Pilot projects are demonstrating the potential of smart metering and smart rates to make the most of supply and demand resources. But as empirical studies show, not all pricing designs are equally...
The changing architecture of demand response in America.
the ratio of peak to off-peak rates is set at 8:1, it would yield a low-end response of 13 percent and a high-end response of 21 percent.
Faced with this variation, the analyst risks making a career-limiting mistake in his or her cost-benefit analysis. This can be exemplified by reviewing the situation in California in the summer of 2002. Seeking to prevent another energy crisis, the California Public Utilities Commission initiated proceedings on DR, advanced metering, and dynamic pricing. Early on, it became clear that the decision to deploy advanced metering was very risky without further information specific to the deployment area. A preliminary cost-benefit analysis using price elasticities from the literature on time-of-use pricing (which ranged from -0.10 to -0.30) carried out for an investor-owned utility showed that such deployment would provide gross benefits ranging from $561 million to $2.6 billion. The cost of AMI, net of operational benefits, was estimated to be $1.08 billion. This suggested that the net benefits would range from a loss of $519 million to a gain of $1,557 million. In other words, the range of net benefits, at some $2 billion, was very wide. 16 A decision then was made to carry out what later became known as the statewide pricing pilot.
Second, the analyst might find himself or herself in the fortunate situation of having a neighboring organization that has done a similar pilot. The analyst simply could borrow the neighbor’s results. However, this approach also entails several risks as the geographic proximity doesn’t necessarily guarantee similar customer characteristics, weather conditions, and service territory characteristics such as rural vs. urban establishments. These factors and many others easily might be different from one utility to the other and influence customers’ price responsiveness and ultimately the net benefits of the AMI deployment.
Third, the analyst could join with analysts at other utilities and seek to develop a national model that explains the inter-pilot variation. Today, while one can theorize on what causes the intrinsic variation in elasticities across pilots, there’s no way to prove or disprove the theories without doing some careful empirical work. Further research that would pool the data across the pilots probably would explain the variation and lead to a meta-analysis model with regional variation that could be used by all those utilities that haven’t yet conducted their own pilots. The cost of developing such a meta-model, while running into six figures, still would cost substantially less than every utility doing its own pilot. Such a project successfully was carried out with data from five time-of-use pricing experiments in the early 1980s by the Electric Power Research Institute (EPRI). Five were carefully chosen from a total of 14 pricing experiments carried out by the Federal Energy Administration, the precursor to the DOE. Two projects were located in California, one in Connecticut, one in North Carolina and one in Wisconsin. The overall project found that once the model accounted for differences in customer appliances and regional weather conditions, an underlying model of customer behavior emerged that could be applied successfully across the country. Again,