UC Berkeley

This dissertation analyzes the energy consumption behavior of residential adopters of solar photovoltaic systems (solar-PV). Based on large data sets from the San Diego region that have been assembled or otherwise acquired by the author, the dissertation quantifies changes in energy consumption after solar-PV installation and determines whether certain household characteristics are correlated with such changes. In doing so, it seeks to answer two related questions: First, "Do residential solar adopters increase or decrease their electricity consumption after they install a solar-PV system?" Assuming that certain categories of residential adopters increase and others decrease, the second question is "Which residential adopters increase and which decrease their consumption and why?"

The database that was used to conduct this analysis includes information about 5,243 residential systems in San Diego Gas & Electric's (SDG&E) service territory installed between January 2007 and December 2010. San Diego is a national leader in the installation of small-scale solar-electric systems, with over 12,000 systems in the region installed as of January 2012, or around 14% of the total number installed in California. The author performed detailed characterization of a significant subset of the solar installations in the San Diego region. Assembled data included technical and economic characteristics of the systems themselves; the solar companies that sold and installed them; individual customer electric utility billing data; metered PV production data for a subgroup of these solar systems; and data about the properties where the systems are located.

Primarily, the author was able to conduct an electricity consumption analysis at the individual household level for 2,410 PV systems installed in SDG&E service territory between January 2007 and December 2010. This analysis was designed to detect changes in electricity consumption from the pre-solar to the post-installation period. To the extent increases are present for some solar adopters, the analysis seeks to determine whether there is a "solar rebound" effect analogous to the "rebound" or "take-back" effect that has been observed and studied within the energy efficiency literature. Similarly, to the extent that electric users may decrease overall consumption after installation of a solar system, the study seeks to explore the possibility that solar adoption is part of a continued effort towards clean energy practices more generally, such as energy efficiency and conservation. In this way, the study seeks to determine whether there is a synergistic effect between solar and decreased consumption, for solar adopters generally or for some subsets therein.

The assembled data allowed testing of various hypotheses that could help explain observed changes in consumption in different households. One variable that was carefully examined was the sizing of the solar system. As part of the study, analysis of 4,355 systems was conducted to determine how each residential solar system was sized with respect to pre-installation energy consumption. Other potentially interesting or explanatory variables for which information was available include total and net costs of the solar system; age of the home; the climate zone (inland or coastal) where the home is located; the home's pre-installation energy consumption; home characteristics such as assessed value and square footage; and the identity of the solar installation contractor.

Aside from extending the literature on the rebound effect to the context of home-based energy generation, this study links to the innovation diffusion literature by focusing on solar "innovators" to understand more about the characteristics that may drive behavior, or conditions under which they also adopt clean energy technologies and practices. The results have clear policy relevance with regard to the development and coordination of policies to promote integration of solar and energy efficiency. Currently several public policies are being developed at various levels of government to encourage both, based on application of the economically rational concept of the "loading order", the California policy that places energy efficiency as the state's highest priority energy resource. However, there has been little study of the interrelationships between them or how these innovations are implemented in practice. This dissertation begins to fill that gap.