ESSAY IN HEALTH ECONOMICS AND INDUSTRIAL ORGANIZATION

Industrial organization focuses on imperfectly competitive markets to understand the behavior of firms and the resulting welfare effects. This is a broad definition as most markets are imperfectly competitive and industrial organization research can then focus on a wide variety of topics. Imperfect competition may be due to many reason. Perfect competition in fact requires: a large number of firms and consumers, free entry and exit, marketability of all goods and service including risk, symmetric information with zero search cost. Moreover the list includes no increasing returns, no externalities, and no collusion. Health care markets are a good example for imperfect competition as generally they violate all requirements included in the previous list. If we focus only on some violation like asymmetric information and no marketability, then health care markets fail in a more clear way than other markets. This justifies the often made claim that the health care market is “different” and implies that any evaluation of its performance must be based on models that explicitly take into account its deviations from the assumption required for perfect competition. The model of perfect competition can still serve as the benchmark of optimal performance, but generally it cannot be used to analyze how health care markets work. For this reason the common thread of this thesis is to analyze health care markets using the theoretical and empirical tools provided by industrial organization. This thesis is composed by three essays. In the first one I am going to propose a theoretical framework to analyze product differentiation with consumers misperception and information disparities. The model is an extension of standard vertical product differentiation (Gabsewictz and Thisse, 1979 and Shaked and Sutton, 1982), where I relax the assumption of perfect information. As I said before asymmetric information is one of the big problem to deal with in health economics. And if products are credence goods, as in case of drugs, many consumers may lack the expertise to ascertain the quality differential with respect to cheaper standard brands, even after purchase. In that case consumers face a risky decision and to the extent they lack information about the true quality differential they may carry out purchase decision according to misperceptions about product quality. In this paper I extend the analysis of Cavaliere (2005) to include the quality choice by firms, when providing higher quality requires a costly effort, and propose to analyse the case of a duopoly with vertically differentiated products with consumers’ misperceptions and information disparities. Consumers are actually split between uninformed and informed consumers. Uninformed consumers are characterized by consumers’ misperceptions as they can underestimate or overestimate the quality differential. As a minimum quality standard is imposed by the Government even uninformed consumers expect that any product sold in the market at least complies with the standard. As low quality can be said to be verifiable, even uninformed consumers can be confident about low quality products: firms are expected to provide at least the minimum quality standard. This last assumption well fits the case of pharmaceutical products. Actually every developed country has a national institution that enforces and verifies drug’s minimum quality standard. The aim of this paper is to shed light on how firm set price and quality when consumers are characterized by asymmetric information and mispercemption obout quality. We do not analyze information decisions by consumers, which are exogenously given, therefore firms follow a Stackelber behavior vis à vis consumers. However we can analyze quality and price competition between firms for the full range of information disparities, i.e. for any split between informed and uninformed consumers that can affect demand functions. Furthermore we distinguish between the case of optimistic misperceptions (uninformed consumers overestimate the quality differential) and the case of pessimistic consumers (uninformed consumers underestimate the quality differential). Competition between firms is represented by a two stage game, in the first stage the two firms compete in qualities, given the market split between informed and uninformed consumers. In the second stage price competition takes place. We will show that both price and quality are strictly depend on asymmetric information as expectation and number of informed consumer affect firm’s choice. For different quality expectations and share of informed consumer we found market failure. In some cases uninformed consumers are cheated by high quality firm when they purchase high quality product, in other cases, for different information level and expectations, adverse selection arises endogenously in the model. The second paper consists in a theoretical model where I analyse incentives for cooperative behaviour when heterogeneous health care providers are faced with regulated prices under yardstick competition. Providers are heterogeneous in the degree to which their interests match to those of the regulator. The basic idea behind yardstick competition is that the price (or price cap) faced by each provider is dependent on the actions of all the other providers (Schleifer, 1985; Laffont and Tirole, 1993). According to Schleifer’s rule, the price each provider faces is based on the costs of all other providers in the industry but not its own. This creates strong incentives for cost control. When there is a large number of providers, this is unlikely to be a problem, mainly because the cost of collusion rises, but even in larger countries, provision might be concentrated among a handful of providers, as is likely for utilities, rail or postal services and for specialist health services, such as bone marrow or lung transplantation. The innovation with respect to the standard model of yardstick competition is the introduction of heterogeneity in the degree to which the provider’s interests correspond to those of the regulator. Because the incentive to collude with other providers will depend on the objectives of the providers, particularly the extent to which their objectives correspond with those of the price-setting regulator. We use “altruism” to describe the behavior of providers whose aims are closely related to those of the regulator and “self-interested” to describe providers whose interests are more divergent from those of the regulator. If we consider the different ownership types in health services this heterogeneity in “altruism” is evident since we observe full public ownership i.e. altruistic providers and full private hospital i.e. self-interested providers. This paper aims then to analyse incentives for collusive behaviour when heterogeneous providers are faced with regulated prices under yardstick competition. We analyse the choice of cost when providers do not collude and when they do, and we consider incentives to defect from the collusion agreement Our results suggest that under the yardstick competition each provider’s choice of cooperative cost is decreasing in the degree of the other provider’s altruism, so a self-interested provider will operate at a lower cost than an altruistic provider. The prospect of defection serves to moderate the chosen level of operating cost. More general results show that collusion is more stable in homogeneous than in heterogeneous markets. The third paper is an empirical analysis where I test the hypotheses of physicians’ altruism and ex-post moral hazard using a large national panel dataset of drug prescription records from Finland. We estimate the probability that doctors prescribe generic versus branded versions of statins for their patients as a function of the shares of the difference in prices between what patients have to pay out of their pocket and what are covered by insurance. The role of physicians and insurance in health care markets has been of interest to economists since the seminal contribution of Arrow (1963). Pioneering the economic analysis of physician behavior in the context of health care, Arrow (1963) noticed that doctors may have motives and objectives that differentiate them from purely profit-maximizing agents. The original ‘ex-post moral hazard’ hypothesis, predicts that health insurance increases the consumption of health care and leads to excessive consumption of services even in a competitive health care market. Ex-post moral hazard has since then been the focus of various empirical and theoretical studies in health economics (see e.g. Feldstein, 1973; Leibowitz, Manning, and Newhouse, 1985; Manning, Newhouse, Duan, Keeler, Leibowitz, and Marquis, 1987; Dranove, 1989; Zweifel and Manning, 2000). We simultaneously test both altruism and ex-post moral hazard in drug prescription behavior using a large national panel of administrative data from Finland. We first develop a theoretical model on physician decision-making, which, in line with Hellerstein (1998) and Lundin (2000), then use a large national panel dataset with all statin prescriptions in Finland between 2003 and 2010 (n=17 858 829 prescriptions) to test the physicians’ altruism and ex-post moral hazard hypotheses, while controlling for a large range of physicians, patients, and drug characteristics. Taking advantage of the panel structure of our national administrative dataset, we directly observe the repeated prescriptions of statins by physicians over time. We find that although the estimated coefficients associated with ex-post moral hazard and altruism are statistically significantly different from zero, their size is very close to zero and the orders of magnitude is smaller than the effects associated with other key explanatory factors. We also find robust and strong evidence of prescription habit-dependency.