s3alfisc

# /// script
# python = ">=3.10"
# dependencies = [
#     "numpy>=1.21.0",
#     "pandas>=1.3.0",
#     "statsmodels>=0.13.0",
#     "pyfixest>=0.11.0",
#     "matplotlib>=3.4.0"
# ]
# ///

s3alfisc

import pyfixest as pf
import pandas as pd
import numpy as np

n_workers = 100
n_firms = 50
n_time_periods = 10

worker_effects = np.random.normal(0, 1, n_workers)
firm_effects = np.random.normal(0, 1, n_firms)

s3alfisc

# Standard Error Defaults in `fixest` and `PyFixest`

`fixest` and `PyFixest` defaults match exactly - the only difference is that with fixed effects, 
the `PyFixest` defaults set `ssc = ssc(adj = "nested")`, while `PyFixest` sets `ssc(adj = "none")`, 
as "nested" is not yet implemented in `PyFixest`. Below are some examples:

## OLS without Fixed Effects

By default, both packages choose iid errors if no particular method for inference is specified.