In this paper, we investigate the impacts of the elimination of primary school fees in Mainland Tanzania. We use 2002 and 2012 census data and conduct a difference-in- differences analysis. Spatial and temporal variation in the implementation process of the reform is generated by distinguishing between cohorts that were exposed to the reform and cohorts that were not, and by considering the intensity of the reform, which is defined based on pre-reform educational outcomes at the gender-district level. We find that exposure to an average of 1.7 years of free primary education has reduced educational inequality by 0.66 standard deviations. We also find that this outcome has been mainly driven by a reform-induced reduction of 6.8 percentage points in the proportion of people who have never attended primary education. The benefits of fee removal have been relatively larger for females compared to males. Therefore, the educational gender gap has been narrowed. Nevertheless, policy- makers should be wary of the reform's limited reach in preventing dropouts and its diminishing effects as time elapses from the year when the reform was rolled out.