Ordinary Least Squares and Sample-Selection Models of Health-Care Demand Monte Carlo Comparison
通过蒙特卡洛模拟，比较了医疗需求估计中两阶段模型与样本选择模型的性能，发现两阶段模型在参数估计均方误差上更优。
This article examines alternative econometric models for health-care demand estimation. The analysis compares the Rand Health Insurance Study two-part model with sample-selection model estimators in a Monte Carlo simulation experiment designed to approximate individual-level health-care demand conditions. The underlying variable distributions are taken from cross- sectional data for a Swiss 1981 population survey. Three sets of error distribution assumptions are examined—bivariate normal, normal logistic, and Cauchy. Despite theoretical concerns with the two-part model, it outperforms the sample-selection model in terms of mean squared error of parameter estimate.