UC Berkeley

Large-scale public capital-investment projects are typically evaluated using standard net-present-value (NPV) analysis. Projects with large, up-front investments, such as dams or energy plants, often are irreversible and have uncertain future benefits. If we expect to reduce uncertainty about these future flows over time, introducing the option to delay the project may increase project value and lead to different decisions. This paper evaluates a pending California water storage and conveyance project that would increase the capacity of an existing dam, the value of which depends upon the future climate state. The research uses a dynamic optimization program with two stochastic state variables to analyze decision making. One state variable is the belief about the future climate state, and the other is the level of green-house-gas (GHG) stocks, both of which change in a random manner. The analysis allows for the possibility of learning more about the true state of the future climate as time goes on. The optimal investment policy differs substantially from that proposed by the policy makers involved in the empirical project.