Tools · SPBoucher
Menu

arch (Python package)

by Kevin Sheppard · Library · also known as arch, bashtage/arch

Kevin Sheppard's Python package for univariate ARCH/GARCH volatility models, bootstraps, multiple-comparison procedures, unit-root and cointegration tests

NCSA licenseFreeOpen sourceLocalSelf-hostable
Visit official site Compare

Link verified Nov 2, 2026

Overview

arch is a Python library by Kevin Sheppard for financial econometrics, released under the University of Illinois/NCSA license. Its core is univariate volatility modelling: mean models (constant, AR, HAR, exogenous regressors) combined with ARCH, GARCH, GJR-GARCH, EGARCH, HARCH and related variance processes and normal, Student-t, skewed-t or GED innovations, estimated by maximum likelihood with analytic forecasts, simulation and bootstrap forecasting. Alongside it ships a bootstrap module (IID, stationary, circular block and others) with confidence-interval helpers, multiple-comparison procedures such as the reality check, SPA and model confidence set, unit-root tests (ADF, KPSS, Phillips-Perron, DF-GLS, Zivot-Andrews), cointegration tests and estimators (Engle-Granger, Phillips-Ouliaris) and long-run covariance estimation.

Version 8.0.0 is current at the time of writing. The package integrates with pandas and statsmodels conventions and is widely used for realized-volatility and risk research on daily or intraday returns. It covers univariate models only; multivariate GARCH, stochastic volatility and high-frequency realized-measure estimators are not included.

Capabilities

  • ARCH, GARCH, GJR-GARCH, EGARCH and HARCH volatility models
  • Mean models with AR, HAR and exogenous terms
  • Analytic, simulation and bootstrap forecasts
  • Bootstrap methods: IID, stationary, circular block
  • Reality check, SPA and model confidence set
  • Unit-root tests: ADF, KPSS, PP, DF-GLS, Zivot-Andrews
  • Cointegration tests and long-run covariance estimation

Typical workflows

  1. GARCH volatility forecasts for risk measures
  2. Model confidence set over competing forecasting models
  3. Stationarity testing before time-series modelling

Strengths

  • Comprehensive univariate volatility toolkit
  • Permissive NCSA licence
  • Consistent with pandas and statsmodels

Limitations

  • Univariate models only
  • No realized-measure or stochastic-volatility estimators
  • Single author-maintained project

Prerequisites

  • Python with NumPy, SciPy, pandas and statsmodels

Who it is for

  • Empirical finance researchers
  • Risk analysts working in Python

Alternatives & comparison

statsmodels lacks GARCH-family models, which arch supplies; EViews and R's rugarch offer comparable volatility modelling in other environments; linearmodels, by the same author, covers panel and IV estimation.

  • EViews

    Commercial Windows econometrics package for time-series estimation, forecasting and macro modelling, sold as single-user, volume, University and free Student Lite editions

    ProprietaryPaid
  • linearmodels

    Python package by Kevin Sheppard adding panel data, instrumental-variable, system (SUR/3SLS) and asset-pricing estimators missing from statsmodels

    Open sourceFree
  • R

    Language and environment for statistical computing and graphics, with the CRAN package archive and a strong base in academic statistics and econometrics

    Open sourceOpen source
  • statsmodels

    Python library for estimating statistical and econometric models with R-style formulas, rich result summaries and tests: OLS, GLM, discrete choice, ARIMA, state space, VAR and more

    Open sourceFree
Compare arch (Python package) with its alternatives →

Sources & verification

Primary sources consulted for this record. Claims are tied to the fields each source supports; the automated link checker re-verifies every URL on a schedule.

  • License
    arch LICENSE.md

    https://github.com/bashtage/arch/blob/main/LICENSE.md

    Supports: license_spdx, source_model

    checked Nov 1, 2026
  • Official site
    arch 8.0.0 — Introduction

    https://bashtage.github.io/arch

    Supports: name, summary, capabilities, vendor

    checked Nov 1, 2026
  • Repository
    bashtage/arch

    https://github.com/bashtage/arch

    Supports: repo_url, capabilities

    checked Nov 1, 2026
Confidence
High
Reviewed
Nov 1, 2026
Links
Verified · Nov 2, 2026

Spotted an error?

Broken link, outdated license, wrong category, or a tool that no longer exists? Tell us; reports are reviewed by the editor.

Apache Parquet

Apache Software Foundation · Data catalogs & formats

In-depth entry

Open columnar file format for analytical data with compression, encodings and statistics, readable by nearly every data engine and library

Apache-2.0Open sourceFreeLocalNov 2, 2026

ClickHouse

ClickHouse, Inc. · Analytical engines & OLAP

In-depth entry

Open-source columnar OLAP database server for fast aggregations over billions of rows, with MergeTree storage, SQL and a managed ClickHouse Cloud

Apache-2.0Open sourceFreeLocalSelf-hostedNov 2, 2026

DuckDB

DuckDB Foundation / DuckDB Labs · Analytical engines & OLAP

In-depth entry

In-process analytical SQL database that queries Parquet, CSV and Arrow data directly from Python, R, Node or the CLI

MITOpen sourceFreeLocalSelf-hostedNov 2, 2026

Massive (formerly Polygon.io)

Massive (Polygon.io, Inc.) · Market & fundamentals data

Market data platform (renamed from Polygon.io in October 2025) with REST and WebSocket APIs, flat files and SQL access for US stocks, options, indices, forex, crypto and futures

ProprietaryPaidEnterpriseNov 2, 2026

Databento

Databento, Inc. · Market & fundamentals data

In-depth entry

Usage-priced market data platform with normalized order-book (MBO/MBP), trades, OHLCV and reference schemas for futures, options and equities, delivered via historical, live and batch APIs in the DBN format

MixedUsage-basedPaidNov 2, 2026

FirstRate Data

FirstRate Data LLC · Market & fundamentals data

In-depth entry

Vendor of historical intraday market data (1-minute to daily bars, tick and options) for US stocks, ETFs, futures, indices, crypto and FX, delivered as zipped CSV bundles under a perpetual licence

ProprietaryPaidAcademicNov 2, 2026