Women Shelters: The Last Resort Option to Escape From Violence

Abstract

This study estimates the causal effect of a European Union-supported women’s shelter project initiated to combat domestic violence in 2012 in Turkey. The paper employs the difference-in-differences methodology by using the micro-level data sets from the national research on domestic violence against women conducted in the years 2008 and 2014. The estimation results show that the project reduced the probability of physical domestic violence by 1.6 percentage points. The decrease in the intensity of physical domestic violence was around 4.6 percentage points. The study utilizes further estimations to detect heterogeneous effects differing by fertility, education, and employment status. The findings indicate that mothers benefited more from the project, with 2 percentage points probability and 6.2 points level reductions. Besides, less-educated or non-working women experienced a larger probability or level reductions in violence, suggesting that the effectiveness of shelters is higher for more disadvantaged groups. Overall, the results show that shelters are reliable policy tools, even in a country with strong male backlash mechanisms.

Keywords: women's shelters, domestic violence, male backlash, household bargaining, difference-in-differences.

JEL Classification: J12, J13, J16, K38.