My dissertation contributes to the macroeconomics of self-fulfilling prophecies. It demonstrates the importance of shocks to beliefs in accounting for aggregate fluctuations. The dissertation consists of three chapters.

The first chapter is a paper joint with Roger E.A. Farmer. We integrate Keynesian economics with general equilibrium theory in a new way. We develop a simple graphical apparatus, the IS-LM-NAC framework, that can be used by policy makers to understand how policy affects the economy. A new element, the No-Arbitrage-Condition (NAC) curve, connects the interest rate to current and expected future values of the stock market and it explains how ‘animal spirits’ influence economic activity. Our framework provides a rich new approach to policy analysis that explains the short-run and long-run effects of policy.

The second chapter studies implications of self-fulfilling beliefs in open economies. Uncovered interest parity states that the carry trade should deliver zero profit, on average. The data robustly reject this hypothesis. In a large sample of countries, high interest rate currencies earn excess returns at short horizons and negative excess returns at longer horizons. I rationalize this observation in a two-country overlapping generations model with complete markets that features multiple dynamic equilibria. Because newborns cannot make decisions about consumption and savings before they are born, there is a set of self-fulfilling beliefs of the currently alive generations about the decisions of the future newborns. I utilize the multiplicity of dynamic equilibria by imposing a structure on the formation of beliefs. Beliefs are self-fulfilling, and shocks to these beliefs generate a large and volatile risk premium that is correlated with the interest rate differential. Changes in uncertainty about beliefs cause a reversal of expected excess returns associated with the current interest differential, similar to the reversal found in the data. I provide empirical evidence in favor of my mechanism and show that persistence of past expectations can account for most of the observed deviation from uncovered interest parity.

The third chapter extends the methodology developed in the first chapter to the open economy framework. I build a model of the eurozone crisis. I study a two-country model in which agents form self-fulfilling beliefs about asset prices. Using the labor market search and matching frictions and abandoning Nash bar- gaining over wage, I create multiple equilibria on the labor market where any unemployment rate can be sustained as a steady state. Self-fulfilling beliefs about the future value of assets select the equilibrium. I show that sudden downward revisions of beliefs (‘animal spirits’) cause stagnation in real economic activity, international financial contagion, and, in the absence of recovery in confidence, permanently high rates of unemployment. High unemployment is viewed as a new steady state, not a temporary deviation from the natural rate of unemployment. Policy aimed at recovering the eurozone needs to trigger optimistic beliefs about the value of assets.

This dissertation explores persistent unemployment dynamics in the U.S., alternative explanations for this phenomenon and potential policy implications. Chapter 1 develops and estimates a general equilibrium rational expectations model with search and multiple equilibria where aggregate shocks have a permanent effect on the unemployment rate. If agents' wealth decreases, the unemployment rate increases for a potentially indefinite period. This makes unemployment rate dynamics path dependent as in Blanchard and Summers (1987). I argue that this feature explains the persistence of the unemployment rate in the U.S. after the Great Recession and over the entire postwar period.

Chapter 2 conducts an empirical exercise to analyze which assumptions prevent a standard model from matching the persistence of the unemployment rate. I do this by using business cycle accounting procedure (Chari et. al. (2007)) on data in wage units (Farmer (2010)). I find that most movements in the unemployment rate are accounted for by the labor supply wedge. In other words, a standard model fits data badly because the assumption that households are on their labor supply is flawed. This finding is consistent with other papers in the literature. It also motivates assumptions that I make in my job market paper.

In Chapter 3, I with Roger Farmer focus on the persistent unemployment dynamics during the Great Depression. We explain the period between 1929 and 1950 within a single model which

is driven by the stock market as a measure of consumer confidence. We document that the stock market measured by the S&P 500 and unemployment were moving closely between 1929 and 1939, but after 1939 the relationship between the two series seems to completely disappear. In particular, the stock market kept falling, but the employment rate recovered to the pre-recession level by 1942. Using our model we analytically study the effects of temporary bond-financed fiscal expansions that are similar to the actual data. We plug the actual data for the stock market and government expenditures from the Great Depression and WWII in the model and show that the model does a good job replicating consumption and unemployment dynamics during this period.

The first chapter proposes a method for solving and estimating linear rational expectations models that exhibit indeterminacy. The method implements an idea of moving expectational errors to the set of fundamental shocks, reducing the number of solutions from infinity to one. This transformation allows one to treat indeterminate models as determinate and, therefore, apply standard solution and estimation methods to them. This chapter provides a simple "rule of thumb," based on a Bayesian model comparison, for identifying expectational errors that generate indeterminacy.

The second chapter reexamines the source of the Great Moderation by estimating New-Keynesian DSGE models with capital accumulation and indeterminacy on U.S. data from 1960 to 2008. It was found that, in contrast to canonical papers, the Federal Reserve's monetary policy rule remained passive in response to inflation before (1960-1979) and after (1982-2008) the Great Moderation. Bayesian model comparisons enable a declaration that, when capital is added, passive monetary policy with indeterminacy provides a better fit to the data in both subperiods. The results of this chapter suggest that during the Great Moderation structural changes were primarily on the demand side of the economy, supporting the idea of financial innovations.

