Cogley and Sargent (2005) in pymc

This repository is a collection of codes and jupyter notebooks that implement the time-varying parameter VAR (TVP-VAR) with stochastic volatility as it was discussed by Cogley and Sargent (2005) and estimate it with the Python package pymc.

There are three main notebooks:

• CogleySargent2005_replication.ipynb: this notebook replicates the paper using the same dataset and prior parameters. By means of a few illustrative figures it compares the pymc posterior sample to the sample drawn by Tim Cogley's Matlab files
• CogleySargent2005_reestimation.ipynb: this notebook reestimates the model using an updated dataset. The updating is accomplished by this notebook (using pandas built-in datareader for FRED)
• Laboratory.ipynb: this notebook does not use real data, it is a laboratory to investigate the "accuracy" of the sampler given that we know the true data generating process. It simulates an artificial dataset (for arbitrary fixed parameter values), which is then used to fit the model (just like we do with real data).

The data folder consists of .csv, .mat and .pickle files storing the posterior samples and the datasets that have been used for the estimation

The generated figures can be found in this folder

Note: the Matlab files in matlab_files_from_cogley were written entirely by Tim Cogley and Thomas Sargent