Structural Macroeconometrics (SMA11)

This is a hands-on course on building and estimating dynamic macroeconomic models (DSGE models). We limit ourselves on perturbation methods and (log-)linear approximations. We first study model solution techniques, then compute and compare model and data moments, and finally estimate the parameter distribution by Bayesian techniques. Other techniques are reviewed too. The course can be passed by a term paper where these methods are applied. The course exercises involve working with the term paper. Therefore, it is highly suggested that a participant is prepared to work with his/her own model.

Course (preliminary) syllabus.

Logistics

Proposed lecture and exercise dates are
May 9, Mon, 11-13 and 14-16; (seminar room 1)
May 10, Tue, 10-12 and 13-15; ; (seminar room 1);
Exercises: 16-18 (computer class 3rd floor)
May 11, Wed, 10-12 and 13-15; (seminar room 3-4)
May 12, Thu, 10-12 and 13-15 (seminar room 3-4).
Exercises:
May 13, Fri, 10-12; (computer class 3rd floor)
May 23, Mon, 10-12; (computer class 3rd floor)
May 25, Wed, 10-12; (computer class 3rd floor)

Course material

My slides! (contains all the course material; unpolished!)

Iris code that I used to compute the spectral density functions shown in the lecture.

Iris Cookbook is very useful to start learning Iris. There might be tiny changes in the command syntax after the new release of Iris. The new release is also much wider in terms of the models and methods.

Exercises

You will work with your own model during the exercise classes. Currently, the idea is that I will start by reviewing coding issues of the exercise theme. After this I will be available to you in helping in practical problems.

Ex 1: We try to code our models to Dynare (I will also review Yada and Iris schemes). You should know your model well and do the following in advance:

Equilibrium conditions (these include decision rules, budget constraints, market equilibrium, etc)
Stationarize your model
Calculate deterministic steady state and try to solve it

I might be able to help you in the above as well.

Ex 2: We compute model moments and talk about data.

You should have your data organized into matrix form at Excel/etc sheets.

Ex 3: We study model and data moments that the D/Y/I produce. For the data moments, we rely mostly on Iris. Download the code.

Ex 4: We set priors and talk about Bayesian estimation:

Term paper

You should return the term paper by the end of November 2011. I am considering the following rules of the game:

My markings will be based on the usual scale.
You should provide theoretical model (not necessarily original); report the most important data and model moments and estimate the model parameters; report MCMC diagnostics, etc.
Those, who are working with purely nonlinear models and are not going to estimate the model parameter values, I ask for data and model moments with particular and thorough emphasis on nonlinearity.
Other rules will follow.

Back to my website http://www.ripatti.net