The third chapter sheds light on a narrow but crucial question in finance: What should be the parameters of a model of the short-term real interest rate? In this chapter, parameters of the real interest rate model are estimated in the broad class of single-factor interest rate diffusion processes on U.S. monthly data. It is shown that the elasticity of interest rate volatility--the relationship between the volatility of changes in the interest rate and its level--plays a crucial role in explaining real interest rate dynamics. The empirical estimates of the elasticity of the real interest rate volatility are found to be about 0.5, much lower than that of the nominal interest rate. These estimates show that the square root process, as in the Cox-Ingersoll-Ross model, provides a good characterization of the short-term real interest rate process.

This dissertation studies how market expectations of systemic bailouts affect credit recoveries, how countries' macro-fiscal conditions change around fiscal distress, and how banking crises and fiscal crises interact with each other and restrict economic recoveries. Chapter 1 is written based on my job market paper. Using daily put options data of U.S. bank holding companies, I measure each bank holding company's exposure to the systemic bailout factor, which is the sensitivity of each bank's out-of-the-money put option price to the variations of sector-wide put option basket-index spreads. I show that low market expectations of the banking sector systemic bailouts played a significant role in the weak bank credit recovery after the subprime crisis. Bank holding companies with higher pre-crisis exposure to the systemic bailout factor experienced larger post-crisis deviations from the pre-crisis bank credit growth trend. Perhaps surprisingly, such pattern is persistent even for banks that are less affected by the post-crisis financial regulations and less exposed to borrowers from the deteriorating sectors. Furthermore, I drill down to the commercial bank subsidiary level data while controlling for parent bank holding company fixed effects. This analysis reveals that commercial bank subsidiaries within the same bank holding company present same credit growth patterns even though they have different exposure to financial regulations and deteriorating sectors. Chapter 2 is coauthored with my colleague at the International Monetary Fund. We present a new database of fiscal crises covering different country groups, including low-income developing countries (LIDCs) that have been mostly ignored in the past. We find countries faced on average two crises since 1970. We also shed some light on policies and economic dynamics around crises. Surprisingly, advanced economies face greater turbulence, with half of them experiencing economic contractions. Fiscal policy is usually procyclical around fiscal crises and we also find that the decline in economic growth is magnified if accompanied by a financial crisis. Chapter 3, which is coauthored with my advisor, Aaron Tornell, studies the interplay between banking crises and fiscal crises when the government exhausts fiscal space to clear up banking sector non-performing loans. Exploiting a newly constructed data set on various fiscal costs of resolving non-performing loans, we find banking crises accompanied with fiscal distress could dampen the positive growth effects related with active non-performing loans resolutions conducted by government. We empirically show that, depending on the timing of banking crises and fiscal crises, such disruptive effects are the result of either lacking fiscal policy space or the bank-sovereign nexus. However, immediate government response to restore fiscal space or regain access to international capital markets could help enhancing medium-term output and credit recoveries even after banking-fiscal twin crises.

My dissertation focuses on the interactions between the conduct of U.S. monetary policy and the expectations formed by households, firms and public institutions about the state of economy. The first two chapters develop new methods that I use in the subsequent chapters to study how expectations formed by economic agents about future economic conditions affect a given economy. The second chapter considers and extends the work in Farmer (2012a) to explain U.S. post-war data, and shows that it outperforms conventional economic theories due to its ability to account for persistent movements in the data. The last chapter explores how the effectiveness of monetary policy changed in the U.S. post-war period, and I provide evidence that since the early 1980's the monetary authority implemented policies that reduced economic uncertainty deriving from unforeseen changes in the expectations about future inflation.

This dissertation studies how financial market volatility or uncertainty in the U.S. economy affects real economic activity both in the U.S. and other open economies. Chapter 1 critically examines a stylized fact about the effects of uncertainty shocks on the U.S. economy. A link between uncertainty and firms' investment, hiring, and production decisions has drawn much attention in contemporary discussions after the 2008 financial crisis. Bloom (2009) showed that uncertainty events, identified by spikes in stock market volatility, triggered immediate falls in output and employment, followed by rapid rebounds. I show that such stock market volatility shocks failed to produce this same pattern of responses after 1983. Chapter 2 studies the effects of risk aversion shocks, measured by increases in the VIX , on emerging market economies (EMEs). By estimating a structural vector autoregression (VAR) model, I find that, although risk aversion shocks do not have much impact on U.S. output, they do have a noticeable impact on the output of EMEs. To explain the contrast between the impact of risk appetite shocks on EMEs and the impact on the U.S. economy, a credit channel is proposed as a propagation mechanism. In the model, an increase in the VIX is translated to a risk-aversion shock that generates a "flight to quality." As international investors pull their money from EMEs, borrowing costs increase and domestic credit falls as a consequence of credit market imperfections. Higher borrowing costs, in turn, lead to a fall in investment that causes a real depreciation and a decline in total output through sectoral linkages. Finally, Chapter 3, which is co-authored with Prakash Loungani, studies the effect of uncertainty shocks on unemployment dynamics by separating out the role of aggregate and sectoral channels. Using SP500 data from the first quarter of 1963 through the third quarter of 2014, we construct a separate index to measure sectoral uncertainty and compare its effects on the unemployment rate with that of aggregate uncertainty in a standard VAR model, augmented by a local projection method. We find that aggregate uncertainty shocks lead to an immediate increase in unemployment, followed by swift reversals. In contrast, sectoral uncertainty shocks have a long-lasting impact on unemployment, with the peak impact occurring after two years. Our findings highlight an additional channel through which uncertainty shocks have persistent effects on unemployment by requiring substantial inter-industry labor reallocation.