In this thesis I presented two essays motivated by the observation that financial crises tend to be accompanied by deeper recessions and slower recoveries, partly due to debt burden (e.g. Reinhart & Rogoff, 2009; Hong and Tornell, 2005; Jordà, et al., 2013). In the first essay I evaluate this claim against the contrasting view that magnitude and persistence of recessions is rather the consequence of bigger and more persistent shocks (Stock & Watson, 2012). To do so, I compute recovery and recession paths through the estimation of impulse responses by local projections methods (Jordà, 2005). I found that the occurrence of financial crises is associated with more severe recessions only if the recession itself is big enough. But this effect disappears when the output loss caused by the recession is below the historical average. More importantly, neither the magnitude of the loss, nor the occurrence of financial crises, nor debt accumulation are associated with sluggish output growth during recoveries. It has also being suggested that expectations prior to the crisis help to determine the magnitude and length of recessions following financial crises (Chauvet and Guo, 2003; Cerra and Saxena, 2008). This and the role of pre-crisis dynamics is not properly reflected in standard DSGE models. In the second essay I account for the effect of pre- crisis dynamics and evaluate whether financial crises are different. To do so, I introduce optimism (in the form of unrealised news about capital quality) in an otherwise standard DSGE model with financial frictions. Under this framework, optimism generates investment–debt / boom-bust cycles accompanied by long recessions. I found that within this framework cycles associated with financial and technology news shocks are different regarding the responses of asset prices and banks’ net worth. Real variables respond similarly to unjustified financial or technological optimism.

This thesis consists of three chapters. In the first chapter, I explore a two-period economy with a three-tier hierarchy in which the principal without full commitment decides how and when to motivate a productive intermediary (agent one) to privately sub-contract and collaborate with another agent (agent two) on a project with uncertain quality. The dynamic moral hazard problem arises due to the agents’ hidden effort choice and the opportunity for future work. Besides free riding, the agent one’s exclusion and over-investment incentives need to be considered due to his private sub-contract option. Both the dominant incentive constraint in the optimal short term contract and the principal’s investment decision depend on the project’s value-cost ratio, the level of synergy in the partnership and the amount of patience. In general, the principal under-invests and stops earlier compared to the first-best outcome. However, there exist scenarios in which agent one always over-invests when the individual work is motivated, and the principal might compromise to motivate a higher effort level by over-investing relative to the static game, especially if the synergy is positive but small and the project’s value-cost ratio is medium. In a two-tier hierarchy, the principal can be weakly better off, but the inefficiency caused by agent one’s private link to the other agent still exists. In the second chapter, I study how a principal motivates an uninformed agent to learn about, and reveal, his quality through private experiments. The principal commits to a reward scheme and she aims to assign the rewards to correspond as closely as possible to the quality of the agent. To get a high reward, the agent experiments privately and discloses the results selectively. I show that the optimal reward scheme features an increasing step function: the initial steps encourage a potential good type agent to continue experiments after early successes; the later steps are designed to deter a bad type agent from over-experimentation after a failure, and the scheme becomes flat when enough successes are reported. If the agent’s incentives to deviate from the intended path of experimentation are weak, a one-step function is optimal: the agent receives a bonus if he reports enough successes; otherwise, he only gets a non-negative compensation. I characterise the conditions where the principal achieves the same efficiency level relative to a public information environment. The third chapter is an extension of the second chapter. I consider a situation in which an uninformed agent persuades a principal for a high reward through costly private experimentation. I show the existence of three types of equilibria as well as their conditions: no-experiment equilibrium, separating equilibria with learning and pooling equilibria with learning. The participation threshold determines the upper bound of the entire set of equilibria, and the over-experimentation determines the boundary between the separating and pooling equilibria with learning. As the agent’s value-cost ratio or prior belief increases, the set of separating equilibria with learning shrinks but the set of pooling equilibria with learning expands. Moreover, when the agent can pre-commit to report a specific number of successes to prove his quality, he tends to commit to a number that is as small as possible.

