What is Behavioral Economics?

For students who don't know much about behavioral economics, I'll start with some definitions.

• It's a branch of microeconomics.
• Neoclassical microeconomics starts from the assumption of rational behavior: optimization with good (but not necessarily perfect or symmetric) information.
• Bounded rationality theories assume some limit of optimization.
• Behavioral economics start from empirical facts about behavior under risk and salience of state. It has a lot of overlap with bounded rationality, but they're not the same idea.
• Nudge refers to the use of behavioral economics ideas to use context (that is irrelevant to optimization) to get people to change behavior.

Doing Behavioral Economics in Shako

• "Nudge" can be researched in Shako style. With respect to empirical work, experimental studies of "nudge" are probably most feasible in Shako.
• Thinking Fast and Slow by Daniel Kahneman (2002 Nobel Prizewinner in Economics, there are good Japanese and (probably) Chinese translations), especially the chapters on "priming".
• Thaler and Sunstein's book Nudge probably has interesting suggestions. But see Data below.
• Prof. Ueichi works on regret, similar problems of testing theories I think. He would be good in AG, and probably would be willing to consult with non-AG students.

Is Prospect Theory "Improved"?

• Prospect theory has two main differences from neoclassical theory. First, it considers prospects, which are the difference between some reference point for the agent and the allocation expected, not the allocation itself, as the consideration for the agent's decision-making. Second, it proposes that consumers worry about losses significantly more than they value gains.
• I wouldn't say that prospect theory is an "improved" model over expected utility. It is different, and has advantages and disadvantages. It has two major problems that remain unsolved, and one "who cares" issue.
• One major problems is a lack of a non-parametric theory (for example, the axiomatic theory of preferences and derivation of utility from that theory)
• More important is a lack of theory of the reference point.
• The "who cares" problem is that expected utility provide a very useful theory of demand that is really hard to empirically disprove, along with a related theory of revealed preference. I don't think prospect theory can do that yet.

Practical Suggestions for Prospect Theory Research

• As far as I know most theoretical work on prospect theory is about bargaining with one-dimensional consumption bundles. What happens if a person is facing a multidimensional bargaining problem? In particular, what if their reference point is also multidimensional?
• What happens if a person is facing multiple bargaining problems, each with its own reference point?
• In Prof. Sawa's stochastic model, the bargaining game doesn't change. What happens if it does?
• Specifically, suppose the players can accumulate capital that changes their outside options?
• Assuming that these are new ideas, all can be researched without advanced mathematics, including extensions of Prof. Sawa's work. Prof. Sawa himself agrees that simulations of stochastic models are a valid research method. But all of them are likely to be possible with simple models.

Data in Shako

• Empirical research with primary data is hard to do in economics, except for very simple constructs (e.g., matrix games or individual decision problems like prospect theory or "nudging") that can be treated individually. These models can be tested with experiments.
• Questionnaire data is easy to get, but good survey instruments are hard to construct. The questionnaire approach also suffers from the problem that most surveys of interest to students have little scientific (i.e., original) content. They are based on well-known models, measuring frequently measured parameters. You can get a degree with effort but professors are rarely happy.
• Economic models at the market, industry, or macro levels generally have good availability of secondary data (prices, quantities, tax rates, and the like).
• Economic models of consumer or firm choice are generally difficult to acquire data for. Firms treat their data as trade secrets, and consumers don't want to spend time and effort to record accurate data, and often are concerned about privacy.