Paper 2013/216

Election Verifiability or Ballot Privacy: Do We Need to Choose?

Edouard Cuvelier, Olivier Pereira, and Thomas Peters

Abstract

We propose a new encryption primitive, \emph{commitment consistent encryption} (CCE), and instances of this primitive that enable building the first universally verifiable voting schemes with a perfectly private audit trail (PPAT) and practical complexity. That is: \begin{myitemize} \item the audit trail that is published for verifying elections guarantees everlasting privacy, and \item the computational load required from the participants is only increased by a small constant factor compared to traditional voting schemes, and is optimal in the sense of Cramer, Gennaro and Schoenmakers~\cite{CGS97}. \end{myitemize} These properties make it possible to introduce election verifiability in large scale elections as a pure benefit, that is, without loss of privacy compared to a non-verifiable scheme and at a similar level of efficiency. We propose different approaches for constructing voting schemes with PPAT from CCE, as well as two efficient CCE constructions: one is tailored for elections with a small number of candidates, while the second is suitable for elections with complex ballots.