Electoral Pitfalls: Why E-Voting is Not the Solution to Our Problems
Jillian Schmidt
The following if a term paper I wrote for which I did extensive research into varying means for voting. Electronic voting, in particular, which will be instituted in some areas for the upcoming Presidential election, is apparently highly flawed, and while we may think we are safe from the kind of controversy the "chads" caused almost four years ago, we are really in for a shock. Here are my findings...
* * *
Due in part to technological advancement, but mostly as a response to the highly contested 2000 Presidential election, electronic means of voting are being explored in order to try and avoid problems with voting tabulation. The ideas of computerized voting at specified polling places, as well as voting from home via the internet have been thrown around, but each presents many issues with regards to the validity of election results. Problems such as hacking, accidental or purposeful improper voting practices at the machines, software reliability, inaccessibility for lower socioeconomic classes, further withdrawal of individuals from society, an inability to identify who truly is casting a vote in an election, and the inability to make local elections truly regional when conducted via the internet are all problems that make e-voting seem like a fruitless attempt at voting reform. However, many feel that the faster tabulation of votes, as well as the removal of human error, makes it a valid and necessary change.
Many forms of electronic voting have been studied and, in some places, implemented already. In fact, “nearly 30 percent of registered voters live in jurisdictions that use [electronic ballots], a jump of 17 percent since the 2000 election” (“Election Reform Briefings: Securing the Vote” 3). These systems are, for the moment, greatly flawed, and many concerns abound as to whether or not these forms should be already in use or ready to install. Because of these flaws and questions, electronic voting should not be implemented as the standard voting practice. Despite the obvious flaws in our current system, as demonstrated in the 2000 Presidential Election, e-voting does not offer a higher level of protection against fraudulent or controversial results, as it is greatly flawed in and of itself.
Paper ballots are the traditional form of voting, and are currently in use in most districts. Paper ballots range from our system in North Carolina, in which a black marker is used to fill in an arrow next to our choice of candidate, to punch card systems, such as the one in place in Florida with which we are all familiar, including a pen-like punch that is pushed through a designated spot next to the chosen candidate. These systems have been in place for many years, and had served, at least until the debacle in Florida, without much controversy. The 2000 election changed all that. Suddenly, the news was flooded with information on hanging chads, little bits of paper still attached where the punch did not go completely through the ballot, pregnant chads, where the punch had been pushed against the ballot but did not puncture the paper, and varying other failures of the paper card system. Many voters also found the ballot’s layout to be confusing, claiming that they had accidentally punched next to the wrong candidate, thus casting a false vote. Numerous hand recounts were done, which were witnessed on television, as countless election officials held up ballots to examine the chads, in hopes of determining what a voter had intended.
As a result of these problems, the Help America Vote Act of 2002 (HAVA) was implemented. It mandated $3.86 billion (“Securing the Vote” 2) to establish a program to provide funds to States to replace punch card voting systems, to establish the Election Assistance Commission to assist in the administration of Federal elections and to otherwise provide assistance with the administration of certain Federal election laws and programs, to establish minimum election administration standards for States and units of local government with responsibility for the administration of Federal elections, and for other purposes. (107th Congress)
Two forms of electronic voting have arisen from this act, voting done in person at a polling place and internet voting. For in-person voting, many companies have created voting machines, but three companies, Diebold, Sequoia, and Election Systems and Software (ES&S), are dominating the market with similar machines (Boutin 122). While each system is slightly different, Paul Boutin explains that all the procedures are similar, and he breaks down the steps of e-voting via the ES&S IVotronic machine (122). Upon entering the polling place, a voter is given, with the ES&S system, a Personal Electronic Ballot, or PEB, which contains in a chip the appropriate ballot. Other systems have similar smart cards, which are given to voters. The PEB is slid into the IVotronic machine, and the voter is then led through the steps necessary to view the choices for each race. Voters are given the opportunity to review their votes before finally pushing the red “Vote” button to finalize their ballots. The votes are stored in triplicate in non-volatile random access memory (NVRAM), so that votes are not lost if power fails. At this point, some voting systems have and others will have soon a paper audit as well. Voters will verify their votes on a print-out which appears under a glass barrier. The voter can accept their votes, or they can reject them and start over. This creates a paper audit trail to help verify votes and assist in the event of a recount or other election dispute (Boutin 122).
