Using Trivariate Copulas to Model Sample Selection and Treatment Effects

Abstract

Simultaneous nonlinear econometric models with discrete outcomes are often difficult to implement. This article considers the use of the copula approach for a model with three jointly determined outcomes. It also deals with the discrete case in which outcomes include a mixture of dichotomous choices and discrete count data. We apply this technique to study self-selection and interdependence between health insurance and health care demand among married couples. The full model consists of a dichotomous choice equation for family insurance and a separate negative binomial equation for each spouse's health care use.

KEY WORDS:

• Count data
• Interdependence
• Medical Expenditure Panel Survey
• Simultaneous equations