Discrete Time Portfolio Management with Transaction Costs

摘要

The purpose of this paper is to present a novel discrete time method that can beused to predict the future behavior of a set of risky assets. Given the predicted behavior of the portfolio the time dependent optimal strategy for portfolio selection can be made whenever it is required. In this work the incurred cost for each transaction is a fixed fraction of the current price of the underlying asset. However, it is straightforward to include the actual costings that are used in practice. The complex nature of the time series sequences of risky assets renders quantitative analysis extremely difficult in general.