Search Options
Home Media Explainers Research & Publications Statistics Monetary Policy The €uro Payments & Markets Careers
Sort by
Francesca Barbiero
Senior Economist · Monetary Policy, Monetary Analysis
Maria Dimou
Economist · Monetary Policy, Monetary Analysis

Credit risk and bank lending conditions

Prepared by Francesca Barbiero and Maria Dimou

Published as part of the ECB Economic Bulletin, Issue 4/2024.

Credit risk has been gradually increasing in recent quarters, but has not reached the levels of deterioration implied by headline measures of bank credit risk based on historical regularities, given a weak economic outlook for the euro area, higher interest rates and rising numbers of bankruptcies. Both non-performing loan ratios and broader measures of credit risk, such as early arrears and underperforming (Stage 2) loans, have been steadily increasing in recent quarters, with some heterogeneity across countries resulting from, for instance, different exposures to more interest rate sensitive sectors like commercial real estate. However, the increase has remained contained overall.[1] In addition, probabilities of default on banks’ balance sheet exposures have barely moved up since the start of the recent monetary policy tightening cycle, despite an increased interest rate burden and a worsened economic outlook.[2] In general, reported default frequencies are below the levels that might be expected based on historical regularities, given the current macroeconomic outlook.[3]

Part of the benign evolution of credit risk on bank balance sheets may be attributable to the fact that, since the start of the tightening cycle, banks have actively reallocated their portfolios towards safer assets (Chart A). Looking at banks’ corporate loan portfolios from the end of 2021, as reported in AnaCredit, the euro area credit register, there was indeed a deterioration in borrower creditworthiness, likely reflecting higher interest rate burdens and the gradual weakening of macroeconomic conditions (Chart A, yellow bars). This would have been expected to prompt a marked deterioration in the loan portfolios of banks. However, on top of the broad-based contraction in the supply of credit since the start of monetary policy tightening, banks have been selectively shifting credit origination towards safer counterparties in their corporate loan portfolios (Chart A, blue bars). This rebalancing has more than compensated for the passive deterioration in the quality of banks’ exposures stemming from changes in the creditworthiness of pre-existing counterparties. As a result, the share of less risky loans in corporate loan portfolios (Chart A, blue circles) actually increased, leading to a lower than expected deterioration in the overall quality of outstanding loans. This is also consistent with the increase in risk aversion reported by banks in the euro area bank lending survey (BLS) since the start of the tightening cycle, paired with increased risk perceptions.[4] The impact of safer and potentially more liquid assets being seen as more attractive also extended to banks rebalancing assets towards securities holdings rather than loans, which is a pattern that largely reflects securitisation activity.[5]

Chart A

Decomposition of developments in the corporate loan portfolios of banks between the fourth quarter of 2021 and the fourth quarter of 2023

(changes in shares of loan volume and contributions, percentage points)

Sources: ECB (AnaCredit) and ECB calculations.
Notes: This chart shows the cumulative quarterly changes in the share of loans belonging to each borrower category from the first quarter of 2022 to the fourth quarter of 2023. The borrower categories reflect mapping between realised default rates and ratings as provided by rating agencies. “Safer borrowers” includes borrowers with a probability of default of up to 5%. “Riskier borrowers” includes borrowers with a probability of default that is higher than 5%. The calculations are based on outstanding amounts of loans granted by euro area banks to euro area residents. The definitions are consistent with Regulation ECB/2021/2 (BSI Regulation). The blue bars represent cumulations of quarterly changes in the shares of loans within each category, calculated while applying the probabilities of default observed in the previous quarter.

