Features of overreactions in the cryptocurrency market

•We examine features of overreactions for cryptocurrency markets.•For this purpose, we propose and model 18 different features.•We use a random forest classification based on all feature combinations.•We also shown that our approach can be exploited by investors.•This trading strategy is superior fo...

Full description

Saved in:
Bibliographic Details
Published in:The Quarterly review of economics and finance 2021-05, Vol.80, p.31-48
Main Authors: Borgards, Oliver, Czudaj, Robert L.
Format: Article
Language:English
Subjects:
College teachers
Crypto-currencies
Cryptocurrency
Mean reversion
Overreaction
Prediction
Random forest
Securities trading
Stock markets
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•We examine features of overreactions for cryptocurrency markets.•For this purpose, we propose and model 18 different features.•We use a random forest classification based on all feature combinations.•We also shown that our approach can be exploited by investors.•This trading strategy is superior for cryptocurrencies compared to the S&P500. This paper examines features of overreactions that are able to enhance the prediction quality for twelve cryptocurrencies compared to the US stock market. For this purpose, we perform random forest classifications on the basis of all feature combinations and a customized performance metric to predict overreactions on interday and various intraday price levels. We find that features describing the price development prior to the overreaction have the highest ability to classify an overreaction for different frequencies, indicating volatility clustering and framing effects. During an overreaction, the duration and the price steadiness are important features describing the overreaction itself. Our findings are largely comparable for cryptocurrencies and the US stock market despite the fact that both markets are fundamentally different. However, the returns of an overreaction trading strategy are superior for cryptocurrencies while those of US stocks are consistently negative due to the different size of their price reversals as the key factor for profitably exploiting our empirical findings. In addition, our results show for all assets and frequencies that the prediction results are slightly higher for positive overreactions compared to negative overreactions.
ISSN:1062-9769
1878-4259
DOI:10.1016/j.qref.2021.01.010