Summary

The 2008 financial crisis highlighted the banking sector’s weaknesses, instability and tendency to take excessive risks that were not adequately mitigated. Understanding the financial sector and assessing its risks is important for many market participants such as supervisors, regulators, shareholders, creditors and borrowers and customers. One way in which both supervisors and banks themselves monitor, manage and mitigate risks is by conducting stress tests. Today, amid what is already a financial crisis, any internal as well as external stress testing of banks is a relevant topic. Six hypotheses were set that focused on the relationship between certain bank-specific risk indicators and the magnitude of the impact of the stress tests on the capital position of banks.

The article focused on the EU-wide stress tests conducted in 2010, 2011, 2014, 2016 and 2018. A total of 153 banks from 22 different countries participated in the stress tests. To test the hypotheses, regression analyses based on both panel and cross-sectional data were used to create econometric models that would describe possible relationships between the variables. The dependent variable was the impact of the stress test on capital, which is reflected through two alternative indicators. Firstly, as the ratio of difference in the level of capitalization (stress vs baseline scenario) to baseline level capitalization. Secondly, as as the ratio of difference in the level of capitalization (stress scenario vs actual capital ratio before the stress test) to the last actual capital ratio in the year preceding the stress test. The bank-specific explanatory variables used were total capital ratio, the share of non performing loans, the efficiency ratio of banks, i.e the cost-income ratio, return on assets, the share of the loan portfolio among total assets and the level of interest bearing liabilities to total assets.

The results of the analysis showed that the EU-wide stress test had a greater impact on the capital of banks with the lowest profitability and low asset quality. In general, the results of this analysis coincide with previous research and studies done on the topic.