It is of deep interest to researchers and practitioners in Information System (IS) to understand the efficacy of the traditional IS and economics theory in modern business environments such as online social networks. In the pursuit to understand such new IS phenomenon and address the gap in extant literature, my dissertation, identifies the strategies that the firms should incorporate in the presence of network effects; i.e., the increases in benefits accrued by a network user with an increase in the number of users, and its impact on consumer behavior. My first essay, challenges the traditional notion that network effects create a strong protective moat for the incumbent in network competition. I show that network effects are over rated in multi-homing setting, where users can co-exist across multiple networks under resource constraints. Over reliance on the strength of network effects by the incumbent firm in multi-homing setting, stems from extant economic theory that is applicable to single homing networks, where users has to choose one of the available networks. The first essay recommends strategies for the level of innovation and its time of delivery that firms should incorporate in order to survive and succeed in multi-homing environment. While the first essay focuses on multi-homing and the strength of network effects, the second essay revisits firm's preemption strategy in single homing setting with network effects, in order to prevent its users from migrating to a new entrant with better technology. I find that, for moderate levels of price and innovation competition, an incumbent with high reputation is better off being non-committal in its preemption. In contrast, committal preemption is apt for other combinations of reputation, innovation and price. While the first two essays focus on the impact of consumer behavior on firm strategies, the third essay delves into the impact of firm strategies on consumer behavior. In particular, I identify identity revelation policies that increase the number of successful transactions and collaborations in a socially embedded marketplace. The results imply that revealing social identities may be detrimental to negotiation and collaboration in a socially embedded marketplace -- a notion that is counter intuitive to networks that are inherently social. / text

The three branches of American government---judicial, legislative, and executive---serve important governmental roles, and present their own interesting political questions. We answer three here. First, what are the differences between judges and politicians, and how does this inform the formers' selection? Second, how do senators behave to satisfy their political preferences and the electorate's? Third, what is the optimal strategy for a candidate in the Electoral College? American states select judges in various ways. In Chapter 1, we analyze "merit selection." Typically, a nonpartisan commission culls applicants for judgeships, and an appointee is selected by the governor. Then, periodically, this judge undergoes a retention election: an up-or-down vote by the state's electorate. We contribute a microeconomic model to analyze these elections. We compare this institution, in welfare terms, to others used to appoint and retain judges. Finally, we analyze a recent and ongoing phenomenon: these elections are transforming from historically rubber stamp formalities into contested, politicized contests. The politicization of issues brought before courts increases the likelihood of judges being ousted. In Chapter 2, we explore the behavior of legislators in the U.S. Senate, and of the voters who elect them. We examine shifts in incumbent senators' espoused political positions over time, as the reelection campaign approaches. We introduce novel game theoretic models of incumbent-challenger interaction. We find, through empirical analysis of senators' roll call votes, that senators moderate their positions over time, as potential reelection approaches. Moreover, this moderation accelerates. This is explained by the behavior of voters: the moderation is mirrored by the attention paid by voters. Also, the identity of an incumbent's challenger plays an important role in the amount of moderation exhibited by the incumbent. In Chapter 3, we consider a highly adaptable game theoretic model of competition in the Electoral College. It takes the form of a repeated game. Candidates make allocation decisions to persuade voters. Candidates get utility from winning office, and disutility from expending resources. We characterize optimal campaign strategy, and present comparative statics. We show, inter alia, that a candidate with an inherent advantage may prefer a longer campaign. / text

published_or_final_version / Industrial and Manufacturing Systems Engineering / Doctoral / Doctor of Philosophy

