Show simple item record

dc.contributor.advisorPiger, Jeremy
dc.contributor.authorGelfer, Sacha
dc.date.accessioned2016-10-27T18:45:32Z
dc.date.available2016-10-27T18:45:32Z
dc.date.issued2016-10-27
dc.identifier.urihttp://hdl.handle.net/1794/20493
dc.description.abstractIn this dissertation I incorporate high dimensional data vectors in estimated Dynamic Stochastic General Equilibrium (DSGE) models, evaluating the labor market dynamics incorporated inside such data vectors, out-of-sample forecasting performance of many models estimated with such data vectors and analytically examining the reduction of macroeconomic volatility that can occur when such data vectors are used in the formation of expectations about the future. The second chapter investigates the extent to which modern DSGE models can produce labor market dynamics in response to a financial crisis that are consistent with the experience of the Great Recession. I estimate two New-Keynesian models, one with and one without financial frictions, in a data-rich environment. I find that negative financial shocks are associated with longer recoveries in real investment, capital-intensive sectors of the labor market and average unemployment duration. I also find the model with a financial accelerator is equipped with better tools to identify the dynamics associated with the Great Recession and its recovery in regard to many labor and financial metrics. The third chapter compares the out-of-sample forecasting performance of the two DSGE models of Chapter II when they are estimated both out of and in a data-rich environment. This chapter finds that many financial time series variance decomposition are significantly better explained using the structural set-up of the New-Keynesian model with financial frictions. DSGE models estimated with high dimensional data vectors significantly out forecast their regularly estimated counterpart in regard to output, investment and consumption growth. Lastly, the use of real-time optimal pool model weighting significantly out-forecasts traditional macroeconomic models as well as an equally weighted weighting scheme in terms of many macroeconomic variables. The fourth chapter examines the role forecasts derived by high dimensional data vectors can have on lowering macroeconomic volatility. Bounded rational agents are introduced into the Chapter II DSGE model with financial frictions and are given the option to use or ignore professionally generated forecasts from a dynamic factor model in their perceived forecasting model. In simulations, I find that professionally generated forecasts can significantly lower the volatility of many macroeconomic variables including inflation and hours worked.en_US
dc.language.isoen_US
dc.publisherUniversity of Oregon
dc.rightsAll Rights Reserved.
dc.subjectAdaptive learningen_US
dc.subjectData-Richen_US
dc.subjectDSGEen_US
dc.subjectDSGE-DFMen_US
dc.subjectFinancial acceleratoren_US
dc.subjectFinancial shocksen_US
dc.titleIncorporating High Dimensional Data Vectors into Structural Macroeconomic Models
dc.typeElectronic Thesis or Dissertation
thesis.degree.namePh.D.
thesis.degree.leveldoctoral
thesis.degree.disciplineDepartment of Economics
thesis.degree.grantorUniversity of Oregon


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record