Improved Coercion-Resistant Electronic Elections through Deniable Re-Voting

In a democracy, it is essential that voters cast their votes independently and freely, without any improper influence. Particularly, mechanisms must be put into place that prevent—or at least severely impede—the coercion of voters. One possible countermeasure to coercion is revoting: after casting a vote under coercion, the voter can re-cast and overwrite her choice. However, revoting is only meaningful as a strategy to evade coercion if the adversary cannot infer whether the voter has modified her choice—revoting needs to be deniable, while still being publicly verifiable. We define the notions of correctness, verifiability, and deniability for a tallying protocol which allows for revoting. We also present a protocol realizing these notions. To the best of our knowledge, our solution is the first to achieve both deniability and public verifiability without asking information about the voter’s previously-cast ballots for revoting. A seemingly competitive line of work, started by the well-known work of Juels, Catalano, and Jakobsson, uses fake credentials as a strategy to evade coercion: the voter presents to the adversary a fake secret for voting. In this work, we extend Juels et al.’s work to achieve deniable revoting. Their solution also allows for revoting, however not deniably. Our solution supports fake credentials as an opt-in property, providing the advantages of both worlds.