This Thesis contributes to the literature of business cycles driven by agents' beliefs. In Chapter 3, we provide novel empirical evidence linking the effects of technology news shocks to uncertainty shocks. Their correlation implies that when financial uncertainty shocks hit the economy, utilization-adjusted total factor productivity (TFP) increases over the medium-term. This leads to an attenuation of the effects on economic activity from news shocks in the short-term and from uncertainty shocks in the medium- term. Supported by these results, we propose an identification strategy to measure the effects of `good uncertainty' shocks and disentangle the importance of technological news, good and bad uncertainties, and ambiguity shocks in explaining business cycle variation. In Chapter 4, I investigate the empirical relationship between agents' responses to future technological changes and the level of uncertainty in the economy. I show that the economic responses to news shocks change substantially over time, and that this dynamic couples with periods of high and low uncertainty. Periods of high uncertainty are characterized by higher positive economic effects of news shocks on output, consumption, investment and real personal income. These results indicate that the continuous updating of agents' expectations about the current and future economic situation operates as a transmission channel for news shocks, amplifying its positive outcomes. Kurmann and Otrok [2013] show that the effects on economic activity from news on future productivity growth are similar to the effects from unexpected changes in the slope of the yield curve. In Chapter 5, I show that these results do not hold in the light of a recent update in the utilization-adjusted TFP series produced by Fernald [2014]. In Chapter 6, I propose a novel method of identifying technological news shocks through instrumental variables based on forecast revisions from the Survey of Professional Forecasters. I construct proxy measures for the slope of the long-run trend of GDP, investment and industrial production, which are strong instruments for recovering the underlying news shock. The procedure has the advantage of relying on information about agents' expectations, instead of the statistical procedures currently used for the news shock identification. By employing a proxy SVAR, I show that news shocks produce substantial effects on impact on GDP and investment. The effects on consumption in the short-run, however, are milder than usually presented by the news shock literature.

In this thesis we derive exact discrete time representation models that correspond to cointegrated systems in continuous time. At the same time, for the parameters of those models, estimation procedures are outlined. The representations are applicable for data observed as both stock or flow variables and with the use of some simulated data, the performance of the estimation procedure is assessed. More importantly, with the aim of analysing the costs, if there are any, of ignoring aggregation in the specification, the results of our estimation procedure are also compared with the ones we would have obtained by applying instead Johansen’s estimation methodology. In the first part (Chapter 2), we detail the analysis for a first- order stochastic differential equation system, as a result, baseline finding are outlined. In the second part (Chapter 3) the analysis is generalized and not only includes higher order specifications in the system but also incorporates deterministic components on it. Finally, in the last part (Chapter 4) of this thesis, three applications of that estimation procedure are presented. In the results, when the system is entirely comprised by stock variables and the specification follows a first order system, both Johansen’s methodology and ours perform very well, with virtually identical estimates and, for the simulated data, improvements as the sample size increases. However, when the variables of interest are flows or the specification follows a higher order system, given that our exact discrete time representation includes moving average components in the error term, Johansen’s estimates show a persistent bias in estimation, consequently, they reflected the cost of ignoring aggregation in the specification.

Chapter 1 is an overview of the thesis in which I explain why work on housing markets merits attention, discuss two broad questions that motivated the research, emphasise the particular avenue I have chosen to pursue, and summarise the new insights to be learned. I also include a short discussion on the methodologies that are used. In Chapter 2, I introduce information heterogeneity into a user-cost house pricing model. I use the model to shed light on two empirical regularities in the housing market: the predictability of housing return and the positive relationship between rent volatility and housing prices. The model also has predictions on overpricing and housing price excess volatility. In Chapter 3, I study a Real Business Cycle model with borrowing constraints and incomplete information. I show that in such an environment noises in signals may have real impacts on the macroeconomy; the effects are induced by learning and amplified and propagated by the collateral effects. Noises may generate sizeable and persistent fluctuations on consumption, credit, asset price, and output. In Chapter 4, I implement a new strategy to identify shocks that drive the co-movements between housing price and consumption. My results show that, in the United Kingdom, productivity shocks and especially news shocks about future productivity explain most of the co-movements. I also show that more than half of the changes in housing price growth were not related to the changes in consumption growth, which casts doubt on the importance of housing wealth effects on consumption.

