This repository shows in detail how to obtain the results from Section 6 of "New Pricing Models, Same Old Phillips Curves?" (Auclert, Rigato, Rognlie, Straub 2023).
The section6.ipynb Jupyter notebook, which you can visualize on GitHub here,
goes through every step of the computation. It:
- Starts with raw price change data in
israeli_data_cleaned.csv, derived from data in Bonomo et al. (2022). - Infers the generalized hazard function
$\Lambda$ and variance of idiosyncratic shocks$\sigma_\epsilon$ that best fit the distribution and frequency of price changes. - Uses
$\Lambda$ and$\sigma_\epsilon$ to calculate the expected price gap functions$E^t(x)$ . - Implements the formula in Proposition 6 of the paper to obtain the pass-through matrix from aggregate nominal marginal costs to prices,
$\Psi$ . - Converts this
$\Psi$ to the generalized Phillips curve matrix$\mathbf{K}$ .
The brief utils.py module contains a few helper functions that are used in the notebook.
If you want to play around with the notebook yourself, you can download a zip of the repository here.
In addition to Python and Jupyter notebook, our code requires the NumPy, SciPy (version 1.7.0+), Pandas, and Matplotlib packages. A good way to obtain these is to install the latest Anaconda distribution.