Association for Public Policy Analysis & Management

Intimate partner violence is a major public health problem in Latin America, and Colombia is one of the countries in the region where it is highly prevalent. Understanding the effects of domestic violence against women is of substantial interest for public policy practitioners and social scientists, as it has large economic and social costs (Carrillo, 1992; Heise et al., 1994; Morrison and Orlando, 1999; UNICEF, 2001) and constrains a woman's choices and involvement in economic activities (Vyas and Watts, 2009; Bhattacharyya et al.,2011).

The goal of this paper is to provide causal estimates of the effect of intimate partner violence on female job exit. In this context, potential endogeneity issues arise due to unobservable characteristics that influence both the probability of getting involved with a violent partner and the probability of females leaving their jobs.

Previous studies on the relationship between intimate partner violence and labor market outcomes in Latin America suggest that women experiencing domestic violence are more likely to work for wages (Agüero, 2013.) Such studies, however, do not account for the potential endogeneity problem aforementioned. To fill this gap in the literature, I estimate the effect of domestic violence on female job exit by using the husband’s experience of domestic violence during childhood as a potential source of exogenous variation. Such events of intergenerational domestic violence affect a woman’s propensity to be a victim of intimate partner violence, but should not have a direct effect on her current labor market status and earnings potential.

Self-reported information on domestic violence was collected in the 2005 Colombian Demographic and Health Survey (DHS.) In this sample of 11,936 women, about 67.3 percent were victims of intimate partner violence. Females declared having experienced any physical (20 percent) or psychological (66 percent) violent incident in the past 12 months. About 31 percent of their husbands were also mistreated during childhood.

In order to estimate the relationship between domestic violence and female job exit, I use a linear probability model treating domestic violence as an exogenous variable. I then use a two-stage least squares specification to provide causal estimates. I define the outcome of interest as a dichotomous variable that equals 1 if a female left her job sometime in the past 12 months and equals 0 if she stayed working. Intimate partner violence is defined as a dichotomous variable that indicates whether the female has been a victim of either physical or psychological domestic violence in the past 12 months. I include wife’s and husband’s characteristics in the set of controls as well as region fixed effects to deal with unobserved heterogeneity.

Using two-stage least squares, preliminary results suggest that any event of domestic violence increases by 20 percentage points the likelihood of job exit, whereas physical and psychological violence increase it by 20.5 and 20.3 percentage points, respectively.