Environmental Preferences and Peer Effects in the Diffusion of Solar - - PowerPoint PPT Presentation

Introduction Data Results Conclusions

Environmental Preferences and Peer Effects in the Diffusion of Solar Photovoltaic Panels

Bryan Bollinger and Kenneth Gillingham June 3, 2011

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Motivation

Policies to promote solar photovoltaic (PV) adoption have been gaining momentum throughout the world, as concerns

• ver global climate change and national security externalities

continue to grow.

In the US: 30% solar energy federal investment tax credit. In California: $3.3 billion 10 year California Solar Initiative (CSI).

Yet we understand very little about the process of diffusion of this technology.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Clustering of Installations: 2001-2003

! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! !! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! !! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

zip code density (pop/mi^2)

0 - 500 500 - 1500 1500 - 3000 3000 - 5000 5000 - 7500 7500 - 10000 10000 - 15000 15000 - 20000 20000 - 30000 30000 - 60000

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Clustering of Installations: 2001-2006

& & & & & & & & & & & & & & & & & & & && & & & & & & & && & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & && & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & && & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & && & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & && & && & & & & & & & & & & & && & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & && & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & && & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & & && & & & & & & && & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & && & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & && & & & & & & & & $ $ $ $ $ $ $ $ $ $$ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $$$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # ### # # # # # # # # # # # # # # # # ## # # # # # # # # # # ### # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # ## # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ### # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ### # # # ## # # ## # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # ## # # # # # # # # # # # # # # # # # # ## # # ## # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # ## # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # ## # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! !! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! !! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !! ! ! !! ! !! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

zip code density (pop/mi^2)

0 - 500 500 - 1500 1500 - 3000 3000 - 5000 5000 - 7500 7500 - 10000 10000 - 15000 15000 - 20000 20000 - 30000 30000 - 60000

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Research Questions

1 Why do we see this clustering in the diffusion of solar?

Environmental preferences? Localized marketing? Social Norms/Peer effects?

2 What factors enhance this clustering effect? Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Maybe it is from a Combination of Factors

Seeking Solar Champions!

Would you like to become a Solar Champion in your SF neighborhood? The San Francisco Department of the Environment and PG&E are offering free solar training for SF citizens interested in helping spread the word about solar in their neighborhoods. What: Neighborhood Solar Champions Training Course When: Check back in 2010 for future class offerings. Classes are held

several times per year on select Saturdays from 9:30-4:00. This includes the 10:00-12:00 Solar Power Basics for Residential Customers class.

Where: PG&E’s Pacific Energy Center, 851 Howard Street San Francisco, CA 94103 Cost: Free! Plus, lunch is provided for participating SF residents Requirements:

• Interest in promoting solar energy in your SF neighborhood
• Commit to giving at least two presentations in your neighborhood a year
• Comfortable with public speaking
• Ability to participate in one full day of training
• No prior solar technical experience necessary (though helpful)

For more information, please contact: Jade Juhl SF Department of the Environment Email: jade.juhl@sfgov.org Tel: (415) 355-3780 Lisa Shell PG&E Email: L1SB@pge.com Tel: (415) 973-0305 We look forward to working with you as a Neighborhood Solar Champion!

Neighborhood Solar Champions

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Possible Identification Concerns

Reasons why it is difficult to identify peer effects: Simultaneity - Manski’s “reflection” problem, where one affects their peers just as the peers affect them. Endogenous Group Formation - Self-selection into a group. Correlated Unobservables - Unobserved variables correlated with zip-code installed base that affect rate of adoption.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Motivation Clustering Questions

Methodological Framework

Our exploration into the determinants of the diffusion of solar PV panels focuses on measuring the effect of previous adoptions on adoptions today. Quasi-experimental evidence using within-zip differences in solar incentives. Hazard model of zip code-level adoption to quantify clustering effects consistent with peer effects. Evidence for contractor-specific clustering effects. Determinants of adoption. Model of adoption at the street-level providing further evidence.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Installation Data Solar Market

Data

We have installation-level data from 2001-2009 for the three investor-owned utility regions (roughly 90% of the market) containing 34,110 residential installations: Type of installation (only CPUC data) Transaction price of the installation (pre- & post-incentive) Size of installation PV installer (contractor) PV manufacturer Zip code of installation Date consumer received solar incentives Date of project completion

