Just read this enticing recent paper by Lant Pritchett
touching upon three different but related issues. Labor migration (from developed to OECD countries) has the potential to improve the incomes of poor families by a factor of 40, when compared to the best possible local policy intervention you can think of. To attain the welfare benefits of the best micro-policy intervention, you would only need to move a very small share of workers from one place to the other (then gains then shared via the tax system). And then, funding for research on macroeconomic policy may have an impact on growth accelerations (or on decelerations the lack of) that is also manifold larger than the best policy intervention you can think of. Bottom-lines 1) Poverty is a place-specific condition, not individual-specific. 2) We have confused development with policies aimed at addressing the consequences of the lack of development.