Using MATLAB to Develop
Macroeconomic Models
Bob Taylor
The MathWorks, Inc.
Introduction
Construct Model
Prepare Data
Identify Model
Calibrate Model
Analysis
2
CONSTRUCT MODEL
Smets-Wouters Model
Linearized Smets-Wouters Model
From Smets-Wouters Structural VARMA to Reduced-Form VAR
Differences with Smets-Wouters Model
3
Smets-Wouters Model
Smets and Wouters (2002)
– Dynamic Stochastic General Equilibrium (DSGE)
model
– Focus is on European economies
– Nonlinear system of equations to model macro and
household economies
– Notable feature is that the money supply is not
observed directly
– Linearized model is a structural Vector Auto-
Regressive Moving Average (VARMA) model
4
Linearized Smets-Wouters Model
Smets and Wouters (2004, 2007)
– Linearized model for the United States economy
– System of 14 equations with a companion system to
model potential output
14 endogenous variables
7 exogenous disturbances
– Equivalent to a structural VARMA(3,2) model with 7
equations and numerous parameter restrictions
– Ignoring restrictions, you can invert the MA lag operator
to obtain an unrestricted reduced-form VAR(3) model
5
