In place of policies that directly price energy, such as a carbon tax, the U.S. and many countries around the world often rely on subsidy programs that encourage consumers to either use cleaner or less energy. But do these subsidies really change people’s behaviors? A new study of a program designed to save energy finds the success of such programs depends greatly on two factors: climate and income. The program was found to have the most success when aimed at low-income households and households who use high amounts of electricity for air conditioning. The program cost just 2.5 cents for every kilowatt-hour reduced for these households. Meanwhile, the program cost significantly more for those households of a higher income bracket and those living in a more moderate climate – almost a dollar per kilowatt-hour reduced.
“While these subsidy programs are politically more appealing than other policies, we spend a lot of money motivating customers to conserve energy and almost no one knows if it’s actually working,” says Koichiro Ito, the author of the study and an assistant professor at the University of Chicago’s Harris School of Public Policy. “At the end of the day, it’s the customers who will pay for these programs in increased electricity prices. So it’s important that we structure these programs right by targeting them at the households who will reap the greatest rewards: low-income households and households in areas with a warm climate.”
The study, which appears in this month’s issue of the American Economic Journal: Economic Policy, looked at California’s 20/20 program, one of the largest programs of its kind in the U.S. The program offered a 20 percent rebate to households that reduced their summer 2005 electricity consumption by 20 percent from what they had used the previous summer. The study found that California’s inland customers, who are generally of a lower income bracket and experience a warmer climate, reduced their energy use by about 5 percent. But there was almost zero impact on coastal customers who are of a higher income bracket and experience a more moderate summer climate. The high cost to extend the program to coastal areas raised the overall cost of the program to 17.5 cents per kilowatt-hour reduced and $381 per ton of carbon dioxide reduced – significantly larger than the $38 per ton that the federal government estimates as the social cost of carbon.
“We can learn from this that people who have high electricity bills from a heavy reliance on air conditioning can more easily change their habits simply by using the air conditioner less,” Ito says. “These same people living in inland California also have a greater motivation to reduce their energy use because they are low income and need the energy rebate more than their coastal counterparts.”
Ito was able to provide a compelling evaluation because he used a rigorous evaluation method so-called a “regression discontinuity design.” In California’s 20/20 rebate program, households were eligible simply if they had been enrolled in the service for 365 days, using a specified cut-off date. All eligible households were then automatically enrolled. Because of these conditions, Ito was able to use this strict eligibility cutoff to compare the energy consumption of households who just barely met the cutoff with those who just barely missed it, creating a treatment and control group that was similar to each other in every other way.
“These conditions made for a strong experiment by which to conduct the analysis,” Ito says. “And so, the primary results – first, that the overall costs of reducing carbon under these subsidy programs’ can be much higher than previously thought, and second, that climate and income matter when considering the cost-effectiveness of the programs – are useful when thinking about the design of similar programs elsewhere.”
For a copy of the study, or to talk with the study’s author, please contact Vicki Ekstrom High at vekstrom@uchicago.edu.