A monetary value for initial information in portfolio optimization (Q1424701)

The authors consider an investor who trades in a financial market so as to maximize his expected utility of wealth at a prespecified time. The investor faces the opportunity to acquire, in addition to the common information flow \(\mathbf{F}\), some extra information \(\mathbf{G}\) at a certain cost. His subjective fair value of this information is defined as the amount of money that he can pay for \(\mathbf{G}\) such that this cost is balanced out by the informational advantage in terms of maximal expected utility. This value is studied for common utility functions in the setting of a complete market modelled by general semimartingales. The assumptions ensure the existence of the so-called martingale preserving probability measure (MPPM) \(\tilde{Q}\) corresponding to a given probability \(Q\). The main property of the MPPM is that it decouples \(\mathbf{F}\) and \(\mathbf{G}\) in such a way that \(\mathbf{F}\)-martingales under \(Q\) remain \(\mathbf{G}\)-martingales under \(\tilde{Q}\). The strong predictable representation property for local martingales is transferred from \(\mathbf{F}\) to the initially enlarged filtration \(\mathbf{G}\). Then, standard duality arguments are applied to solve the utility maximization problem in a general complete model when the initial information is non-trivial. Combination with martingale representation results permits to derive the utility indifference value for common utility functions. At last, closed form expressions for this value are provided in an Itô's process model where the additional information consists of a noisy signal about the terminal stock price.