Abstract
The great financial crisis forced authorities to rethink the approach to regulation of the banking system held until then. The result is a new supervisory framework in which the stability of the system as a whole is tested through the so-called stress tests. Authorities design a severe macroeconomic scenario with which to test the resilience of the system; then, aiming to fill information asymmetries and restore market confidence, results are publicly disclosed and follow-up actions are implemented. The aim of this study is to assess the efficacy of such supervisory toolkit under several points of view. Firstly, through a qualitative assessment of the procedures so far conducted along with a review of the so-called consumption side analysis (i.e. the review of the literature focusing on the market reactions due to the release of stress tests outcomes) we shed lights on the best stress test practice. More precisely, we point out that the presence of a government backstop mechanism is crucial for the test to sooth panic and reduce systemic risk, being this one the real difference between useful and useless procedures. Institutional follow-up plans make sure and immediate the coverage of capital gaps, hence reducing the systemic risk. Secondly, we assess the informational content of the 2014 Comprehensive Assessment, the first macroprudential practice conducted before the ECB took the direct supervision of the largest European banks. We assess how much of the capital curtailments can be explained by accounting, regulatory or public data providing a ceiling to the amount of new information disclosed through the exercise in both its components, Asset Quality Review and Stress Test. Then, we complete our analysis by implementing an event study on the release date: rather than disclosing new idiosyncratic information on each bank under scrutiny, we find that the market reaction observed may be due to the inference process made by market participants. In other words, from the analysis of the CA findings the public inferred the risk factors that the regulator fears the most, adjusting stock prices accordingly. Thirdly and lastly, we propose a third line of research in the field of the macroprudential supervision: rather than focusing on either the production or consumption side analyses, we investigate the net effect of the macroprudential supervision on systemic risk. By applying the Kalman filter, thanks to a dynamic estimation of the Fama-French factors, we obtain a time-varying indicator of the riskiness of each bank. Even if we take into account the impact of the business cycle on such indicator, we find that the riskiness of banks participating to the SCAP declined soon afterwards. Such result appears even more interesting if we compare the above-mentioned betas (and the one of the banking system as a whole) with the ones of the other American industries: while the former experiences a marked decline, the latter increases the value (or remains constant) during the two years after the disclosure. Once again such result confirms the importance of an institutional backstop mechanism.