« ann arbor is coming again | メイン | hating Bayesian »


why U-shaped?


Labor market shocks and marriage duration
by Ioana Marinescu

Abstract: Are couples who have been together longer more likely to stick together in the face of adversity? This paper uses monthly data from the Survey of Income and Program Participation to investigate the impact on relationship duration of labor market shocks such as layoff, discharge for cause and disability. Using a Cox proportional hazard model, I show that generally these labor market shocks significantly increase the separation hazard. A man’s getting fired has the largest impact on separation: it significantly increases the monthly hazard of separation by 38.5%, which is equivalent to increasing the annual probability of separation to 4.3% from its average of 3.1%. I then show that the impact of labor market shocks typically initially decreases and then increases with relationship duration. Assuming that a labor market shock is a signal of the value of the relationship to one’s partner, the initial decrease in the effect can be explained by the fact that, as one learns more about one’s partner, any new information is less likely to change one’s evaluation. The subsequent increase in the impact of labor market shocks on the probability of separation can be explained by the fact that relationships change over time such that, at longer relationship durations, there is a relatively higher proportion of relationships perceived to be mediocre; thus, adding a labor market shock to the existing dissatisfaction can cause one of the partners to end the relationship.