StataFlow 1.0.0

StataFlow: A Python econometrics toolkit aligned with Stata 17

StataFlow

A Python econometrics toolkit that reproduces Stata 17 results with high precision.

---

from stataflow.compat.stata import reghdfe

result = reghdfe(df, y="lwage", x=["exper", "edu"],
                 absorb="firm_id year_id", vce="cluster", cluster="firm_id")
result.display()

---

Features

• 14 Stata commands in Python: regress, reghdfe, ivregress 2sls, ivreghdfe, logit, probit, poisson, ppmlhdfe, did_imputation, eventstudyinteract, csdid, rdrobust, areg, xtreg_fe
• Stata-style regression table: result.display() produces a formatted output matching Stata's layout
• High-dimensional fixed effects: MAP iterative absorption handles 10K+ FE levels without memory issues; individual slope absorption (absorb(firm_id##c.time))
• Driscoll-Kraay panel HAC: time-series-robust standard errors with Bartlett kernel autocorrelation correction
• Instrumental variables: 2SLS, GMM2S, and LIML estimators with weak-instrument diagnostics (Kleibergen-Paap F + Stock-Yogo critical values)
• Binary, count, and PPML models: Logit, Probit, Poisson, and PPML with HDFE
• Causal inference: DID (BJS imputation, Sun-Abraham, Callaway-Sant'Anna) with doubly-robust methods; Regression Discontinuity with 11 bandwidth selectors
• Stata-compatible syntax: factor variables (i.group##c.post), analytic weights, multiple FEs
• Validated against Stata 17: every public capability has field-level Python-Stata dual-run evidence

Installation

pip install StataFlow

Python 3.10+ required. Dependencies: NumPy, pandas, SciPy.

Quick Start

Stata-compatible API

import pandas as pd
from stataflow.compat.stata import regress, reghdfe, logit, ivregress_2sls, ppmlhdfe

# OLS with robust standard errors
result = regress(df, y="wage", x=["edu", "exper"], vce="robust")
result.display()

# High-dimensional fixed effects
result = reghdfe(
    df, y="wage", x=["edu", "exper"],
    absorb="firm_id year_id", vce="cluster", cluster="industry"
)

# Logit
result = logit(df, y="inlf", x=["nwifeinc", "educ", "exper"])
result.display()

# 2SLS with LIML
result = ivregress_2sls(
    df, y="lwage", x_exog=["edu"], x_endog=["exper"],
    instruments=["age", "kidslt6"], vce="robust"
)

# PPML with HDFE
result = ppmlhdfe(
    df, y="trade", x=["lndist", "contig", "fta"],
    absorb=["exporter", "importer", "year"], vce="cluster", cluster="exporter"
)

Native Python API

from stataflow import OLS, AbsorbingOLS, Logit

model = OLS(data=df, y="wage", x=["edu", "exper"])
result = model.fit(vce="robust")
result.display()

Using results

# Stata-style table
result.display()
result.display(show_ci=True)  # with confidence intervals

# Programmatic access
for c in result.coefficients:
    print(f"{c.name}: b={c.beta:.6f}, se={c.std_err:.6f}, t={c.t_stat:.2f}")

print(f"R² = {result.fit.r2:.4f}, N = {result.sample.nobs}")

Supported Models

Family	Available via	Estimators & VCE
Linear	regress, areg, xtreg_fe, reghdfe	OLS with ols / robust (HC1) / cluster (1-way, 2-way) / dkraay (panel HAC)
IV	ivregress_2sls, ivreghdfe	2SLS, GMM2S, LIML (with Fuller), first-stage diagnostics, weak-IV tests
Binary / Count	logit, probit, poisson	MLE with ols / robust / cluster
PPML + HDFE	ppmlhdfe	IRLS with ols / robust / cluster, separation detection, eform
DID	did_imputation, csdid, eventstudyinteract	BJS imputation, Callaway-Sant'Anna (reg + DR), Sun-Abraham IW
RDD	rdrobust	Sharp / Fuzzy RD, 11 MSE+CER bandwidth selectors, cluster/nncluster VCE

Documentation

• User Guide — full tutorial and concept guide (中文: 用户手册)
• Cookbook — copy-pasteable recipes for common tasks (中文: 中文 Cookbook)
• Examples — runnable demo scripts

Running Tests

# Unit and integration tests
pytest tests/ -v --ignore=tests/golden/

# Golden dual-run tests (require local Stata 17)
pytest tests/golden/ -v

License

This project is licensed under the MIT License. See LICENSE for details.