How long does it take to build a nuclear power plant? A non-parametric event history approach with P-splines

Abstract

Governments deciding to use nuclear energy as part of their country׳s energy mix are faced with long-term planning efforts and huge investments. As nuclear power plants constitute one of the socially and politically most contested technologies, the question arises, which time horizons companies as well as politicians have to consider for the accomplishment and grid-connection of individual and whole fleets of reactors. Unfortunately, there are no large-N studies investigating the time for completion of such large-scale projects. For the first time, we statistically explain the duration of the construction of all initiated nuclear plant projects so far. Based on the International Atomic Energy׳s comprehensive Power Reactor Information System (PRIS) we assess the impact of demographic, economic, and political preconditions of a country, at the same time accounting for different types of reactor technologies. To account for non-linear relationships, we apply non-parametric survival models with P-splines. A main result of our analysis is that time of connection to grid increases over the years indicating increased societal sensibilities, respect for higher security standards, and increased project complexities. The Harrisburg and the Chernobyl disaster did not induce a separate additional delaying effect.