Application of stochastic programming to management of cash flows with FX exposure

Abstract

In this thesis we formulate a model for foreign exchange (FX) exposure management and multi-currency cash management taking into consideration random fluctuations of exchange rates and net revenues of a multinational firm (MNF). The central decision model used in this thesis is a scenario-based stochastic programming (SP) recourse model. A critical review of alternative scenario generation methods is given followed by analysis of some desirable properties of the scenario tree. The application of matching statistical moments of a probability distribution to generate a multiperiod scenario tree for our problem is described in detail. A four-stage SP decision model is formulated using the random parameter values. This model evaluates currency / cash flows hedging strategies, which provide rolling decisions on the size and timing of the forward positions. We compute an efficient frontier from which an investor can choose an optimal strategy according to his risk and return preferences. The flexibility of the SP model allows an investor to analyse alternative risk-return trading strategies. The model decisions are investigated by making comparisons with decisions based purely on the expected value problem. The investigation shows that there is a considerable improvement to the "spot only" strategy and provides insight into how these decisions are made. The contributions of the thesis are summarised below. (i) The FX forward scenario trees are derived using an arbitrage-free pricing strategy and is in line with modem principles of finance. (ii) Use of the SP model and forward contracts as a tool for hedging decisions is novel. (iii) In particular smoothing of the effects in exchange rates and the smoothing of account receivables are examples of innovative modelling approaches for FX management.