skip to navigation
skip to content

arch 1.0

ARCH for Python

Latest Version: 4.3.1


This is a work-in-progress for ARCH and related models, written in Python (and Cython)

What is this repository for?

  • Mean models
  • Constant mean
  • Heterogeneous Autoregression (HAR)
  • Autoregression (AR)
  • Zero mean
  • Models with and without exogensou regressors
  • Volatility models
  • ARCH
  • EWMA/RiskMetrics
  • Distributions
  • Normal
  • Student’s T


See the example notebook for a more complete overview.

import datetime as dt
import as web
st = dt.datetime(1990,1,1)
en = dt.datetime(2014,1,1)
data = web.get_data_yahoo('^FTSE', start=st, end=en)
returns = 100 * data['Adj Close'].pct_change().dropna()

from arch import arch_model
am = arch_model(returns)
res =


Documentation is hosted on read the docs


  • NumPy (1.7+)
  • SciPy (0.12+)
  • Pandas (0.14+)
  • statsmodels (0.5+)
  • matplotlib (1.3+)

Installing * Cython (0.20+) * nose (For tests) * sphinx (to build docs) * sphinx-napoleon (to build docs)


Setup does not verify requirements. Please ensure these are installed.


pip install git+git://


Anaconda builds are not currently available for OSX.

conda install -c arch


With a compiler

If you are comfortable compiling binaries on Windows:

pip install git+git://

No Compiler

All binary code is backed by a pure Python implementation. Compiling can be skipped using the flag --no-binary

pip install git+git:// --install-option "--no-binary"

Note that it isn’t possible to run the test suite will fail if installed with --no-binary


conda install -c arch

More about ARCH

More information about ARCH and related models is available in the notes and research available at Kevin Sheppard’s site.

File Type Py Version Uploaded on Size
arch-1.0-cp27-none-win_amd64.whl (md5) Python Wheel 2.7 2014-09-03 102KB
arch-1.0-cp33-none-win_amd64.whl (md5) Python Wheel 3.3 2014-09-03 99KB
arch-1.0.tar.bz2 (md5) Source 2014-09-03 33KB (md5) MS Windows installer 2.7 2014-09-03 322KB (md5) MS Windows installer 3.3 2014-09-03 317KB (md5) Source 2014-09-03 47KB