Economic Stimulus from Public Health Programs: Externalities from Mass AIDS Treatment Provision in South Africa

Abstract

Return on investment in international development assistance has attracted greater focus in recent years. Government programs, particularly public health programs, are likely to indirectly affect community members beyond the intended direct beneficiaries via externalities that could be many times larger than the direct benefits. This study evaluates the direct and indirect impact of the provision of mass AIDS treatment on labor market outcomes, using biomarker data to separately identify the impact by HIV status. We used rich data from a demographic surveillance site in rural South Africa with a 99% survey response rate. We leveraged differences in access to antiretroviral therapy (ART) that are uncorrelated with unobservables, and we used a set of rigorous causal inference methods, including machine learning (lasso), to estimate a plausibly causal impact. As ART access scaled up between 2004 and 2011, employment increased by 8.5 percentage points for HIV-infected individuals who were within 2 km from the nearest ART clinic compared to those more than 5 km away. We found sizable increases of 6.3 percentage points (22%) among HIV-uninfected individuals who had no HIV-infected household members. Labor force participation decreased over this period, which suggests that labor demand plays an important role in driving the observed employment increases. Our results demonstrate that the economic benefits of ART are broadly distributed, operate via channels outside the household, and have properties of a public good. Investments in health-related human capital may have important stimulus effects on local economies that should be considered alongside conventional economic policy.