Paper 2019/822

Surveying definitions of coercion resistance

Ben Smyth

Abstract

We explore definitions of coercion resistance in the computational model of cryptography; discovering all but one are too weak (i.e., satisfiable by voting systems that are not coercion resistant) and the other is too strong (i.e., unsatisfiable by voting systems that are). Hence, we show that coercion resistance has not been adequately formalised. Our results cast doubt over the security of voting systems that have been proven secure with respect to those definitions; a new definition is necessary.