Combining time-varying and dynamic multi-period optimal hedging models

This paper presents an effective way of combining two distinct approaches used in the hedging literature--dynamic programming (DP) and time-series (GARCH) econometrics. Theoretically consistent yet realistic and tractable models are developed for traders interested in hedging a portfolio. Results from a bootstrapping experiment used to construct confidence bands around the competing portfolios suggest that, whereas DP--GARCH outperforms the GARCH approach, they are statistically equivalent to the OLS approach when the markets are stable. Traders may achieve significant gains, however, by adopting the DP--GARCH model rather than the OLS approach when markets are volatile. Copyright 2002, Oxford University Press.