Bayesian Analysis of Structural Effects in an OrderedEquation System

We describe a simulation-based algorithm for Bayesian estimation of structuraleffects in models where the outcome of interest and an endogenous treatment variableare ordered. Our algorithm makes use of a reparameterization, suggested by Nandramand Chen (1996) and Li and Tobias (2005) in the context of a single equation ordered-probit model, which significantly improves the mixing of the standard Gibbs sampler.We illustrate the improvements afforded by this new algorithm in a generated dataexperiment, examine the implications of \defult" priors for the reparameterized modelparameters on priors for the \structural" coeficients, and also make use of our methodsin an empirical application. Specifically, we take data from the National LongitudinalSurvey of Youth (NLSY) and investigate the impact of maternal alcohol consumptionon early infant health. Our results show clear evidence that the health outcomes ofinfants whose mothers drink while pregnant are worse than the outcomes of infantswhose mothers never consumed alcohol while pregnant. In addition, the estimatedparameters clearly suggest the need to control for the endogeneity of maternal alcoholconsumption.[...]