Social Norms and Energy Conservation

This paper evaluates a series of programs, run by a company called OPOWER, to send Home Energy Report letters to residential utility customers comparing their energy use to that of their neighbors. Using data from randomized natural …eld experiments at 600,000 treatment and control households across the United States, I estimate that the programs reduce energy consumption by 1.1 to 2.8 percent relative to baseline. The program provides additional evidence that non-price interventions can substantially and cost e¤ectively change consumer behavior: the e¤ect is equivalent to that of a short-run electricity price increase of 17 to 28 percent, and the cost e¤ectiveness compares favorably to that of traditional energy conservation programs. I show that there is substantial treatment e¤ect heterogeneity and implement an optimal variable selection algorithm to illustrate how "pro…ling," or using a statistical decision rule to target the program at households with higher expected Conditional Average Treatment E¤ects, could further improve cost e¤ectiveness by 43 percent. JEL Codes: C44, D12, L94, Q41. Keywords: Social norms, energy demand, randomized …eld experiments, pro…ling. — — — — — — — — — — — — — — — — — — — — — — — — — — — — I thank, without implicating, Ian Ayres, Bob Cialdini, Tyler Curtis, Rajeev Dehejia, Kenneth Gillingham, Larry Goulder, Michael Greenstone, Matt Harding, Kosuke Imai, Seema Jayachandran, Karthik Kalyanaraman, Alex Kaufman, Ogi Kavazovic, Alex Laskey, Aprajit Mahajan, Sendhil Mullainathan, Dave Rapson, Todd Rogers, Eldar Sha…r, Joe Shapiro, Lan Shi, Marc Solomon, Dmitry Taubinsky, and seminar participants at Harvard, Stanford, and the Environmental Defense Fund for helpful conversations related to this project. This is a revision of the original July 2009 working paper (Allcott 2009).