Abstract Background The frequent occurrence of medicine stockouts represents a significant obstacle to tuberculosis control in South Africa. Stockouts can lead to treatment alterations or interruptions, which can impact treatment outcomes. This study investigates the determinants and effects of TB drug stockouts and whether poorer districts are disproportionately affected. Methods TB stockout data, health system indicators and TB treatment outcomes at the district level were extracted from the District Health Barometer for the years 2011, 2012 and 2013. Poverty terciles were constructed using the Census 2011 data to investigate whether stockouts and poor treatment outcomes were more prevalent in more impoverished districts. Fixed-effects regressions were used to estimate the effects of TB stockouts on TB treatment outcomes. Results TB stockouts occurred in all provinces but varied across provinces and years. Regression analysis showed a significant association between district per capita income and stockouts: a 10% rise in income was associated with an 8.50% decline in stockout proportions. In terms of consequences, after controlling for unobserved time invariant heterogeneity between districts, a 10% rise in TB drug stockouts was found to lower the cure rate by 2.10% (p