A Generalized Heckman Model With Varying Sample Selection Bias and Dispersion Parameters

Many proposals have emerged as alternatives to the Heckman selection model, mainly address non-robustness of its normal assumption. The 2001 Medical Expenditure Panel Survey data is often used illustrate this model. In paper, we propose a generalization sample model by allowing bias and dispersion parameters depend on covariates. We show that may be due assumption constant parameter rather than normality Our proposed methodology allows us understand which covariates are important explain phenomenon only form conclusions about presence. explore inferential aspects maximum likelihood estimators (MLEs) for our generalized More specifically, satisfies some regularity conditions such it ensures consistency asymptotic MLEs. Proper score residuals models provided, adequacy addressed. Simulated results presented check finite-sample behavior verify consequences not considering varying parameters. analyzing medical expenditure suitable drawn using approach coherent with findings from prior literature. Moreover, identify relevant presence in dataset.