Now that the voter’s job is complete, the polling officials take over. Approximately every hour, “election judges manually add the totals from each machine to make sure the number of votes matches the number of voters who have come in” (Boutin 123). After the polls close, officials print the results from each machine, store the results to a master PEB, and send the results via an encrypted phone line to a computer at election headquarters. The paper printouts and PEBs are also delivered in person to headquarters. Eventually, all the records of the votes – the paper printouts, the triplicate records on the voting machines, and the PEBs – are destroyed as part of the “final security check …[preventing] the vote from being tampered with while it’s in storage” (Boutin 123).
Dialogues regarding instituting internet voting systems have also taken place. Voting would take place over a secure website and encrypted connection, particularly allowing absentee voters overseas, but also the general American public, to vote more easily in the upcoming elections. Over half of all American households now have a personal computer, meaning that access is becoming more widespread. It would be useful for people who are stuck at home or at work who are unable to get away before the polls close, and many believe it will encourage more young people, who are considered particularly tech-savvy, to get involved in politics (Alvarez).
For e-voting systems in polling places, there are many arguments that make their implementation questionable. One of the major arguments is the financial issue. HAVA brought on corporate competition for contracts with the states, which are worth millions (Oakley). Diebold, for instance, holds a contract for over $53 million with the state of Georgia alone (“Frequently Asked Questions…”). These contracts are most often held by the three largest manufacturers, who, according to Freddie and John Oakley, “worked hard to make sure that states and their counties were required to make huge purchases of new voting equipment” from them, leaving little market room for small companies (“Can America trust electronic voting?”). The Oakleys worry that “the major voting-systems manufacturers and vendors wield close to monopoly power, having swallowed up most of their smaller competitors. This makes the major manufacturers and vendors the equivalent of public utilities, offering vital services virtually free of competition” (“Can America trust…”). Diebold and ES&S together, for instance, comprise approximately 80% of US electronic voting systems (Fitrakis). The Oakleys add that the machines are “funded in part by Congress but subject to virtually no congressional standards of quality” (“Can America trust…”), meaning that states could be buying machines with little to no testing and no guarantee of their security and effectiveness. Without competition, the potential for taxpayers’ money to be going to overpriced, ineffective machines makes this argument alone worth considering in regards to electronic voting.
Another reliability questions regards the impartiality of the manufacturers. Diebold, the third largest manufacturer, is a known Bush supporter. The company’s president, Walden O’Dell, said in an August 2003 Ohio Republican fundraising letter that he is “committed to helping Ohio deliver its electoral votes to the president next year” (Oakley). While O’Dell has since retracted this statement, the PEW Charitable Trusts assert that “without open-source codes detailing how machines collect and count votes, the opportunities for mischief are ripe” (“Securing the Vote” 3).
In a school board election in Fairfax County, Virginia, the electronic machines used, according to the CBS Evening News report chronicling the error, “simply failed to count an unknown number of votes for Republican … candidate Rita Thompson” (“Securing the Vote” 9). While there is no evidence that the machine’s manufacturer, which was unidentified in the literature available on the error, purposely set the machine to discriminate against Republican voters, it does in fact beg the question of reliability and fairness within the machines’ software.
With this revelation comes the concern about what the computer really records, regardless of the manufacturers’ biases. Earlier this year, a special election in south Florida, using e-voting machines, found that in a race determined by only a dozen votes, 134 ballots were recorded as being blank (“Securing the Vote” 9). While part of this may be related to voter-error, the fact that the race came down to so few votes shows that these missing votes were critical to the outcome of the election.
