Expansion of the public sector and redistributive policies may reduce income inequality, but formal tests suffer from the problem of endogeneity of government size with respect to the distribution of income. Studying 30 European countries over the period 2004-2015, we apply instrumental variable estimation techniques to identify a causal relationship between income inequality and government size, measured as the government expenditure share in GDP. Using a novel instrument – the number of political parties in the ruling coalition – we find that accounting for the possible endogeneity of government size increases the magnitude of the estimated negative effects. Our findings thus suggest that much of the literature underestimates the true role of the government in attenuating income inequality. The estimated relationship between income inequality and government size persists in a series of robustness checks.