By Isabelle Michalski-Roland and Alexander Prior.
Policymakers have debated SME finance for decades. While bank lending remains the main source of external finance for most SMEs, obtaining credit can be difficult for some firms. Lending decisions often rely on financial track records and collateral, which can disadvantage younger firms and businesses with fewer tangible assets. Financing frictions also appear to be unevenly distributed. Firms in rural or less densely populated regions may face greater barriers to bank finance. UK evidence indicates that innovative SMEs in less accessible regions are more likely to encounter rejection and to be discouraged from applying for credit for fear of rejection, while other research identifies a location-related penalty in the allocation and pricing of SME loans (Lee and Brown (2017) and Cowling et al (2020)).
One potential response to these financing frictions has been the growing adoption of financial technology by banks. Britain’s largest high-street lenders have increasingly partnered with fintech firms to streamline lending processes and expand digital capabilities. Using data on more than 25,000 unsecured SME loans, this article investigates whether bank-fintech partnerships expand credit supply and improve access to finance in underserved markets. We find that fintech helps banks expand and cheapen SME lending, narrowing urban-rural lending gaps, but it does not solve the financing challenges of higher-risk firms.
Fintech adoption allows banks to make larger and cheaper loans to SMEs
The study matches loan-level data from Experian with firm characteristics from Bureau van Dijk’s (BvD’s) FAME database and bank-level data from the Bank of England, covering SME loans originated by the ‘Big five banks’ (HSBC, Barclays, Lloyds Banking Group, NatWest Group, and Santander UK) between 2018 and 2024.
To measure fintech adoption, we construct a novel measure of a bank’s ‘fintech intensity’, defined as the cumulative number of partnerships with fintech developers focused on SME lending technologies. We identify these partnerships from CB Insights press release announcements and classify them manually by function. Between 2011 and 2024, we identify 50 partnerships between fintech developers and the ‘Big five banks’, of which 12 relate specifically to SME lending technologies. Lending technologies are the third most common category of partnership after payments and client operations (Chart 1).
Chart 1: Partnership types by function
Number of partnerships between fintech developers and the ‘Big five banks’, by function, 2011–24
Footnotes
- Notes: Categories are not mutually exclusive, so the bars sum to more than 50 distinct relationships.
- Sources: CB Insights and authors’ calculations.
Partnership activity was minimal in 2011–14 but accelerated from 2015 onwards (Chart 2).
Chart 2: Fintech intensity over time
Cumulative number of partnerships between fintech developers and the ‘Big five banks’, total versus SME lending-related, 2011–24
Footnotes
- Sources: CB Insights and authors’ calculations.
The headline results are striking. A one-unit increase in fintech intensity is associated on average with a 5.4% increase in loan volumes and a 21.1% reduction in interest rate spreads at origination. The identification exploits variation in the timing of individual banks’ fintech partnerships, controlling for region, industry, bank, and year-month fixed effects to isolate supply-side effects from demand-side heterogeneity. The results are robust to more granular fixed effects (bank and combined region-industry-time fixed effects) as in Degryse et al (2019). On a mean loan of £203,000, the volume effect implies roughly £11,000 in additional credit per loan; on a median loan of £22,000, it implies roughly £1,200. On a mean annualised spread of 10.3%, a 21.1% reduction in the monthly spread is equivalent to approximately 225 basis points per additional fintech partnership, and on a median annualised spread of 6.3%, to approximately 135 basis points. Our findings are qualitatively related to Core and De Marco (2024), who show that banks with stronger IT capabilities provide more and cheaper guaranteed loans to small businesses.
Fintech acts as an ‘equaliser’ across urban and rural areas
A key question is whether fintech helps reduce geographic disparities in access to finance. We find evidence of an urban-rural credit gap: SMEs located in low-population-density areas receive loans that are around 14% smaller than otherwise comparable firms in more densely populated regions. To test whether fintech adoption mitigates this disparity, we interact fintech intensity with a rurality indicator equal to one for regions in the lowest quartile of population density. The results suggest that it does. While fintech adoption is associated with roughly a 5% increase in loan volumes in urban areas, the corresponding effect exceeds 13% in rural areas (Chart 3). This suggests that digital lending technologies disproportionately expand credit in places where financing constraints are most pronounced. In this sense, fintech acts as an urban-rural equaliser, narrowing existing geographic disparities in access to credit. By contrast, we find no comparable differential effect on spreads, implying that the rural impact operates primarily through greater credit availability rather than lower borrowing costs.
Chart 3: Rural areas see the largest gains in lending volumes
Impact (in per cent) on lending volumes at origination, urban versus rural areas
Footnotes
- Notes: Estimated percentage increase in loan volumes at origination per unit of fintech intensity. Estimates and their 95% confidence intervals.
- Sources: Bank of England, BvD, CB Insights, Experian and authors’ calculations.
The benefits of fintech accrue mainly to lower-risk firms
Borrower risk is a key dimension along which the effects of fintech adoption vary. We classify firms by quartile of their QuiScore, a proprietary credit risk indicator provided by BvD FAME. It ranges from 0 to 100, where higher values denote lower credit risk and stronger financial health. We use the QuiScore at loan origination as our measure of borrower creditworthiness and classify firms by quartiles of the score distribution. We denote the riskier tail of the distribution (first quartile) as high-risk borrowers. For loan volumes, the effects are concentrated exclusively among lower-risk SMEs. Borrowers above the 25th percentile of the QuiScore distribution experience a 6.4% increase in loan volumes. For high-risk firms below the 25th percentile the coefficient is economically small and statistically insignificant.
