General Description
The second EABCN training school will be a five-day
course on methodologies for the analysis of the business cycle.
Professor Lucrezia Reichlin, from the Universite
Libre de Bruxelles, will instruct the course in collaboration
with Domenico Giannone. It is primarily aimed at participants
in the Euro Area Business Cycle Network.
The course will consist of five days lectures
(approximately 6 hours per day). The objective is to introduce
the students to the topic and enable them to apply the methods
to empirical data. For each topic, there will be lecturing and
applications on the computer using Matlab 6.5. Class notes and
codes will be made available before the beginning of the course.
Introduction. Motivation for factor analysis
in economic time series: overview.
Lecture 1. The traditional factor model
in the static and dynamic case.
Here we will discuss identification and estimation for a given
size of the cross-section n (small panels). Topics covered are:
identification, relation between maximum likelihood estimation
and principal component estimation, quasi maximum likelihood,
the kalman filter.
Lecture 2. Unobserved component models,
trend-cycle decompositions and filtering macroeconomic time series.
Dynamic principal components.
We will illustrate examples of unobserved component models as
particular cases of factor models and analyse them in population.
In particular, we will analyse filtering implied by the extraction
of the components for different univariate and multivariate examples
and introduce dynamic principal components in this context.
Lecture 3. Large panels: the static
and dynamic factor model. Illustration of alternative estimation
techniques proposed by the recent literature.
We will analyse the factor model for n large and explain how results
discussed in Lecture 1 change in this case. We illustrate the
alternative estimation techniques proposed by the recent literature
and discuss relative performances for stylized empirical examples.
Lecture 4. Applications: coincident
and leading indicators, core inflation, regional and national
cycles.
Lecture 5. Applications: forecasting
and structural identification.
We illustrate existing applications in the literature and show
how to reproduce those applications on the basis of data sets
that students are encouraged to propose.
Administrative Information
The course will take place at the National Bank
of Belgium and participants will be invited to make their own
arrangements regarding their accommodation and meals. Lunches
will be provided by the NBB. For your convenience CEPR has reserved
a block of rooms and participants should contact the hotel directly
when their application has been accepted.
Hotel: N.H. Atlanta
Rate: €135
Telephone Number: +32 2 217 01 20
Contact: Ms Trivi Ashta
Please quote BNB-GROUPE 20 AU 24/09 when you call.
A maximum of two attendees can be admitted from
each institution. To apply to attemnd this course, please contact
the EABCN Secretariat.
EABCN and CEPR gratefully acknowledge
the generous provision of facilities by the National Bank of Belgium
for this course. |
