Where solar goes: political context and the composition of rooftop deployment

Radical right parties oppose climate legislation, and where they are strong that opposition is legible in votes and in national policy output. Whether it reaches the point at which climate policy becomes physical, in the form of a panel on a roof or a turbine in a field, is a separate question, and one the electoral record cannot answer. Decentralized deployment is the product of many private decisions taken inside local jurisdictions, and political context could bear on those decisions through authorities with something to withhold, through owners who do not want what is on offer, or through the absence of the actors, capital and competence a project has to assemble. These channels carry different implications and are routinely conflated, because the outcome they produce is the same one. This paper separates them by changing the question. Asking whether conservative places deploy less is hard to identify and uninformative about mechanism: a jurisdiction's total deployment reflects its building stock, its electricity prices, its subsidies and its politics jointly. By contrast, the composition of deployment varies within a place: which buildings receive installations, where in the jurisdiction they sit, and at what scale. It can therefore be compared while holding the place constant, and the candidate channels imply different patterns across those comparisons. Switzerland allows that design to be built. Municipalities are divided into land-use zones by a harmonized federal typology, a national cadastre estimates solar potential for ten million individual roof surfaces, and rooftop installations meeting stated technical conditions require notification rather than a building permit, which fixes one channel by law rather than by assumption. Installed capacity in each municipality-by-zone cell is measured against the roof potential available at the relevant scale, with municipality fixed effects absorbing everything constant within a place, so that what is identified is a differential across zones. Three size classes are estimated jointly with matched denominators, under two outcome measures. Deployment on industrial-zone buildings is markedly lower where support for the radical right is higher, at small and middle scale, and is not lower at large scale; the difference between the two is the paper's central quantity, and it remains distinguishable from zero in thirty-three of thirty-seven specifications. The shape is a step rather than a gradient, with the transition falling between 50 and 150 kilowatts. The six-category zoning variable carries its information in one cut, industrial against everything else, rather than in a graded typology. Neither municipal gatekeeping nor a general disposition against environmental policy survives. What remains are two accounts that imply the same distribution, and a case design capable of separating them.