Graduate Center Masterclasses

What this course is about:

The study of nonlinearities permeates modern macroeconomics.
This course will introduce tools to estimate the possibly nonlinear effects of various (possibly non-Gaussian) economic shocks. The tools/models we will cover come in two classes:

1. Models where nonlinearities and non-Gaussianity help to identify shocks
2. Models where identification assumptions (such as instruments) are borrowed from linear models.

We will study both approaches in detail. Whenever feasible, we will look at example codes (written mostly in Matlab).

Prerequisites:

I will assume knowledge of linear time series models and some Bayesian econometrics - i.e. I will use terminology such as ‘Gibbs Sampling’ or ‘MCMC’ without first deriving these algorithms or proving under what conditions they work. I’ll be happy to provide background reading on these topics.

Course Outline:

We will have 6 classes. Below I list the topics for each class and some representative literature.

1. Identification of nonlinear Effects via Direct Estimation of Moving Average Models
   Barnichon and Matthes (2018), Barnichon et al. (2022),Barnichon et al. (2022)
2. Tail risks
   Adrian et al. (2019), Loria et al. (2019)
3. Identification via Non-Gaussianity
   Lewis (2021), Lewis (2024), Jaroci´nski (2021)
4. Time-Varying Parameter VARs
   Cogley and Sargent (2002), Primiceri (2005), Del Negro and Primiceri (2015), Amir-Ahmadi et al. (2020)
5. Instruments and Time-Varying Parameters
   Mumtaz and Petrova (2023), Amir-Ahmadi et al. (2023)
6. Time Variation as a Direct Function of Observables
   Gargiulo et al. (2024), Kim et al. (2021)

*** Please find the references in the Masterclass Description - Download below ***