This thesis consists of four essays on finance, learning, and macroeconomics. The first essay studies whether learning can explain why the standard consumption-based asset pricing model produces large pricing errors for U.S. equity returns. I prove that under learning standard moment conditions need not hold in finite samples, leading to pricing errors. Simulations show that learning can generate quantitatively realistic pricing errors and a substantial equity risk premium. I find that a model with learning is not rejected in the data, producing pricing errors that are statistically indistinguishable from zero. The second essay (co-authored with Anna Mikusheva) studies the properties of the common impulse response function matching estimator (IRFME) in settings with many parameters. We prove that the common IRFME is consistent and asymptotically normal only when the horizon of IRFs being matched grows slowly enough. We use simulations to evaluate the performance of the common IRFME in a practical example, and we compare it with an infrequently used bias corrected approach, based on indirect inferences. Our findings suggest that the common IRFME performs poorly in situations where the sample size is not much larger than the horizon of IRFs being matched, and in those situations, the bias corrected approach with bootstrapped standard errors performs better. The third essay (co-authored with Ricardo Caballero) documents that, in contrast with their widely perceived excess return, popular carry trade strategies yield low systemicrisk- adjusted returns. In contrast, hedging the carry with exchange rate options produces large returns that are not a compensation for systemic risk. We show that this result stems from the fact that the corresponding portfolio of exchange rate options provides a cheap form of systemic insurance. The fourth essay shows that the documented overbidding in pay-as-you-go auctions relative to a static model can be explained by the presence of a small subset of aggressive bidders. I argue that aggressive bidding can be rational if users are able to form reputations that deter future competition, and I present empirical evidence that this is the case. In auctions without any aggressive bidders, there is no evidence of overbidding in PAYGA.

This dissertation focuses on learning and expectations formation in Macroeconomics and Finance and the role of information production in shaping macroeconomic fluctuations. The first chapter provides a theory of information production to explain two features of modern business cycles. In my theory information is produced along two dimensions, a pro-cyclical quantitative margin and a counter-cyclical qualitative margin, that generates both slow recoveries and episodes of "rational exuberance" where optimistic booms tend to end in crises. The second chapter provides supporting evidence for the proposed cyclical variation in private information production using term loan data in the United States. Finally, the third chapter documents biases in terms of over-optimism and overconfidence in forecasts of real GDP growth from the survey of professional forecasters in the United States.

(Cont.) the release of information until the long-term results of the firm are realized. In equilibrium, when the belief about the aggregate state is high, managers will be given short-term incentives, delaying the release of information. When the belief about the aggregate state is low, long-term incentives will be prevalent and information will be released without delay. This produces asymmetric learning dynamics for the economy, with gradual booms and rapid recessions. In a boom the belief about the aggregate state increases, information is pushed off into the future, and learning is slow. In a recession the belief is falling, triggering a switch to long-term incentives, that brings forward the release of information and accelerates learning. Chapter 2 presents a model of corporate misreporting in an environment where investors have heterogeneous beliefs and short sale constraints. The disagreement between investors provides a motive for agents who start a firm to limit the amount of information which it releases to the public so as to sponsor speculation over its value. This incentive to limit information is stronger when the heterogeneity of beliefs among investors is stronger. Investors also learn about a firm's expected profitability from the information released by other firms in the industry. I show that this creates a strategic complementarily in the precision of information released by each firm. This can give rise to multiple equilibria: one in which all firms release precise reports and one in which their reports are inaccurate ...

This volume focuses on current issues of debate in the area of modern macroeconomics and money, written from (a broadly interpreted) post Keynesian perspective. The papers connect with Philip Arestis' contributions to macroeconomics and money, and pay tribute to his distinguished career.

This dissertation presents two chapters on empirical models of macroeconomics and finance, and one chapter on a theoretical model for conducting monetary policy. The first chapter applies machine learning algorithms to construct non-parametric, nonlinear predictions of mortgage loan default. I compile a large dataset with over 20 million loan observations from Fannie Mae and Freddie Mac, for the period 2001-2016 at the quarterly frequency. Different machine learning algorithms are applied to predict in sample (training sample), and to forecast out-of-sample (testing data). I find that the forecast performance of nonlinear and non-parametric algorithms are substantially better than the traditional logit model. Additionally, machine learning algorithms allow identification of the predictive power of specific variables. The results indicate that loan age is the most important predictor of loan default before and after the 2008 financial crisis. However, I find that market loan- to-value is the most effective predictor of mortgage loan default during the recent financial crisis. Finally, I use machine learning to formulate risk-based capital stress tests for Fannie Mae and Freddie Mac under different scenarios. I forecast their mortgage credit losses and associated capital needs during the financial crises. The results obtained are more accurate than those from the Federal Housing Enterprise Oversights (OFHEO), and other existing stress test studies. In the second, and third chapters, I tested the effectiveness of Monetary policy by empirical, and theoretical models. With the severity of the 2008 financial crisis, and apparent inefficacy of traditional monetary and fiscal policies, the Federal Reserve together with the U.S. government introduced unconventional policy measures. The Large Scale Asset Purchase (LSAP) and Troubled Asset Relief Program (TARP) are some of these policies introduced by the Federal Reserve and Department of Treasury. While these policies may have been important in preventing a deepening of the financial crisis and laying the foundation for the economic recovery, there were collateral effects on bank profitability. In this chapter, I study the impact of both the LSAP and TARP programs on banks? profit and risk taking using a large panel. The results indicate that these programs had a positive effect on banks? profit (Chapter 2). In chapter three, I use a small-scale DSGE model for the economy of Iran to analyze monetary policy. The model is extended to include housing and oil sectors. The model is adapted for the peculiarities of Iran's Central Bank, which uses money supply as a function of oil income and production growth. I study the reaction function of the model to technology, oil, and monetary shocks in this specific Iranian monetary policy framework. The results show that monetary shocks has only nominal effect on inflation but not on the real sector such as investment, consumption, or production. Also, positive oil income shocks lead to an increase in inflation instead of an increase in production.