Working Paper
Poverty Measures and Anti-Poverty Policy with an Egalitarian Constraint

Bourguignon and Fields (‘Poverty Measures and Anti-Poverty Policy’) and Gangopadhyay and Subramanian (‘Optimal Budgetary Intervention in Poverty Alleviation Schemes’) have derived optimal budgetary rules for the redress of poverty through direct income transfers when poverty is measured by the Foster, Greer and Thorbecke Pα class of indices in the context of a constrained optimization exercise which one may call the ‘canonical problem’. For the stock of poverty measures most commonly in use, the canonical problem yields one of only three types of optimal solution which Bourguignon and Fields call, respectively, a ‘type-r’, a ‘type-p’ and a ‘mixed-type’ policy. The authors suggest that other types of policy—such as a ‘proportionality rule’ in which the budget is allocated in proportion to a poor person’s share in the aggregate poverty deficit—could be considered optimal in certain circumstances. This paper explores the sorts of circumstances—meaning combinations of poverty indices and different types of ‘egalitarian’ restrictions on optimal transfers— under which ‘unconventional’ policies (including the Proportionality Rule) could emerge as optimal budgetary allocations.