Do energy retrofits work? Evidence from commercial and residential buildings in Phoenix

There is an ongoing debate whether energy efficiency retrofits save energy in practice. This paper quantifies the energy savings of the Energize Phoenix program in Phoenix, Arizona. Impacts of retrofits are analyzed using pre-post treatment billing data from January 2008 to April 2013, covering 201 residential buildings and 636 commercial buildings. There are five types of retrofits for residential buildings and six types for commercial buildings. Empirical results from fixed effects panel regression models indicate monthly energy savings of 12% for commercial buildings and 8% for residential buildings. The realized energy savings are 30–50% lower than those predicted by engineering models. We have also separated the energy savings provided by individual retrofits and retrofit bundles, and estimate how the effectiveness of retrofits varies based on building attributes, all of which should be considered when making energy efficiency decisions. Learning effects are also observed in this program, which means as the post-retrofit time increases, there is a possibility for improved learning and control of energy use.