systemgmmkit 0.5.9

Generic panel-data econometrics workflow helpers for FE, RE, IV/2SLS, and Difference/System GMM in Python.

systemgmmkit

---

systemgmmkit 0.5.8.dev0

systemgmmkit is a Python package for panel-data econometrics that provides a unified workflow from baseline linear models to advanced dynamic-panel GMM estimation.

The package combines:

• Ordinary Least Squares (OLS)
• Pooled OLS
• Fixed Effects
• Random Effects
• Panel IV / 2SLS
• Difference GMM
• System GMM
• Diagnostics
• Post-estimation
• Reproducible reporting

within a single framework.

The package is designed for applied empirical research in economics, finance, management, operations, public policy, political economy, industrial organization, productivity analysis, and other panel-data settings.

---

Why systemgmmkit?

Many econometric workflows require researchers to combine multiple packages to obtain:

• baseline estimation;
• panel estimators;
• instrumental-variable estimation;
• dynamic-panel estimation;
• diagnostics;
• post-estimation;
• publication-ready reporting.

systemgmmkit aims to provide these capabilities through a consistent and reproducible API.

The package is built around four principles:

1. Explicit model specification.
2. Reproducible workflows.
3. Transparent diagnostics.
4. Verification against established reference implementations.

---

What's New in 0.5.8

Version 0.5.8.dev0 introduces the first public post-estimation framework and completes the baseline linear-model workflow.

New public APIs include:

• OLSSpec
• PooledOLSSpec
• run_ols()
• run_pooled_ols()
• predict()
• fitted_values()
• residuals()
• vcov()
• confint()
• lincom()
• wald_test()
• marginal_effects()

These additions establish the foundation for future advanced post-estimation capabilities while maintaining compatibility with the broader panel-estimation framework.

---

Core Estimator Coverage

Linear Models

• Ordinary Least Squares
• Robust OLS
• Pooled OLS
• Clustered OLS

Panel Models

• One-Way Fixed Effects
• Two-Way Fixed Effects
• Random Effects
• Panel IV / 2SLS

Dynamic Panel Models

• Difference GMM
• System GMM
• One-Step Estimation
• Two-Step Estimation
• Windmeijer-Corrected Standard Errors
• Collapsed Instruments
• Restricted Lag Windows

Reporting

• Markdown Export
• CSV Export
• LaTeX Export
• Structured Result Objects
• Integration with universal-output-hub

---

Verification Philosophy

Verification is a core design principle of systemgmmkit.

Whenever practical, estimators are benchmarked against established Stata implementations including:

• regress
• xtreg
• ivregress
• xtabond2
• xtdpdgmm

Benchmark scripts, comparison workflows, and validation artifacts are maintained within the repository.

The objective is not merely to produce estimates, but to provide transparent evidence that estimates match trusted reference implementations under maintained benchmark specifications.

---

Current Validation Status

The major estimation paths currently exposed through the public API have either:

1. been directly benchmarked against Stata reference implementations; or
2. have dedicated comparison workflows maintained within the repository.

Verified Components

Component	Status
OLS	PASS_STATA_PARITY
Robust OLS	PASS_STATA_PARITY
Clustered OLS	PASS_STATA_PARITY
Confidence Intervals	PASS_STATA_PARITY
lincom	PASS_STATA_PARITY
Wald / F Tests	PASS_STATA_PARITY
Fixed Effects	PASS_STATA_COMPARISON
Random Effects	PASS_STATA_COMPARISON
Panel IV / 2SLS	PASS_STATA_COMPARISON
Difference GMM	PASS_XTABOND2_PARITY
System GMM	PASS_XTABOND2_PARITY
Windmeijer Standard Errors	PASS_XTABOND2_PARITY
Hansen Diagnostics	PASS_XTABOND2_PARITY
Sargan Diagnostics	PASS_XTABOND2_PARITY
AR(1) Diagnostics	PASS_XTABOND2_PARITY
AR(2) Diagnostics	PASS_XTABOND2_PARITY

---

Verified OLS Benchmark

The OLS and pooled OLS implementations have been verified against Stata using a real FD001 panel-data benchmark.

