[pt] Esta tese de doutorado compõe-se de três artigos, sendo dois em finanças empíricas e um em microestrutura de mercado. O primeiro artigo estuda de que forma movimentos nas curvas de juros futuros em Reais e Dólares Americanos negociados na BM&F estariam relacionados com duas medidas de prêmio de risco cambial, uma à priori, calculada com base nas expectativas de variação cambial três meses à frente apuradas pelo Focus-BC, e outra à posteriori, calculada sobre a variação cambial efetiva realizada nos mesmos três meses. Os resultados mostram que movimentos da curva de DI parecem mais correlacionados com a variação cambial efetiva do que com as expectativas coletadas entre os agentes. O segundo artigo é uma variação do modelo de Ang e Piazzesi (2003), e investiga a contribuição do mercado de câmbio sobre o prêmio a termo na curva de juros futuros em Reais no Brasil. Usa-se uma UIP no lugar de uma Regra de Taylor para modelar a dinâmica da taxa de curto prazo, o que nos permite substituir as variáveis macro usuais de inflação e produto pela expectativa de variação cambial e prêmio de risco cambial na especificação do prêmio a termo na curva. O terceiro artigo propõe um modelo de mercado interdealer em três estágios onde o processo de revelação de informação é modelado como um sinal ruidoso e invertido de forma seqüencial nos dois estágios de negociação no mercado inter-dealer que se seguem à transação inicial. As simulações realizadas sugerem que a diversificação de risco na economia diminui quanto maior a precisão do sinal nos dois estágios. / [en] In this thesis we discuss two empirical essays in finance and one in market microstructure. The first article studies the joint dynamics of the two most liquid term structure of interest rates traded at BM&F, one in Brazilian reais and the other in US dollars, and two currency risk premia measures. One currency risk premia measure is obtained using currency expectation surveys conducted by the Central Bank of Brazil, while the other will be residual from the three month forward premium traded each day and the effective currency observed on the liquidation date three months after. Results show that the term structures will explain some of the realized currency risk premia observed three months after. We see this as an evidence in favor of information in the curves more correlated to the effective currency movement in three months than the expected devaluation. The second article proposes and extension of the framework introduced by Ang and Piazzesi (2003) to accommodate a no- arbitrage term structure model with macro factors. We replace the usual inflation and output macro factors for two currency variables, the expected currency devaluation and the currency risk premia. Results here show a better fit when compared to existing models estimated for Brazil. The third article proposes an inter-dealer market model in three stages, where disclosure of information is modeled by noisy informative signals. Simulations show that dealers better informed will play strategically to avoid revealing information and the risk-sharing in the economy will be lower when we increase the precision of the informative signals.

Market microstructure studies the process of exchanging assets underexplicit trading rules. With algorithmic trading and high-frequencytrading, modern financial markets have seen profound changes in marketmicrostructure in the last 5 to 10 years. As a result, previously establishedmethods in the field of market microstructure becomes oftenfaulty or insufficient. Machine learning and, in particular, reinforcementlearning has become more ubiquitous in both finance and otherfields today with applications in trading and optimal execution. This thesisuses reinforcement learning to understand market microstructureby simulating a stock market based on NASDAQ Nordics and trainingmarket maker agents on this stock market. Simulations are run on both a dealer market and a limit orderbook marketdifferentiating it from previous studies. Using DQN and PPO algorithmson these simulated environments, where stochastic optimal controltheory has been mainly used before. The market maker agents successfullyreproduce stylized facts in historical trade data from each simulation,such as mean reverting prices and absence of linear autocorrelationsin price changes as well as beating random policies employed on thesemarkets with a positive profit & loss of maximum 200%. Other tradingdynamics in real-world markets have also been exhibited via theagents interactions, mainly: bid-ask spread clustering, optimal inventorymanagement, declining spreads and independence of inventory and spreads, indicating that using reinforcement learning with PPO and DQN arerelevant choices when modelling market microstructure. / Marknadens mikrostruktur studerar hur utbytet av finansiella tillgångar sker enligt explicita regler. Algoritmisk och högfrekvenshandel har förändrat moderna finansmarknaders strukturer under de senaste 5 till 10 åren. Detta har även påverkat pålitligheten hos tidigare använda metoder från exempelvis ekonometri för att studera marknadens mikrostruktur. Maskininlärning och Reinforcement Learning har blivit mer populära, med många olika användningsområden både inom finans och andra fält. Inom finansfältet har dessa typer av metoder använts främst inom handel och optimal exekvering av ordrar. I denna uppsats kombineras både Reinforcement Learning och marknadens mikrostruktur, för att simulera en aktiemarknad baserad på NASDAQ i Norden. Där tränas market maker - agenter via Reinforcement Learning med målet att förstå marknadens mikrostruktur som uppstår via agenternas interaktioner. I denna uppsats utvärderas och testas agenterna på en dealer – marknad tillsammans med en limit - orderbok. Vilket särskiljer denna studie tillsammans med de två algoritmerna DQN och PPO från tidigare studier. Främst har stokastisk optimering använts för liknande problem i tidigare studier. Agenterna lyckas framgångsrikt med att återskapa egenskaper hos finansiella tidsserier som återgång till medelvärdet och avsaknad av linjär autokorrelation. Agenterna lyckas också med att vinna över slumpmässiga strategier, med maximal vinst på 200%. Slutgiltigen lyckas även agenterna med att visa annan handelsdynamik som förväntas ske på en verklig marknad. Huvudsakligen: kluster av spreads, optimal hantering av aktielager och en minskning av spreads under simuleringarna. Detta visar att Reinforcement Learning med PPO eller DQN är relevanta val vid modellering av marknadens mikrostruktur.