| Abstract: | We make the point that preferences for democracy are positively correlated with level of education. This correlation is robust even after controlling for a range of personal characteristics, including country of residence, income, age, or using different definitions of preferences for democracy or using instrumental variables. Interestingly, the results hold across countries with different level of democracy. We use data from World Values Surveys and show that our results are consistent with a simple theoretical model in which education makes political accountability easier. |
