Price Elasticity Estimates Under Homothetic Separability

This Note shows that the methodology used in many analyses of the peak-load pricing experiments is seriously flawed. The reported results predict considerably larger response to time-of-day prices than can be expected from a correctly specified model. The reasons for this lie in the use of utility functions in which subsets of goods are assumed to be separable from one another, and in the assumption that the utility functions for each group are "homothetic," or homogenous. The results of two analyses of time-of-day experiments are replicated, using a set of random numbers instead of the actual experimental data. The resulting "price elasticity estimates" are not significantly different from the ones published in the literature. Thus it is concluded that price elasticities estimated from time-of-day experimental data with demand curves from a homothetically separable utility function are mostly assumption and are only insignificantly influenced by the data.