Lecture Notes
Financial Frictions in General Equilibrium.
The goal of the financial frictions in general equilibrium lecture notes is the study of the several mechanisms financial and macro economists have developed to map financial market disturbances into aggregate fluctuations.
The lecture notes may depreciate intertemporally because these are not updated regularly.
This is a link to the reading list for the macro financial lecture notes.
1. Financial Frictions before the Flood: The lecture notes start by reviewing dynamic optimizing monetary models. The review is built around the first five chapters of Carl Walsh's graduate money textbook. These chapters are complemented by lectures notes (linked here) that study several monetary models. The notes focus on the Fisher equation, the expected inflation effect, and the liquidity effect using dynamic optimizing monetary models. The next lectures add a financial friction to a monetary model, as in the liquidity effects model of Robert Lucas and Tim Fuerst and in a somewhat similar model studied by David Gordon and Eric Leeper. This motivates turning to a bit of history of thought by discussing models of financial frictions developed by William Brainard and James Tobin and Karl Brunner and Allan Meltzer. The next part of the course concludes with an introduction to costly state verification in macro models (see work by Stephen Williamson and Ben Bernanke, Mark Gertler, and Simon Gilchrist), the collateral constraints model of Nobu Kiyotaki and John Moore, and the liquidity constraint-credit shock model of Bengt Holmström and Jean Tirole.
2. Financial Frictions and the Real Economy: The second set of lecture notes begins with the liquidity preference model of Franklin Allen and Douglas Gale. The liquidity preference model is used to motivate models of bank runs, liquidity constraints, and financial fragility. With this background, students are asked to study the impact of different classes of financial frictions on the dynamics of real business cycle models. The RBC models presented to students are developed by Urban Jermann and Vincenzo Quadrini, Saki Bigio, Shouyong Shi, and Markus Brunnermeier and Yuliy Sannikov.
3. After the Flood: Financial Frictions and Central Banks: The third part of the lecture notes integrate interbank markets, private banks, and central banks into the theories and models of parts 1 and 2. Relevant papers include ones by Franklin Allen, Elena Carletti, and Douglas Gale, Stephen Williamson, Marvin Goodfriend and Ben McCallum, Mark Gertler and Peter Karadi, Mark Gertler, Nobu Kiyotaki, and Albert Queralto, Emmanuel Farhri and Jean Tirole, and Frederic Boissay, Fabrice Collard, and Frank Smets.
Empirical and Computational Methods.
The goal of the empirical and computational lecture notes is to provide students with tools to study and evaluate monetary policy. The lecture notes may depreciate intertemporally because these are not updated regularly.
This is the link to the reading list for the lecture notes.
Part I: Bayesian Methods to Estimate Linearized New Keynesian DSGE Models
The lecture notes begin with a review of the methods used to solve and estimate a medium scale new Keynesian (NK-)DSGE models. However, as a prelude, there are lecture notes (linked here) discussing the tools often used in macroeconomic research to measure trend and cycle along with the application of these moments to judge the fit of linearized DSGE models to sample data. As for NK-DSGE models, several relevant papers are by Lawrence Christiano, Marty Eichenbaum, and Charles Evans, Marco Del Negro, Frank Schorfheide, Frank Smets, and Rafael Wouters, and Jesús Fernández-Villaverde, Juan Francisco Rubio-Ramírez, and Frank Schorfheide. The review covers the optimality and equilibrium conditions of a NK-DSGE model, stochastically detrends these conditions, and building the model's steady state using the stochastically detrended optimality and equilibrium conditions. The stochastically detrended optimality and equilibrium conditions are the basis for constructing a linearized approximate solution of the NK-DSGE model, which concludes the first part of these notes. The next section of these notes reviews the Kalman filter and Kalman smoother drawing on material from James Hamilton's time series textbook. The Kalman filter engages the solution of the linear approximate solution of the NK-DSGE model to construct its likelihood. Next, the lecture notes discuss the choices often made for the priors of the parameters of the NK-DSGE model. A short introduction to Markov chain Monte Carlo (MCMC) methods is followed by a similar brief review of the Metropolis-Hasting (MH) simulator. The lecture notes conclude by applying a generic MH-MCMC sampler to the linear approximate solution of the NK-DSGE model and its likelihood.