Economics is the study of the allocation of resources. Since Arrow's Fundamental Welfare Theorems, we know that competitive-markets achieve Pareto allocations when governments correct market failures. Thus, it has largely been the mission of economists to serve as 'Market Engineers': To identify and quantify market failures so the government can implement Pareto-improving policy (make everyone better without making anyone worse). Do Pareto- improving policies get implemented? How does policy become implemented? Achieving a Pareto efficient allocation of a nation's resources requires studying the implementation of policy, and therefore studying the allocation of political resources that influence policy. Policy implementation begins with the electoral process. In this dissertation, I use auction analysis, econometrics, and game theory to study political resource allocations in the electoral process. This dissertation consists of three research papers: Finance-Augmented Median-Voter Model, Vote Empirics, and Colonel Blotto Strategies. The Finance-Augmented Median-Voter Model postulates that candidates' campaign expenditures are bids in a first-price asymmetric all-pay auction in order to explain campaign expenditure behavior. Vote Empirics empirically analyzes the impacts of campaign expenditures, incumbency status, and district voter registration statistics on observed vote-share results from the 2004 congressional election. Colonel Blotto Strategies postulates that parties' campaign allocations across congressional districts may be a version of the classic Col Blotto game from Game Theory. While some equilibrium strategies and equilibrium payoffs have been identified, this paper completely characterizes players' optimal strategies. In total, this dissertation solves candidates' optimal campaign expenditure strategies when campaign expenditures are bids in an all-pay auction. The analysis demonstrates the need for understanding exactly the impacts of various factors, including strategic expenditures, on final vote results. The research uses econometric techniques to identify the effects. Last, the research derives the complete characterization of Col Blotto strategies. Discussed extensions provide testable predictions for cross-district Party contributions. I present this research not as a final statement to the literature, but in hopes that future research will continue its explanation of political resource allocation. An even greater hope is that in time this literature will be used to identify optimal "policy-influencing policies"; constitutional election policies that provide for the implementation of Pareto-improving government policies. / text

In this dissertation, I present three essays which examine questions in the field of public economics using a game theoretic approach, and I derive hopeful results and helpful rules for international negotiation. In my first chapter, I examine minimum participation constraints. In the presence of heterogeneity, a minimum participation (MP) clause in a public goods arrangement can serve as a device to create a more homogeneous group. When coalitions are restricted in what they can bargain over, exclusion of some agents from the bargaining process can be Pareto improving. This paper gives a general set of sufficient conditions for such an exclusion result to hold, and presents examples of when exclusion does, and does not, improve upon unanimity. In the second chapter, I discuss the problem of determining which externality situations merit international cooperation. I create a general framework of linearized parameters to examine a general externality problem, and then I provide the sufficient conditions for a parameter to move non-cooperative and cooperative solutions in opposite directions under certain circumstances. I argue that situations which behave in this manner and which have a higher parameter value have more benefit to cooperation through the increased range in actions to bargain over. The third chapter extends upon the second chapter and applies the framework developed to an externality problem. I present a particular story of correlation in fish growth and a corresponding model which gives an example of an increasing action gap. I describe the method of use of the framework, and using the linearized parameters developed in the second chapter, I attempt to show the divergence of non-cooperative and cooperative actions in this setting, demonstrating the need for negotiation among sovereign entities. / text

Previous studies of the influence of personality on behaviour in experimental games have provided conflicting and inconclusive results. The present investigation was designed to search on a broad front for personality correlates of behaviour in a two-person bargaining game, the one used being a derivation of the Deutsch and Krauss Trucking Game. Five personality tests, covering fifty-three personality traits, were administered to 192 undergraduate students attending courses at The University of Stirling, and from these the experimental groups were randomly chosen, the only constraint being the sex of the subjects. The tests were The Sixteen Personality Factor Questionnaire, The Guilford/Zimmerman Temperament Survey, The Study of Values Test, The Edwards Personal Preference Schedule, and The Test of Social Insight. The trucking game was played for 30 trials by two groups of subjects, each containing 24 male dyads and 24 female dyads, under two experimental conditions: Condition I, where subjects had access to full information regarding the other's payoffs, and Condition II, where only incomplete information of the other's payoffs was available. It was hypothesized that behaviour in the game would be influenced by (i) amount of information available about the payoffs of the other; (ii) sex of the players (comparing single-sexed dyads); and (iii) players' personality. No differences due to either amount of information available about the other's payoffs, or sex of the players, were found. An analysis of the data provided by the combined experimental groups, however, successfully located indications of personality effects on behaviour in the game, as measured by total joint payoff summed over 30 trials, total time taken, the number of concessions made to the other player, and first strategy-choice on individual trials. The personality variables concerned were Emotional Stability and Radicalism/Conservatism, (Factors C and Ql of The Sixteen Personality Factor Questionnaire); Personal Relations, (Factor P of The Guilford/Zimmerman Temperament Survey); Theoretical Value, (T scale of The Study of Values Test); Exhibition, ('exh' variable of The Edwards Personal Preference Schedule); and Cooperativeness, (Scale III of The Test of Social Insight). It is suggested that the relationship of these personality variables to game-playing behaviour should be the subject of further investigation.

