A note on the hedging effectiveness of GARCH models

This note compares the hedging effectiveness of the conventional hedge ratio and time-varying conditional hedge ratios (of which GARCH ratio is a special case). It is shown that, in large sample cases, the conventional hedge ratio provides the best performance. For small sample cases, a sufficiently large variation in the conditional variance of the futures return is required to produce the opposite result. The result is due to the fact that the hedging effectiveness measure is based upon the unconditional variance; meanwhile, the conventional hedge ratio minimizes the unconditional variance and the conditional hedge ratio aims at minimizing the conditional variance.