Explorations in the Economics of Choice and Chance

This award funds research in economic theory. The researcher plans four distinct projects. The first models an individual who is searching across possible options one-by-one. The goal is to understand the best way to search when the individual already knows that there is a specified number of options and has some information about options. This project will help us understand the value of providing information to people before they search through structured search lists. Applications include optimal design of web searches and optimal design of job/ hiring searches. The second project builds and analyzes a model of medical testing, to determine optimal testing protocols when definitive clinical tests are more expensive than less informative alternatives. The third project will develop a modified model of contagions that accounts for evasive human behavior that increasingly slows the contagion as it grows more prevalent. The model could contribute to more accurate epidemiological forecasting. The final project will look at demand for risky investments, and in particular investments that are similar to rollover lotteries. The goal is to better understand why these investments are rapidly growing and how best to design them. This project contains four methodologically related sub-projects on the economics of choice and chance. The first research project will develop a sequential model of search in environments with finitely many options and prior knowledge of the options. This is an essential innovation on search theory, since rarely does one search in the stereotypical informational vacuum. This project rather explores search that transpires in smartly ordered lists, as offered in web search contexts. Recall optimally occurs, with older options more often recalled. Early work suggests an answer to a long open question: one expects to search longer when the prize distribution is more disperse. The second project will enrich a standard epidemiological model with a random pairwise matching population game where individuals choose how vigilantly to avoid acquiring it. The unique Nash equilibrium will yield a behavioral twist on the SIR Model. We will estimate this parametric model with data from the 2009 H1N1 influenza pandemic, and show that allowing for human ingenuity facing an ongoing contagion improves the explanation of the data. The third project will explore how should one proceed with costly but informative medical tests before acting. Our main qualitative conjecture, motivated by numerical simulations, will determine when one transitions from low cost to high cost medical testing, depending on the testing precision or medical urgency. The main conjecture of the project upends an insight from a well-known 2001 paper on the optimal level of experimentation that assumed continuous time. The final project will explore an implicit market cleared by probabilities. This proposes a simple economic explanation for lottery gambling that is consistent with the larger set of facts revealed by rollover lotteries: People gamble because they derive utility both from the money gains (varying across people), and the gambling experience. The results are applicable to a range of investments with risky payoffs. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.