Economics equates rationality with the satisfaction of a set of axioms. We discuss these axioms and the associated empirical evidence. Topics discussed include completeness and transitivity of preferences, limited attention, overconfidence, and whether humans can successfully construct the full sampling distributions. We then consider procedural rationality and the unreasonable cognitive requirements for solving simple dynamic programming problems. We examine alternatives to mathematical optimization in which people use simple rules of thumb (heuristics) to make decisions. We highlight the heuristics and biases program which clearly shows that the “as if” assumption in neoclassical economics is routinely violated. These heuristics include the representativeness heuristic, the gambler’s fallacy, the hot hand fallacy, the conjunction fallacy, the availability heuristic, the affect heuristic, the anchoring heuristic, base rate underweighting in Bayes’ law, conservatism from underweighting the likelihood of a sample, hindsight-bias, confirmation-bias, false consensus, and regression to the mean. Incentives do not eliminate biases. We also discuss the “great rationality debate.”

This chapter first describes the generic situations of risk, subjective uncertainty, ambiguity, and true uncertainty. It then outlines the main decision theory under risk in neoclassical economics, expected utility theory (EU). We outline EU’s axiomatic structure and highlight the independence axiom that has often been rejected by the evidence, as exemplified by the Allais paradox. The two main drawbacks of EU are that it does not allow for individuals who (i) derive utility from changes in outcomes relative to a reference point (although this is consistent with it's underlying axioms) and (i) weight probabilities in a nonlinear manner. Other violations of EU are also considered; these include description invariance and preference reversals.

The introductory chapter sets the stage for the rest of the book. It begins by tracing the historical development of behavioral economics, then giving a working definition of the subject, followed by a discussion of the scope of behavioral economics. In particular, the various kinds of data that form the basis of inferences in behavioral economics are discussed. These include observational data, experimental lab data, field data, survey data, and neuroeconomic data. It is argued that no particular source of data holds hierarchical sway over any of the other sources of data. All sources of data have their pros and cons, and we need all of them to build a more rounded understanding of human behavior. We lay great emphasis on the scientific method and describe best practice in the natural sciences, following the Popperian tradition. It is vital that students get themselves well versed in these ideas to put them in the right frame of mind for taking a course in behavioral economics.

Chapter 2 relaxes two features of EU: linear probability weighting and utility defined over final wealth levels. Our main focus is on prospect theory (PT). The salient features of PT are reference dependence (utilities are derived from changes in wealth relative to a reference point), loss aversion (losses bite more than equivalent gains), and inverse S-shaped probability weighting functions that replace the “linearity of probabilities” in EU. Under PT, attitudes to risk are determined “jointly” by the shapes of the utility function and the probability weighting function, giving rise to a four-fold pattern of risk attitudes. We also give an exposition of rank dependent utility. We consider several applications of prospect theory. These include, exchange disparities; optimal contracts; tax evasion puzzles; backward bending labor supply curves; attitudes towards low probability events, and loss aversion among professional golf players. Close primate relatives also exhibit loss aversion; thus, loss aversion precedes the evolutionary separation of humans from close primate relatives. We also apply PT to the Ellsberg paradox.

Extensive empirical evidence demonstrates that humans possess other-regarding preferences, that is, they care about the well-being of others, in addition to their own. Humans are predisposed to cooperate with each other, but they are conditional cooperators; they respond to kindness with kindness and unkindness with unkindness. The intentionality of actions by others is important in judging the unkindness of their actions. This chapter explores the evidence on human sociality, using several experimental games such as: the ultimatum and dictator games, the trust and the gift exchange games, and the public goods game with and without punishments. The external validity of lab experiments is also considered. Other topics include the evolutionary origins of preferences and the behavioral differences between WEIRD and non-WEIRD societies. We consider the role of human morality and the aversion to lying and breaking promises. Even when lies cannot be discovered, a significant fraction of people chooses to remain honest, while others tell partial lies and only a few lie maximally. We also outline social identity theory, whereby humans treat ingroup members more favorably relative to outgroup members.

In this chapter we consider a variety of non-optimizing mental shortcuts that people use to make sense of the complex world around them. We consider mental models; narrow bracketing and broad bracketing; the diversification heuristic; and the difference between transaction utility and acquisition utility. We then consider mental accounting, whereby people form different mental categories for different sources of incomes/expenditures and money is not fungible across them. This explains diverse phenomena, for example, why people prepay for vacations and post-pay for consumer durables. We consider projection-bias, the tendency to project our current preferences in inferring our future tastes and actions. This is followed by a discussion of Herbert Simon’s aspiration adaptation theory, which gives a boundedly rational approach to decision making. It highlights the procedural aspects of decisions. The final topic in this chapter is behavioral finance. We give a statement of efficient markets hypothesis (EMH), which is the cornerstone of modern finance. We show that EMH is rejected by the evidence. We also give a brief note on corporate finance.

In neoclassical economics, choices reflect considered judgment, so one can use revealed preference arguments for welfare judgments. The evidence from behavioral economics calls this view into question. This chapter considers the proposed modifications to neoclassical welfare economics when people might not be the best judges of their actions. The generic form of paternalism in behavioral economics is soft paternalism that nudges the choices of boundedly rational individuals in a beneficial direction, but the choices of rational individuals are minimally affected. We consider several forms of soft paternalism: libertarian paternalism, asymmetric paternalism, and light paternalism. Our main focus is on libertarian paternalism, and we consider the evidence from several forms of nudges as well as the importance of distinguishing between partial and general equilibrium effects of any policy intervention. We consider the inevitability of the choice architect and the concept of informed consent. We critically evaluate the criticisms of this approach, for example, those by the contractarian approach. We consider applications to sin goods and to the foundations of behavioral public economics.

This chapter covers selected topics in behavioral game theory. We begin with the winner’s curse in common value auctions where players overbid in auctions and make losses. We then consider standard bargaining models such as the Nash bargaining outcome and alternating offers bargaining; the evidence does not support the latter. We introduce evolutionary game theory and the main equilibrium concept, evolutionary stable strategies (ESS). This is followed by a description of replicator dynamics that gives rise to ESS, and a study of its properties. A range of examples illustrate the underlying concepts. We then consider stochastic dynamics generated by an underlying source of persistent randomness and introduce a stochastically stable steady state. Finally, we consider psychological game theory in which beliefs directly enter the utility function allowing for a rigorous modelling of emotions such as guilt and shame.