Hedging large risks reduces the transaction costs

Abstract

As soon as one accepts to abandon the zero-risk paradigm of Black-Scholes, very interesting issues concerning risk control arise because different definitions of the risk become unequivalent. Optimal hedges then depend on the quantity one wishes to minimize. We show that a definition of the risk more sensitive to the extreme events generically leads to a decrease both of the probability of extreme losses and of the sensitivity of the hedge on the price of the underlying (the `Gamma'). Therefore, the transaction costs and the impact of hedging on the price dynamics of the underlying are reduced.