Coherent estimations for residential photovoltaic uptake in Germany including spatial spillover effects

The share of solar energy in German electricity generation has increased strongly over recent years. This is largely due to guaranteed feed-in tariffs together with decreasing prices for solar panels. Residential PV systems play a decisive part providing households with a possibility to contribute to the Energiewende and benefit from the use of renewable energy. Their regional distribution varies distinctly across Germany implying different requirements in distribution grids as well as uneven utilization of national policy measures. Our paper focusses on the spatial diffusion of roof mounted PV systems and the underlying drivers in Germany. We extend previous findings not only by including additional explanatory variables but also by considering cross-regional spillover using spatial econometric models. Estimation results show that spatial dependence is a relevant determinant for explaining regional clusters of PV adoption. Recurrent visual perception or peer-effects might explain spatial autocorrelation as potential adopters follow decisions by actors in the proximity. Another reason for spatial dependence might be a concentration of craft skills or solar initiatives, which leads to an accelerated diffusion in a region and its surroundings. Whereas the first explanation corresponds to the specification of a spatial lag model, the latter is in line with a spatial error specification. However, our results indicate that although spatial lag is present, spatial dependence in the residuals has higher explanatory power. Hence, we suppose that spatial spillover is not mainly driven by social imitation but by unobserved regional characteristics. Notably, high values for solar radiation, the share of detached houses, electricity demand and inverse population density of a region favour the PV uptake.