Essays in behavioural economics

Abstract

The thesis consists of three stand-alone essays. Defaults are influential, cheap to change, and therefore of great interest to policymakers. However, it is still unclear what explains their influence. Optimal Defaults and Uncertainty presents a model in which uncertainty contributes to default inertia: decision makers may be content to stick with the default and avoid the costs of learning their optimal decision. The socially optimal default policy I find differs significantly from optimal policy in models where procrastination alone drives default inertia. I show that alternative policy measures may be more effective in improving welfare, and so the effectiveness of defaults may be more limited than previous models suggest. In Screening Salient Thinkers, I explore a model of second-degree price discrimination in which consumers with context-dependent preferences choose from a menu of price-quality bundles. Specifically, the range of prices and qualities in the menu determines the weight that consumers give to the two attributes when they evaluate bundles. ‘Focusing thinkers’ place more weight on the attribute that varies the most within the menu; for ‘relative thinkers’ the opposite is true. The monopolist exploits both types of bounded rationality. In the focusing case the cost of asymmetric information is directly reduced; with relative thinkers the monopolist can use a ‘decoy good’ to extract higher revenues from all consumers. Finally How Long Is Now? explores an important degree of freedom in models of present-biased preferences: when does the present end and the future begin? First I present evidence that illustrates how economists have used this degree of freedom to explain behaviour in a variety of different contexts. Second, using a novel, between-subjects experimental design, I test a hypothesis that endogenises the cut-off between the present and the future: the ‘as soon as possible’ effect. The effect predicts that the soonest option in a menu fixes the present horizon and implies a time-specific form of menu dependence. The experimental data collected does not support the hypothesis and this result appears robust to a number of analytical approaches.