Diffusion of Residential Solar Power Systems : A Dynamic Discrete Choice Approach

Solar electricity generation is a strategic component in the energy portfolio of many countries to replace fossil fuels and reduce green-house gas emissions. Installations of solar photovoltaic (PV) systems by residential homes in the United States have grown rapidly in the past decade, driven by decreasing costs of PV systems and government incentives. However, PV systems currently still produce only a small fraction of the total electricity generated in the U.S. So, it is important to study how federal and local incentives can be made more effective to achieve the goal of widespread residential adoption of PV systems. We study the diffusion of PV systems by modeling the adoption decisions of individual households who are forward-looking and consider installing PV systems as an investment with its future economic benefits. Moreover, the space-time dynamics of installations reveal a clear pattern that a household’s adoption decision is strongly influenced by the adopters in their neighborhood. We propose a dynamic discrete choice model in which a household considers the inter-temporal trade-off between installing the system now and waiting by analyzing the current and future return on investment and peer effects. Our model incorporates unobserved household heterogeneity in addition to segmenting households by their home values and geographic locations. We collect a novel data set with detailed information on PV adoptions in Austin, Texas to estimate the model parameters. For inference, we use a Bayesian method to overcome the computational burden. Our estimation reveals a strong peer effect in a household’s adoption decision. Our structural model permits counterfactual analysis to generate insights to design incentive programs that are effective in accelerating the PV diffusion process. As an illustration, counterfactual policy simulations for local rebate programs demonstrate how policy-makers can achieve higher adoption rates with a lower budget