Herding behavior and contagion in the cryptocurrency market

Abstract

This study aimed to analyze herding behavior and contagion phenomena in the cryptocurrency market. We selected 50 of the most liquid and capitalized currencies in the period from March 2015 to November 2018 (daily data). The methodology used for detecting herding behavior comprised adaptations of the cross-sectional absolute deviation (CSAD) and cross-sectional standard deviation (CSSD) tests, as well as Hwang and Salmon’s (2004) model. For the contagion effect, we utilized adaptations of Forbes and Rigobon’s (2002) (FR) test, and FR test extensions based on the comoments of Fry, Martin, and Tang (2010) and Fry-McKibbin and Hsiao (2018). The results of using the CSSD test and Hwang and Salmon’s (2004) state space model revealed herding behavior, demonstrating extreme periods of adverse herd behavior. As regards the contagion effect, the modified FR test and its extensions with comoments were able to identify the Bitcoin contagion in other currencies in almost all cases.