How strong is the evidence that inequality reduces people's political engagement? Previous studies have found that people discuss politics with friends less often in less equal countries, consistent with the thesis. However, it could be that political disengagement leads to higher inequality, rather than inequality leading to disengagement. This paper therefore re-examines the thesis that inequality causes disengagement, using two novel empirical strategies, and multilevel models fitted to data from all five waves of the World/European Values Surveys. First, I assess whether the relationship between inequality and political discussion holds not just cross-sectionally (across countries at a given point in time) but also longitudinally. In other words, have countries where inequality has been rising more than average also experienced larger-than-average declines in political discussion? Second, I use an instrument for inequality to examine whether the correlation between inequality and discussion holds even when any reverse-causality is ruled out. The instrument, which comes from Easterly (2007), is the suitability of a country's land area for growing wheat relative to sugar. (Wheat production is typically undertaken by smallholders, whereas sugar and other tropical cash crops have historically been grown on large plantations, leading to the concentration of land ownership and thus income.) I find mixed evidence for a relationship over time, but a clear correlation between the wheat-sugar ratio and lower rates of political discussion among friends. Overall, the results presented here are consistent with the argument that inequality depresses engagement.
