Economists can't always conduct controlled experiments to test hypotheses about public policy, but sometimes experiments occur naturally. New York City electric rates provide an example. About 1.75 million New York apartments have metered electricity, paying an average of 21 cents per kilowatt hour, among the highest rates in the nation. At the same time, about 250,000 older apartments have unlimited, unmetered electric power included in the rent. A recent article by Sam Dolnick in The New York Times compares electricity use in the two types of apartments. The NYC experiment can be used to test three a priori hypotheses about the elasticity of demand for electricity.
First, demand might be completely inelastic. Although this hypothesis would not appeal to most economists, it reflects a line of argument often found in popular discussions. How many times have you heard someone say something like this: "Raising electric rates wouldn't help conservation. Rich people can afford to use as much as they want regardless of the price. And poor people would still need electricity to live, so raising the price would just make them poorer."
Second, demand might have constant elasticity. Because of their convenient mathematical form, constant elasticity curves are often used in empirical studies. Because they never intersect the axes, they imply that the quantity demanded would be unlimited if the price fell to zero.
Third, the demand curve might be linear. Textbooks are full of linear demand curves. Some people joke that is because lazy authors find them easier to draw. A linear demand curve implies that demand is finite even when the good is free.
Observations from the New York experiment shed light on these hypotheses. Demand is greater unmetered apartments. Some people do use electricity very wastefully, for example, running their air conditioners full blast while they are at work so that their cats will be comfortable. Still, demand does not increase without limit. Electric use is only about 30 percent higher in unmetered apartments than metered ones.
Taken together, these observations allow us to reject the hypotheses of perfect inelasticity and constant elasticity. Looks like the lazy old linear demand curve provides the best fit after all. And yes, skeptic, charging for electricity does provide an effective incentive for conservation.
Follow this link to download a free set of classroom-ready slides about metered vs. unmetered electricity in New York.