Since the mid-1990s worldwide efforts were undertaken to improve the effectiveness of financial supervision, through modifications in the architecture and governance. Did these improvements mitigate the 2008–2009 Crisis? This paper brings the first systematic analysis of the role of three main efforts: consolidation in supervision, decreasing central bank involvement and improving supervisory governance. The analysis employs a new and complex database on supervisory architecture and governance for 102 countries and uses two new indicators to evaluate the supervisory regime: the Financial Supervision Herfindahl Hirschman (FSHH) and the Central Bank Supervisor Share (CBSS) Indexes. The empirical tests allow us to disentangle the relative effects of the supervisory regimes on macroeconomic resilience. We conclude that two supervisory features—supervisory consolidation and supervisory governance—were negatively correlated with resilience, while central bank involvement in supervision did not have any significant impact. Our results show that the conditions under which micro-features of the supervisory design produce automatically macro-optimal outcomes are far from identified, and consequently contradict what was the generally accepted view before the crisis.