This paper uses household data to test whether the determinants of demand for education have
changed during the Tanzanian government’s push for Universal Primary Education (UPE) in
the 2000s. Drawing on the existing theoretical and empirical literature, we test three main
hypotheses. First, we test whether demand for education is driven by the direct and
opportunity costs of education. We find that the abolition of school fees has led to an increase
in enrolment within agricultural households. However, wealth is still a significant determinant
of demand, indicating that structural differences in educational attainment have remained
largely intact over the period. In line with other studies, we find opportunity costs of sending
children to school to be important, yet their impact is modest. Second, the paper estimates
returns to education which is then used as an explanatory factor in the demand for education
estimations. The paper finds strong indications that returns do not determine demand. Third,
we test the importance of preferences in determining demand and find that educational
choices are affected by the views held by others within the community, which is supportive of
the importance of social norms in determining educational choices. However, after the
introduction of UPE, the size of this impact halves, suggesting that the push has reduced the
influence of previous social norms on the demand for education.