In this short paper some computational aspects of dynamic programming (DP) applied to the real option problem of the dual fuel boiler in (Kulatilaka, 1993) are examined. A Gauss code has been written, and it is reported in the text, for both pedagogical and practical purposes. As a matter of fact, understanding how the code works helps to delve better into the mechanics of the DP algorithm. On the other hand, the reported code is an hands tool for those who want to apply the (Kulatilaka, 1993) model to real business situations. Since the estimation of the stochastic process driving the state variable of the system is crucial for a good application of the model, four different econometric methods have been applied to the estimation of Ornstein - Uhlenbeck parameters, namely OLS for the (Kulatilaka, 1993) and the (Dixit, Pindyck 1994) specifications, GMM and MLLE for (Roncalli, 1999) approach. Several comparative statics exercises have been performed on real data on oil, coal, natural gas and the boiler characteristics.
Read full abstract