Rural amplification likewise operates only for lower-risk borrowers, whose loan volumes in low-density areas rise by roughly 16% (Chart 4). High-risk firms see no volume benefit regardless of location. Moreover, the rural amplification effect increases monotonically across the QuiScore distribution: from 7.1% for borrowers in the second quartile (Q2), to 8.0% in Q3, to 13.3% in Q4. Spreads tell a more nuanced story. Both lower-risk and higher-risk SMEs experience some reduction in borrowing costs, but the magnitude is larger for lower-risk borrowers: roughly 22% against 17% for higher-risk. Geographic amplification in spreads is absent for both groups.
Chart 4: The benefits are concentrated among lower-risk borrowers
Impact (in per cent) on lending volumes at origination, urban versus rural areas, for low risk and high risk borrowers
Footnotes
- Notes: Estimated percentage increase in loan volumes per unit of fintech intensity. Estimates and their 95% confidence intervals. Higher-risk borrowers are defined as those in the first quartile of the QuiScore distribution (the risky tail). Lower-risk borrowers are defined as those above the 25th percentile of the distribution.
- Sources: Bank of England, BvD, CB Insights, Experian and authors’ calculations.
Fintech expands lending through lower costs, not better screening
Why does fintech adoption increase lending? The literature points to several possible explanations. Among those are better screening and cost reduction. We build a model that lets the data distinguish between the two because it makes different predictions for risky borrowers.
Under the better-screening view, fintech tools help banks identify borrower quality more accurately. Lower-risk firms, previously pooled with riskier borrowers, gain access to more credit at lower rates when screening precision improves. Higher-risk firms, by contrast, move the other way: they lose the benefit of pooling and receive less credit at higher rates. Under the cost-reduction view, fintech makes lending cheaper rather than more informative. Digital lending platforms can process firms with a strong hard-data footprint quickly and at low cost using information such as filed accounts, payment records, and transaction data. Because firms with a hard-data footprint are disproportionately lower-risk, the gains land mostly on them, especially in areas that were previously expensive to serve. For riskier and more opaque borrowers, where banks still rely heavily on relationship lending and soft information, the effect should be limited. The key difference between the two channels is, therefore, what happens to higher-risk firms. Better screening predicts less lending and higher borrowing costs. Cost reduction predicts little change. That is what allows the data to identify the dominant channel.
The UK evidence points firmly to cost reduction. The lending gains are concentrated among lower-risk firms, while higher-risk firms see no increase in loan volumes. The rural lending boost is also confined to lower-risk borrowers. Taken together, these findings suggest that fintech partnerships are making SME lending cheaper and easier to deliver, rather than fundamentally improving banks’ ability to assess borrower risk. The partnerships in our sample involve cloud-based credit platforms, software-as-a-service underwriting tools and digital process automation. These technologies help banks process applications more quickly and at lower cost, but they do not typically provide new information about borrowers. Instead, they allow banks to make more efficient use of the information they already collect. Recent Bank of England research has shown that SME lending is less profitable than lending to larger corporates, in part because of higher operating costs. Our findings suggest that fintech adoption helps address exactly this challenge.
A different story emerges when new sources of borrower data are combined with advanced analytics. Chen et al (2025) studied a Chinese bank that first adopted machine-learning models and then separately integrated rich alternative data, including transaction records, VAT invoices and unstructured text. The machine-learning upgrade alone generated only modest improvements in screening accuracy. It was the addition of new borrower information that more than doubled the gains in accuracy. This distinction helps explain our results. The fintech partnerships used by UK banks today appear to reduce the cost of processing existing information, rather than substantially improving the information available for credit assessment. As a result, fintech complements traditional relationship lending rather than replacing it, particularly for borrowers whose creditworthiness remains difficult to assess using standard data.
What this means for the growth mandate
The credit expansion is concentrated among lower-risk borrowers, suggesting that fintech adoption has primarily supported lending growth among firms with stronger credit profiles. The amplified rural effects suggest fintech can partially address the persistent geographical inequalities in UK SME finance. As the Government’s Financial Services Growth and Competitiveness Strategy prioritises fintech development, these results indicate where the returns are likely highest.
However, higher-risk SMEs see no increase in loan volumes. Many of them are among the most credit-constrained because they are informationally opaque and require relationship-based assessment. One important question is whether the higher-risk firms that do not benefit from banks’ fintech adoption are also firms with strong growth prospects. Our analysis does not directly examine firm growth, focusing instead on conventional measures of credit risk. Nevertheless, high-growth firms are often perceived as riskier because they are younger, expand rapidly, or lack established credit histories. To the extent that growth potential and measured risk overlap, our findings suggest that digitalisation may not automatically channel credit towards all high-potential firms. While fintech adoption appears to expand lending to lower-risk SMEs, complementary policies may still be needed to support firms that combine strong growth opportunities with greater credit risk.
Looking ahead, if cost reduction is the dominant channel, the frontier of fintech’s impact may depend less on algorithmic sophistication and more on the data environment. For traditional banks to unlock the information-precision channel for currently underserved, opaque (and therefore potentially a priori higher-risk) SMEs, the binding constraint may be access to richer borrower data rather than better algorithms. Whether banks can acquire such data, through partnerships, regulatory frameworks, or their own digital ecosystems, is an important open question.
Open finance has the potential to be part of the answer. By expanding access to borrower data, it could enable fintech tools not only to reduce costs but also to make more accurate credit assessments, particularly for firms that are currently difficult to evaluate using traditional information sources.
Bank Insights articles do not necessarily represent the views of the Bank of England’s policy committee members.