To add to this very interesting discussion, I think Sen and Dreze were more "puzzled" by the fact that Bangladesh spends less public money on health and education (than its neighbor to the west) and yet has better health and education outcomes. To some of us, this is not a puzzle. Bangladesh uses a multiplicity of service providers, including NGOs, the private sector and religious organizations (in the case of schools) to deliver services. In some cases (such as secondary education), financing is still done by the state. There is some evidence that these arrangements yield better efficiency, either because providers are more accountable to the state (through explicit contracts) or because of intrinsic motivation of the service providers (as in the case of some NGOs). The link to growth is that Sen-Dreze and others believe, often implicitly, that higher growth leads to more public resources which, in turn, should yield better social outcomes. But if the second link in this chain is broken, then growth is largely irrelevant, except for--as Mushtaq points out--the effect on the demand for health and education.