It is commonly believed that Islamic political representation leads to lower women’s rights in societies. However, there is a lack of causal studies on the relationship between Islamic rule and outcomes such as education, female political participation, adolescent marriages, etc. The first attempt to capture the causal impact of local Islamic rule has recently been made by Erik Meyersson, who aims at analysing the effect of democratically elected Islamic political on education and related outcomes in Turkey. Contrary to the common belief, he shows that the arrival of the Islamic party in municipalities increased the female secular high school education, and this effect remains persistent up to 17 years after, and also shows that Islamic rule decreased the adolescent marriages and increased female political participation in municipal council in 2009.
Even though Turkey is a secular republic, there has been an increase in the Islamic political representation in democratic process since 1990s. And in 1994, pro-Islamic Refah (Welfare) party came in second place at local elections. The party got 20 per cent of the votes at national level, and at local elections, it got 12 per cent of the mayor seats. Therefore, Turkey is a useful testing ground to look for the causal effects of Islamic rule.
It is not easy to isolate the causal effect of Islamic rule on outcomes such as education. The reason behind is the endogeneity of identity of the mayor and the characteristics of the voter. In the municipalities where female high school enrolment and attainment are lower, people may be more likely to vote for an Islamic party. Electing a mayor is not a random process, and unobserved factors could lead to less education for females as well as electing an Islamic mayor. Therefore, using regression analysis may give biased estimates.  To solve this identification issue, Meyersson uses a regression discontinuity (RD) design, a method to estimate causal effects, to study the effect of having an Islamic mayor on outcomes such as education, adolescent marriages etc. using new data set of 2600 Turkish municipalities.  He compares the municipalities where the Islamic party barely won or lost the elections. The idea behind is that these municipalities should be similar in terms of observable and unobservable characteristics. Therefore, this can allow us to compare municipalities for which having an Islamic mayor can be considered as random. 
Meyersson shows that female participation in both politics and the economy increased after the Islamic political control. There are positive effects on female secular high school completion, enrolment in high school over a period of six years. Mandatory primary schooling and proxies for religious types of schooling remains unaffected. When it comes to men, we see insignificant and smaller effects. In the longer run, the author shows that the results are persistent for 17 years after the elections in 1994, and having an Islamic mayor decreased the number of adolescent marriages for female. The paper also shows that female participation in politics has increased in the elections in 2009.
The author explains these results not by Islamic identity of the parties, but rather with barriers to enter to the education for women in the secular education system. In Turkey, participation for high school education is voluntary and under the secular system rules. These rules are headscarf ban, mixed classes and secular curriculum. The author argues that the positive effect of Islamic political control is due to their political actions to handle these barriers.
The Islamic (Refah) party was against the headscarf ban, and it refused to apply this rule in the municipalities it was in power (ECHR(2003)). Besides, the party agreed to take some help from religious foundations to increase religious education facilities for people who are religious conservatives. Indeed author claims and shows that in the places where people are poorer and more conservative, which results in higher barriers to enter, the causal effects of electing Islamic party on education (female high school) are larger. He also shows that Islamic mayor had cooperation with religious foundations to provide them religious educational facilities where women often wear the headscarf, can pray in praying rooms, etc. ; evidence shows that parents were more willing to send their daughters to school. These effects are smaller and less precise for men. This is consistent with the fact that in the secular system, women are more affected by the barriers to enter than men.
These results suggest that under specific circumstances, the election of Islamic politicians can lead to positive effects. The effect of having an Islamic mayor is valid for the regions where people had lower income and were religiously conservative which leads higher probability to vote for the Islamic party. And yet, the effect is weak for the wealthier and higher educated communities. Last but not least, the paper emphasizes that the positive effects of Islamic party on poor and pious cannot be separated from the barriers to enter .
. Meyersson, E. Islamic Rule, Empowerment of the Poor and Pious, Econometrica, Vol.82, No.1 (January,2014), 229-269
. The use of the regression discontinuity (RD) design has been discussed on a previous post: https://poresp.wordpress.com/2014/10/05/groundwater-and-rural-poverty-in-india/
. For information on the regression discontinuity (RD) design: Imbens,G & Lemiuex,T. Regression discontinuity designs: A guide to practice, Journal of Econometrics, Vol.142
. Lee,D & Lemieux,T. Regression discontinuity designs in economics, Journal of Economic Literature, 48(2)(2010), 281-355
. ECHR: European Court of Human Rights