In this course, participants receive a general introduction to:
- methods of identification and global sensitivity analysis,
- their DYNARE implementation (identification toolbox and global sensitivity analysis toolbox)
- their application to Dynamic Stochastic General Equilibrium (DSGE) macroeconomic models
- techniques for improving the speed of estimations in DYNARE
The course will be open to a maximum of 30 qualified and selected participants.
Please apply online
Deadline: September 30, 2020.
Successful applicants will receive a confirmation email by no later than October 16, 2020.
Identification analysis and global sensitivity analysis (GSA) can help us in answering these questions:
- What is the stability region of a DSGE model? Which structural parameters mainly drive the unique saddle-path solution of DSGE models?
- What drives the sign and magnitude of impulse response functions? What parameters are mostly responsible for cross-correlations and auto-correlations of endogenous variables?
- Which structural parameters mainly drive the fit of observables (e. g., GDP, and consumption)? Is there any trade-off?
- Is it possible to determine identifiable and non-identifiable parameters before starting the estimation? Is it possible to assess ex-ante the strength of the identification of parameters in an estimated DSGE model?
- DYNARE Introduction
- Global Sensitivity Analysis Introduction
- Monte Carlo Filtering
- Kalman Filtering in DYNARE
- Identification analysis of DSGE models
Although the course is mainly addressed to DYNARE users, it will provide a general introduction for estimating models to DYNARE.
- Roberta Cardani (JRC)
- Olga Croitorov (JRC)
- Fabio Di Dio (JRC)
- Lorenzo Frattarolo (JRC)
- Massimo Giovannini (JRC)
- Stefan Hohberger (JRC)
- Marco Ratto (JRC)
The course is organised by the Joint Research Centre (JRC) in Ispra (Italy) and it will be held “virtually” through an online platform.
This is a workshop with hands-on sessions; each participant is required to have installed on his/her laptop computer MATLAB version ranging from 7.9 (R2009b) to 9.8 (R2020a).
As an alternative to MATLAB, it is possible to use GNU Octave versions 5.2.0 (under Windows) and 4.4.1 (under macOS).
For any questions, please email: JRC-MACRO-GSA-COURSE@ec.europa.eu