Computational Tools and Macroeconomic Applications

By Lawrence J. Christiano and Jesús Fernández-Villaverde




The purpose of this course is to provide an overview of computational methods for macroeconomics and international finance. The course will emphasize methods that have proven to be useful in practice. We will show participants how the methods discussed are implemented in Dynare, which is free software for solving and simulating equilibrium models in MATLAB. Dynare software for several economic models will be distributed and participants will experiment with the code as part of the lectures. The technical points (e.g., about multiplicity or non-existence of model solutions) will always be illustrated using examples that are of fundamental economic interest. Some familiarity with basic macroeconomic models and with MATLAB will be assumed. A streaming video of the lectures is available here.



Handouts for Christiano’s Part of the Course

1)    Overview of perturbation and projection methods (see Judd’s text book and Christiano-Fisher, JEDC, 2000). 

(A more extensive version of these notes appears here.)

2)    Solution and analysis of simple New Keynesian model (Christiano, Trabandt and Walentin chapter in 2011 Handbook of Monetary Economics, Friedman and Woodford).

3)    Dynare exercise (zip file)

a)    Dynare syntax

b)    Applications

i)         Rationale for the Taylor principle.

ii)       Potential conflict between working capital channel and Taylor principle.

iii)      Sensitivity of dynamic response of inflation and output gap to persistence of shock.

iv)      In examples when Taylor rule moves the interest rate in the right direction, does not move it far enough.

v)      The Taylor rule may push the interest rate in the wrong direction when there are news shocks.

4)    A DSGE Model with Financial Frictions: Risk Shocks.

5)    Computing Ramsey Optimal Policy by Linearization.



Handouts for Fernández-Villaverde’s Part of the Course

6)    Why Non-linear/Non-Gaussian DSGE Models?

7)    Perturbation Methods.

8)    Projection Methods.

9)    Filtering and Likelihood Estimation.

10)          Heterogeneous Agents Models.

11)          Examples: Code Perturbation and Code Projection.