To What Electricity Price Do Consumers Respond? Residential Demand Elasticity Under Increasing-Block Pricing

It is straightforward to evaluate how a perfectly-optimizing, perfectly-informed customer will respond to a non-linear price schedule, but such a customer is rare. In the common case of increasing-block pricing of water and electricity, consumers do not know what marginal price they face during a billing period, because they do not know what demand shocks will occur during the period. If consumers instead set optimal behavioral rules based on the distribution of possible marginal prices they will face (e.g., turn off lights in unused rooms, set the A/C to 74 degrees, replace incandescent lights with CFLs), their consumption will not exhibit discrete responses to the discrete jumps in the price schedule. Using data from a large electricity utility, I show that the empirical distribution of consumption quantities is not consistent with consumers accurately knowing and responding to the marginal price they will face. I then estimate the price elasticity of demand using a panel of household observations at two-year intervals, identifying elasticity from the changes in the increasing-block price schedule. The results suggest that most consumers are probably responding to the expected marginal price or even less precise information about what marginal price they will face. The results are difficult to reconcile with the common approach of estimating demand elasticity as a function of the ex post marginal price that the consumer faces along an increasing-block price structure. 1 Director, University of California Energy Institute (www.ucei.org); E.T. Grether Professor of Business Economics and Public Policy, Haas School of Business, University of California, Berkeley (faculty.haas.berkeley.edu/borenste); and Research Associate of the National Bureau of Economic Research (www.nber.org). Email: [email protected]. I am grateful to Koichiro Ito for excellent research assistance and very helpful comments, and to Lucas Davis, Joe Farrell, Stephen Holland, Matt Kahn, Sheila Olmstead and seminar participants at MIT, UC Santa Cruz, the UC Energy Institute, Yale and Berkeley-Stanford IOfest ’08 for very helpful comments and discussions.