The global financial crisis has highlighted the importance of early identification of weak banks. Problems that are identified late typically cost more to solve. In the EU, authorities have agreed to move towards a more centralised prudential system, an important argument being that the banking sectors of EU members have become increasingly interlinked as a result of financial integration.
But given the importance of this topic, there is a surprising lack of rigorous cross-country analysis of determinants of bank distress in the EU. Most of the bank distress studies focus on the US, which had numerous bank failures that provide a rich data set for a “forensic” examination of the determinants of distress (e.g., Wheelock and Wilson, 2000).
New evidence on EU bank distress
There are country studies outside the US, but so far no study has analysed determinants of bank distress in the EU as a whole. The aim of our analysis (Poghosyan and Cihak 2009) is to fill this gap using a comprehensive database of bank defaults in the EU that covers both the global financial crisis and over a decade leading up to it.
We compile a unique data set on distress in EU banks, using two main sources of information. The first source is Bureau Van Dijk’s BankScope database, from which we extract financial data on 5708 banks in the EU countries between 1996 and 2007. The second source is a hand-collected dataset on bank distress extracted from the NewsPlus database powered by Factiva, a Dow Jones company. Underlying the NewsPlus searches is the notion that a bank is in distress when it becomes subject to negative reports in the media.
The NewsPlus searches were performed for each of the banks, and for each year, using a combination of the bank name and keywords designed to capture references to failing banks (such as “rescue,” “bailout,” or “financial support”). When a search for a bank led to some hits, we examined websites of relevant supervisory authorities and media reports in more detail, to confirm that the keywords indeed related to this bank.
Using this strategy, we identified 79 distress events for 54 EU banks since 1997. The number of banks is smaller than the number of distress events, since some banks experienced multiple distress events over time.
In line with economic theory, our results show that better capitalised and more profitable banks are less likely to experience distress in the upcoming year. Similarly, the probability of distress is inversely related to asset quality. Assuming that the higher loan loss provision profile implies a riskier loan portfolio, the positive sign for this variable indicates that the probability of bank distress is influenced by the deterioration of the loan portfolio.
One feature of the results that may seem surprising at first sight is the lack of a liquidity variable. We have included liquidity ratios in our more detailed results, but they did not come out significant. This mostly reflects the fact that our model focuses on longer-term factors of distress. Liquidity problems usually erupt over a much shorter period, and there are sometimes only days between the start of liquidity drain and a failure or intervention. But the liquidity problems are often a manifestation of underlying longer-term issues with the institution’s financial health, which can be captured by the other variables in the regression.
The baseline results also suggest that market discipline exerted by depositors does play a significant role. In line with evidence for some other countries (e.g., Kraft and Galac 2007), we find that banks that “bargain for resurrection” in difficult times by increasing their deposit rates are more likely to experience financial distress in the following year.
Another interesting finding is that the occurrence of financial distress in a bank significantly increases the likelihood of distress in its peers. In our study, this is captured by the significantly positive coefficient on the “contagion dummy” (which equals one for banks with a similar size or a similar balance sheet structure as a failed bank).
Additional variables and robustness checks
We augment the baseline model by introducing several additional control variables. We evaluate whether stock markets impose discipline on banks as suggested by one of the Basel II Accord pillars. Previous research on this topic has led to ambiguous results.
- Papers studying US banks (e.g., Flannery, 1998) tend to find that stock market indicators have a useful predictive content for identifying financial distress.
- Research on other countries, however, is generally less conclusive (e.g., Bongini et al., 2002).
We also introduce the ratio of stock indices for EU banks relative to the FTSE-100 market index and find a significant positive association between this variable and bank distress in the next period.
We then explore the effect of wholesale funding on the likelihood of bank distress. Wholesale lenders tend to be jitterier in the event of financial turbulence, and banks more vulnerable to sudden withdrawals (Huang and Ratnovski, 2008). Recent evidence, such as Northern Rock, provides examples of runs by retail depositors preceded by runs by wholesale lenders. A model with the share of wholesale financing in bank total liabilities as an additional explanatory variable suggests that banks relying more heavily on wholesale financing are more likely to experience distress than banks that are mostly financed by retail depositors.
