Targeting Effectiveness of Social Transfer Programs in Botswana: Means-Tested versus Categorical and Self-Selected Instruments

We will evaluate the targeting effectiveness of fifteen social transfer programs in Botswana using data from the Botswana Multi-topic Household Survey 2015/16. Results reveal that, with one exception, these programs provide minimal coverage of the poor. Further, these programs are largely ineffective in targeting the poor and inclusion errors allow large leakages to the nonpoor. Benefit incidence analysis reveals that most social assistance and asset transfer programs and a public works program are progressive and pro-poor. Programs for building human capital through financing tertiary education are progressive but not pro-poor, suggesting inequality in access to higher education. Because education is a pathway out of poverty, this may contribute to intergenerational transmission of poverty. Means-tested programs do not necessarily target the poor better than categorical and self-targeted programs. Thus, eligibility criteria may not be strictly enforced during selection of beneficiaries for major means-tested programs. Reforms are needed to improve targeting effectiveness and minimize program leakages.