Computer agents are required to make appropriate decisions quickly and efficiently. As the environments in which they act become increasingly complex, efficient decision-making becomes significantly more challenging. This thesis examines the positive ways in which human emotions influence people’s ability to make good decisions in complex, uncertain contexts, and develops computational analogues of these beneficial functions, demonstrating their usefulness in agent decision-making and communication. For decision-making by a single agent in large-scale environments with stochasticity and high uncertainty, the thesis presents GRUE (Goal Re-prioritization Using Emotion), a decision-making technique that deploys emotion-inspired computational operators to dynamically re-prioritize the agent’s goals. In two complex domains, GRUE is shown to result in improved agent performance over many existing techniques. Agents working in groups benefit from communicating and sharing information that would otherwise be unobservable. The thesis defines an affective signaling mechanism, inspired by the beneficial communicative functions of human emotion, that increases coordination. In two studies, agents using the mechanism are shown to make faster and more accurate inferences than agents that do not signal, resulting in improved performance. Moreover, affective signals confer performance increases equivalent to those achieved by broadcasting agents’ entire private state information. Emotions are also useful signals in agents’ interactions with people, influencing people’s perceptions of them. A computer-human negotiation study is presented, in which virtual agents expressed emotion. Agents whose emotion expressions matched their negotiation strategy were perceived as more trustworthy, and they were more likely to be selected for future interactions. In addition, to address similar limitations in strategic environments, this thesis uses the theory of reasoning patters in complex game-theoretic settings. An algorithm is presented that speeds up equilibrium computation in certain classes of games. For Bayesian games, with and without a common prior, the thesis also discusses a novel graphical formalism that allows agents’ possibly inconsistent beliefs to be succinctly represented, and for reasoning patterns to be defined in such games. Finally, the thesis presents a technique for generating advice from a game’s reasoning patterns for human decision-makers, and demonstrates empirically that such advice helps people make better decisions in a complex game. / Engineering and Applied Sciences

Essays in this dissertation cover three topics in behavioral economics: social preferences, ambiguity aversion and self-control. The first essay, based on work with Aurelie Ouss, studies the behavior of individuals making decisions to punish norm violators. It addresses two types of questions. First, what parameters affect these punishment decisions? Second, what do outcomes look like when these decisions are aggregated? Experimental data show that individual punishment decisions appear to respond to individual cost and not necessarily social cost. Additionally, individuals appear not to take the probability that violators will be apprehended into account. Finally, punishment by others does not act as a perfect substitute for own punishment. These combined effects mean that aggregate levels of punishment rarely resemble those in line with commonly used benchmarks such as optimal deterrence. The second essay, based on work with Uma Karmarkar, studies how information affects valuation of ambiguous financial prospects. Experimental results show that across several domains individual valuations appear to react much more strongly to favorable information than unfavorable information. Additional studies indicate that this effect is driven by two mechanisms. The first is a bias towards the integration of favorable information. The second is an effect of ambiguity aversion, individuals appear to be averse to subjective ignorance and so unfavorable information has a positive component: it removes some of this uncertainty. The final essay looks at how dual-self (Fudenberg-Levine (2006)) decision makers can use commitment contracts to combat self-control problems and implement long-run optimal behavior. The main results show that both stick contracts, which levy a fine when an individual gives in to a temptation, and carrot contracts, which give rewards for resisting, can simulate binding commitments. However, carrots have several advantages over sticks. Sticks create a temptation to cancel the contract, carrots are less vulnerable to trembles and finally carrots allow for more flexibility. / Economics

