Description

This presentation provides an overview of herding behavior in financial markets, with a particular focus on cryptocurrencies. Herding occurs when investors imitate the actions of others rather than relying on their own information, leading to excessive co-movement of asset prices. The discussion contrasts rational market benchmarks, such as the Efficient Market Hypothesis (EMH) and the Capital Asset Pricing Model (CAPM), with behavioral deviations caused by herding. The evolution of herding measurement is reviewed, from the Cross-Sectional Standard Deviation (CSSD) of Christie and Huang (1995) to the Cross-Sectional Absolute Deviation (CSAD) approach of Chang et al. (2000). Particular emphasis is placed on CSAD as a robust measure of return dispersion under extreme market conditions. Empirical applications are conducted using CRIX, CoinGecko, and Kaggle cryptocurrency datasets. Log returns are analyzed together with CSSD and CSAD dynamics to identify periods of investor convergence. Regression-based herding tests reveal mixed evidence across datasets. Significant herding effects are detected in the CRIX and Kaggle samples, whereas recent CoinGecko data do not provide statistically significant evidence of herding. The results suggest that longer historical samples exhibit stronger convergence among investors, while recent cryptocurrency markets appear more heterogeneous. The study highlights the importance of behavioral effects in digital asset markets and points toward future research on sentiment-driven herding and robust Huber-loss-based measures.