Replication Instructions for: "Tail Risk Exposure and the Cross-Section of Expected Stock Returns"

This repository contains the Jupyter notebooks required to replicate the results in the paper "Tail Risk Exposure and the Cross-Section of Expected Stock Returns", along with additional results presented in the supplemental appendices. Each figure and table is generated and plotted in different notebooks with step-by-step instructions.

Repository Structure

The repository consists of a series of notebooks, numbered in order of execution, but each can be run independently. However, some analyses (such as 6_Summary_Statistics, 7_Portfolio_Analysis, and 8_Regression_Analysis) require the final sample to generate results.

Notebooks:

• 1_Simulations.ipynb

  • Simulates data for analysis.

• 2_Import_Factor_Data.ipynb

  • Imports common factor data from external sources.

• 3_Estimators_Functions.ipynb

  • Defines functions for risk factor estimation based on daily return data.

• 4_Process_Daily_Returns_Factors.ipynb

  • Downloads and computes risk factors (daily return-based) using WRDS data.

• 5_Process_Monthly_Returns_Factors.ipynb

  • Downloads and computes risk factors (monthly return-based) from WRDS data and saves the final sample.

• 6_Summary_Statistics.ipynb

  • Produces summary statistics tables, correlation tables, persistence measures, and trends of Tail Risk Exposure variables.

• 7_Portfolio_Analysis.ipynb

  • Generates results tables and figures for univariate and bivariate portfolio sorting strategies.

• 8_Regression_Analysis.ipynb

  • Computes and presents Fama-MacBeth regression results tables.

Additional Scripts:

• estimation_mixture.py
  • Implements the methodology for estimating Tail Risk Exposure, following Chabi-Yo et al. (2018)

Requirements

WRDS Access

The notebooks 4_Process_Daily_Returns_Factors.ipynb and 5_Process_Monthly_Returns_Factors.ipynb require access to a WRDS account with API access. Users must insert their WRDS credentials in these notebooks:

os.environ['WRDS_USERNAME'] = '...'
os.environ['WRDS_PASSWORD'] = '...'

Python Dependencies

The following Python libraries are required:

pip install pandas numpy matplotlib ipython wrds scipy statsmodels linearmodels pycop pandarallel

• pandas
• numpy
• matplotlib
• IPython
• wrds
• scipy
• statsmodels
• linearmodels
• pycop
• pandarallel (highly recommended for parallel processing, but not mandatory)