Being careful about the potential for endogeneity bias, I find robust evidence that "institutions for private property" share a more fundamental relationship with health expenditures than does national income. This research should interest a wide audience. First, health scholars may be interested in its relatively careful estimate of income's relationship to health spending. Second, institutions and commitment scholars should be interested in its evidence of institutions' primacy in a heretofore overlooked, but theoretically and substantively attractive, application. Finally, policy entrepreneurs may find important the implication that reforming governance structures can be more productive than is directly funding health services. Copyright Blackwell Publishing Ltd 2005.