Portfolio insurance and delta hedging involves transaction costs that are recognized in finance theory but have not been studied in many empirical applications. Econometric time series models can be used to forecast the number of recompositions of an insured portfolio, and the time between each recomposition. We forecast these variables using modified autoregressive conditional Poisson models (ACP) and autoregressive conditional duration models (ACD) The model are successful in capturing the series’ autocorrelation, and provide sensible estimates of the transaction cost associated with the recompositions.
