Murat K. Munkin

Bayesian analysis of a self-selection model with multiple outcomes using simulation-based estimation: an application to the demand for healthcare

Abstract This paper studies a self-selection model with discrete and continuous outcomes and a treatment variable. The treatment variable is endogenous to the two outcome variables. The approach of the paper is fully parametric and Bayesian. The Bayes factor is calculated with the Savage–Dickey density ratio and used for model selection. The model is applied to two different micro data sets, the 1987–1988 National Medical Expenditure Survey and the 1996 Medical Expenditure Panel Survey. The paper studies the effect of managed care and fee-for-service type of private insurance on the demand for healthcare. It also compares the effects of private insurance and Medicaid in covering health care expenses of elderly Americans.

Private Insurance, Selection, and Health Care Use: A Bayesian Analysis of a Roy-Type Model

Observed differences in medical utilization between the privately insured and uninsured reflect the combined effects of self-selection and insurance incentives (moral hazard). This article provides a Bayesian framework for decomposing the disparity into incentive and selection components. The effect of self-selection in private insurance on the number of doctor visits is estimated using a multiyear sample of the U.S. adult non-Medicare population obtained from the Medical Expenditure Panel Survey. We use a flexible econometric framework based on the “Roy model” and develop a Markov chain Monte Carlo algorithm. We estimate the distribution of treatment effects and find strong evidence indicating selection, which accounts, on average, for 50% or more of the observed disparity in doctor visits.

Disentangling incentives effects of insurance coverage from adverse selection in the case of drug expenditure: a finite mixture approach

This paper takes a finite mixture approach to model heterogeneity in incentive and selection effects of drug coverage on total drug expenditure among the Medicare elderly US population. Evidence is found that the positive drug expenditures of the elderly population can be decomposed into two groups different in the identified selection effects and interpreted as relatively healthy with lower average expenditures and relatively unhealthy with higher average expenditures, accounting for approximately 25 and 75% of the population, respectively. Adverse selection into drug insurance appears to be strong for the higher expenditure component and weak for the lower expenditure group.

Demand for cigarettes: a mixed binary-ordered probit approach

This study analyses the demand for cigarettes fitting observed zero outcomes with a trivariate model consisting of an equation for the starting smoking decision, an equation for the quitting decision, and an equation that models the level of cigarettes consumed. Five competing specifications are considered to explain level, with the ordered probit, which accommodates pile-ups of counts in the dependent variable, providing the best fit. Marginal effects of explanatory variables are calculated providing strong evidence of race and gender differences in consumption patterns. The estimated marginal effects are robust to alternative categorizations of the level of cigarettes.

A Bayesian analysis of the OPES model with a nonparametric component: An application to dental insurance and dental care

This paper analyzes the effect of dental insurance on utilization of general dentist services by adult US population aged from 25 to 64 years using the ordered probit model with endogenous selection. Our econometric framework accommodates endogeneity of insurance and the ordered nature of the measure of dental utilization. The study finds strong evidence of endogeneity of dental insurance to utilization and identifies interesting patterns of nonlinear dependencies between the dental insurance status and individuals age and income. The calculated average treatment effect supports the claim of adverse selection into the treated (insured) state and indicates a strong positive incentives effect of dental insurance.

Simulating the contribution of a biospecimen and clinical data repository in a phase II clinical trial: A value of information analysis.

The potential contributions of a centralized data warehouse or repository in clinical research include the expedited accrual of subjects for phase II trials. Understanding the contribution of data warehouses that integrate clinical, biospecimen, and molecular data for the conduct of clinical trials is essential to inform private and public decisions on resource allocation and investment. We conducted a value of information analysis using data from recent trials at the Moffitt Cancer Center and simulated the potential reductions in trial size due to possible alternative scenarios of expedited accrual. In this study, we compared alternative data sets using a single model to assess value of information. Our findings suggest that the reductions in trial size range from 0% to 43%, depending on the amount of censoring in overall survival. The ability to expedite the accrual of patients for clinical trial studies using large data repositories that store data on inclusion/exclusion criteria and response to standard of care therapies demonstrated significant improvement in reducing the number of subjects needed to achieve similar end-results, as evaluated using value of information analysis with a limited number of parameters and a parsimonious model of overall survival.

The MCMC and SML estimation of a self-selection model with two outcomes

A self-selection model with discrete and continuous outcomes and a treatment variable is considered. The treatment variable is endogenous to the two outcome variables. Two estimation procedures are proposed and compared. The first estimation approach is Bayesian and uses the Markov Chain Monte Carlo (MCMC) methods. It constructs stationary Markov chains that converge to the posterior distribution of the parameters of the mode. The second one is a full information maximum likelihood approach, using the simulated maximum likelihood (SML). estimator. Both methods are tested on a numerical example.