Spot pricing of public utility services

This thesis analyzes how public utility prices should be changed over time and space. Earlier static and non spatial models of public utility pricing emerge as special cases of the theory developed here. Electricity is emphasized although the models can be used for natural gas and other public utilities. If the transactions costs of price changes were zero, optimal prices should be changed continuously as supply and demand conditions change. Such prices are referred to as "full spot" prices. The full spot price of electricity at any point and time depends on total demand, availability of generating units, short run marginal operating costs of generators, the spatial locations of all supplies and demands, and the configuration of the transmission and distribution system. Since all of these are stochastic, so are full spot prices. Optimal "wheeling charges", i.e. price differences between points, also vary stochastically. In practise actual prices must be changed discreetly, and are therefore only approximations to full spot prices. Price changes are of two basic types. Predetermined price changes are adequate to respond to anticipated changes in conditions. Price recalculations are needed to respond to unanticipated changes. The optimal timing and mix of recalculations and predetermined changes depend on: the transactions costs of each type of change; the stochastic and deterministic rates of change of full spot prices; and the ability of customers and suppliers to change behavior in response to different price patterns. Conventional time-of-use rates recalculate prices systematically only at occasional rate hearings, and change them only a few times a day. Such prices deviate greatly from full spot prices, for many utilities. The thesis models customer behavior under full spot and other time varying prices, and discusses the types of customers likely to get the largest benefits from full spot pricing. The final chapter simulates behavior by four customers under six rates, from flat prices calculated a year in advance, to full spot pricing. The gross social welfare benefits of full spot pricing, before transactions costs, are three to ten times the benefits of conventional time-of-use pricing. The gross benefits of full spot pricing are less than ten percent of the customers' total energy costs. For small customers this may not be enough to counterbalance higher :ransactions costs from spot pricing. The thesis calculates "breakeven sizes" for each customer type such that larger customers should be on full spot pricing. These are on the order of one megawatt, for the customers and utility system modeled. The thesis suggests that, for many large customers and independent power producers, conventional time-of-use rates are dominated by properly calculated prices which change every hour and are recalculated at least daily. For some other customers, rates of intermediate sophistication are best. Thesis Supervisor: Richard Schmalensee Title: Professor of Applied Economics