Syllabus

Last updated: 28 January 2022
The content in this Jupyter Book is subject to change.

Course description

This is a course in time series econometrics with a focus on the models and methods used in empirical macroeconomics. The emphasis of the course will be on applying these tools in practice, rather than on than underlying econometric theory. To that end, we will try to implement everything covered in lectures by writing computer programs. The language of instruction used in the course is Python. No previous programming experience with Python is assumed. However, students are expected to quickly familiarize themselves with the basics of the language by working through the material and completing the exercises that will be provided. The course will be taught using a mixture of online lectures and hands-on programming sessions.

Course outline

Please note that the following outline is preliminary and may be modified if needed.

  • Introduction to time series analysis, basic concepts and models

    • time series vs cross-section

    • stationarity and ergodicity

    • white noise, martingales, martingale difference, AR, MA processes

    • autocovariance, autocorrelation and partial autocorrelation

  • Review of the multivariate normal distribution

    • marginal, conditional, joint distributions

    • useful properties

  • Review of maximum likelihood estimation

    • log-likelihood function, score, Hessian

    • identification

    • optimization

    • asymptotic properties

    • quasi-maximum likelihood

  • ARMA models

    • estimation

    • model selection

    • forecasting

  • Nonstationarity

    • Unit root tests

    • ARIMA models

  • Conditional heteroskedasticity

    • ARCH, GARCH models

  • VAR models

    • estimation

    • forecasting

  • State Space models

    • Kalman filter

    • identification

    • estimation

    • forecasting

  • Granger causality

    • unconditional

    • conditional

  • TBD other topics:

    • structural models

    • spectral analysis

    • regime switching models

    • Bayesian methods

Textbooks and other readings

There is no required textbook for this course. All necessary material (slides, lecture notes and programming demos) will be posted on the course web site. There are several freely available online books that I will sometimes make reference to, but none of them is mandatory:

Important course information

Class meetings will be held on Zoom. You will need a microphone and webcam (you need not always have your webcam on, but you will want to be able to be visible at times).

Course goals

  • Obtain a basic knowledge of time series theory and methodology.

  • Learn how to analyze time series data by exploratory analysis, model identification and fitting.

  • Learn how to use Python to perform such analysis, by using existing libraries, as well as by writing own code to implement specific time series tasks.

  • Learn useful research tools and practices.

Grading

Grades will be determined as follows: homework assignments (55%), midterm project (15%), final project (30%).

Late Submissions

No.