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Installation Data Solar Market

What has been Happening: New Installations

4000 6000 8000 10000 12000 14000

Installations

total residential 2000 4000 6000 8000 10000 12000 14000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Installations

total residential

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Installation Data Solar Market

What has been Happening: Solar PV Prices

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Installation Data Solar Market

Time Between Adoptions in a Zip Code, 2001-2009

Fraction 200 400 600 800 1000 time since last residential installation in zip

(a) Time

Fraction 2 4 6 8 log time since last residential installation in zip

(b) Log Time

Figure: Time and log time between installations within a zip code

Average time between adoptions: 64 days, Median: 21 days.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

California’s Utility Regions Administer the CSI Individually

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Basic Idea of Quasi-experiment

We exploit the changing incentive amounts within zip codes across the utility district boundaries.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Quasi-Experimental Evidence

Dependent variable: log(time between adoptions) PG&E and SCE border SCE and SDG&E border second region

• 0.925

2.044 (0.707) (0.616) after first region changes step

• 0.009
• 0.027

(0.592) (0.645) after both regions change step 0.000 1.126 (0.000) (0.641) second region x after first region changes step

• 2.475
• 1.061

(0.605) (0.705) second region x after both regions change step

• 0.968
• 1.817

(0.541) (0.589) Zip Code Controls Y Y Month Controls Y Y R-squared 0.603 0.376 N 72 213 Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Hazard Model

We assume the adoption of solar PV systems follows a non-homogenous Poisson process (i.e., a Cox or doubly-stochastic Poisson process) with the following hazard rate: λz(t) = λz(0) exp(Xztβ + f (z, t, ǫzt), (1)

Xzt includes the market and time varying explanatory variables. f (z, t, ǫzt)) includes market and time control variables and a stochastic term capturing the unobserved heterogeneity.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Estimation Equation

The waiting time between events is exponentially distributed. We can rewrite the expression for log time between adoptions as: log(∆tzt) = Xztβ + ηzt + ǫzt. (2) ηzt are zip, time, or time-zip fixed effects.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Zip Code Hazard Model Estimation Results

Dependent variable: log(time between adoptions) (1) (2) (3) (4) log zip installed base

• 0.509
• 0.955
• 0.953
• 0.948

(0.019) (0.036) (0.037) (0.037) SCE x log zip code installed base

• 0.013
• 0.012

(0.031) (0.031) CCSE x log zip code installed base

• 0.015
• 0.020

(0.038) (0.038) previous installation price

• 0.001

(0.007) Zip FE (ηz ) Y Y Y Y Month Controls (ηt) N Y Y Y Incentive Step Controls Y Y Y Y R-squared 0.151 0.147 0.152 0.153 N 31,613 29,172 29,172 28,404 Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Contractor-Specific Results

Dependent variable: log(time between adoptions) (1) log contractor zip installed base

• 1.213

(0.059) log competitors zip installed base 0.127 (0.038) Zip-Contractor FE Y Month Controls Y Incentive Step Controls Y R-squared 0.132 N 18,039 Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

The Effect of Demographics

We interact demographics with zip-code installed base to find: Increases Peer Effects: Median home value. Decreases Peer Effects: % driving hybrids and % pop white.

At least 10% significance for each. We include zip code FE in these specifications.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

The Effect of Demographics on Adoption

We regress the zip code fixed effects on demographics to see how adoption is influenced by demographic factors: Increases Adoption: % car registrations that are hybrids, population, household size, % population who are white, % population with college degrees, % population older than 65. Decreases Adoption: % population who are male, % population who carpool.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions Quasi-Experiment Model of Diffusion Zip Code Analysis Results Street-Level Analysis

Street-Level Analysis

Use street-year-month panel of 2001-2007 CEC address data. Use a linear probability model. Dependent variable is a dummy for whether there is an installation on the street in a given month. Find that both whether there has been an installation on the street in a previous month and the cumulative number of installations in the zip code influence the probability there will be an adoption in a street in a given month.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Main Findings

There is clear clustering behavior in the diffusion of solar PV. This localized clustering appears to stem from several factors:

Environmental preferences. Peer effects. Marketing taking advantage of peer effects.