A reallocation of bank portfolios towards safer assets could reflect their attempts to contain the cost of credit risk by avoiding a sharper increase in loan loss provisions and broader measures of credit risk. After reaching record levels in the aftermath of the global financial crisis and the sovereign debt crisis, the cost of credit risk has remained contained over recent years (Chart B, green area).[6] As a result of increased supervisory pressure, banks have undertaken costly efforts to clean up their balance sheets and faced continuous scrutiny of their risk management practices. Costs related to credit risk, in combination with other operational inefficiencies, contributed to low profitability levels, high cost of equity and a lack of capital distributions in the years before the recent tightening cycle, and especially during past crises. Therefore, pronounced risk-off attitudes would be consistent with banks’ efforts to keep down the cost of credit risk in order to attain profitability levels that would enable them to continue to distribute capital to their investors.[7] Currently, as a result of these adaptive banking strategies, loan loss provisions and broader measures of credit risk remain contained despite renewed scrutiny from supervisors and markets. For the same reason, loan loss provisions and measures of credit risk in bank balance sheets have partially lost their ability to fully reflect the level at which the tightening of monetary policy has been affecting firms and households. This suggests that a more holistic view of monetary policy transmission to the real economy is warranted. It likely also reflects the fact that firm and household balance sheets were in a favourable state at the start of the tightening cycle, which, coupled with strong profitability and employment, may have prevented a greater deterioration in creditworthiness.

Chart B

Lending rate to non-financial corporations and its components

(percentages per annum)

Sources: ECB (balance sheet items, MFI interest rate statistics), Bloomberg, Moody’s and ECB calculations.
Notes: This chart decomposes the realised lending rate to non-financial corporations (blue line) into contributions from bank cost components. The residual between the realised lending rate and the various cost components identifies a measure of intermediation margin. The costs of deposits, bank bonds and money market and ECB borrowing are expressed as spreads vis-à-vis the base rate, i.e. the three-year overnight index swap rate (black line), weighted by their respective importance in banks’ funding mixes. The latest observations are for December 2023.

The empirical evidence suggests that banks became more prudent in the allocation of credit as a result of regulatory and supervisory considerations (Chart C, panel a). Heightened supervisory scrutiny compels banks to adopt a more risk-averse stance in their lending activities.[8] As regulatory pressure mounts, banks become increasingly motivated to comply with capital adequacy standards and mitigate their exposure to risky assets. Consequently, banks strategically reallocate credit towards safer borrowers or less risky assets in order to adhere to regulatory requirements and supervisory pressure, and to avoid unwarranted increases in loan loss provisioning needs ex post. The role of regulatory and supervisory pressure is captured via banks’ replies to the BLS question on the impact on credit standards of supervisory or regulatory actions in the previous 12 months. Empirical analysis suggests that banks that reported a tightening impact from supervisory or regulatory actions significantly reduced lending to riskier borrowers compared with lending to safer borrowers over the horizon considered.[9] Specifically, a bank tightening credit standards because of supervisory and regulatory pressure to the same extent as the average bank between 2021 and 2023 (cumulated net percentage of 20%), reduced lending to riskier borrowers by 3 percentage points, which compares with an overall 9.8% decline in loan volume for the sample of riskier borrowers over the horizon considered. This supports the hypothesis that, as a consequence of ongoing scrutiny by prudential authorities, banks may have been prudent in their allocation of credit to borrowers with weak repayment prospects, notwithstanding the fact that other factors, like a lower demand for loans, may also have played an important role for these borrowers.

Chart C

Empirical evidence on factors driving allocation of credit supply

a) Impact of regulatory/supervisory pressure

(percentage points, ex ante loan volumes)

b) Impact of ex ante capital buffers and excess liquidity

(percentage points, ex ante loan volumes)

