Skip to content

fipelle/replication-hasenzagl-et-al-2020

Repository files navigation

A Model of the Fed's View on Inflation

This repository contains the source code for replicating the results in the paper:

Hasenzagl, T., Pellegrino, F., Reichlin, L., & Ricco, G. (2020). A Model of the Fed's View on Inflation.

If you have any questions, comments, or suggestions please create a new issue or email the authors.

Code structure

The main directory is organized as follows:

  • annex_global_data: Contains a directory with the data files, tc_mwg.jl, and iis_charts.ipynb for the model with global variables. To estimate this model use these files instead of the files with the same names in the data and code directories. The dataset for the global model includes the Baltic Dry Index (BDI) which is available here: https://www.balticexchange.com/en/index.html.
  • code_main: Contains all of the Julia code necessary for replication.
    • The Metropolis-Within-Gibbs subdirectory contains the code for the Metropolis-Within-Gibbs algorithm.
  • csv_output: Used for storing the .csv output files.
  • data: Contains the data used in the estimation. The data is saved in .csv and .xlsx files.
  • docs: Contains the paper and online appendix.
  • img: Used for storing the output figures.

The code is written in Julia 1.6.4 (https://julialang.org/).

The code uses a number of Julia packages. All necessary packages can be installed using the import_packages.jl script. To do so, start Julia and use the following command at the Julia REPL prompt:

julia> include("import_packages.jl")

Running the code

The main file is user_main.jl. This script runs the following exercises:

  • The in-sample estimation is run by setting run_type=1 in user_main.jl.
  • The conditional forecasting exercise is run by setting run_type=2 and specifying the start date of the forecasting exercise, and the conditioning variables and time periods. Note that the paper does not include a conditional forecasting exercise.
  • The out-of-sample forecasting exercise is run by setting run_type=3 and specifying the start date of the forecasting exercise.

After choosing the run_type run the script by starting Julia and using the following command at the Julia REPL prompt:

julia> include("user_main.jl")

Figures and Tables

The figures and tables are created in two Jupyter (https://jupyter.org/) notebooks:

  • iis_charts.ipynb: creates all figures relating to the in-sample estimation.
  • oos_charts.ipynb: creates all figures relating to the out-of-sample forecasting exercise and the RMSE of the trend-cycle model relative to the RMSE of a random walk with drift.

Citation

If you are using any part of the code for academic work (including, but not limited to, conference and peer-reviewed papers), please cite using the following bibtex code:

@misc{hasenzagl2020inflation,
    title={A Model of the Fed's View on Inflation},
    author={Hasenzagl, Thomas and Pellegrino, Filippo and Reichlin, Lucrezia and Ricco, Giovanni},
    year={2020},
    eprint={2006.14110},
    archivePrefix={arXiv},
    primaryClass={econ.EM}
}