Portfolio managers use index futures for a variety of reasons. Regardless of their motivation, they will keep a close eye on the relation between the index futures returns and their stock-portfolio returns. Whenever this relation is perceived to have changed, the manager will decide whether it is worthwhile to rebalance the index futures-portfolio mix accordingly. Exact measures as to when and how much rebalancing should occur have not yet been established. This article proposes a dynamic hedging algorithm based on a reverse order CUSUM-squared (ROC) testing procedure, first discussed in M. H. Pesaran and A. Timmermann (2002). A comparison with standard alternatives (naïve, expanding, EWLS, and rolling estimation windows)The authors are grateful for comments and suggestions from participants and discussants (in particular Kin Lam, Jay Muthuswamy, and Robert Webb) at the 15th Annual Asia-Pacific Futures Research Symposium, the ESAM2004 and QMF2004 conferences, the Q-Group Meeting in Melbourne, and seminar participants at Macquarie University and Monash University.