Abstract:
We study variation of peer effects in rooftop photovoltaic adoption by households. Our investigation employs geocoded data on all potential adopters and on all grid-connected photovoltaic systems set up in Germany through 2010. We construct an individual measure of peer effects for each potential adopter. For identification, we exploit exogenous variation in two dimensions of photovoltaic system roof appropriateness of neighbors: their inclination and their orientation. Using discrete choice models with panel data, we find evidence for causal peer effects. However, the impact of one previously installed PV system on current adoption decreases over time. We also show that visible PV systems cause an increase in the odds of installing which is up to three times higher in comparison to all PV systems. At rural locations visibility may be less important, which indicates that word-of-mouth communication plays a stronger role.