Part II: VARs, Bayesian VARs, and Structural VARs
1. Introduction to Vector Autoregressions: These lecture notes introduce student to VARs, begin a discussion of VARs and fundamentals, describe several methods to compute impulse response functions (IRFs) and forecast error variance decomposition (FEVDs), and presents a Bayesian algorithm to calculate error bands for IRFs and FEVDs conditional on a just-identified recursive ordering. Instructions are also provided for computing symmetric and asymmetric errors bands corrected for serial correlation using advice found in the classic 1999 paper by Chris Sims and Tao Zha. Also, there is discussion of two algorithms developed by José Luis Montiel Olea and Mikkel Plagborg-Møller that calculate simultaneous or sup-t confidence bands.
2. Bayesian VARs, Priors, and Identification: Next, the lecture notes discuss priors for BVARs, short and long run restrictions to construct structural VARs, and critiques of SVARs. The discussion of short run identifying restrictions includes the classic 1998 paper by Chris Sims and Tao Zha on priors and estimation of BVARs given short run non-recursive restrictions. The Blanchard-Quah (BQ) decomposition is the source of the long run restrictions. There is a short digression on the Beveridge-Nelson decomposition to motivate the BQ decomposition. Jon Faust and Eric Leeper develop a critique of long run identifying restrictions. Sign restricted VARs are reviewed beginning with papers by Jon Faust and Harald Uhlig. Atsushi Inoue and Lutz Kilian, Raffaella Giacomini and Toru Kitagawa, and Jonas Arias, Juan Francisco Rubio-Ramírez, and Dan Waggoner update posterior analysis of, study robustified priors for, and propose new samplers to estimate sign restricted VARs. The lecture notes next turn to proxy VARs, which use instrumental variables (IV) to identify SVARs. This is an old story as noted by Adrian Pagan and studied by Pierre-Daniel Sarte and Robert King and Mark Watson. Proxy VARs are discussed by Valerie Ramey while the problem of inference in the presence of weak instruments, as described by Charles Nelson and Dick Startz, is implemented for SVARs by José Luis Montiel Olea, Jim Stock, and Mark Watson. Jim Stock and Mark Watson also provide a framework to analyze the assumptions and condition necessary to estimate proxy VARs and SVARs identified by IV. These notes finish by grounding critiques of SVARs in recent papers by Jesús Fernández-Villaverde, Juan Francisco Rubio-Ramírez, Tom Sargent, and Mark Watson, Juan Francisco Rubio-Ramírez, Dan Waggoner, and Tao Zha, and Christiane Baumeister and James Hamilton.
3. BVARs and Monetary Policy Evaluation: Monetary policy evaluation is the last section of these lectures. Several approaches are presented to identify monetary policy VARs. Among these are papers by Alan Blinder and Ben Bernanke, Chris Sims, Steve Strongin, Ben Bernanke and Ilian Mihov, Lawrence Christiano, Marty Eichenbaum, and Charles Evans, Markku Lanne and Helmut Lutkepohl, Dave Gordan and Eric Leeper, Eric Leeper and Jennifer Roush, Jon Faust, Eric Swanson, and Jonathan Wright, and Eric Leeper and Tao Zha. The focus is on the connection between SVAR identifying restrictions and economic models and monetary theory. The notes conclude by presenting students with tools to estimate Markov-switching and time-varying parameter VARs with stochastic volatility developed by Giorgio Primiceri, Chris Sims, Dan Waggoner, and Tao Zha, Giorgio Primiceri and Marco Del Negro, and Fabio Canova and Fernando Pérez Forero.