Bank lending is a key channel through which monetary policy affects the real economy. This column explores how the effects of monetary policy on credit are shaped by bank specialisation in borrower industries and size categories. Analysing granular loan data reveals that such specialisation is widespread among euro area banks. Moreover, higher specialisation is typically associated with more favourable lending conditions. Most importantly, specialised banks tend to partially insulate their preferred borrower groups from the consequences of monetary policy. These findings suggest that bank specialisation plays a meaningful role for the aggregate and distributional consequences of monetary policy.
The post-Covid inflation surge has brought conventional interest rate policy back to the forefront of central banking, including at the ECB. In this context, the credit channel is widely regarded as an important mechanism in the transmission of monetary policy. The idea is that changes in policy rates affect real economic activity indirectly by affecting the availability of bank credit to households and firms. Understanding how exactly the availability of corporate credit responds to policy rate changes is therefore essential for assessing the full implications of monetary policy.
In this context, much of the existing literature has considered bank and firm characteristics separately, for instance examining the role of bank liquidity (Dell’Ariccia et al. 2017) or firm default risk (Anderson and Cesa-Bianchi 2024). In a recent paper, I instead analyse the interaction between banks and their borrowers, focusing on bank specialisation as a novel feature affecting the transmission of monetary policy (Kuhmann 2025). Loan concentration has previously been identified as a key feature of bank lending in the US (Blickle et al. forthcoming) and explored in various contexts across different euro area countries (e.g. De Jonghe et al. 2020). My paper is the first to study how bank specialisation interacts with monetary policy transmission.
Bank specialisation is defined as the difference between the weight of a borrower group in a bank’s loan portfolio and that group’s weight in the overall economy. It therefore measures the extent of a bank’s over-exposure to particular borrower types. The analysis draws on comprehensive credit register data from AnaCredit, covering all loans to euro area firms exceeding €25,000 between mid-2020 and mid-2024.
Bank specialisation in the euro area
Bank specialisation by borrower industry and size is a widespread phenomenon among euro area banks. Figure 1 contains box plots illustrating the distribution of specialisation values for each bank’s three most preferred categories. Specialisation is generally positive for Top-1 and Top-2 categories, which implies that nearly all banks specialise in one or two specific industries or size groups. This also means that almost no banks in my sample are fully diversified; that is, few have lending shares in their preferred category that align with the economy-wide share. This pattern persists when specialisation is weighted by loan amounts and is also observed among the largest banks in the sample. Specialisation is therefore not limited to small or niche institutions but is a broader characteristic of euro area banking.
Figure 1 Bank specialisation in top categories
Note: Distribution across banks of specialisation values in the respective top categories in 07/2020 and 09/2024. Panel (a) represents specialisation in top industries and panel (b) in top size categories.
Beyond this core finding, the paper documents several additional patterns associated with bank specialisation in the euro area. Specifically, there is substantial heterogeneity in the importance of specialised lenders across borrower industries and size categories. Moreover, larger banks tend to be somewhat less specialised. Finally, specialisation is broadly associated with lower interest rates, larger loan amounts, longer maturities, and higher collateral shares. This is consistent with what Blickle et al. (forthcoming) document in the US context and suggests that specialisation allows for more effective screening and monitoring of borrowers.
Specialisation and monetary policy
How does this pronounced specialisation affect the transmission of monetary policy to interest rates and credit volumes? To address this question, the paper employs a local projections-instrumental variable framework using data aggregated to the bank-industry and bank-size category level. Crucially, the estimation specifications include interaction terms of policy rate changes with different measures of specialisation. Changes in the policy rate are instrumented with high-frequency identified monetary policy shocks based on Altavilla et al. (2019).
The interaction coefficients capture how the degree of specialisation alters the responses of interest rates and credit volumes to changes in monetary policy. Figures 2 and 3 illustrate these marginal effects on rates and credit, respectively.
Specialisation is measured using two binary indicators at the bank-category level. The first dummy variable indicates whether a bank’s specialisation value for a category falls within the highest quartile of the within-country distribution (left columns). The second dummy variable indicates whether a category is the bank’s most preferred one (right columns). In the latter specification, coefficients are estimated controlling for the value of specialisation to isolate the additional effect of being the top category.
Figure 2 shows that in response to an exogenous 25 basis point policy hike, banks raise interest rates up to two basis points less for borrowers in industries or size categories where they are highly specialised (left column). Holding the level of specialisation constant, there is a significant additional dampening effect of being in a bank’s top industry or size category (right column).
Figure 2 Monetary policy effects on interest rates
Note: Dynamic consequences of an exogenous 25 basis point increase in the policy rate on interest rates. Based on data aggregated at bank-industry (top row) and bank-size category (bottom row) level. Column one shows the interaction of monetary policy with the top quartile dummy of specialisation and column two the interaction with the top category dummy. Grey bars represent 68% and 90% confidence intervals based on clustered standard errors by month.
Specialisation is even more relevant for the responses of credit volumes. Specifically, credit among borrowers in the highest industry or size category specialisation quartile falls by around one percentage point less in response to a monetary policy hike (Figure 3, left column). This is a sizeable marginal effect compared to the average credit reduction of around 2% (not shown in the figure). The right column in Figure 3 shows that there is an additional 0.5 to 1 percentage point effect of being in a bank’s most preferred industry or size category.
Figure 3 Monetary policy effects on credit amount
Note: Dynamic consequences of an exogenous 25 basis point increase in the policy rate on log credit amount. Based on data aggregated at bank-industry (top row) and bank-size category (bottom row) level. Column one shows the interaction of monetary policy with the top quartile dummy of specialisation and column two the interaction with the top category dummy. Grey bars represent 68% and 90% confidence intervals based on clustered standard errors by month.
These results suggest that banks insulate borrowers from industries and size categories in which they specialise from monetary policy-induced changes in interest rates and credit. This behaviour is consistent with De Jonghe et al. (2020) who find that banks reallocate credit towards industries where they specialise in response to adverse funding shocks. My findings point to similar behaviour in the context of monetary policy transmission.
Additional results
The baseline results suggest that firms are partially insulated from the consequences of monetary policy when borrowing from specialised banks. The paper explores two extensions to this finding.
First, analysing the ECB’s 2022-23 tightening cycle indicates that the marginal effects of specialisation on interest rates are primarily driven by differences in the treatment of new borrowers (the extensive margin). In contrast, the marginal effects on credit are primarily due to differences among existing borrowers (the intensive margin). Second, industry-level local projections show that monetary policy has a weaker effect on credit in industries that are dominated by specialised banks. Moreover, the share of credit provided by banks that specialise in a given industry increases after contractionary monetary policy. This is consistent with the baseline finding that monetary tightening induces banks to effectively reallocate credit towards industries in which they specialise.
Conclusion
Bank specialisation is highly prevalent among euro area banks and influences the transmission of monetary policy to corporate credit. Specifically, banks adjust interest rates and lending volumes less strongly for borrowers in industries and size categories where they specialise.
As a result, some firms may be less exposed to the effects of monetary policy simply because they belong to a borrower group where specialisation is particularly pronounced. More broadly, monetary policy may be less effective when the most relevant industries or size categories in an economy are characterised by intense specialisation. These findings underscore the importance of considering bank-firm interactions and banking structure when evaluating the transmission of monetary policy.
Source : VOXeu
