The Quarterly Review of Economics and Einance, Vol. 38, Special Issue, 1998, pages 755-769 Copyright 0 1998 ‘Ikustees of the University of Illinois All rights of reproduction in any form reserved ISSN 1062.9769 -Optimal Timing of a Mine Expansion: Implementing a Real Options Model GONZALO CORTAZAR JAIME CASASSUS Pontificia Universidad Catblica de Chile We present the results of implementing a real options m&l for valuing an investment project that expands production capacity and/or modifies unit costs of a cvpper mine. Th model and its implementation addresses the three requirements we find necessary to increase the use of the real option methodology by the practitioner community: a user-acceptable stochastic model for commod- ity prices with mean reversion, a customized real asset model which includes the main managerial jlexibilities of opening-closing production or delaying investments, and a user-fkiendly computer implementation. A case study shows that a signajkant fraction of investment value may be due to th4 flexibility of delaying investment, value that decreases o.s co#er prices increase. Critical investment prices are analyzed. Corporate finance decisions are primarily made to create value. This principle implies for capital budgeting that only positive net present value projects should be undertaken. Even though this principle seems clear and simple, it is difftcult to use under uncertainty for three main reasons. First, if cash flows are volatile we must determine equilibrium risk premiums to add to the risk free interest rates for discounting expected cash flows. This requires a rather involved procedure that includes using some equilibrium model, like the Capital Asset Pricing Model, with results very dependent on its assumptions. Real investments have the additional difficulty of having to do this computation specifically for each one, ’ in order to take into consideration its operating leverage. Second, good managers react to changing conditions, expanding successful projects and closing or delaying unprofitable ones. This induces a nonlinear cash flow that makes computing expected cash flows a much more difftcult task than simply evaluating cash flows at the expected value of all uncertain vari- ables. 755