Here is the abstract from my new paper with Selim Gulesci:

We assess the religious and social impacts of female schooling in Turkey using a law change in compulsory schooling. The law, implemented in 1998, bound individuals born after a specific date to 8 years of schooling while those born earlier could drop out after 5 years. This allows the implementation of a Regression Discontinuity Design and the estimation of meaningful causal estimates of schooling. Using a dataset of married Turkish women in 2008, we find large reducing effects of a year of schooling on expressions of religiosity, such as the habit of wearing a headscarf, attending Qur’anic courses, and regular prayer. Parallel to these, we also document a partial empowerment effect, whereby women are more likely to make marriage decisions themselves, less likely to marry under the legal age, and to experience better household characteristics. A noteworthy non-result is the lack of clear effects on female labor force participation. On one hand, we show that returns to schooling in terms of women’s status and living conditions may be substantial even when labor-related returns are not. In particular, our results are consistent with education allowing social mobility out of religiously conservative environments; with women more independently choosing richer, and more educated husbands outside the family circle. On the other hand, however, we also document the absence of commensurate impacts for the country’s large ethnolinguistic minorities. An evaluation of the education reform thus needs to weigh its average empowering effects against increased inequality across ethnolinguistic groups.