Moral hazard and the demand for physician services: First lessons from a French natural experiment

Abstract

This paper presents empirical results on moral hazard in demand for medical care, using a longitudinal dataset on 4578 individuals followed during two years. The dataset contains two subgroups, one for which a copayment rate of 10% for physician visits was introduced in 1994, and an other for which no change occurred during the period of the study. This enables us to use these data as coming from a controlled natural experiment. We test if the number of visits per agent was modified by this copayment rate. The data reject the hypothesis for office visits, but do not for home visits. This suggests that there is moral hazard in demand for some physician services, but also that when non-monetary costs are important, small changes in monetary price may not induce any major change in behavior

Introduction

The magnitude and growth rate of health expenditures have been a major concern for most developed economies during the last decades. A frequently encountered view is that excessive expenditures are a consequence of the existence of large-scale health insurance schemes, resulting in a divorce between the amount consumed by each agent and the cost she actually bears. According to this `moral hazard' explanation, a natural solution would be to impose deductibles (or copayments), in order to provide adequate incentives to cost reduction.

The potential benefits of such a policy have, however, been vividly debated. The key aspect of the discussion is, in essence, an empirical one: to what extent can the introduction of a deductible be expected to significantly reduce each agent's level of health expenditures? If moral hazard is an important feature of medical consumption, then not only deductible are socially acceptable, but they improve welfare. In the opposite case, however, imposing to everyone some minimum deductible is inefficient, because it reduces the scope of mutually beneficial insurance contracts without any gain in terms of aggregate risk.

A large literature has been devoted to the analysis of demand for medical care in general, and physician services in particuliar. In this paper, we contribute to this general line of research by analyzing the results of a French `natural experiment'. In 1993, public health insurance reduced its cost-sharing rates for most ambulatory services. Most French households, however, are, in addition, partly or fully covered by a supplementary, private insurance contract. Some of the firms providing these contracts decided to compensate for this reduction in public coverage by extending their guarantees and increasing the premium; others, on the contrary, seized this opportunity to redesign their contracts by introducing a copayment rate. We follow two samples of employees belonging to these two classes, and investigate the differences in reactions due to the reform.

The structure of the paper is as follows. In Section 2, we briefly review the conclusions of contract theory and the kind of tests they suggest. In Section 3, we present the French system of health insurance, and describe the characteristics of our dataset. In Section 4, we present some first results, and discuss these results in the last section.