Peer Effects and Learning by Doing in the Diffusion of Solar

The solar photovoltaic (PV) industry in the United States has been the recipient of substantial production subsidies at the federal and state level, often motivated both by environmental externalities and dynamic spillovers from learning-by-doing in the diffusion of the new technology. Understanding the nature of this diffusion and documenting the empirical extent of these spillovers are crucial for assessing the benefits of solar production subsidies. Using a rich dataset on solar PV installations in California, we find evidence of peer effects that depend on the installed base at the zip-code level, suggesting the presence of social learning or snob effects for consumers. We then estimate a structural model of supply, using the zip-code level installed base as a demand shifter to instrument for the quantity being installed at any given time. We find evidence of highly non-appropriable contractor LBD at the regional level. These results suggest that there is an economic basis for a production subsidy on solar in California, even if the current subsidies may still be overly generous. ∗The authors would like to thank the participants in the Stanford Economics Department Environmental Reading Group, the Stanford MS&E Policy and Economics Research Roundtable, the UC-Berkeley ARE seminar series, and faculty and students from the Stanford Graduate School of Business for their valuable comments and insights. All remaining errors are solely the responsibility of the authors.