On average, higher per capita income comes with lower corruption levels. Yet, countries like Mexico, Libya and Saudi Arabia are relatively wealthy but experience comparatively high corruption levels. Simultaneously, countries like Madagascar or Mozambique (in the 1990s) combine poor economic development with a low level of corruption. I propose that the two most common variables in corruption research – wealth and democracy – are mutually conditional: economic development brings about a larger (and stronger) middle class that demands public goods from the government. However, citizens’ ability to influence governmental decision-making varies by political regime type. In democracies, citizens are, on average, more successful in demanding goods from the government than in autocracies. Using a large-N approach (up to 139 countries, 1984–2006), the analysis finds robust empirical support for the proposed conditional effect.