Social Assistance Programs and Household Welfare in Eswatini

Abstract

Eswatini has notably high levels of poverty and inequality. Recurrent, negative shocks are an important contributing factor. This study assesses the performance of the largest social assistance programs in Eswatini, based on 2016/17 national household survey data. It examines the coverage rates of these programs, and their incidence and effectiveness in reducing poverty and inequality. The study also examines the association between program participation and negative shocks reported by households, in particular, drought and food price shocks associated with the 2015-2016 El Niño event. Across programs, benefits are concentrated among poor households. However, the performance of programs in reducing poverty and inequality tends to be limited because of low intended or actual benefit levels and shortfalls in intended or actual coverage of the poor. Households that receive program benefits are more likely to report a drought shock. Except in the case of emergency food aid, which is provided ex post, we interpret this pattern to indicate that programs tend to provide ex-ante coverage to those vulnerable to this shock. At a minimum, enhancing the performance of programs in addressing poverty, inequality, and the adverse effects of shocks would require that actual benefit levels equal intended levels (for example, by procuring sufficient food commodities to meet the needs of the school feeding program) and that intended benefit levels are fully aligned with program aims (for example, by providing grant amounts to schools that are large enough to allow for tuition-free government secondary education for orphaned and vulnerable children). Absent greater budgetary allocations to programs, addressing these benefit-related disconnects may require improving the targeting of select program benefits to poorer households such as by using a proxy means test. We simulate the effects of programs on poverty and inequality reduction from such hypothetical reforms.