Abstract
This thesis consists of two parts. The first part includes two essays that deal with the pharmaceutical market, and one essay that looks at strategic incentives that arise in optimal treatment involving untried drugs. The second part, consisting of two essays, examines some implications of altruism. Part I: Two of the essays (joint with Mats Ekelund) are empirical studies of the pharmaceutical market in Sweden. We consider all New Chemical Entities (NCEs) introduced in Sweden between 1987 and 1997. In the first essay, we examine drug pricing in the price regulated Swedish market and compare the results with a previous study of the US market, where no such regulation exists. Similar to the US study, we find that relative launch prices are positively correlated with the degree of therapeutic advance. In contrast to the US study, the presence of substitutes has a negligible effect on both launch prices and price dynamics. In the second essay, we consider the empirical relation between therapeutic advance and market shares. We use a model of horizontal and vertical product differentiation to derive a hypothesis that is tested on the NCE data. Vertically differentiated drugs on average gain larger market shares and command higher prices than horizontally differentiated drugs. Moreover, as a general rule competing substitutes have less influence on the former than on the latter. In the third essay, we develop a simple model of strategic interaction in which two agents learn about a common payoff relevant parameter. The motivating example considers two physicians who choose between two treatments, one of which has an unknown success rate. The physicians learn about the unknown treatment by prescribing it (experimenting). We contrast two information scenarios, one in which the physicians can observe the outcomes of their own treatments only, and the other in which they also can observe the outcomes from the other physician’s treatments. The pure equilibria entail an efficient amount of experimentation in both scenarios. However, strong free riding effects arise in the latter case. These are likely to cause Pareto dominated outcomes in which learning is completely thwarted. Part II: The fourth essay (joint with Jörgen W. Weibull) examines the behavior on insurance markets in a large economy when individuals have altruistic concerns for others’ welfare. The main question we address is whether strategic incentives to free ride on others’ altruism can cause insurance market failure. We also study the interaction between altruism and the adverse selection effects that arise when there is asymmetric information about the individuals’ loss probabilities. We find that if the individuals differ in their risk, and if the individual risks are observable by insurers, the degree of altruism must be (perhaps unrealistically) high in order to cause market failure. A more complex pattern is found in the case of asymmetric information: low levels of altruism increase the number of equilibria (compared to the case without altruism), while high levels of altruism cause complete market failure. The fifth essay (joint with Magnus Johannesson) also considers behavior consistentwith preferences for others’ welfare. We are concerned with how subjects allocate moneybetween themselves and others in a dictator game experiment. Deviations from the standard game theoretic prediction of the outcome in this game have been observed in numerous experiments. One possible explanation for this behavior is that individuals have altruistic concerns for others; another explanation is that individuals are motivated by reciprocity. We perform a standard double blind procedure and another design in which anonymity is guaranteed between dictators and recipients, thus removing any remaining reciprocity from the standard procedure. We could not reject the null hypothesis of no difference between the experimental groups in the two procedures. We interpret this finding as evidence of other-regarding behavior not motivated by reciprocity.