Incentive effects in the demand for health care: a bivariate panel count data estimation

Abstract

This paper contributes in three dimensions to the literature on health care demand. First, it features the first application of a bivariate random effects estimator in a count data setting, to permit the efficient estimation of this type of model with panel data. Second, it provides an innovative test of adverse selection and confirms that high-risk individuals are more likely to acquire supplemental add-on insurance. Third, the estimations yield that in accordance with the theory of moral hazard, we observe a much lower frequency of doctor visits among the self-employed, and among mothers of small children. Copyright © 2002 John Wiley & Sons, Ltd.

What belongs where? Variable selection for zero-inflated count models with an application to the demand for health care

“…We illustrate the proposed methods by applying them to a sub-sample of a data set originally used by Riphahn et al (2003) to analyse the demand for health care. 5 The data set stems from the German Socio-Economic Panel Study (SOEP, see Wagner et al, 2007) and is an unbalanced panel of 7,293 individual families observed from one to seven times.…”

“…Variable descriptions along with summary statistics are given in Table 1. 6 Whereas Riphahn et al (2003) estimate a bivariate model and consider both the number of doctor visits and the number of hospital visits as dependent variables, we focus on analysing only the number of doctor visits in the last three months. Figure 1 shows a histogram of the dependent variable.…”

“…Thus, SSVS chooses a very parsimonious model specification. Riphahn et al (2003) were particularly interested in the role of the choice of private insurance on health care demand. We see that neither the dummy variable indicating public health insurance (public) nor the dummy variable for add-on insurance (addon) are included in the two equations.…”

Fixed and Random Effects Models for Count Data

“…This would correspond to a mixed negative binomial model. [The model used in Riphahn (2003).] If it is assumed that ε it has the G(θ,θ) distribution assumed in Section 2 and u i has a normal distribution with mean zero and standard deviation σ, then we obtain a "true" random effects NB model that parallels the model developed earlier.…”

Risk Selection under Public Health Insurance with Opt‐Out

“…Geil et al (1997) and Riphahn et al (2003), also using data from the SOEP, estimate count data models and provide further evidence in favour of the identifying assumption.…”