Do commodity investors herd? Evidence from a time-varying stochastic volatility model

Commodities markets due to their unique characteristics that are they exhibit negative correlation with returns of traditional asset classes and are among the few assets that offer protection from the effects of inflation have recently garnered investors' attention especially through the development of commodity index financial products. This financialization process that started in the early 2000s and escalated after 2004 has precipitated price comovements among various types of commodities creating a proper setting for the examination of herding behavior. Employing a comprehensive dataset of investable commodities indices we examine the existence of herding behavior via static and time varying models. Our findings reveal no evidence of herding behavior according to static model. However, when rolling window analysis is in place significant anti-herding behavior is detected. These behavioral patterns are corroborated through a time varying stochastic volatility model. Our results contain significant implications for investors, commodities producers and policy makers.