Customer Strategies for Responding to Day-Ahead Market Hourly Electricity Pricing
Real-time pricing (RTP) has been advocated as an economically efficient means to send price signals to customers to promote demand response (DR) (Borenstein 2002, Borenstein 2005, Ruff 2002). However, limited information exists that can be used to judge how effectively RTP actually induces DR, particularly in the context of restructured electricity markets. This report describes the second phase of a study of how large, non-residential customers' adapted to default-service day-ahead hourly pricing. The customers are located in upstate New York and served under Niagara Mohawk, A National Grid Company (NMPC)'s SC-3A rate class. The SC-3A tariff is a type of RTP that provides firm, day-ahead notice of hourly varying prices indexed to New York Independent System Operator (NYISO) day-ahead market prices. The study was funded by the California Energy Commission (CEC)'s PIER program through the Demand Response Research Center (DRRC). NMPC's is the first and longest-running default-service RTP tariff implemented in the context of retail competition. The mix of NMPC's large customers exposed to day-ahead hourly prices is roughly 30% industrial, 25% commercial and 45% institutional. They have faced periods of high prices during the study period (2000-2004), thereby providing an opportunity to assess their response to volatile hourly prices. The nature of the SC-3A default service attracted competitive retailers offering a wide array of pricing and hedging options, and customers could also participate in demand response programs implemented by NYISO. The first phase of this study examined SC-3A customers' satisfaction, hedging choices and price response through in-depth customer market research and a Constant Elasticity of Substitution (CES) demand model (Goldman et al. 2004). This second phase was undertaken to answer questions that remained unresolved and to quantify price response to a higher level of granularity. We accomplished these objectives with a second customer survey and interview effort, which resulted in a higher, 76% response rate, and the adoption of the more flexible Generalized Leontief (GL) demand model, which allows us to analyze customer response under a range of conditions (e.g. at different nominal prices) and to determine the distribution of individual customers' response.