Central Bank supervision is one of the pillars of capital regulation. Based on a unique database built using supervision data from the Central Bank of Brazil, we evaluate the effectiveness of the Central Bank's supervision over banks given the Central Bank's proprietary credit rating and signaling requests for higher capital buffers. We also examine the main determinants of capital buffer management in addition to supervision. We find evidence that (i) Brazilian Central Bank supervision imposes excess capital buffer needs on banks, especially small and midsize banks; (ii) market discipline may play no role in driving capital ratios; and (iii) the business cycle has a negative influence on bank capital cushions, suggesting pro-cyclical capital management. We conclude that supervision plays a major role in markets where market discipline is weak and for smaller banks which act on pro-cyclical way.

In this study, we test for the possible contagion effects of the 10-year Greek government bond yield. We first employ the well-documented adjusted correlation coefficient of Forbes and Rigobon (2002) and then we estimate an exponential generalized autoregressive conditional heteroskedasticity model extended for volatility spillovers. Finally, we propose an extension of the corrected Dynamic Conditional Correlation (cDCC) model, which allows for structural breaks in the correlation dynamics. The suggested cDCC specification provides a natural testing framework for the correlation contagion hypothesis. Compared with other similar approaches, the proposed structural break cDCC approach allows for consistent inferences. The results do not confirm any contagious effects stemming from the 10-year Greek sovereign bond.