Pilferage from opaque food subsidy programs: Theory and evidence

Theft rates from subsidized food programs vary greatly and strongly influence program efficiency. Unfortunately, the determinants of these variations remain understudied because the agencies that run these programs seldom publicize the allocations of subsidized food to local markets. We develop a theoretical model of pilferage which predicts that: (i) pilferage from opaque programs is likely to rise more than proportionately with per capita food allocations; (ii) pilferage of inferior goods may be lower in poorer communities; (iii) pilferage rates need not rise as price subsidies are increased; and (iv) pilferage may rise as the relative quality of subsidized food is reduced. A comprehensive literature review and new estimates of pilferage across regions of the Philippines validates these predictions. Our finding, that around 48% of the subsidized rice went missing, is robust to new tests for sampling and recall error. Our policy discussion encourages geographic over administrative targeting, greater transparency in food allocations, and the use of realistic quotas.