, including: 1) historical evidence that indicates SMEs have a higher probability of default than larger firms – Table A, 2) many small businesses rely heavily on a single market or supplier, amplifying operational risk, 3) SMEs generally maintain thinner capital buffers and less liquidity, reducing their capacity to absorb losses, and 4) smaller firms typically provide less frequent financial reporting that may mean issues take time to surface and complicates credit risk assessment. Large banks may be able to reduce impairment rates on SME lending by using advanced digitalisation and big‑data analytics, including using payment data, to improve real‑time risk assessment, enhance borrower monitoring, and enable more accurate forward‑looking credit decisions. This is something that smaller banks that specialise in SME lending say they are able to do.

Operating costs are also an important driver of the RoE gap. These are typically higher for SMEs relative to larger corporates as most underwriting, compliance, monitoring, and servicing costs are incurred per borrower rather than per pound lent. A larger corporate typically borrows a larger amount relative to a SME, which means these costs are lower given the size of the loan. There is scope to further reduce these costs through automation and digitalisation of processes. Although it is unlikely that operating costs could be reduced to the same level as large corporate loans given the smaller typical SME loan size.

Loan approval processes can also be resource intensive for business owners. These assess the financial health of the borrower, their credit history, and sometimes the owners’ personal credit rating. They often require a clear revenue record and a strong asset base to use as collateral. These requirements can be a barrier for early-stage or fast-growing businesses that reinvest heavily, have little physical assets, and may not yet be profitable. Although, schemes like the British Business Bank’s Growth Guarantee Scheme, that provides a guarantee on a percentage of eligible loans, can help de-risk SME loans and make them more attractive to lenders.

There’s the potential for large banks to reduce costs and impairments and boost RoE on SME lending through investment in automation and digitalisation.

The lower typical RoE that the average large commercial bank makes on SME lending, and other factors mentioned above, mean there can often be less incentive to lend to SMEs relative to other product lines. In addition, given their business models, commercial banks are typically not the primary growth funders of high potential growth scale-ups that are growing rapidly but are not yet cash-flow stable, something discussed in Unlocking growth: what can the literature tell us about what’s holding back high-growth firms?. While banks are part of the broader funding ecosystem, and provide banking services to all SMEs, their role is often limited to the more established end of the SME population. They provide some of the cheapest debt available, but this comes with risk filters that means only businesses with established revenue streams and tangible collateral tend to benefit.

Taking steps to lower delivery and monitoring costs would be instrumental in improving the economics of SME lending. SME loans are typically smaller, more heterogeneous and more labour‑intensive to assess, which raises operating costs per‑loan and depresses returns. Investment in digitalisation, automation and data‑driven credit processes can materially reduce these costs by streamlining underwriting, improving risk assessment, and enabling more efficient ongoing monitoring. Standardised data, automated decisioning and early‑warning tools can also help manage default risk and impairment rates. These supply-side improvements could be complemented by addressing demand-side constraints highlighted in the literature that include borrower discouragement, limited awareness of available finance options, and informational frictions that deter some SMEs from applying for bank finance in the first place. Together, policies that improve banks’ cost efficiency and risk management while reducing application frictions and improving transparency for borrowers could help bring forward productive investment, raising risk‑adjusted returns on SME lending while maintaining robust and proportionate prudential safeguards.

Share your thoughts with us at Bankinsights@bankofengland.co.uk