Electronic voting has been promised as the solution to the 'hanging chad' problems seen in the last presidential election. Of course, anyone who reads Greg Palast's work knows that 'hanging chad' was irrelevant but it made for good head lines. A much more serious threat to the integrity of our voting system are new electronic voting systems that do not include a paper trail. These systems, known as Direct Recording Electronic (DRE), take one's vote, usually through a touch screen and then record the result in their memory. If that memory or the software that processes it is tampered with or if the software has a bug, there is almost no way to detect it.
In the past voter fraud required a large organization to obtain ballots, fill them out and then stuff them into ballot boxes. The cost of stealing an election, the difficulty of pulling it off and the likelihood of getting caught all made it relatively unattractive to execute voter fraud on a wide scale. DREs take care of all three problems. A small group of people at relatively little expense can steal an entire election with almost no chance of being caught.
There have already been numerous problems with DREs. Everything ranging from Diebold, one of the major DRE vendors, having such deeply lousy security that they had their entire source code available on an unprotected FTP server. Some professors analyzed the code and concluded that the code was so badly designed and insecure that they would have failed any undergraduate who had turned in anything like it. In addition there are multiple cases of votes going missing (see question 3.4). For more examples of DRE problems see here.
Thankfully there is a trivial solution to the DRE's security problems that allows us both the benefits of electronic voting without the downsides – require a paper trail. When someone votes require that a printer produce a ballot that the voter can review and confirm and then place into a secure box. The votes can be counted electronically but if there is any question about the election then the boxes can be opened and the votes recounted.
Today there already is a technology that allows this – optical scan ballots. These are a more advanced version of the old 'fill in the circle with a #2 pencil' forms that you may have used in school. One is given an optical scan ballot that one then fills in using a pen. The ballot is fed to the voting machine which looks at the ballot. If there are any errors then the ballot is rejected (unless you live in a poor black neighborhood – see the question "Is this the only way votes were stolen?" in this article). If the ballot is valid then its results are tabulated and reported electronically. If there are any questions about the ballot then one can always rerun the original ballots. The system is cheap, it's reliable and it's security more than passes the 'good enough' test.
However every time someone tries to get a law or rule passed that requires a paper trail the DRE companies release armies of lobbyists to strike the law/rule down. Why? I suspect the real answer is found not in conspiracy theories but in economics.
With an optical scan system each precinct usually has for or five voter booths, typically folding tables with partitions, where people fill out their ballots. There is then one optical scanning machine where everyone submits their ballot. The optical scanner electronically records the results of the vote for quick tabulation and also drops the ballot into a locked box for safe keeping and later recount if there should be a problem.
With a touch screen DRE each of those four or five booths needs to have a computer with a touch screen in it. Even if optical scanning machines cost twice as much as a single DRE machine this still means that the up front costs for a DRE system at at least twice those of an optical scanning system. That is, rather than having four or five folding tables and a single optical scanner, one has to have four or five touch screen DREs. For cash strapped cities/states having to spend twice the cost of an optical scanning system in order to get a DRE system is a big deal.
Where DREs have an advantage is in ongoing costs. With optical scanning systems one has to print up a bunch of ballots, usually in multiple languages, and each precinct often needs its own unique ballot, get the ballots distributed to all the right places and pray like crazy that nothing changes at the last minute and that no mistakes have been made otherwise you have to pulp the whole run and do a new printing. Furthermore, after the voting is finished you have to go around, collect all the boxes, organize all that paper and file it in a way that will allow it to be recounted later and make sure it is all securely stored. All of which is expensive.
DREs have none of these problems. With a DRE there is no paper to distribute or collect and mistakes or changes are trivial to make. So while a DRE may cost more up front it has lots of savings over the long run. DREs can also offer a more interactive user experience that can pro-actively identify problems in ballots ("You forgot to vote for president, please go back and make a choice", "I'm sorry, you have already selected a senator, would you like to change your choice?", etc.) and can be made much more accessible to folks like the blind.
So DRE manufacturers can argue that while DREs may cost a lot more up front, they will save a lot of money in the long run.
Now imagine trying to sell DREs with a paper trail. First, each DRE now has to have a printer (more expense), paper (more expense) and that paper has to be collected, organized and securely stored after the election. Once DREs are required to have paper trails their long term cost advantages are reduced considerably.
While it is unfortunate that making our voting system robustly secure introduces a cost disadvantage to DRE manufacturers nevertheless I would rather have a secure voting system than a pretty screen to touch. This is why I signed the petition from Greg Palast and Martin Luther King III asking Ashcroft for more secure voting. This is why I signed VerifiedVoting's resolution asking Congress to mandate paper trails. And this is why I called Washington to talk with various Congress critters in order to get them to support HR 2239.
Recently, however, VerifiedVoting.org became a non-profit and started accepting donations. I was very happy to see this because I have been tracking the activities of the founder of VerifiedVoting, Professor Dill, for several months now and over that time I have seen just how effective he has been in working with a large group of people to get public attention to the problem and getting volunteers organized to support efforts to require paper trails. Knowing what great work VerifiedVoting has done I was happy to make a donation.
So please, donate early, donate often and make sure your freedom insurance is up-to-date.