<p>This study examines several emerging topics in strategic communication and information design. The first chapter studies the role of verification in persuasion and its interaction with commitment in a Bayesian persuasion framework in which the sender is not fully bound by the committed plan. Both theoretically and experimentally, we demonstrate that making verification easier can significantly improve information transmission when commitment is low, but its effect is limited when commitment is high. However, empirically receivers do not respond as strongly as predicted by theory, which is consistent with base-rate neglect and conservatism. On the other hand, senders generally anticipate receivers' actions and best respond to the empirical behavior of receivers. We provide empirical implications for cases where verification is instrumental in improving information transmission and where it is not. </p> <p><br></p> <p>The second and third chapters are based on joint work with Dr. Collin Raymond. In the second chapter, we investigate how increasing the complexity of the message space in the presence of limited memory can reduce misrepresentation in strategic communication. We enrich a standard cheap talk game so that senders must communicate not just a payoff-relevant state, but also payoff-irrelevant attributes correlated with the state.  We show that increasing the set of attributes that may need to be reported (i.e., the complexity of the game) improves the amount of information transmitted in equilibrium. Our findings demonstrate that the reporting of redundant information may induce equilibria that feature improved outcomes compared to simpler, more direct reporting systems, and point out the importance of complexity when trying to induce truthful information revelation.</p> <p><br></p> <p>In the third chapter, we analyze some extensions on the effect of complexity. We present experimental evidence which shows that  too much of an increase in complexity leads to a reversal of those gains. Limited memory on the part of players, as well as the relative complexity faced by senders and receivers, drives these changes, and individuals experience cognitive costs when dealing with complex environments that they are willing to pay to avoid. </p> <p><br></p>

Submitted by Caio Paes Leme Lorecchio (caio.lorecchio@gmail.com) on 2017-06-07T16:01:31Z No. of bitstreams: 1 Bad Reputation with Rating Systems.pdf: 556649 bytes, checksum: 01e7c1222ac8b0ed3258d714f7adc3b2 (MD5) / Rejected by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br), reason: Boa tarde Caio, Por favor, tirar o acento da palavra Getulio em todas as páginas, Agradecimento/Resumo/Abstract em caixa alta, letra padrão ABNT ou APA (capa.....) e tirar a linha do cabeçalho conforme conversamos. Abs. Suzi 3799-7876 on 2017-06-07T18:39:31Z (GMT) / Submitted by Caio Paes Leme Lorecchio (caio.lorecchio@gmail.com) on 2017-06-07T19:58:22Z No. of bitstreams: 1 Bad Reputation with Rating Systems.pdf: 373018 bytes, checksum: 6855d5f3fc595b138d084679ef3eeabe (MD5) / Approved for entry into archive by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br) on 2017-06-08T11:43:14Z (GMT) No. of bitstreams: 1 Bad Reputation with Rating Systems.pdf: 373018 bytes, checksum: 6855d5f3fc595b138d084679ef3eeabe (MD5) / Made available in DSpace on 2017-06-08T12:52:00Z (GMT). No. of bitstreams: 1 Bad Reputation with Rating Systems.pdf: 373018 bytes, checksum: 6855d5f3fc595b138d084679ef3eeabe (MD5) Previous issue date: 2017-05-11 / Este trabalho analisa um modelo de má reputação com sistemas de rating como uma forma particular de memória limitada. Em cada período, um cliente preocupado apenas com ganhos correntes escolhe se contrata ou não um especialista. O cliente compreende as regras de transição do sistema, mas observa apenas a realização de um rating (uma nota) que carrega informação sobre o provável tipo de especialista para tomar a decisão de contrato. Um especialista do tipo estratégico escolhe prover ou não o tratamento correto quando contratado e um especialista do tipo ruim sempre oferece o tratamento mais caro, independentemente do problema observado. Quando clientes observam todo o histórico de interacões, um especialista estratégico ou tem fortes incentivos para oferecer o tratamento barato (quando o tratamento correto seria o mais caro) ou eventualmente a crença no mercado de que ele é do tipo ruim é suficientemente grande para que deixe de ser contratado. Quando clientes possuem apenas o sistema de rating como fonte de informação, este trabalho demonstra que não apenas é possível evitar esse efeito negativo, como também é possível aumentar os ganhos de equilíbrio em comparação à ausência de qualquer sistema informacional. Ademais, este trabalho desenha os sistemas ótimos do ponto de vista tanto do cliente quando do especialista para todas as crenças iniciais, discutindo como eles diferem em um sistema de dois estados e quando há ganhos de eficiência. / We study a bad reputation model with rating system as a special form of limited memory. At each period, a myopic customer knowing the rules of the system but observing only a current public realization of a finite set of states uses this information to infer expert's type and take hiring decisions. A strategic expert chooses whether or not to provide correct treatment whenever hired and a bad (committed) expert always proposes an expensive treatment. With full memory, a patient expert cannot refrain from gaining reputation of being bad or lying to separate herself from a bad type. With rating systems, we show that it is possible not only to overcome bad reputation effect, but generate higher equilibrium outcomes relative to trivial information censoring (no memory at all). We characterize optimal systems from customer and strategic expert's point of view in a two-state setting for all prior beliefs and show how they differ and when a rating system can bring efficiency to experts' markets.

