Robert Rosenthal - US grants

Game theory has become an indispensable tool for research activities in many branches of economics in recent years. In applied micro fields especially, where the behavior of the individual is often the focus of study, models of noncooperative games have become the primary models, most particularly when imperfect competition is an important element of the situation under study. The fact is, however, that the standard models of noncooperative game theory do not describe real-world behavior very well. The goal of this project is to build game models which more accurately describe and predict the behavior of real-world participants in economic activities. The models developed by this project include: (1) Evolutionary Modelling; (2) Models of Bounded Rationality for Game Situations: and (3) New Models that incorporate observed behavior such as a stick-with-winners philosophy associated with much trial-and-error or avoidance of situations in which unsuccessful outcomes have come about in the past. In the standard approaches to game-theoretic modelling, players are assumed to be completely rational and assumed to assume that their fellow players are completely rational. For applied purposes, these assumptions seem dubious, especially given the experimental evidence compiled over the years on both one- player and many-player games. This research project aims at constructing and developing game-theoretic models in which players are not assumed to behave in such completely-rational ways, with the goal being more realistic descriptive and normative tools for use in applications, especially economic applications. There is already a literature on the subject of modelling games with players of incomplete rationality, but the subject seems to be resistant to breakthroughs so far. The research pursues several approaches that could lead to breakthroughs. One is to continue the line of research in which strategy selection comes about as the result of evolutionary forces acting so as to favor those strategies which do relatively well in large populations. Another direction is to continue work on models in which players are assumed to act in particular nonoptimal ways. In both of these classes of models, behavior can be generated which, in certain examples, is both intuitively plausible and qualitatively different from the standard game- theoretic solutions prescribed.

A key assumption in traditional game theory is that all players are completely rational. This sometimes leads to absurd conclusions. One part of this project is to continue exploring the implications of more evolution-based hypotheses about behavior in games. Within economic theory, three problems are addressed. First, recent strategic-game models of the problems associated with the massive debt owed by certain sovereign nations to private commercial banks have yielded provocative but counterfactual predictions. The goal here is to see whether the results change. Second, in some economic games a counterintutive theoretical result has emerged: that more competition on one side of a market leads to a price change in the wrong direction. The intention here is to explore the general class of market games in which this phenomenon arises. Third, adopting the view that government decision-making is a complicated ongoing bargaining process involving competing and overlapping constituencies, bargains struck at one point in time necessarily have implications for what will confront future bargainers; yet there is no way that compromises can be agreed to by the as-yet-nonexistent players. The aim here is to explore the extent of the social inefficiencies that inevitably result. The research on evolution and bounded rationality in games is especially exciting. One of the major problems with bounded rationality is modeling the limitations on rationality in a way that is not arbitrary or only descriptive. In this project economic agents use a variety of rules of thumb for bounded rationality reasons. These rules are applied across a number of different social situations and evolutionary processes. The frequencies of the specific rules found in a society are derived. For instance, one might imagine that in a society in which the rule of thumb "avoid risky investments" was widely practiced, the prospects for high payoff (at least temporarily) from more aggressive strategies might be better on average than in other societies, although not in all situations. This in turn might lead to an invasion of such strategies and the subsequent relative success in turn of more conservative strategies. This project develops analytic tools that permit us to determine under what conditions a steady state with a blend of behaviors comes about and under what conditions perpetual periodic waves occur.

One of the characteristics of standard modelling in decision theory, game theory and economic theory has been the postulate that all the people in the models are completely rational; they have well-specified objectives, unlimited memories, and unlimited abilities to compute without using up time or other resources. In recent years discontent with this postulate has been growing within economics because of the cumulative weight of years of data from experiments consistently showing that even in very simple decision problems and games the theories do not predict the actions of subjects very well. This project explores alternatives to the standard optimization model. In these alternatives, each decision-making situation is a pattern recognition problem in which the individual decision maker decides which of a small set of decision making tools ("rules of thumb") or actions to use. The project addresses such questions as how did the individual's set of rules and method of assigning rules to decision making problems come to be and how do they evolve. More specifically, the project examines two broad classes of models. In the first class there are single decision-maker models in which the decision maker's choices influence the path of some stochastic process about which the decision maker has little information but which affects his well-being. The effect of certain crude rules of thumb on the long-term survival of an individual is explored in a variety of contexts. The second class builds on previous work by the investigator on bounded rationality models in which individuals use rules of thumb to conserve on decision-making costs and to overcome imperfect reasoning and computing abilities. A population of individuals interacts over time using crude rules of thumb in a variety of settings. Here the focus is on what profiles of rules can survive over time as poorer performing rules are weeded out. The models used in the latter part of the project are adapted from the literature on anonymous random matching games.

