The last decades have witnessed an extraordinary surge of interest in utilizing large datasets for macroeconomic forecasting. This has also been partially fueled by the advent of machine learning methods for applications in economics and finance.

This course offers a review of some of the latest methods and most important techniques for forecasting in macroeconomic applications from the “classic” time series literature and the “modern” machine learning literature. Importantly, the course also reviews state-of-the-art techniques for out-of-sample forecast evaluation. These methods will be applied in two macroeconomic forecasting horse races to assess which methodologies are more advantageous in real-world applications. After this course, which includes both theory sessions and hands-on exercises in practical sessions, you will be able to apply these methods to your own research.

This course has two main objectives:

• First, to review state-of-the-art techniques for forecasting in macro using large datasets from both the time series and machine learning literature.
• Second, to showcase the strengths and benefits of these different methodologies for prediction using two real-world applications: Forecasting policy-relevant variables using the FRED-MD database and downside risk (Growth-at-Risk) for the US.

Learn state-of-the-art techniques for macroeconomic forecasting

The course includes theory sessions (10 hours) and practical sessions (10 hours). In the practical sessions, you will work in R to implement the techniques and methodologies covered in the theory lectures

Additional class materials will be provided before and during the sessions. The instructors will also be available throughout the course to discuss your individual research ideas and projects.

You will participate in an active and fruitful environment with international colleagues without incurring high costs, thanks to the delivery of this course, which will be taught online but will be live and interactive. Sessions will be recorded and videos will be available for a month once the course has finished.

Learn how to apply modern Time Series Methods to your own research