Summary
The lectures provide an applied presentation of standard forecasting models which are typically employed in policymaking institutions to forecast macroeconomic time series. Lectures are followed by practical sessions with hands-on computational exercises on MATLAB software.

Get the Syllabus

Installing MATLAB: Dauphine provides a free licence for students. Please use your own laptop if you can with MATLAB installed. Please have Datafeed, Optimization and Econometrics toolboxes installed. You will need to tick them in the package list during the installation of MATLAB.

Real time macro-data on MATLAB


This is a must-read as all of the lectures will rely on DBnomics. The latter is a database aggregator which can be queried directly from MATLAB. Please read carefully this following note on DBnomics and download as well the MATLAB function which allows to get real time macro data:
DBnomics.pdf a small guide to use DBnomics on MATLAB. call_dbnomics.m MATLAB function to query DBnomics.

Introduction to MATLAB programming

This introduction should take no more of four hours of in class teaching.

Lecture 1: Introduction to MATLAB Programming
Objectives of the Lecture:

  • Utilize the basic mathematical operations;
  • Get know the general purpose commands of Matlab;
  • Manipulate matrix calculus;

Materials:

chapter1.pdf Handout
Lecture 2: Plotting
Objectives of the Lecture:

  • UPlotting 2D and 3D graphs;
  • Define the curve shape, titles and legends.

Materials:

chapter2.pdf Handout
Lecture 3: Loops and Conditional Statements
Objectives of the Lecture:

  • Mastering the use of conditional statements;
  • Coding loops with both while and for.
Materials:

chapter3.pdf Handout
Lecture 4: Functions
Objectives of the Lecture:

  • Write a function with n-input arguments;
  • Code a function n-output variables;
Materials:

chapter4.pdf Handout

Handouts list

Handout 2: The AR(p) Benchmark model
Objectives of the Lecture:

  • Constructing a likelihood function;
  • Estimating an AR(p) model;
  • Performing forecasting analysis: both in-sample and out-sample;
  • Finding the optimal lag using likelihood comparison test;
  • Understanding the basis of the state-space representation.

Materials:

chapter2.pdf Handout
f_logLk.m likelihood function (MATLAB file)
Random_AR1.m Estimating an AR(1) model on fake data (MATLAB file)
Forecasting_AR1.m Forecasting fake data with an AR(1) model (MATLAB file)
Forecasting_ARP.m Forecasting fake data with an AR(P) model (MATLAB file)
Handout 3: Vecto Auto-Regressive (VAR) models
Objectives of the Lecture:

  • Building and estimate a VAR Model;
  • Performing business cycles analysis;
  • Imposing restrictions on the VAR coefficients;
  • Performing forecasting.

Materials:

chapter3.pdf Handout
estim_NKPC.m The NK curve in VAR model (MATLAB file)
SW JEP 2001 Vector Autoregressions paper

Toy models list

Here is the list of available models that are necessary for your examination. You will need to use one these toy models, and compare its forecasting hability with a benchmark VAR and AR models.