A major component of tax administration reform in sub-Saharan Africa for the last 30 years has been the creation of semi-autonomous revenue authorities (SARAs). The effects of their creation on revenue performance have been much debated, although there are only a few quantitative studies.

The core argument of this paper is that existing research suggesting diverse and contradictory outcomes has not taken account of trends in revenue performance in the years before the establishment of SARAs. Our analysis, based on 46 countries between 1980–2015, provides no robust evidence that SARAs induce an increase in revenue performance. This does not imply that SARAs may not provide benefits for tax collection, but they do not demonstrably increase (or decrease) revenue collected.