Time-variations in herding behavior: Evidence from a Markov switching SUR model

This paper aims at testing for time-variations in herd behavior in stock markets. In particular, we analyze how investors' behavior differs between times of market turmoil and tranquil trading periods. Thereby, we take into account herding within a certain market as well as international spillovers in herd formation. Our evidence for the US and the Euroarea suggests that, during periods of high volatility, deviations from rational asset pricing are more persistent and spillovers between the markets are substantially amplified. In general, our findings show that during periods of crisis, like the recent global financial crisis and the period after the dot.com bubble bursting, stock prices are much more driven by behavioral effects compared to tranquil times.