Macroeconometrics (Y1)

The goal of the course is to provide an introduction to the methods of modern applied macroeconometrics. The different approaches currently used in applied work are reviewed, including the basics of the empirical dynamic stochastic general equilibrium models. The prerequisite is sufficient knowledge of multivariate time series analysis (at least, S3. Advanced Econometrics or equivalent) and advanced macroeconomic theory. In addition these, Money and monetary theory (Y1), Open economy macroeconomics would be an asset.

Course syllabus.


Lectures (by Professor Antti Ripatti): Lectures  (21 hours) Professor Antti Ripatti: 13.1. - 24.2.2014 on Mondays at 12-15, Place: Economicum, seminar room 1 (Arkadiankatu 7).

N.B.!  No lectures on Monday 17.2.2014; an extra lecture on Thursday 27.2.2014 at 9-12, Place: Economicum, seminar room 1. 

Exercises (by Doctoral student Min Zhu):  Economicum, the seminar room 2 (Arkadiankatu 7), at 12-14
  1. Jan 22
  2. Feb 5
  3. Feb 19
  4. Mar 5

Phd students will not participate exercise classes, see below!

Course material

The slides (updated 12 Jan) come in two forms:
The content is These are still unpolished: so update/refresh often. The first page tells the date of the version. I try to update them jointly.


Exercises form 25 % of the course evaluation. At least 50 % of the assignments has to be completed (showing good effort) to be able to pass the course.  You should email your work to the instructor (min . zhu [at] helsinki . fi) in advance of each exercise class or (printed) at the beginning of each session. In computational exercises you should email/print the source code and the output of the computer program (plots are typically not necessary).
  1. The first problem set  
  2. The second problem set 
  3. The third problem set (complete set). The mdata.mat file (use the Save As command to download it). If you plan to use R program (that is a good idea if you know it), the data can be got from the Excel file (the variables of interest are YER, PCD, STN). If you use Matlab/Octave, you may utilize the hpfilter.m and penta2.m files to compute the HP-filtered series (you also need LeSage Econometrics toolbox, see the link below). I have not, yet, found any easy-to-use spectral toolboxes that work nicely in Octave.
  4. The fourth problem set aims in estimating the basic New-Keynesian model. Any reasonable attempt with explanation why it fails is accepted.

PhD Students

The following setting applies to PhD students:
    1. : Code the model using Dynare or Iris toolbox. Compute impulse responses. Related Iris tutorial
    2. : Collect the relevant data and compute the data moments we study at the lectures. Related Iris tutorial 
    3. : Estimate the posterior of the parameters, and study MCMC diagnostics. Look at Iris tutorial and the System Priors paper below.


Choose Matlab if you have access to it. Otherwise use Octave. They are available in all computer platforms. Matlab may be found from University linuxes. Apply an account from IT services. You may also use your own personal computer as X11 server (over the fast network). Then the linux matlab would look like the one on the PC. You need to install the libraries (dynare, iris) to linuxes though.

If you google "matlab tutorial" you get tons of very good tutorials (an example). Pick one and spend a weekend to explore and to type the examples. (Octave is enough here!)

There are two widely used libraries for validating DSGE models: Iris-toolbox and Dynare. Iris-toolbox is better suited for this course, but it does not work on Octave. It is also a bit more demanding than Dynare. Hence, I suggest that PhD students use Iris-toolbox and undergrads Dynare.

  • available at Computer class in Economicum (plus the libraries below) and University unixes.
  • Install libraries: Dynare and Iris-toolbox
  • Iris-toolbox is the most comprehensive choice and contains functions/methods to compute all the stuff that is covered by the course.
  • 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.
  • Open source: download, windows download, Matlab like IDEs
  • Libraries: Dynare
  • You may rely on some free libraries in computing data moments
    • LeSage's spatial econometrics contains useful functions, but not all.
    • Standard Octave libraries such as Statistics, Econometrics and Signal prcessing toolbox contain code that you need.
    • Signal processing toolbox should contain standard primitives to compute spectral stuff. Please, let me know if you find better option. (Spectran could be an option, but I have not tested it.)

Go to my homepage