Economists and financial analysts have begun to recognise the importance of the actions of other agents in the decision-making process. Herding is the deliberate mimicking of the decisions of other agents. Examples of mimicry range from the choice of restaurant, fash-ion and financial market participants, to academic research. Herding may conjure negative images of irrational agents sheepishly following the actions of others, but such actions can be rational under asymmetric information and uncertainty. This paper uses futures position data in nine different markets of the Commodity Futures Trading Commission (CFTC) to provide a direct test of herding behaviour, namely the extent to which small traders mimic the posi-tions of large speculators. Evidence consistent with herding among small traders is found for the Canadian dollar, British pound, gold, S&P 500 and Nikkei 225 futures. Consistent with survey-based results on technical analysis, the positions are significantly correlated with both current and past market returns. Using various time-varying volatility models to accommodate conditional heteroskedasticity, the empirical results are found to be robust to alternative models and methods of estimation. When a test of causality-in-variance is used to analyse if volatility among small traders spills over into spot markets, it is found that spillovers occur only with Nikkei 225 futures. The policy implications of the findings are also discussed.