Why would we use something with no well-defined security model and a known polynomial-time attack when we have provably security distributed voting systems? As with digital cash, cryptographers have devoted a lot of time to researching secure voting protocols.