Herding behaviour and price convergence clubs in cryptocurrencies during bull and bear markets

Abstract

This paper sets out to explore whether convergence and herding phenomena exist for digital currencies. Daily data cover a large spectrum of cryptocurrencies in separate bull and bear periods. Empirical estimations for detecting club convergence and clustering are performed by the methodology proposed by Phillips and Shu (2007, 2009). Econometric outcomes reveal preliminary evidence of powerful herding behaviour. The lowest of large-cap digital currencies attract mainly cryptocurrencies which are about in the middle of the large-cap and medium-cap categories whereas the highest-cap cryptocurrencies are parts of convergence clubs with mostly large-cap or purely medium-cap digital currencies during bear markets. Notably, segmentation is higher during bear markets as clusters are formed around more numerous cryptocurrencies than during bull markets. Convergence is stronger during flourishing periods. Secondary herding is also realized among pairs of clubs. Our findings enable investors to better diversify their portfolios and ameliorate their risk-return trade-off during extreme events. © 2021 Elsevier B.V.