Skip to main content

Implements high-frequency financial econometrics tools.

Project description

Table of Contents

  1. HFFE: High-Frequency Financial Econometrics
    1. Introduction
    2. Stocks
    3. Options
    4. Under development
    5. Redo introduction
    6. Add examples on how to use
    7. Add examples on how to use
    8. Change file to
    9. Add citations
    10. Sample data
    11. [] Change internals to pandas dataframe
    12. []
    13. [] Truncated variance estimator
    14. [] Local volatility estimator
    15. [] Confidence intervals

HFFE: High-Frequency Financial Econometrics


This package implements some tools of the high-frequency financial econometrics literature. The HFFE package provides a class Stock. This class can be used in two ways:

  1. Instantiate the class and provide stock data. The instance __init__ method will automatically compute geometric returns, realized and bipower variances, and separate diffusive from jump returns using the stock data supplied.
  2. Use the class methods by themselves as needed. All of the methods in Stock are static methods and can be used without instantiating the class.

The class can be instantiated via the default constructor by supplying an iterable containing prices (floats) and an iterable containing date stamps (in the format YYYYMMDD for year, month and day). The default constructor also takes an optional iterable containing time stamps (in the format HHMM for hour and minute or HHMMSS for hour, minute and second). The data is assumed to be rectangular, meaning that each day contains the same number of price observations.

Prices are assumed to be observed (N) times per day, at the same discrete intervals. For example, if prices are sampled every 5 minutes starting from 9:30 AM and finishing at 4:00 PM, then we will have (N=79) price observations per day. Prices are assumed to be observed for (T) total days. In each of these (T) days we have (N) price observations. Geometric returns (log-returns) are computed for each of the days and for each of the discrete sampling intervals. However, overnight returns are not considered, and the packaged only focuses on the analysis of intraday returns. If there are (N=79) price observations each day, then we will have (n=78) returns each day.


from hffe import Stock
from random import normalvariate as randn
# generate some fake data
prices = np.array([10.0 + randn(0, 1) for _ in range(10)])
dates = ['20181112']*len(prices)
times = [f'093{i}' for i in range(10)]
# instantiate class
stock = Stock(prices, dates, times)
# at this point we can access the returns, variance measures and
# obtain the diffusive returns separated from the jump returns
print(f'Number of prices per day: {["prices"]}\n'
  f'Number of days: {["days"]}')
# print measures of variance
print(f'Realized Variance: {stock.RV}\n'
  f'Annualized Realized Variance: {(252*stock.RV)**0.5}\n'
  f'Bipower Variance: {stock.BV}\n'
  f'Annualized Bipower Variance: {(252*stock.BV)**0.5}')



Under development

TODO Redo introduction

TODO Add examples on how to use

TODO Add examples on how to use

TODO Change file to

TODO Add citations

  1. TODO To volatility estimators

  2. TODO To jump threshold and separtion

  3. TODO To time-of-day factor

TODO Sample data

  1. TODO Sample stock data

  2. TODO Sample SPX options data

TODO [] Change internals to pandas dataframe


TODO [] Truncated variance estimator

TODO [] Local volatility estimator

TODO [] Confidence intervals

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for hffe, version 0.0.13
Filename, size File type Python version Upload date Hashes
Filename, size hffe-0.0.13-py3-none-any.whl (28.7 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size hffe-0.0.13.tar.gz (14.7 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page