Retrofit and New-Build Installations of Residential Solar PV: Fukushima as a Natural Experiment

This study investigates how adoption of residential solar PV systems responds to changes in the economic value of adoption caused particularly by electricity price fluctuations. We shed light on the importance of a house(hold) characteristic that has been largely overlooked in the literature: the distinction between retrofit installations and new-build installations. To identify the effect of electricity prices on solar PV adoption, we regard the 2011 Fukushima nuclear accident and subsequent shutdown of nuclear power plants in Japan as a natural experiment that resulted in large (temporal and regional) exogenous variations in electricity
generation costs. Using Japanese data for 2009–2014, we find a downward bias of 40–50% in the estimated effect of electricity prices if they are not instrumented with the cost variations. Our estimation shows that retrofit installations are 2–2.5 times more responsive to changes in electricity prices than new-build installations. An important policy implication of these contrasting responses is that financial incentive programs such as feed-in tariffs can be more cost-effective if they target more at owners of existing homes than at owners of new-build homes because retrofitting decisions are more likely to be induced by such incentives.