Domestic Violence: A Nonrandom Affair

In this paper, we develop and estimate a model of violence between romantically linked men and women.Physical violence is viewed as both a source of direct gratification and as an instrument for controlling the victim's behavior.Our model is a Stackleberg type model in which the assailant maximizes expected utility subject to the stochastic reaction function of the victim.Our model is estimated by a bounded-influence regression technique because the process generating violence appears to lead to a heavy-tailed error distribution.Our empirical results suggest that increases in the assailants (i.e. the male's) income serve to increase violence, while increases in the proportion of the year that he is employed serve to decrease violence.Further, the employment effect is larger than the income effect.y way of contrast, our results suggest that the effect of a change in the female's employment or income depends heavily on her economic status relative to the male's.Finally,we find that improvements in the female's opportunites outside the relationship significantly reduce the level of violence.