Using a unique and comprehensive dataset, this paper develops and uses three distinct methods to quantify the risk of a systemic failure in the global banking system. We examine a sample of 334 banks (representing 80% of global bank equity) in 28 countries around 6 global financial crises (such as the Asian and Russian crises and September 11, 2001), and show that these crises did not create large probabilities of global financial system failure. First, we show that cumulative negative abnormal returns for the subset of banks not directly exposed to a negative shock (unexposed banks) rarely exceed a few percent. Second, we use structural models to obtain more precise point estimates of the likelihood of systemic failure. These estimates suggest that systemic risk is limited even during major financial crises. For example, maximum likelihood estimation of bank failure probabilities implied by equity prices suggests the Asian crisis induced less than a 1% increase in the probability of systemic failure. Third, we also obtain estimates of systemic risk implied by equity option prices of U.S. and European banks. The largest values are obtained for the Russian crisis and September 11 and these show increases in estimated average default probabilities of only around 1-2%. Taken together our results suggest statistically significant, but economically small, increases in systemic risk around even the worst financial crises of the last 10 years. Although policy responses are endogenous, the low estimated probabilities suggest that the distress of central bankers, regulators and politicians about the events we study may be overstated, and that current policy responses to financial crises and the existing institutional framework may be adequate to handle major macroeconomic events.