Herding takes place when individuals tend to rely on the consensus opinion and past trades rather than on fundamental asset pricing. In this study, we focus on 101 Chilean stocks over the period 1990-2009. Contrary to empirical evidence for the United States and international markets (e.g., Christie & Huang, 1995; Demirer & Kutan, 2006; Gleason, Lee & Mathur 2003; Gleason, Mathur & Peterson, 2004), we conclude that herding in individual stocks is more likely during extreme down markets, after controlling for trading volume. Our results also suggest an inverse relationship between volatility and trading volume, which contradicts the one-factor mixture-of-distribution hypothesis (MDH). Such an inverse association appears statistically more significant when choosing more extreme return thresholds. © 2010 de Cladea.