| Households' decisions to purchase solar photovoltaic systems are characterized by large upfront costs, differentiated products, and uncertainties about future government subsidies. This study analyzes the interplay of these factors using a dynamic discrete choice model. I use a newly assembled dataset, that covers all installations from 2002 through 2006 in California at the household level, to estimate demand for solar panel installations. I find that across the distribution of housing values, households are price elastic with respect to both temporary and permanent changes in price. Also, I find that elasticities vary across the distribution of housing values. The marginal effect of technological innovation is significant and positive with respect to the probability of purchase. I find that a 1% increase in the efficiency rate increases the probability of purchase by 6.4%. This is result is compounded by the fact that efficiency rates increase 30% over the sample period. Through counterfactual simulations, I show that in the absence of government subsidies 49.5% of all purchases would not have occurred. Additionally, over 70% of the total reduction in market capacity when subsidies are removed is directly attributable to larger capacity installations. Lastly, I find no evidence that household behavior is affected by the uncertainty associated with future subsidy regimes. |
