Land-contingent poverty alleviation schemes

Abstract

Target groups for rural poverty alleviation schemes in South Asia and elsewhere have often been identified in terms of their landholding class, with landless or near landless households being the intended beneficiaries. This paper offers a quantitative approach to the analysis and design of land-contingent poverty alleviation schemes. An application to Bangladesh suggests that, when income is generally unobserved, landholding class is a good instrument for targeting poverty relief, in that it is strongly correlated with the informationally feasible allocations which minimize aggregate poverty for a given budget. It is, however, far from being a perfect instrument. Even under otherwise ideal conditions, the maximum feasible reduction in poverty which is attainable by relying solely on land-based targeting is substantially less than that attainable with perfect income information. Indeed, the poverty alleviation impact of domestically financed and revenue-neutral transfers across 10 landholding classes in rural Bangladesh could also be achieved without targeting by simply giving every household a lump-sum payment equivalent to 10% of mean rural income. Political restrictions on tax powers or errors made in administration will further reduce the gains from land-contingent targeting. Greater impact on poverty for a given budgetary cost will require more precise targeting within landholding classes.