To test whether bank risks are converging across EU countries, we have estimated models with random effects, both at the individual bank level and at the country level. The key result is that the standard deviation of the random intercept is insignificant at the country level, implying that the EU countries are relatively homogenous in terms of the bank baseline hazard after accounting for the explanatory variables. These findings lend support for establishing common benchmark criteria for banking sectors across the EU countries (see for example De Larosiere, 2009).
To illustrate the economic impact of individual financial ratios, Figure 1 shows the marginal impact at the sample mean for the three variables that were found to have a significant impact on bank the probability of distress: capitalisation, asset quality, and earnings. Comparison of marginal effects across these three determinants suggests that the probability of distress is somewhat more responsive to changes in asset quality and earnings relative to changes in capitalisation. This finding highlights the importance of asset quality and earnings as compared with bank capitalisation for early identification of weak banks.
Figure 1. Marginal effects of capitalisation and asset quality on probability of distress
The impact of variables entering the baseline specification is qualitatively similar across various specifications, providing an indication of robustness of the baseline model. To further assess the reliability of the results, we employed a battery of additional robustness checks with respect to (i) failures in small banks being underreported; (ii) bank distress being affected by common shocks (such as shocks to the euro-dollar exchange rate); (iii) banks being subject to a “stigma effect”, struggling to improve their reputation and preventing repeated incidences of distress; and (iv) effect of dominance of certain countries and certain organisational structures on bank distress. Finally, we have experimented with different sub-periods of the total sample, concluding that although the 2008 distress was different from anything seen in the recent past, much of the mechanics identified in the longer-term data remained in place during the recent crisis.
Using a unique data set, we find that recent financial integration policies in the EU have led to convergence of bank risks across EU members. This could justify a closer coordination of financial supervision in Europe.
Examining the importance of capitalisation and market discipline as two main pillars of the Basel II Accord, we find a significant effect of capitalisation on bank distress. But, its economic impact is lower than the impact of asset quality and earnings. We argue that these two variables should be taken into account in addition to bank capitalisation when designing pan-European benchmarks of sound banking conduct. We also find solid evidence on the importance of market discipline in the EU banking (on the side of both depositors and financial market participants), justifying the importance of transparency and dissemination of information for more efficient EU-wide banking regulation.
Among other results, we provide empirical evidence suggesting the importance of contagion effects in the EU banking. Furthermore, we show that banks operating in more concentrated banking sectors are more likely to experience bank distress relative to banks operating in less concentrated markets. Lastly, we show that bank hazards increase with a higher share of wholesale funding, in line with recent theoretical models emphasising this channel of financial vulnerability.
Further work in this area would benefit from the creation of a unified database of supervisory data in the EU. Country supervisors have access to more detailed indicators than what is publicly available, such as bank exposures to individual sectors and various breakdowns of data by maturity, currency, and performance. Such information, currently unavailable in a single database for the EU, could be used to improve our understanding of the main factors influencing soundness of banks.
Disclaimer: The views presented here are those of the authors and not necessarily those of the International Monetary Fund.
Bongini, Paola., Luc Laeven and Giovanni Majnoni (2002), “How Good is the Market at Assessing Bank Fragility? A Horse Race Between Different Indicators”, Journal of Banking and Finance 26: 1011–31.
De Larosiere, Jacques (2009), Report of the High-Level Group on Financial Supervision in the EU, (European Commission, Brussels).
Flannery, Mark J (1998), “Using Market Information in Prudential Bank Supervision: A Review of the US Empirical Evidence”, Journal of Money, Credit and Banking 30(3): 273-302.
Huang, Rocco and Lev Ratnovski (2007), “The Dark Side of Bank Wholesale Funding”, Federal Reserve Bank of Philadelphia Working Paper 09-3, December.
Kraft, Evan and Tomislav Galac (2007), “Deposit Interest Rates, Asset Risk and Bank Failure in Croatia”, Journal of Financial Stability, 2:312-336.
Poghosyan, Tigran and Martin Cihak (2009), “Distress in European Banks: An Analysis Based on a New Dataset“, IMF Working Paper 09/9.
Wheelock, David and Paul Wilson (2000), “Why Do Banks Disappear: The Determinants of US Bank Failures and Acquisitions”, Review of Economics and Statistics, 82(1):127-138.