Essays on Business Cycles

The topic of my dissertation is to understand the sources of business cycles. In particular, using structural estimation, I quantitatively investigate different types of shocks that propagate within a country (Chapter One) and that cause business cycle comovement across countries (Chapter Two and Three). In the first chapter, Wataru Miyamoto and I propose the use of data on expectations to identify the role of news shocks in business cycles. News shocks are defined as information about future fundamentals that agents learn in advance. Our approach exploits the fact that news shocks cause agents to adjust their expectations about the future even when current fundamentals are not affected. Using data on expectations, we estimate a dynamic, stochastic, general equilibrium model that incorporates news shocks for the U.S. between 1955Q1 and 2006Q4 using Bayesian estimation. We find that the contribution of news shocks to output is about half of that estimated without data on expectations. The precision of the estimated role of news shocks also greatly improves when data on expectations are used. Although news shocks are important in explaining the 1980 recession and the 1993-94 boom, they do not explain much of other business cycles in our sample. Moreover, the contribution of news shocks to explaining short run fluctuations is negligible. These results arise because data on expectations show that changes in expectations are not large and do not resemble actual movements of output. Therefore, news shocks cannot be the main driver of business cycles. Chapters Two and Three focus on the driving forces of business cycles in open economies. We start Chapter Two with an observation that business cycles are strongly correlated across countries. We document that this pattern is also true for small open economies between 1900 and 2006 using a novel data set for 17 small developed and developing countries. Furthermore, we provide a new evidence about the role of common shocks in business cycles for small open economies in a structural estimation of a real small open economy model featuring a realistic debt adjustment cost and common shocks. We find that common shocks are a primary source of business cycles, explaining nearly 50\% of output fluctuations over the last 100 years in small open economies. The estimated common shocks capture important historical episodes such as the Great depression, the two World Wars and the two oil price shocks. Moreover, these common shocks are important for not only small developed countries but also developing countries. We point out the importance of our structural approach in identifying several types of common shocks and their sizable role in small open economies. The reduced form dynamic factor model approach in the previous literature, which often assumes one type of common component, would predict only a third of the contribution estimated in the structural model. Chapter Three further our understanding of the business cycle comovement across countries by investigating the transmission mechanism of shocks across countries. Our reading of the literature indicates that even though business cycles are correlated across countries, existing models are not able to generate substantial transmission through international trade. To the extent that business cycles are correlated across countries, it is because shocks are correlated across countries. We show that the nature of such transmission depends fundamentally on the features determining the responsiveness of labor supply and labor demand to international relative prices. We augment a standard international macroeconomic model to incorporate three key features: a weak short run wealth effect on labor supply, variable capital utilization, and imported intermediate inputs for production. This model can generate large and significant endogenous transmission of technology shocks through international trade. We demonstrate this by estimating the model using data for Canada and the United States. We find that this model can account for the substantial transmission of permanent U.S. technology shocks to Canadian aggregate variables such as output and hours documented in a structural vector autoregression. Transmission through international trade is found to explain the majority of the business cycle comovement between the United States and Canada while exogenous correlation of technology shocks is not important.