In Chapter 1, we estimate the causal effect of homeownership on employment using a regression discontinuity design that exploits an arbitrary threshold arising from a homeownership program that assigns a house to low-income families in Chile. We establish that homeownership decreases employment by between 3.95 and 5.60 percentage points. These results contrast with previous non-experimental literature, which has often found positive effects. Our findings seem to be driven by children of the heads of households not entering the labor market, rather than workers being motivated to leave their job. We also find that residential stability and neighbourhood quality are unlikely to drive the effects, contrary to what has been proposed by previous theoretical papers. Chapter 2 studies the effect of homeownership on the academic achievements of children in the household, using a regression discontinuity design that exploits an arbitrary threshold arising from a voucher-based homeownership assistance program in Chile. Despite the fact that the homeownership program substantially increases the quality of the homes in which students live, I do not find that it affects their test scores. In a subgroup analysis, I find that homeownership decreases the test scores of elementary school students by 0.16 to 0.18 standard deviations. These effects may be due to the fact that, when receiving a voucher, many families cease to live with a hosting family, who are often close relatives (e.g. grandparents), and begin living in their own house. This seems to suggest that students experience a decrease in learning support that was previously provided to them by those close relatives. My results contrast with previous studies, which have often found positive effects of homeownership on students' academic achievements. In Chapter 3, I exploit a plausibly exogenous variation in the characteristics of principals to explore their effectiveness in improving school outcomes. Using a difference-in-differences approach, I find that principals appointed under the reform tend to be younger, less experienced, and more highly educated. Drawing from a panel dataset of teacher responses, I observe that the new principals improve the general climate in their schools by decreasing violence and expanding community engagement. On the other hand, they do not improve teacher-monitoring practices, teachers' pedagogical methods, or students' test results. A plausible explanation for these results is the lack of positive or negative incentives given to principals based on the performance of employees in their schools. Evidence in this paper suggests that, in certain institutional settings, school principals do not seem to be as relevant as is often assumed.

Fairness and efficiency are two classical and connected topics in economics. They have become well known, perhaps due to Adam Smith’s two influential works: The Theory of Moral Sentiments (1759), which highlights a concern for fairness concern as part of morality, and The Wealth of Nations (1776), which underlines a concern for efficiency. However, during the rapid development of economics, fairness has received disproportionately less attention than efficiency. As a result, many people, including some economists, have incorrectly understood that economics as a subject no longer cares about fairness. The primary objective of this thesis is to dispel this misperception. We would argue here that, similar to efficiency, fairness is an important factor for both social and individual decisions, and sometimes its effect can be determined. Written in a three-paper format, this thesis explores fairness from three different angles. These angles cover the broad areas of how theoretical economists model fairness in social choice theory, how the general public perceive distributive fairness, and how people implement their fairness norms in making real-life donations. This multidimensional exploration is believed to be crucial to a comprehensive understanding of fairness.