9409090 Rosenthal This research will focus on three related topics. In the first project, game-thoeretic models are developed which focus on incentive schemes to implement desirable public good allocations in situations where information is incomplete for some social groups. In the second project, models are developed which deal with the interacting dynamics between two social groups. The goal is to explore under what conditions would opportunism drives out cooperation if individual members of one group behave in an opportunistic manner against the individuals of the other group. The third project develops models of communication and coordination. The aim is to assess the effects of lags in the flow of economic information across groups in the absence of market price signals. The study is of significant interest because it would enrich our understanding of how society is organized and the different ways individual behavior respond to that organization.

Project Abstract Despite their widespread application, there are well known examples in which economic theory, decision theory, and game theory fail to describe human behavior realistically. Since these theories are used in public policy analysis, the prescriptions of such analyses can be questioned on this ground. This research program aims to further understanding of the conditions under which the standard theories fail and to develop new theories for use in such conditions. The research is divided into four projects. The first of these is a collection of laboratory experiments concerning two person zero sum games. This is the class of games for which the theory of games was initially invented, and for which it might be expected to match reality best. Past laboratory experiments on such games have indicated that there are senses in which the theory describes reality well and senses in which it does not. The experiments in this project are designed to further clarify of this issue with the eventual purpose of contributing toward a modified theory that fits reality better. The second project is concerned with the further development of a novel theory in which economic actors are assumed to be involved in such a complicated world that they have no choice but to engage in either imitative or experimental modes of behavior. The analysis of an earlier prototype mathematical model produced long run patterns that were delicately balanced between stability and instability. Since it is important to know when an economic system is stable, in this project variations on the theoretical prototype are studied in order to see when the instabilities are robust and whether such model instabilities are related to instabilities in real economies. The work involves both mathematical analysis and computer simulations. The third project is an attempt to merge the classical theory of games with the more recently developed evolutionary approach to game theory. The two approaches appear to describe reality well in different circumstances; the hope is that a hybrid theory might do better than the two theories separately. Project four stays within the confines of classical game theory. It is an attempt to generalize and see what is behind a peculiar and poorly understood pattern of equilibrium behavior that has turned up in a diverse collection of applications: from models of auctions to models of labor market screening. In these models, individuals randomize their actions in order to keep them secret from opponents, but they do so in ways that are internally highly structured.

? DESCRIPTION (provided by applicant): Less than 50% of cardiac arrest (CA) survivors exhibit good neurologic outcome, emphasizing the need for new neuroprotective strategies in addition to meticulous management of temperature. Our research performed with a canine model of CA and resuscitation (ROSC) demonstrated neuroprotection with oximetry-guided normoxic resuscitation compared to the previously standard practice of hyperoxic resuscitation. These results contributed to a major change in AHA/ACLS guidelines for CA/ROSC; i.e., minimize ventilatory O2, maintaining hemoglobin oxygen saturation >94%. While these procedures can be safely used in-hospital for CA/ROSC, the risk of hypoxia associated with rapidly lowering inspired O2 makes this paradigm dangerous in pre-hospital resuscitation. In light of these limitations, our primary aim is to determine the level of O2 inspired during the firt 2 hr of critical care in a hospital setting that optimizes neurologic outcome following pre-hospita resuscitation. We hypothesize that in contrast to the benefit of normoxia during early resuscitation, maintenance of moderate hyperoxemia at the period following the initial reperfusion-induced free radical surge, and prior to the onset of inflammation, will improve clinical outcome. Our related, albeit independent secondary aim is to test the hypothesis that inflammation, oxidative stress, and brain mitochondrial dysfunction contribute substantially to post-ischemic brain injury. Comparisons will be made of neurologic, histologic and biochemical outcomes following normoxic, mildly hyperoxic, and severely hyperoxic ventilation and in the absence or presence of sulforaphane-induced expression of cytoprotective genes whose products protect against these injury mechanisms. Methods of approach include use of our highly clinically relevant canine model of CA/ROSC for short-term outcomes, and a rat CA and resuscitation model for long-term outcomes. Additional comparisons between males and females will enhance potential for clinical translation and detect any sexually dimorphic mechanisms of brain injury and responses to different O2 levels or sulforaphane treatment. Translational outcome measures include advanced histopathology and neurobehavioral tests. Mechanistic outcomes include measurements of mitochondrial bioenergetics, cerebral metabolism of 13C-labeled glucose, proton NMR of energy metabolite levels, inflammatory microglial activation, and markers of oxidative stress. Relevance: Results from our studies will provide fresh new insight into the levels of inspired O2 used in a hospital setting that result in best neurologic outcome after out-of-hospital CA/ROSC. These experiments will also determine if treatment with sulforaphane after resuscitation further improves neurologic function, based on stimulated expression of cytoprotective gene products that inhibit oxidative stress, inflammation, and mitochondrial dysfunction. Either approach toward neuroprotection could be safely translated to clinical trials, eventually improving the quality of life experienced by the hundreds of thousands who survive CA each year.