Bidding evaluation and strategies: An application in the competitive electric power industry

Since the passage of the Public Utilities Regulatory Policies Act of 1978, competitive bidding has become widespread. As the electric system becomes more decentralized, it is important that efficient production be based on truthful energy cost information and that the system maintains optimal diversity of plant types. A two-dimensional auction model is developed in this dissertation. In this model, a bidder submits both fixed and variable payment bids, with the latter depending on how much electricity is produced. The dispatch period of a unit is modeled as depending on the energy bid. The bidder possesses a privately known technology curve that trades off capacity costs against energy costs. The bidder chooses an electric generation technology, a capacity bid and an energy bid based on its own private information. The utility chooses the winning bid based on a pre-announced Evaluation Function. The solution concept used is a symmetric Nash Equilibrium. We show that an appropriate Evaluation Function can help achieve revelation of true energy costs and an optimal choice of technology. The information premium depends on the locus of the technology choice in the capacity-energy-cost domain.;We then add uncertainty about the dispatch schedule to the analysis. The utility also provides a guaranteed dispatch schedule. We investigate two forms of evaluation criteria and contract. In the first, we assume that the utility dispatches the units according to the submitted energy bid and will "pay-and-not-take" if the realized dispatch schedule falls short of the guaranteed one. In the second, we assume that the utility will exercise the "take" option even in "bad" states. Results from the analysis of these two contracts show that providing a dispatch guarantee cannot ensure truthful revelation of energy costs. The technology choice in both cases is operationally efficient but socially sub-optimal. We also find that the utility can reduce the expected contract payment by affecting the technology choice locus.