arch
ARCH for Python
Decision gist · record as of 2026-08-14
Yes. arch is a mature, actively maintained library (1551 GitHub stars, top 5000 PyPI by downloads) with no known vulnerabilities, permissive licensing, and a focused feature set for financial econometrics. Install friction is moderate but manageable; prebuilt wheels are available for all major platforms and Python versions. Recommended for anyone doing volatility modeling, unit root testing, or bootstrap-based inference on financial time series.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.10+.
- Optional Numba acceleration requires setting ARCH_NO_BINARY=1 environment variable and installing from source.
- Medium install friction due to compiled components (Cython/Numba).
License · maintenance · safety
NCSA (unclear) — Licensed under NCSA, which is permissive and compatible with most commercial and open-source projects. No notable restrictions on use or redistribution.
last release 2025-10-21 (297 days) · last repo commit 2026-08-10 · 1,551 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 1,059,093 downloads/mo, #4,423 on PyPI
Alternatives
Verify before relying
pip install arch
from arch import arch_model
import pandas as pd
returns = pd.Series([...]) # your returns data
am = arch_model(returns)
res = am.fit()- Whether Numba acceleration provides meaningful performance gains for typical use cases
- Specific volatility model types beyond ARCH, GARCH, TARCH, EGARCH, and EWMA that are implemented
What it is and what it does
arch is a Python library for financial econometrics, specializing in volatility modeling and time-series analysis. It implements ARCH and GARCH models (and variants like TARCH, EGARCH, EWMA) for capturing changing volatility in financial returns, alongside classical econometric tests like Augmented Dickey-Fuller, Phillips-Perron, and KPSS for unit roots. The package also provides cointegration testing (Engle-Granger, Phillips-Ouliaris), multiple comparison procedures (SPA, StepM, MCS), and bootstrap methods (IID, stationary, block-based) for confidence interval construction and resampling-based inference.
The library is built on numpy, pandas, scipy, and statsmodels, with optional Cython and Numba acceleration for performance. It targets researchers, practitioners in finance and insurance, and quantitative analysts who need to model time-varying volatility, test for stationarity, or construct robust confidence intervals for financial statistics. The package is actively maintained, supports modern Python versions (3.10–3.13), and has been in development since 2014.
Use it for
- Model conditional volatility in stock returns using GARCH to forecast risk and value-at-risk
- Test whether economic time series have unit roots before fitting regression models
- Estimate long-run covariance matrices for Newey-West corrected standard errors in econometric models
- Bootstrap confidence intervals for Sharpe ratios or other portfolio performance metrics
- Test for cointegration between pairs of financial assets to identify mean-reversion trading signals
- Compare predictive models using multiple comparison procedures to avoid data-snooping bias
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
arch is a mature, actively maintained library (1551 GitHub stars, top 5000 PyPI by downloads) with no known vulnerabilities, permissive licensing, and a focused feature set for financial econometrics. Install friction is moderate but manageable; prebuilt wheels are available for all major platforms and Python versions. Recommended for anyone doing volatility modeling, unit root testing, or bootstrap-based inference on financial time series.
Install
arch on PyPI
Before you install
Medium install friction due to compiled components (Cython/Numba). Actively maintained with recent releases; supports Python 3.10–3.13 with prebuilt wheels across major platforms. Five runtime dependencies (numpy, pandas, scipy, statsmodels, packaging) are standard in data science environments.
Requires Python 3.10+. Optional Numba acceleration requires setting ARCH_NO_BINARY=1 environment variable and installing from source.
License in practice
Licensed under NCSA, which is permissive and compatible with most commercial and open-source projects. No notable restrictions on use or redistribution.
Quickstart
pip install arch
from arch import arch_model
import pandas as pd
returns = pd.Series([...]) # your returns data
am = arch_model(returns)
res = am.fit()Verify before relying
- Whether Numba acceleration provides meaningful performance gains for typical use cases
- Specific volatility model types beyond ARCH, GARCH, TARCH, EGARCH, and EWMA that are implemented
Package facts
| License | NCSA unclear |
| Python support | Supports the current Python release >=3.10 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 5 packagesnumpypandasscipystatsmodelspackaging |
| Maintenance | Actively maintained 297 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 1,059,093 / month, #4,423 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableIntended Audience :: End Users/DesktopIntended Audience :: Financial and Insurance IndustryIntended Audience :: Science/ResearchOperating System :: MacOS :: MacOS XOperating System :: Microsoft :: WindowsOperating System :: POSIXProgramming Language :: CythonProgramming Language :: PythonProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Topic :: Scientific/Engineering |
Evidence: arch-8.0.0-cp310-cp310-macosx_10_9_x86_64.whl; arch-8.0.0-cp310-cp310-macosx_11_0_arm64.whl; arch-8.0.0-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; arch-8.0.0-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl; arch-8.0.0-cp310-cp310-musllinux_1_2_x86_64.whl; arch-8.0.0-cp310-cp310-win_amd64.whl; arch-8.0.0-cp311-cp311-macosx_10_9_x86_64.whl; arch-8.0.0-cp311-cp311-macosx_11_0_arm64.whl; arch-8.0.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; arch-8.0.0-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl; arch-8.0.0-cp311-cp311-musllinux_1_2_x86_64.whl; arch-8.0.0-cp311-cp311-win_amd64.whl; arch-8.0.0-cp312-cp312-macosx_10_13_x86_64.whl; arch-8.0.0-cp312-cp312-macosx_11_0_arm64.whl; arch-8.0.0-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; arch-8.0.0-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl; arch-8.0.0-cp312-cp312-musllinux_1_2_x86_64.whl; arch-8.0.0-cp312-cp312-win_amd64.whl; arch-8.0.0-cp313-cp313-macosx_10_13_x86_64.whl; arch-8.0.0-cp313-cp313-macosx_11_0_arm64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “ARCH GARCH volatility modeling”
- archProvides ARCH/GARCH volatility models, unit root tests, cointegration…
- spandrel-extra-archesRegisters additional PyTorch model architectures with spandrel's…
- py-lets-be-rationalComputes implied volatility from option prices using Peter Jaeckel's…
Give your agent the search over MCP, or paste the wish link into any chat.
More Scientific/Engineering packages
NumPy provides an N-dimensional array object and a comprehensive suite of mathematical, linear algebra, Fourier transform, and random number functions for scientific computing in Python.
pandas provides fast, flexible data structures (Series and DataFrame) for loading, cleaning, transforming, and analyzing labeled or relational data in Python.
scipy provides numerical algorithms for mathematics, science, and engineering—including optimization, integration, linear algebra, Fourier transforms, signal and image processing, and ODE solvers—built on numpy arrays.
scikit-learn provides a comprehensive Python library for supervised and unsupervised machine learning, including classification, regression, clustering, dimensionality reduction, and model evaluation tools built on NumPy and SciPy.
Install it if you need to train, evaluate, or deploy supervised or unsupervised learning models.
dill extends Python's pickle module to serialize and deserialize a much wider range of Python objects, including functions, lambdas, classes, and interpreter sessions, to byte streams for storage or network transmission.
Multiprocess is an enhanced fork of Python's standard multiprocessing library that uses dill for better serialization, allowing you to spawn processes with a threading-like API and share complex objects between them.
Install it if you use multiprocessing and encounter pickle serialization limits with lambdas or complex objects.
See also bootstrapped · linearmodels · volatility3 · py-lets-be-rational · gstools · spreg · quantstats · statsmodels · ta · pyportfolioopt