Expert Commentary

Peer effects in the diffusion of solar photovoltaic panels

2012 study by New York University and Yale University on possible peer-influence effects in the installion of solar panels within different ZIP codes.

A home solar panel system is a consumer product like few others. Because the alternative choice — electricity from the grid — is almost universally available, relatively inexpensive and has no up-front cost, potential buyers of solar systems require other incentives. These can range from government rebates to credit for energy sold back to the grid to the warm feeling one gets for doing a good deed — in this case, for the environment.

A 2012 study by New York University’s Stern School of Business and the Yale School of Forestry and Environmental Studies indicates that another factor may be at work as well: peer effects, or the influence that others can have on our behavior, even without direct communication. The study, “Peer Effects in the Diffusion of Solar Photovoltaic Panels,” was based on analysis of a dataset of 79,000 residential solar system installations in California between January 2001 and August 2011.

The data show interesting patterns, as the authors relate: “More densely populated ZIP codes tend to have more installations, yet there are densely populated ZIP codes with few installations, and less densely populated ones with many installations,” indicating that a peer effect could be at work.

Results of the study include:

  • At the ZIP code level, for every 1% increase in the number of installations, there is a 1.15% increase in the average daily solar-panel adoption rate.
  • At the street level, a 1% increase in installations leads to a 9.28% increase in the adoption rate — eight times the effect of a ZIP code-level installation effect.
  • No evidence was found that the increase in adoption rates was due to the particular solar system company being used — say, one contractor working a particular area. “This suggests that if the peer effect has an information component, it does not in itself lead to more installations by the same contractor.”
  • High income and home value were not strong predictors of solar panel installation. “ZIP codes with a higher median income, higher proportion of people aged 20-45 and more than 65, as well as ZIP codes with higher valued homes exhibit lower adoption rates conditional on the other covariates.”
  • Significantly higher adoption rate of solar panels took place in ZIP codes with greater populations and above-average percentages of male, white, college educated, hybrid-vehicle owners.
  • The impact of “peer effects” within like-minded or homogeneous communities increased the adoption of solar panel technology by 10.5%. For heterogeneous or more mixed communities, peer effects accounted for an increase in overall uptake of 14.3%.

“Geographic clustering appears to occur at both a ZIP code and neighborhood level,” the authors conclude, “and does not simply match the population density or the ‘greenness’ of the ZIP code.” This has implications for the marketing of solar-power systems as well as for the structure of government programs that promote them.

Keywords: technology, renewable energy, California

About The Author