Bank stress tests: implications on accounting discretion, transparency and market discipline

Since 2009, regulators worldwide have conducted large-scale stress tests to reveal systemically important banks’ soundness to financial markets. Regulators aim to enforce market discipline that penalises excessive risk-taking and requires banks to operate more responsibly leading to financial stability. In this thesis, I contribute to this current debate by empirically analysing the implications of bank stress tests on three important aspects, namely accounting discretion, transparency and market discipline. First, based on a unique accounting dataset of stress-tested and untested European banks, I reveal that the accounting information of stress-tested banks is affected by stress tests. In particular, stress tests incentivise bank managers to exercise accounting discretion over loan loss provisions to manage both capital and earnings. The results suggest that stress tests exacerbate discretionary behaviour with the purpose of passing stress tests and conveying a sound picture of the bank’s financial condition to regulators and market participants. Second, examining a unique textual dataset of stress-tested European banks, I find that stress tests incentivise banks to enrich their textual narratives utilising certain stress test terms that I call ‘stress test sentiment’. This effect may specifically apply to newly than regularly stress-tested banks. Importantly, banks seem to compensate an increased stress test sentiment using a more positive disclosure tone; this may obfuscate market players, as market measures indicate lower information asymmetry and more analyst coverage. Third, based on a dataset of European and U.S. stress-tested banks, I show that stress tests do promote market discipline in both positive and negative directions as well as in short- and long-term event windows. In Europe, bank fundamentals are improved in terms of reduced bank risk-taking and funding structure, whilst the U.S. results are inconclusive. However, stress tests also tend to exacerbate negative performance of weaker institutions, due to market discipline, which could unintentionally compromise financial stability. In summary, this thesis provides novel results on stress tests that might be of interest to policymakers and regulators. I conclude that stress tests are an important addition that increases regulatory awareness and can enhance financial stability. However, I also show that stress tests may lead to unintended drawbacks on bank accounting practice and market discipline. Therefore, stress tests must be paired with a carefully executed disclosure policy to be a more effective regulatory tool.