A Länder-based Forecast of the 2017 German Bundestag Election

When elections are distant, polls are poor predictors. Too few voters are paying attention and too much can change before election day. Structural models can establish baseline expectations but suffer from high uncertainty and underspecification imposed by small samples. We present an early forecast of the 2017 Bundestag election results for individual parties that leverages economic and political data as well as state parliament (Landtag) election results in the German states (Länder) to sidestep these shortcomings. A linear random effectst model provides our estimates. Länder elections are dispersed over the calendar and offer the advantage of capturing both actual voter preferences and new political issues. We argue that this approach offers a promising method for early forecasts when polls are not informative.