What should the World Bank optimally do with the US$10 to $20 billion it can loan each year? Has it, in fact, done what is optimal? These two questions, one theoretical and one empirical, have been around for a long time and remain controversial in both academic and policy circles. This study seeks to contribute to the debate by suggesting a simple framework within which to measure the World Bank against an optimal international public financier for development. It goes on to argue that a careful treatment of the empirical evidence on Bank lending strongly contradicts optimal behavior under different assumptions. The evidence, in fact, rejects any notion that the Bank has substituted for private capital or that it has successfully catalyzed private development finance. These questions of fact are separate from the normative issues of whether to "mend" or "end" the current system of multilateral development finance.