We analyze the effect of turnout requirements in referenda in the context of a group turnout model. We show that a participation quorum requirement may reduce the turnout so severely that it generates a "quorum paradox": In equilibrium, the expected turnout exceeds the participation quorum only if this requirement is not imposed. Furthermore, a participation quorum does not necessarily imply a bias for the status quo. We also show that in order to induce a given expected turnout and avoid the quorum paradox, the quorum should be set at a level that is lower than half the target. Finally, we argue that a super majority requirement to overturn the status quo is never equivalent to a participation quorum. (JEL: D72) (c) 2010 by the European Economic Association.