Benchmark model:

risk ~ degradation_index + sensor_mean_z + pc2 + op_setting1 + op_setting2

Panel structure:

entity = unit
time = cycle

Observed agreement:

Metric	Result
Maximum coefficient difference	4.64e-14
Maximum standard-error difference	2.04e-14

These differences represent machine-precision agreement with Stata.

---

Installation

Stable release:

pip install systemgmmkit

Development version:

pip install git+https://github.com/Akanom/systemgmmkit.git

Local development installation:

pip install -e ".[dev,all]"

---

Quick Start

Ordinary Least Squares

from systemgmmkit import OLSSpec, run_ols

spec = OLSSpec(
    dependent="y",
    regressors=["x1", "x2"],
    controls=["z1", "z2"],
    covariance="robust",
)

result = run_ols(spec, df)

print(result.summary_frame())

Equivalent Stata command:

regress y x1 x2 z1 z2, vce(robust)

---

Pooled OLS

from systemgmmkit import PooledOLSSpec, run_pooled_ols

spec = PooledOLSSpec(
    dependent="y",
    regressors=["x1", "x2"],
    controls=["z1"],
    covariance="clustered",
)

result = run_pooled_ols(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata command:

regress y x1 x2 z1, vce(cluster firm_id)

---

Fixed Effects

result = run_fixed_effects(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata:

xtreg y x1 x2, fe

---

Random Effects

result = run_random_effects(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata:

xtreg y x1 x2, re

---

Panel IV / 2SLS

result = run_panel_2sls(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata:

ivregress 2sls

---

Difference GMM

result = run_difference_gmm(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata:

xtabond2

---

System GMM

result = run_system_gmm(
    spec,
    df,
    entity="firm_id",
    time="year",
)

Equivalent Stata:

xtabond2

---

Post-Estimation

The 0.5.8 development cycle introduces the first public post-estimation framework.

Import:

from systemgmmkit import (
    predict,
    fitted_values,
    residuals,
    vcov,
    confint,
    lincom,
    wald_test,
    marginal_effects,
)

---

Predictions

Equivalent Stata:

predict yhat

Python:

pred = predict(result)

or

pred = result.predict()

---

Fitted Values

fit = fitted_values(result)

---

Residuals

Equivalent Stata:

predict ehat, residuals

Python:

resid = residuals(result)

---

Variance-Covariance Matrix

V = vcov(result)

---

Confidence Intervals

ci = confint(result)

---

Linear Combinations

Equivalent Stata:

lincom x1 + x2

Python:

lincom(
    result,
    {
        "x1": 1,
        "x2": 1,
    },
)

Returns:

• estimate
• standard error
• statistic
• p-value
• confidence interval

---

Wald Tests

Equivalent Stata:

test x1 x2

Python:

wald_test(
    result,
    R=[
        [0,1,0],
        [0,0,1],
    ],
)

Returns:

• Wald statistic
• degrees of freedom
• p-value

---

Marginal Effects

marginal_effects(result)

For linear estimators, marginal effects correspond to estimated slopes.

---

Reporting and Export

Results can be exported to:

• Markdown
• CSV
• LaTeX

and integrated with:

• universal-output-hub
• publication pipelines
• reproducible research workflows

---

Recommended Reporting Practice

For empirical research, report:

• estimator;
• specification;
• covariance estimator;
• instrument count;
• AR diagnostics;
• Hansen diagnostics;
• backend;
• package version.

This improves transparency and reproducibility.

---

Citation

If you use systemgmmkit in academic work, please cite:

Akanbi, Oluwajuwon Mayomi.

systemgmmkit:
Panel Data Econometrics and Dynamic GMM Workflows in Python.

Version 0.5.8.

https://github.com/Akanom/systemgmmkit

BibTeX:

@software{akanbi_systemgmmkit,
  author = {Akanbi, Oluwajuwon Mayomi},
  title = {systemgmmkit: Panel Data Econometrics and Dynamic GMM Workflows in Python},
  year = {2026},
  url = {https://github.com/Akanom/systemgmmkit}
}

Replace with DOI information once a software paper or Zenodo archive is published.

---

License

MIT License.

See LICENSE for details.