pytwoway 0.0.3

Estimate two way fixed effect labor models

pytwoway

Two way fixed effect models for labor in python

Full documentation can be found here.

Quick start:

To install from pip:

pip install pytwoway

To run using command line interface:

pytw --my-config config.txt --akm --cre

Example config.txt:

data = file.csv
filetype = csv
col_dict = "{'fid': 'your_firmid_col', 'wid': 'your_workerid_col', 'year': 'your_year_col', 'comp': 'your_compensation_col'}"

To run in Python:

• If you have data

from pytwoway import twfe_network
tn = twfe_network.twfe_network
# Create twfe object
tw_net = tn.twfe_network(data, formatting, col_dict)
# Convert long data into event study data (not necessary if the data is already in event study format):
tw_net.refactor_es()
# Run the bias-corrected AKM estimator:
tw_net.run_akm_corrected(user_akm)
# Cluster firms based on their wage CDFs (required for the CRE estimator)
tw_net.cluster(user_cluster)
# Run the CRE estimator
tw_net.run_cre(user_cre)

• If you want to simulate data

from pytwoway import sim_twfe_network
sn = sim_twfe_network.sim_twfe_network
# Create simulated twfe object
stw_net = sn(sim_params)
# Generate data
sim_data = stw_net.sim_network()

• If you want to run Monte Carlo on simulated data

from pytwoway import sim_twfe_network
sn = sim_twfe_network.sim_twfe_network
# Create simulated twfe object
stw_net = sn(sim_params)
# Run Monte Carlo
stw_net.twfe_monte_carlo(N, ncore, akm_params, cre_params, cluster_params)
# Plot results
stw_net.plot_monte_carlo()