Most of Indonesia’s banking system collapsed during the 1997–98 financial and economic crisis. We estimate that the net cost to taxpayers of the government’s blanket guarantee of banks’ liabilities, issued in February 1998, is about 40 per cent of annual GDP. Large banks fared worse in the crisis than small ones and state banks fared worse than private ones. Despite this, and despite the fact that bank capital turned out to have been inadequate, the government reduced the capital requirements for all banks, transferred the assets of closed banks, together with the lowest quality loans of those that were recapitalized, to a state-owned holding company, and thus excluded the private sector from participating in the process of liquidating these assets. The government offered to recapitalize several banks jointly with the private sector, but participation was restricted to the former owners, and even they could only participate on very unfavorable terms. As a result, too many banks were closed, too many nationalized and several were unnecessarily merged. We propose a more market oriented approach that would have strengthened banks by raising capital requirements and also minimized fiscal costs by auctioning those that failed to meet these requirements. In the case of insolvent banks, bidders should have been invited to submit tenders for taking over both their assets and liabilities. In all cases, bidders should have been able to choose between liquidating banks and keeping them operational, after injecting enough cash to meet the new capital adequacy requirements.