Around the world, people talk about 'corrupt cultures', implying a predisposition for a group of people to behave in corrupt ways. Measures of corruption are in fact strongly correlated with 'cultural variables' such as strong family ties, the traditional end of the World Values Survey’s tradition-rational dimension, the survival end of the WVS’s survival-expressive dimension, individualism, and power distance. The decision to be corrupt involves both cultural norms and a calculation of risks and rewards. A kind of n-person prisoners’ dilemma can result, where a bribing equilibrium results even if hypothetically all officials wish they were impartial and efficient and all people paying bribes wish they didn’t have to. The practical challenge is not to change cultural values and beliefs, but instead to disrupt corrupt equilibriums and alter the risk-reward calculations of bribe-takers and bribe-makers. Fortunately, theory and examples provide guidance and inspiration – even when we can only shrug our shoulders when asked how to engineer cultural change in a broader sense.