Sequential Investments in Power Plants: A Structural Estimation Approach
Abstract
In this thesis we develop a structural estimation model for sequential investment decisions emerging in real options applications. The optimization methodology, Mathematical Program with Equilibrium Constraints (MPEC), is used to solve our structural model in order to reduce computational complexity. We show, using a generated stylized dataset, that the estimated parameters resemble the equivalent parameters in the theoretical real options model. The model's ability to forecast behavior of real firms is investigated with data from U.S. coal-fired power plant investments. Our analysis reveals that investments in coal-fired power plants are complex with both economic and non-economic factors influencing decisions. While our structural model only accounts for expected dark spreads, we find that a more elaborate model is required to capture the complexity of power plant investments. Finally, we suggest further work that will improve the models ability to conduct counterfactual policy analysis.