Deriving a forecast model for European election turnout

Turnout is a key indicator in European Parliament elections in the absence of a direct executive outcome. Forecasting turnout is an important exercise in requiring the identification of a parsimonious model with good lead time from the array of structural, demographic, and attitudinal variables employed in rich explanatory models of turnout, and simultaneously minimizing prediction error. Building on a series of regression models using aggregate data, this article explores the applicability of such an approach to turnout in the EU-27 countries and considers the explanatory added value that deriving such a forecast model can also provide.