PyFixest 0.7.3

• Fixes a bug in the computation of confidence intervals that lead to wildly incorrect CIs.
• Introduces a confint() method for Fixest objects.

PyFixest 0.7.0

Potentially Breaking Changes

• for Fixest objects, the returned dict containing all estimated models is now called all_fitted_models instead of res_model.

Internal Refactoring

• This release provides quite a bit of internal refactoring - overall, the code should now be much easier to read & hopefully be more maintainable.
• Additionally, it cleans up the FixestFormulaParser class.

New Post-Processing Features

v 0.7.0

• adds a Fixest.fetch_model() method to retrieve individual models out of Fixest
• adds post estimation routines for objects of type Feols: coef(), se(), confint,tidy(), etc

After estimating multiple models, it is now possible to fetch individual models (objects of type Feols) from an object of type Fixest via the fetch_model() method:

from pyfixest import Fixest
from pyfixest.utils import get_data
import pandas as pd
data = get_data()

fixest = Fixest(data)
fixest.feols('Y ~ X1 | csw(X2,X3)', vcov = 'iid')

mod = fixest.fetch_model(1)
# Model:  Y~X1|X2+X3
mod
# <pyfixest.feols.Feols at 0x2da61c41b10>

mod.tidy()

#	        Estimate	Std. Error	t value	Pr(>|t|)	confint_lower	confint_upper
#coefnames						
# X1	        0.03975	   0.120714	0.329288	0.741973	-0.03218	0.047319

v0.6.6

• adds support for Python 3.11
• makes wildboottest.py optional in order to do so (wildboottest depends on numba, which does not support python 3.12)

v0.6: IV Support

Version 0.6 introduces support for IV estimation via 2SLS.

• Syntax for IV estimation unambiguously follows fixest' syntax: you can specify IV estimation via a formula with depvar ~ exog.vars | fixed effect vars | endog.var ~ instruments. If no fixed effects is specified, just drop the second part of the formula.
• It is possible to estimate over-identified models, but not possible to estimate models with more than one endogenous variable.
• It is also not (yet) possible to combine IV estimation with multiple estimation syntax.

from pyfixest import Fixest
from pyfixest.utils import get_data

data = get_data()
data.head()

fixest = Fixest(data = data)
fixest.feols("Y ~ 1 | X2 | X1 ~ Z1 + Z2")

v0.5

IV Support

• Includes very basic support for IV estimation with one endogenous variable and one instrument.
• Inference: iid, HC1-3, CRV1.

Syntax for IV:

import pyfixest as pf
import numpy as np
from pyfixest.utils import get_data

data = get_data()

fixest = pf.Fixest(data = data)
fixest.feols("Y~X1 | csw0(X2, X3) | X1 ~ Z1", vcov = {'CRV1':'group_id'})
fixest.summary()

# ###
# 
# Model:  IV
# Dep. var.:  Y 
# Inference:  {'CRV1': 'group_id'}
# Observations:  1998
# 
#            Estimate  Std. Error    t value     Pr(>|t|)
# Intercept -3.941293    0.221354 -17.805377 2.442491e-15
#        X1 -0.265817    0.261940  -1.014803 3.203217e-01
# ---
# ###
# ...

What's Changed

• add codecov support by @s3alfisc in #69
• update docs + explicitly drop Intercept based on column name by @s3alfisc in #71
• update inference log in vcov method closing #61 by @s3alfisc in #73
• IV Functionality by @s3alfisc in #74

Full Changelog: v0.4.0...v0.5

v0.4.0

• introduces a ssc() function to control small sample corrections
• by default, the small sample correction defaults for iid, HC1 and CRV1 inference now match fixest defaults
• defaults for inference when the vcov argument is not specified now matches fixest as well. Prior, the vcov argument was required
• multiple smaller bug fixes
• implements CRV3 inference with and without fixed effects
• implements wild (cluster) bootstrap inference via wildboottest

version 0.3.0

Supported Features:

• Basic support for OLS.
• HC and CRV inference
• Multiple estimations via sw and csw syntax.
• Regression interaction via i().
• on-the-fly SEs
• tidy, summary, iplot methods

Overview of Features: link