This thesis consists of three essays in the field of information economics. The first essay studies manipulation of information by partisan media. The recent increase in partisan media has generated interest in what drives media outlets to become more partisan. I develop a model to study the role of diffusion of information by word of mouth. In the model, a media outlet designs an information policy, which specifies the level of partisan slant in the outlet’s news reports. The news spread via a communication chain in a population of agents with heterogeneous preferences. The slant has an impact on whether the agents find the news credible and on their incentives to pass the news to others. The analysis elucidates how partisanship of media can be driven by political polarisation of the public and by the tendency of people to interact with people with similar political views. The second essay, co-authored by Jakub Redlicki, investigates falsification of scientific evidence by interest groups. We analyse a game between a biased sender (an interest group) and a decision maker (a policy maker) where the former can falsify scientific evidence at a cost. The sender observes scientific evidence and knows that it will also be observed by the decision maker unless he falsifies it. If he falsifies, then there is a chance that the decision maker observes the falsified evidence rather than the true scientific evidence. First, we investigate the decision maker’s incentives to privately acquire independent evidence, which not only provides additional information to her but can also strengthen or weaken the sender’s falsification effort. Second, we analyse the decision maker’s incentives to acquire information from the sender. The third essay analyses competition between interest groups for access to a policy maker. I study a model of lobbying in which two privately-informed experts (e.g., interest groups) with opposite goals compete for the opportunity to communicate with a policy maker. The main objective is to analyse the benefits which competition for access brings to the policy maker as opposed to hiring an expert in advance. I show that competition for access is advantageous in that it provides the policy maker with some information about the expert who did not gain access and gives the experts an incentive to invest in their communication skills. On the other hand, hiring an expert in advance allows the policy maker to use a monetary reward to incentivise the expert to invest more in his communication skills.

Diese Dissertation besteht aus drei unabhängigen Artikeln in dem Forschungsfeld der Informationsökonomik. Ein wiederkehrendes Motiv in allen drei Artikeln ist die ambivalente Rolle von privater Information. In Kontrast zur klassischen Entscheidungstheorie, in der mehr Informationen Individuen niemals schlechter stellt, analysiere ich drei verschiedene Umgebungen, in denen mehr Konsumenteninformation die Konsumentenrente verringern kann. / This dissertation comprises three independent chapters in the field of information economics. The recurrent theme of all three chapters is the ambiguous role of information: While in standard decision theory additional information enables individuals to weakly increase utility through making better choices, I analyze three di erent environments in which more information to consumers may actually be detrimental to consumer utility.