The dissertation is made up of three papers on the political economy of corruption and rent-seeking. Two of the papers make use of the historical experience of Britain to illustrate the theoretical points being made. The first paper shows that eighteenth-century Britain displayed patterns of corruption similar to those of developing countries today. To explain anti-corruption reforms, the paper develops a model in which the political elite is split between government officials and asset-owners. Government officials can act in one of two regimes: a corrupt one in which they are free to maximize their income from the provision of government goods, and one in which a regulated system leaves no room for individual profit maximization. Faced with a change in the level of demand for government goods, officials become able to extract rents at a level that leads to other members of the elite voting to enact reforms. The logic of the model is tested using a new dataset of members of the House of Commons and its main implications are validated. The second paper develops a model of how the British political class came to give up its power to extract rents from the economy between the 1810s and the 1850s. The key of the explanation lies in understanding the bargaining process between economic agents who seek permission to engage in economic activity and a legislature that can grant such permissions. The third paper analyzes the distributive effects of corrupt interactions between government officials and citizens. Corruption is modeled as a solution to an allocation problem for a generic government good G. Beyond a transfer from citizens to the government, corruption redistributes welfare towards "insiders" who share some natural connection to the government and to other insiders. Corruption also redistributes welfare towards those who are skilled in imposing negative externalities, and encourages the imposition of such negative externalities. / Government

Unit-demand auctions have been well studied with applications in several areas. In this dissertation, we discuss new variants of the unit-demand auction that are motivated by practical applications. We design mechanisms for these variants that have strong properties related to truthfulness, efficiency, scalability, and privacy. The main contributions of this dissertation can be divided into two parts. In the first part, we introduce a new variant of the classic sealed-bid unit-demand auction in which each item is associated with a put option; the put option of an item gives the seller the right to sell the item at a specified strike price to a specified bidder, regardless of market conditions. We motivate our unit-demand auction setting by discussing applications to the reassignment of leases, and to the design of multi-round auctions. For the classic sealed-bid unit-demand framework, the VCG mechanism provides a truthful auction with strong associated guarantees, including efficiency and envy-freedom. For an item in our auction, the strike price of the associated put imposes a lower bound on the auction price. Due to these lower bound constraints on auction prices, we find that the VCG mechanism is not suitable for our setting. Instead, our work draws on two fundamental techniques, one from the realm of mechanism design for numerical preferences -- the dynamic unit-demand approximate auction of Demange, Gale, and Sotomayor -- and one from the realm of mechanism design for ordinal preferences -- the Top Trading Cycles algorithm -- to obtain a natural auction that satisfies the lower bound constraints on auction prices. While we cannot, in general, achieve either efficiency or envy-freedom in our setting, our auction achieves suitably relaxed versions of these properties. For example, this auction is envy-free for all bidders who do not acquire an item via the exercise of a put. We provide a polynomial time implementation of this auction. By breaking ties in an appropriate manner, we are able to prove that this auction is truthful. In the second part, we specify rules for a dynamic unit-demand auction that supports arbitrary bid revision. In each round, the dynamic auction takes a tentative allocation and pricing as part of the input, and allows each bidder -- including a tentatively allocated bidder -- to submit an arbitrary unit-demand bid. Each round of our dynamic auction is implemented via a single application of the sealed-bid unit-demand auction proposed in the first part. We show that our dynamic auction satisfies strong properties related to truthfulness and efficiency. Using a certain privacy preservation property of each round of the auction, we show that the overall dynamic auction is highly resistant to shilling. We present a fast algorithm for implementing the proposed auction. Using this algorithm, the amortized cost of processing each bidding operation is upper bounded by the complexity of solving a single-source shortest paths problem on a graph with nonnegative edge weights and a node for each item in the auction. We also propose a dynamic price adjustment scheme that discourages sniping by providing bidders with incentives to bid early in the auction. / text