Enough is enough? Analyzing the relationship between economic inequality and support for democracy

Although rising economic inequality is linked to declining support for democracy in multiple theoretical arguments and the public discourse alike, the empirical evidence for their linkage is mixed. We identify five main theoretical approaches that link economic inequality to democratic support, which are based on, respectively,: a) performance evaluations, b) the redistributive function of democracy, c) political inequalities, d) social and psychological dysfunctions, and e) fairness and expectations. We argue that while the majority of these arguments indeed predict some negative effect of inequality on democratic support, they can also reasonably be interpreted to imply that 1) long-term trends in inequality have consequences that are distinct from those of contemporaneous observations of inequality, 2) effects of inequality may not be linear, but rather increasing in the level of inequality, and 3) the impact of changes in inequality over time may depend on the level of inequality. We test these predictions using country-level data running from 1990 to 2022. Given the inconsistency of prior evidence, and the high degree of model-dependence implied by the choice of included countries, the choice of measurements of economic inequality, and the choice of measurements of democratic support, we compare our estimates across a range of country sample and inequality measurements, and between democratic support and democratic satisfaction. We find no evidence for long-term trend effects or non-linear effects of inequality, nor do we find evidence that the effect of changes inequality are dependent on the level of inequality. We do find reasonably robust evidence for a negative within-country effect of contemporaneous levels of income inequality on democratic support. Higher average income inequality also appears correlated with lower average democratic support, but we are unable to rule out that this is a spurious effect, in particular with respect to confounding due to difference in GDP.