Estimating Peak and Off-Peak Marginal Costs for an Electric Power System: An Ex Ante Approach

This paper proposes and implements an ex ante approach to estimating thermal electric power system long-run average and peak and off-peak marginal costs subject to pollution emission constraints. Mixed-integer programming is used as a framework for a static, cost-minimizing model of system capacity and operating costs incurred in meeting loads that vary with time and geographic location. The model assumes multiple generating technologies can be used to meet system loads. Transmission and losses are explicitly considered in this transshipment mathematical program. The model is implemented for the system of the New York State Electric and Gas Corp. Results show that peak marginal costs at a particular location in the system are three to four times system average costs and can be more than ten times as great as base load marginal costs. Marginal costs during a particular demand period vary by as much as 30 percent from one load center to another in the system. System average total costs increase (with output) when new steam plants are all coal-fired, but decrease when they are nuclear-fueled. When the most restrictive pollution standards are imposed, system average total costs are approximately 25 percent higher than when no standards are imposed. These results provide substantial empirical evidence in support of the case for time-specific rates based on marginal costs.