Topics in Macroeconometrics aim is to endow students with a working knowledge of the modern econometric methods used in macroeconomics and, to some extent, in finance.
Topics range from classic time series concepts such as linear univariate and multivariate processes (ARMA, VAR) to techniques that have only recently entered the applied macroeconomist' toolbox (Bayesian VAR and Factor Models).
Students need to be familiar with econometric theory at the level of Advanced Econometrics (246PP).
The main references for this courses are:
Brockwell, Peter J. and Richard A. Davis, Introduction to Time Series and Forecasting, Springer, 2002
Enders, Walter, Applied econometric time series. John Wiley & Sons, 2008
Hamilton, James D., Time Series Analysis, Princeton University Press, 2005
The level of these textbook goes from (relatively) easy (Brockwell and Davis) to difficult (Hamilton). The level of the class is probably in between the level of these two books. Enders' book has many applications of many time series concepts introduced during lectures.
I will provide handouts to highlight challenging aspects of a topic or present material not covered by the references.
Other references useful for specific part of the course will be posted here!
The class has is own Teams channel. I will use it extensively for important communication.
Day | Time | Classroom |
---|---|---|
Tuesday | 14:45 | P1 |
Wednesday | 12:15 | P1 |
Lectures will be streamed live on Teams. A link to the stream will be made available on the [Teams channel] a few hours before the start of each class. Lectures will not be recorded, and in-person should be considered the standard way of attendance
The emphasis of the course is on methods and the analysis of real dataset. The best way to understand time series concepts is to write lots of code implementing the various techniques discussed in class on real macro e financial data. The computer language used in this course is Julia.
The final grade depends on a series of homework assignments.
From Cross-Section to Time Series: asymptotic theory under serial correlation
Stationary Process
Linear processes
The Wold representation theorem
ARMA processes: estimation and forecasting
ARIMA models for non-stationary time series
Multivariate Time Series
Vector Auto-Regressions (VAR)
Structural VARs: identification
Structural impulse responses
The Bayesian paradigm
Likelihood, prior, and posterior
Bayesian computations
Applications: Bayesian VAR
Factor models and High Dimensional Econometrics
Principal components
Dynamic factor models
Ridge and Lasso