Abstract: This paper estimates the structural parameters of a dynamic model where parents with one child periodically decide whether or not their child uses various mental health services. In this model, mental health services improve a child's mental health (which parents care about), however, mental health services may be costly to the parents both in terms of utility and household consumption. Using a panel data set collected as part of the Fort Bragg Mental Health Demonstration, we estimate the model with a maximum likelihood procedure that accounts for unobservable differences in mental health endowments of children and population heterogeneity in parental preferences and in the effectiveness of mental health services. We estimate that parents experience relatively high disutility when a child uses mental health services, implying parents enroll their children in mental health services only if these services have multi-period effects on their child's mental health. Correspondingly, we find that outpatient and inpatient mental health services have permanent effects on a child's mental health. We conclude that the improvement over time of the mental health of the children in our data is, in a large part, the outcome of forward looking parents choosing to increase their child's mental health.

Full paper (206 KB PDF) | Full paper (1692 KB Postscript)