Explaining voting behavior in the Gothenburg congestion tax referendum

The Gothenburg congestion tax was introduced in 2013 and later subjected to a consultative referendum where the citizens, despite getting first-hand experience with the scheme, rejected it. This article explains voting behavior in the referendum using both self-expressed motives and five nested models to test various explanations suggested in previous research. Drawing on an extensive longitudinal study, we conclude first that although a majority voted against the tax in the referendum, attitudinal preferences have become more positive since its introduction – supporting previous findings and hypothesis of familiarity effects. Second, we present a model for voting behavior that explains significant portions of the variance, concluding that it is not the outcomes of the charges that are important, but rather if the charges are in line with basic values, if the uses of the revenues (in this case, infrastructure investments) are supported, and if the institutions and processes introducing the charges are perceived as legitimate, trustworthy, and responsive. The article ends with general policy recommendations on the basis of these findings.