This thesis introduces three novel competitive games that fill the gaps between games of perfect information and games of imperfect information. Each game has a common underlying structure with small, but crucial, differences. Concretely, each game has trials in which dominance should be respected and trials where trickery attempts are possible. Furthermore, we focus on risk attitude (Envelope Game); explore reasoning processes through verbal protocols (Transfer Game); and assess the effects of additional information (Suitcase Game). Behavioural experiments show that even these simple games are cognitively very challenging and that behaviour often deviates from the predictions of popular frameworks. The main contributions from this thesis are (a) the creation of three novel games that help fill the gaps between perfect information and imperfect information; and (b) the exploration of these games and their implications. Findings from the first experiment indicate a linear relationship between the willingness to transfer value from Option A to Option B and a higher initial value for Option A. We also found that decision times for player one reflect which choices he contemplates whilst decision time for player two does not relate to her choices. Finally, most participants are assessed as risk averse. When larger amounts are involved a risk averse player one more strongly desires to transfer value compared with a non-averse player one; but we do not find any behavioural differences for player two. From our second experiment we learn that participants are often not consistent in their reasoning and behaviour across trials. Despite the simplicity of the game we observe many violations of dominance. Furthermore, participants do not strongly adhere to a specific framework. Using verbal protocols we learn about the reasoning that is used to make decisions. This procedure also identified a weakness of the design: participants often consider small amounts irrelevant (since they barely affect payoffs). Our third experiment focuses on the effect of additional knowledge. We find evidence that equal divisions are made more frequently when additional knowledge is provided and that participants attempt to trick their opponents. Furthermore, we explore whether heuristics can explain behaviour since frameworks are often too precise or make ‘random behaviour’ predictions.

Chapter 1 reviews the literature about the bayesian persuasion. It first describes two main approaches to bayesian persuasion: concavification approach and information design approach. Next I consider some extensions to the basic model of bayesian persuasion, like competition between different senders, privately informed receiver and dynamic bayesian persuasion. Some other contributions reviewed include costly bayesian persuasion and bayesian persuasion when receiver’s optimal action is only a function of an expected state. Chapter 2 deals with two-dimensional bayesian persuasion. In this chapter I investigate a model when the receiver has to make two decisions. I am interested in optimal signal structures for the sender. I describe the upper bound of sender’s payoff in terms of his payoff when only marginal distributions of two dimensions are known. Completely characterise optimal simultaneous and sequential signal structures, when each dimension has binary states. This approach extends concavification approach to bigger state space, than explored in previous contributions to bayesian persuasion. Finally I characterise optimal sequential signal structure when there are three states for each dimension. In chapter 3 I investigate together with my co-author the effect of absence of common knowledge on the outcomes of coordination games in a laboratory experiment. In our experiment, around 76% of the subjects have chosen the payoff-dominant equilibrium strategy despite the absence of common knowledge. However, 9% of the players had first-order beliefs that lead to coordination failure and another 9% exhibited coordination failure due to higher-order beliefs.

In Chapter 1 we document how pension investments by individuals in the Chilean social security system are influenced by portfolio recommendations of Happy and Loaded, a pension advice firm. Following H&L's recommendations about which of five portfolios to invest in, investors shift amounts that often exceed 20% of portfolios value and 1.3% of Chilean annual GDP, in a week. We uncover what drives investment recommendations, the resulting return consequences for the Chilean stock market and social security portfolios, and the characteristics of followers and their investment outcomes. Paradoxically, investors who followed H&L's advice would have earned more by sticking with their original portfolio over time, regardless of the portfolio selected. These findings provide a cautionary tale for the design of privatized social security systems. In chapter 2 we study the value of public information in a stochastic pure exchange economy where agents trade assets in financial markets to reallocate risk, and a subset of those agents face a mandatory savings constraint. As the mandatory savings constraint depends on equilibrium prices, changes in information may allow a Planner that faces the same constraints as the agents in terms of the available assets, information, and savings constraints, to obtain Pareto improvements relative to the equilibrium without information. Changes in information cause the posteriors to change, thus affecting equilibrium prices and shifting the constraints that the Planner has to satisfy. We provide conditions for the arrival of new information before trading to obtain ex-post and ex-ante welfare improvements relative to the initial equilibrium without information. The reaction of prices to the arrival of new information is key in our analysis. We relate the value of information in exchange economies with the literature on Bayesian persuasion.