We investigate the effectiveness of initiating deposit insurance at the outset of a banking crisis. Using a conjoint analysis approach that allows us to consider the simultaneous impact of multiple deposit insurance attributes and various counterfactuals, we ask a multinational sample of respondents how they would view hypothetical account profiles following the failure of a large competing bank. Previous experience matters: respondents from countries without explicit deposit insurance exhibit greater withdrawal risk, suggesting that the introduction of deposit insurance during a crisis may be only partially successful in preventing bank runs. They also impose a higher deposit interest rate premium. Having a long-term bank relationship reduces withdrawal risk, as does the absence of co-insurance.