A Bad Peace or a Good War: A Structural Estimation Model of Spousal Conflict and Divorce

The optimal balance between keeping marriages intact, despite sustained conflict, or allowing for divorce continues to be a subject of policy debate, even after years of changes to divorce laws. To understand the tradeoifs, I construct a structural game theoretic model of spousal interaction with information asymmetries which may generate Pareto inferior outcomes, including inefficient conflict or divorce. Models with conflict as an equilibrium outcome have not been analyzed before in the household bargaining literature. The structural parameters are estimated using data from the National Survey of Families and Households, a nationally representative panel of 13,000 households with unique questions on conflict and spousal beliefs about the outside options. The model fits the data well and has good out - of - sample predictive properties. The estimation results imply that a large majority of spouses are deeply hurt by conflict and think that their own post - divorce opportunities are poor. Exaggerating a partner's disutility from conflict and underestimating the outside prospects generate inefficient negotiation outcomes in the model. The estimated parameters are mostly in line with intuition. For instance, marital heterogamy indicators, such as the difference in spousal ages, tend to have a negative impact on the value of marriage relative to divorce. Mandatory separation periods before a divorce decree can be granted adversely affect the outside options. Stronger child support enforcement generally makes divorce more attractive to wives and less attractive to husbands, but the effect varies with educational attainment. I simulate several policy changes and find that the elimination of the separation requirements increases the fraction of divorced couples by 8.4 percent and is a weak deterrent to conflict. Perfect child support enforcement reduces the incidence of divorce and conflict by as much as 9.2 and 18.4 percent, respectively.

Note: Abstract extracted from PDF file via OCR