Revisiting housing and populism: the politics of wealth and housing discontent

Research on housing and populism has produced conflicting findings – evidence suggests declining housing affordability can have both a negative and positive association with populist support. Focusing on the UK, this paper argues these findings can be reconciled by decomposing the heterogeneous effects of housing affordability into distinct, but related, housing crises. Furthermore, it argues these crises will be related to political preferences through the interaction between geography and tenure. The paper hypothesises that homeowners in areas of high house prices will be less likely to hold populist political preferences due to a wealth effect that legitimises the existing socio-political system. By contrast, social renters in unaffordable areas will be more likely to support populism due to their higher likelihood of experiencing housing insecurity and as the scarcity of affordable housing will provide a conducive context in which to scapegoat immigrants. Analysis using data on local housing conditions and political preferences finds broad support for these hypotheses. The analysis finds that social renting in an area of widespread housing insecurity is positively associated with both opposition to immigration and voting for the Reform Party. It also finds that homeownership does negatively interact with local house prices in predicting attitudes to immigration, but the effect is to attenuate the positive association between house prices and opposition to immigration. This contradicts previous studies which may have been subject to omitted variable bias. No evidence is found for an interaction between homeownership and affordability in predicting support for the Reform Party. The paper contributes to an emergent research agenda around housing discontent and populism. It concludes with reflections on the future direction of this research agenda, suggesting there is value in measuring housing insecurity at the individual level, rather than as a contextual variable, and interrogating causality.