|
I examine economic design issues in the realm of dynamic organ allocation for transplantation and behavioral market design/contract theory.The second and third chapters focus on two issues in the design of the U.S. deceased-donor organ allocation system, which represents the majority of transplants performed in the U.S. In contrast with the dominant trend in Economics, which pays attention almost exclusively to kidney allocation, I analyze two market failures in allocating hearts and livers in the U.S. The fourth chapter, on the other hand, provides an information theoretical perspective on recent findings documenting failures to play dominant strategies in dominant strategy mechanisms. In the second chapter, I study a long-standing controversy on the principles for allocating organs for transplantation from a dynamic contracting perspective. This approach is motivated by the fact that in the allocation of hearts and livers, the U.S. transplantation authority has recurrently faced a pervasive problem of asymmetric information about transplant candidates' medical urgency. I investigate the optimal design of prioritization rules under different social welfare functions while taking patients' incentives to misrepresent medical needs into account. I find that the current prioritization scheme is not incentive compatible. Moreover, I deliver a way of incentivizing truth-telling by prioritizing patients using their medical urgency histories, which is remarkably different from how the latter are used in practice. Despite screening is possible using promises of organ consumption in the future, when the social objective is to prioritize medically urgent patients, its optimality is ambiguous. In some parameter regions, pooling is preferred. In sharp contrast, when social preferences are utilitarian, pooling is preferred most of the time. In coauthorship with Brett Gordon and Nitin Katariya, in the third chapter, we study a previously unknown market failure in allocating deceased donor livers due to the reliance on measures of candidates' medical urgency that vary over time. It is infeasible to collect this information in real-time because blood tests are necessary to measure medical urgency. To ensure allocations are made using updated information, the U.S. transplantation authority requires transplant centers to submit candidates' information according to a schedule. Using a unique data set provided by the U.S.~transplantation authority, we find that in 7,502 of 19,295 liver allocations, at least one transplant candidate was prevented from getting an organ offer because the actual recipient was ranked higher thanks to the latter having outdated information in the national system. We find that among the most urgent patients between 2002 and 2013, this friction generated the loss of the equivalent of 2.7% of the total life years gained through transplantation. Studying the timing of candidate reports, we find empirical patterns consistent with a set of strategic responses to the existing update schedule, which have an ambiguous effect on welfare. Although transplant centers have an incentive to quickly report any candidate whose condition is deteriorating, which helps to alleviate the informational friction, they should prefer to delay reporting candidates with improving conditions to maximize these candidates' chances of getting livers. We discuss the feasibility of simple policy changes that could help to mitigate the strategic delay in reporting patients' scores. The fourth chapter studies a general issue in market and mechanism design from a behavioral perspective. It is well-known that, in mechanism design applications, participants often fail to pin down their optimal strategy, even if there is a dominant strategy to play. In collaboration with Daniel Martin, we study the link between misperception of payoff generating rules and play of dominant strategies in the Becker-DeGroot-Marschak mechanism, a single-agent decision problem widely used to elicit subject valuations. Using a rational inattention framework, we show that a decrease in the cost of processing information increases the frequency of play of the BDM dominant strategy. Building on the recent literature on contingent thinking, we conjecture that the difficulty of determining the dominant strategy is related to the easiness of understanding the payoff generating rule. In an experimental setting, we study the impact of framing payoffs contingency-by-contingency by comparing behavior under versions of the BDM with and without this framing, which is intuitively more straightforward. We find that this frame substantially reduces choice mistakes. We determine how much better-informed subjects are about contingent payoffs under this frame using both stated beliefs and an information-theoretic model. The treatment effect is comparable to the impact of learning with feedback. The contingent framing also reduces information gaps across demographic groups.
|
