The puzzle of the AfD’s spatial voting pattern: Mechanisms linking local context and individual populist radical right vote in Germany

Between 2013 and 2017, Germany experienced a considerable increase in populist radical right voting. However, this increase was not uniform across German regions. While the gap in the party’s election result becomes most visible between the former FDG and West Germany, there are differences within the Eastern and the Western part of the country, too. Apart from that, the common explanation of the rural-urban divide in populist radical right voting does not fully apply to the German case as there are rural places in Northern Germany that have very low rates of AfD voters whereas the party was able to succeed in the metropolitan Ruhr area. The left-behind approach is neither fruitful: The most prominent stronghold of the AfD – Erzgebirge/Saxony – is the economically best performing federal state in the East. Germany therefore represents an interesting case for the question of spatial patterns of populist radical right voting behaviour. While scholars already know very much about macro- and micro-level factors for populist radical right voting, the mechanisms linking both together are largely unknown. The paper therefore aims to explain the link between local context and individual vote choice for the AfD and thereby understand the spatial variation of AfD results. To explain the puzzle, the paper proposes a two-step theoretical argumentation: First, it argues for a modified understanding of local contextual effects. Not only the present and immediate surrounding of an individual is relevant but also how the local context performs in comparison with other contexts and how it did perform in the past. Therefore, spatial disparities and local change have to be taken into account when thinking about contextual effects. Second, the assumption of a general effect of local contexts on individual vote choices needs to be reconsidered: Individuals are different, contexts are different, thus, contexts may have different impacts on different individuals and voters may react differently on their environments. Hence, the local context acts through different mechanisms, which may vary between regions and individuals. In order to explain the spatial variation of the AfD's election results, therefore, a differentiated approach is needed: On the one hand, local contextual determinants need to be considered in relation to situations in other places and in relation to the past; on the other hand, a differentiated understanding of the mechanisms is necessary, since contextual situations can have different effects on different individuals in different places. The paper combines individual-level data and contextual-level data on neighbourhood-, county- and regional level to account for the uncertainty about the (ir-)relevance of different geographical levels. Multi-level regression models with cross-level interactions and structural equation models are conducted to answer the research question. By using unique geo-referenced small-scale data, the paper seeks to fill the research gap regarding mechanisms that link contextual effects to the individual level to gain a better understanding of spatial patterns of the AfD vote.