South Africa is facing one of the largest HIV epidemics in the world and has been since the start of this millennium. Around 2004 the government started the roll-out of a nationwide scheme of free antiretroviral treatment, which treats HIV. Although these medications are proven to be efficacious in medical trials, we estimate the causal impact and effectiveness of this roll-out on (HIV-related) mortality in South Africa. We have used both district and provincial level data on (HIV-related) mortality, ARV usage, HIV prevalence and poverty. Data sources include Stats SA, the Thembisa Model, the District Health Information System and the Multidimensional Poverty Index. We have used a Fixed Effects model in 3 age categories (<15 years old, 15-49 years old and 50+ years old) to estimate the causal effect of ARV usage on annual mortality, in order to control for any unobserved time-invariant heterogeneity between districts and provinces.

Through our estimations, it becomes apparent that ARV usage has had a significant negative effect on overall and HIV-related mortality for those under 50 years old, on both a district level and provincial level. Most importantly, from our provincial analysis we find that a 1 percentage point increase in the ARV usage rate in children under 15 years old decreases their HIV-related mortality rate by 0.318 percentage points. Additionally, we find in our district level analysis that 1 percentage point increase in the overall ARV usage rate decreases the mortality rate in 15 to 49 year old by 0.2 percentage points. In those over 50 years old, the results are more ambiguous. From our findings, it can be said that the roll-out of free antiretroviral treatment has had a negative effect on the annual (HIV-related) mortality in those under 50 years in particular. Therefore, they suggest that in addition to being efficacious, free provision of ART is also effective. This strengthens the belief that the ART roll-out has had a positive effect on South African’s health and should therefore be continued and potentially expanded. This research adds to the existing research on the positive effects of free ART provision in South Africa.