Sources: ECB (individual balance sheet items, AnaCredit), BLS, ECB supervisory reporting and ECB calculations.
Notes: The chart displays the coefficients from the regression: Loan growthb,f=Riskierf+ β1Saferf×Xb+β2Riskierf ×Xb +Yb+Zf+γILS+ϵb,f, where Loan growthb,f is the (log) change in loan volume at the bank-firm level between December 2021 and September 2023. Saferf and Riskierf are complementary dummy variables equal to 1 if the ex ante probability of default of the borrower is above (below) 5% respectively. Xb  is a variable equal to 1 if supervisory or regulatory actions in the previous 12 months had a tightening impact on credit standards (panel a) or the ex ante capital buffer or the excess liquidity over asset ratio (panel b). The regression includes bank and firm controls (Yb and Zf), such as the size of the bank and the age of the firm, as well as industry-location-size fixed effects (γILS). Standard errors are clustered at the bank level. The coefficients are rescaled by the cumulated net tightening of credit standards from regulatory/supervisory pressure over the period 2021-23, averaged across loans to small and medium-sized enterprises and large corporates, as reported in the BLS.

By contrast, bank balance sheet strength and absorption capacity before the tightening cycle does not seem to have significantly affected the qualitative allocation of credit (Chart C, panel b). Ample capital and liquidity buffers affect the way monetary policy is transmitted and can interact with banks’ credit risk management.[10] At the start of the recent tightening cycle, banks held large levels of liquidity and their capital buffers significantly exceeded the regulatory requirements. This was the result of years of stringent regulatory and supervisory scrutiny, which rendered the euro area banking system more resilient to shocks such as the banking turmoil of March 2023. However, the ex ante levels of capital buffers and excess liquidity ratios do not appear to be significant in explaining the shifts in the allocation of credit towards safer and riskier borrowers over the tightening cycle, whereas regulatory and supervisory pressure does seem to play a significant role, as discussed above.

  1. See, for example, the series for non-performing loans and for Stage 2 loans in the ECB’s supervisory banking statistics.

  2. See, for example, the box entitled “Corporate vulnerabilities as reported by firms in the SAFE”, Economic Bulletin, Issue 1, ECB, 2024.

  3. See also af Jochnick, Kerstin “The single supervisor ten years on: experience and way forward”, LBBW Fixed Income Forum, Frankfurt, 13 March 2024, and “Euro area banking sector”, Financial Stability Review, ECB, May 2024.

  4. See “The euro area bank lending survey – first quarter of 2024”, ECB, April 2024.

  5. The rebalancing of bank assets towards securities from the end of 2021 was not associated with changes in the overall risk profile of the loan books.

  6. In this context, the cost of credit risk is the sum of capital charges and expected (credit) losses.

  7. See the box entitled “Banks’ capital distributions and implications for monetary policy”, Economic Bulletin, Issue 6, ECB, 2023.

  8. Altavilla, C., Boucinha, M., Peydró, J. L., and Smets, F., “Banking supervision, monetary policy and risk taking: big data evidence from 15 credit registers”, Working Paper Series, No 2349, ECB, January 2020.

  9. To test for alternative explanations that could be underpinning bank behaviour when it comes to credit risk, we constructed a bank-firm dataset by linking AnaCredit with banks’ individual replies to the BLS, along with balance sheet positions from individual balance sheet item statistics. We focused on the change in loan volume at the bank-firm pair level between December 2021, before the tightening cycle, and September 2023. To characterise factors that affected the change in the relative allocation of credit to ex ante safer or riskier borrowers over this period, we employed a standard empirical approach that allowed us to include single-bank firms when identifying bank credit supply shocks by assuming that shocks affect firms within the same industry, location and size classification in a similar way (see Degryse, H., De Jonghe, O., Jakovljević, S., Mulier, K., and Schepens, G., “Identifying credit supply shocks with bank-firm data: Methods and applications”, Journal of Financial Intermediation, Vol. 40, October 2019.) We then specify a cross-sectional regression model, where we include bank and firm controls, such as the size of the bank and the age of the firm, as well as industry-location-size fixed effects as described above.

  10. In particular, reallocation of credit from riskier to safer borrowers would be expected to be less prevalent among banks with higher capital and excess liquidity levels, in keeping with Gambacorta, L. and Shin, H.S., “Why bank capital matters for monetary policy”, Journal of Financial Intermediation, Vol. 35, July 2018, pp. 17-29.