Machine Learning in Futures Markets
In this paper, we demonstrate how a well-established machine learning-based statistical arbitrage strategy can be successfully transferred from equity to futures markets. First, we preprocess futures time series comprised of front months to render them suitable for our returns-based trading framewor...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-03-01
|
Series: | Journal of Risk and Financial Management |
Subjects: | |
Online Access: | https://www.mdpi.com/1911-8074/14/3/119 |
id |
doaj-22cd9747149f4c9e879cc7a0e44a54e9 |
---|---|
record_format |
Article |
spelling |
doaj-22cd9747149f4c9e879cc7a0e44a54e92021-03-14T00:00:16ZengMDPI AGJournal of Risk and Financial Management1911-80661911-80742021-03-011411911910.3390/jrfm14030119Machine Learning in Futures MarketsFabian Waldow0Matthias Schnaubelt1Christopher Krauss2Thomas Günter Fischer3Department of Statistics and Econometrics, University of Erlangen-Nürnberg, 90403 Nürnberg, GermanyDepartment of Statistics and Econometrics, University of Erlangen-Nürnberg, 90403 Nürnberg, GermanyDepartment of Statistics and Econometrics, University of Erlangen-Nürnberg, 90403 Nürnberg, GermanyDepartment of Statistics and Econometrics, University of Erlangen-Nürnberg, 90403 Nürnberg, GermanyIn this paper, we demonstrate how a well-established machine learning-based statistical arbitrage strategy can be successfully transferred from equity to futures markets. First, we preprocess futures time series comprised of front months to render them suitable for our returns-based trading framework and compile a data set comprised of 60 futures covering nearly 10 trading years. Next, we train several machine learning models to predict whether the <i>h</i>-day-ahead return of each future out- or underperforms the corresponding cross-sectional median return. Finally, we enter long/short positions for the top/flop-<i>k</i> futures for a duration of <i>h</i> days and assess the financial performance of the resulting portfolio in an out-of-sample testing period. Thereby, we find the machine learning models to yield statistically significant out-of-sample break-even transaction costs of 6.3 bp—a clear challenge to the semi-strong form of market efficiency. Finally, we discuss sources of profitability and the robustness of our findings.https://www.mdpi.com/1911-8074/14/3/119statistical arbitragefutures marketsmachine learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Fabian Waldow Matthias Schnaubelt Christopher Krauss Thomas Günter Fischer |
spellingShingle |
Fabian Waldow Matthias Schnaubelt Christopher Krauss Thomas Günter Fischer Machine Learning in Futures Markets Journal of Risk and Financial Management statistical arbitrage futures markets machine learning |
author_facet |
Fabian Waldow Matthias Schnaubelt Christopher Krauss Thomas Günter Fischer |
author_sort |
Fabian Waldow |
title |
Machine Learning in Futures Markets |
title_short |
Machine Learning in Futures Markets |
title_full |
Machine Learning in Futures Markets |
title_fullStr |
Machine Learning in Futures Markets |
title_full_unstemmed |
Machine Learning in Futures Markets |
title_sort |
machine learning in futures markets |
publisher |
MDPI AG |
series |
Journal of Risk and Financial Management |
issn |
1911-8066 1911-8074 |
publishDate |
2021-03-01 |
description |
In this paper, we demonstrate how a well-established machine learning-based statistical arbitrage strategy can be successfully transferred from equity to futures markets. First, we preprocess futures time series comprised of front months to render them suitable for our returns-based trading framework and compile a data set comprised of 60 futures covering nearly 10 trading years. Next, we train several machine learning models to predict whether the <i>h</i>-day-ahead return of each future out- or underperforms the corresponding cross-sectional median return. Finally, we enter long/short positions for the top/flop-<i>k</i> futures for a duration of <i>h</i> days and assess the financial performance of the resulting portfolio in an out-of-sample testing period. Thereby, we find the machine learning models to yield statistically significant out-of-sample break-even transaction costs of 6.3 bp—a clear challenge to the semi-strong form of market efficiency. Finally, we discuss sources of profitability and the robustness of our findings. |
topic |
statistical arbitrage futures markets machine learning |
url |
https://www.mdpi.com/1911-8074/14/3/119 |
work_keys_str_mv |
AT fabianwaldow machinelearninginfuturesmarkets AT matthiasschnaubelt machinelearninginfuturesmarkets AT christopherkrauss machinelearninginfuturesmarkets AT thomasgunterfischer machinelearninginfuturesmarkets |
_version_ |
1724221894476955648 |