This paper uses a conjoint survey experiment fielded in the US, Australia, Chile, and Argentina to develop and test the compensatory theory of tax fairness, which states that higher taxes on the rich can be used to compensate for other benefits unequally granted by the state. Drawing on social psychology, this paper argues that evidence of preferential treatment by the state violates well-established fairness principles and shows, experimentally, that it leads to taxation to restore equality in crisis times, irrespective of wealth and across a variety of settings. The paper makes two important contributions: it provides the first direct, causal evidence of the importance of compensatory arguments for tax preferences and presents unconfounded estimates of the effect of more established fairness considerations as benchmarks against which to compare the importance of compensatory arguments.