Committing to coal and gas: long-term contracts, regulation, and fuel switching in power generation

Abstract

Fuel switching in the electricity sector has important economic and environmental consequences. In the United States, the increased supply of gas during the last decade has led to substantial switching in the short term. Fuel switching is constrained, however, by the existing infrastructure. The power generation infrastructure, in turn, represents commitments to specific sources of energy over the long term. This dissertation explores fuel contracts as the link between short-term price response and long-term plant investments. Contracting choices enable power plant investments that are relationship-specific, often regulated, and face uncertainty. Many power plants are subject to both hold-up in investment and cost-of-service regulation. I find that capital bias is robust when considering either irreversibility or hold-up due to the uncertain arrival of an outside option. For sunk capital, the rental rate is inappropriate for determining capital bias. Instead, capital bias depends on the regulated rate of return, discount rate, and depreciation schedule. If policies such as emissions regulations increase fuel-switching flexibility, this can lead to capital bias. Cost-of-service regulation can shorten the duration of a long-term contract. From the firm's perspective, the existing literature provides limited guidance when bargaining and writing contracts for fuel procurement. I develop a stochastic programming framework to optimize long-term contracting decisions under both endogenous and exogenous sources of hold-up risk. These typically include policy changes, price shocks, availability of fuel, and volatility in derived demand. For price risks, the optimal contract duration is the moment when the expected benefits of the contract are just outweighed by the expected opportunity costs of remaining in the contract. I prove that imposing early renegotiation costs decreases contract duration. Finally, I provide an empirical approach to show how coal contracts can limit short-term fuel switching in power production. During the era prior to shale gas and electricity market deregulation, I do not find evidence that gas generation substituted for coal in response to fuel price changes. However, I do find evidence that coal plant operations are constrained by fuel contracts. As the min-take commitment to coal increases, changes to annual coal plant output decrease. My conclusions are robust in spite of bias due to the selective reporting of proprietary coal delivery contracts by utilities.