Peer effects appear to be economically significant: a 1% increase in installations in a zip code leads to a roughly 1% decrease in the time to the next adoption. Greater environmental preferences decrease peer effects.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Thank you for your time! Comments and suggestions are welcome.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Related Literature

Kahn & Vaughn (2009) - Clustering of hybrids and LEED buildings. Just a small sample of the literature on peer effects: Education - Hoxby (2000), Sacerdote (2001), Cipollone & Pellizzari (2007), Duflo et al. (2008), Ammermueller & Pischke (2009), De Giorgi et al. (2010). Criminal activity - Glaeser et al. (1996), Bayer et al. (2009). Retirement plans - Duflo & Saez (2003). Welfare participation - Bertrand et al. (2000). Agricultural technology adoption - Foster & Rosenzweig (1995), Conley & Udry (2010).

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Histogram of Residential Prices ($/W)

.1 .2 .3 .4 Density 5 10 15 20 price ($/W) Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Histogram of Residential System Sizes (kW)

.1 .2 .3 Density 5 10 15 20 size (kW) Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Residential Installations in a Zip Code, 2001-2009

.01 .02 .03 .04 Density 100 200 300 400 Number of installations in zip code Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Key Installation Data

In our examination of peer effects, we are only looking at the 34,110 residential installations which run to completion (46,331 initiated).

Table: Residential

Variable Mean

• Std. Dev.

Min. Max. size (kW) 4.873 2.752 0.119 48.3 price ($/W) 8.395 1.451 3.018 14.997 N 36,112

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Demographic Variables

We add zip-code level demographic variables from Sourcebook America and American Factfinder:

Population, Household size, Median income, % pop male, % pop white, % pop college degrees % pop 20-45, % pop 65+, % pop who drive to work, % pop who carpool % pop using public transit, % pop who work at home or walk to work, % pop with over a 30 min commute Number of owner occupied homes (1000s), Median value owner occupied home (millions) Home loan (spending index), Home repair (spending index)

Finally, we have hybrid registrations by zip code from R.L. Polk.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Robustness Check: Quarter-Zip FE

Dependent variable: log(time between adoptions) (1) (2) (3) log zip installed base

• 1.375
• 1.331
• 1.334

(0.173) (0.178) (0.180) SCE x log zip code installed base

• 0.198
• 0.201

(0.171) (0.184) CCSE x log zip code installed base

• 0.051
• 0.051

(0.188) (0.197) previous installation price

• 0.000

(0.007) Zip-quarter FE Y Y Y Incentive Step Controls Y Y Y R-squared 0.072 0.075 0.076 N 18,936 18,936 18,602 Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Address Data for the CEC Subsample

For the CEC subsample (2001-2007), we have the actual addresses at which the installations occurred. We create a panel where each

• bservation is a street-month, where streets are divided at each zip

code border.

Table: Summary statistics for street-level analysis

previous installations no yes Total no new installation 1,326,465 52,990 1,379,455 new installation 17,644 3,018 20,662 Total 1,344,109 56,008 1,400,117

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Number of Installations On a Street

.2 .4 .6 .8 Fraction 2 4 6 8 10 number of installations per street Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: Model

Use street-year-month panel of 2001-2007 CEC address data. Use a linear probability model. Dependent variable is a dummy for whether there is an installation on the street in a given month. Key independent variables are whether there has been an installation on the street in a previous month, and the cumulative number of installations in the zip code.

Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV

Introduction Data Results Conclusions

Appendix: LPM Street-Level Results

Dependent variable: Adoption occurs on that street (1) (2) (3) (4) previous installation 0.037 0.037 0.037 0.159 (0.001) (0.001) (0.001) (0.033) zip code installed base (100s) 0.019 3.397 4.557 (0.001) (0.024) (0.861) zip code installed base squared

• 0.009
• 1.094
• 0.827

(0.000) (0.013) (0.237) Zip FE (ηz ) Y Y N N Month Controls Y Y N N Zip-Month FE N N Y Y Street Controls N N N Y R-squared 0.012 0.012 0.112 0.728 N 1,400,117 1,400,117 1,400,117 4,345 Bryan Bollinger and Kenneth Gillingham Peer Effects in the Diffusion of Solar PV