The hope is that a paper audit trail can help to avoid this type of error. Mandated in several states, paper audits must be added to touch-screen machines. The machine would print a receipt after a voter has completed the ballot, and the voter can accept it if the votes are correct, sending the receipt to a secure container for storage, or reject it and begin again (Boutin 125). While this seems like it would circumvent the problems discussed earlier, few states have mandates to install such security measures, and only a few of California’s precincts have it in place now. Even with a paper printout used to certify votes at the end of the day, it is no assurance of the validity of the election’s results. If a voting machine fails, is hacked, or is programmed to be biased, the paper printout at the end of the day will show “whatever report of the votes cast that the machine was programmed to show us” (Oakley).
There are also questions as to the safety of the machines from tampering and hackers. Diebold’s company website was hacked in March 2003, and internal documents from the company were stolen and posted online (Fitrakis). Sequoia Voting Systems was also hacked, and its code was posted on the internet, which is ironic, given the company’s claim of being “tamperproof” (Oakley). “The code reveals something else: [Sequoia’s voting technology] relies heavily on Microsoft components,” the Oakleys say, which they note is “particularly vulnerable to hacking.” Another smaller manufacturer, VoteHere, Inc., had their systems hacked in 2003, an attack in which internal documents and software plans were accessed (“Site of electronic voting firm hacked”). While no actual voting machines were accessed, the question is raised as to whether or not a manufacturer can ensure the safety of its machines, if it cannot guarantee the safety of its sensitive documents.
Other safety and reliability issues regarding software and the machines themselves abound. In the March California primary election, issues with the smart card readers, which identify that a valid voter is about to begin his ballot, “kept 36 precincts in San Diego from opening on time” (Boutin 121-22). Also in San Diego, a power fluctuation in the touch-screen voting machines caused start-up screens to pop up which election officials had never seen before, even though they had been trained for many hours, thus causing delays (“Securing the Vote” 9). Also in California, Orange County poll workers who were still not familiar with the new voting system accidentally gave voters the ballots for the wrong districts, invalidating at least 7000 votes (Boutin 122). Another election in Fairfax County, VA, had problems after the polls closed. Many of the 233 precincts in the county tried to send their results to the election center at the same time, tying up the line for many hours. Election officials in many precincts gave up and hand-delivered the results. An unrelated software error delayed many results for up to 21 hours. Even more troublesome, many voters claimed that several machines crashed and deleted votes while they were in the booths. On top of all these problems, election officials took 10 voting machines to be repaired off-site and then returned them to service with voting data stored inside, which was a violation of the state law (Boutin 124). These incidences and other similar ones have been so numerous that California election committee leaders have asked their secretary of state to decertify the machines so that they will not be in use this November (Boutin 122).
Security is also an issue. While “most experts agree that it’s highly unlikely a hacker could walk into a polling place and throw an election,” seeing as how voting machine manufacturers cannot protect their own websites and computers, it seems doubtful that these assurances are valid (Boutin 122). Since very few machines have a physical paper trail, it leaves a hole open that could be exploited (Boutin 122). An independent consulting firm called RABA Technologies ran a test in a Maryland precinct with a simulated vote, and found that if someone were to be left alone with the machines, the locks on the outside of the machines could be picked in about ten seconds, giving access to the memory cards holding the votes. Also exposed was a keyboard jack, which, when a standard keyboard was plugged in, enabled the team to overwrite election results without leaving any evidence (Boutin 125). These are some frightening examples of the vulnerability of electronic voting machines.
Internet voting also carries many of its own risks. One major concern is that of spoofing; that is, hackers creating identical websites to which voters would be diverted. They would erroneously cast their votes there, thus having no actual votes placed in an election. Similarly, there are concerns of denial of service attacks (DoS) which act to choke out networks by aiming at such targets as corporate firewalls (“Denial of Service Attacks…”) “They attempt to overwhelm the processing capability or bandwidth capacity of the device” (“Denial…”). Skilled computer users who wanted to throw a wrench in the voting process could conceivably clog the networks surrounding the election, rendering it virtually useless.
A major question is how the truthfulness of the voting would be assured. At a polling place, even though one’s vote is recorded anonymously, one must first show identification in order to prove that he is in fact the person whom he says he is and that he is registered to vote. Over the internet, it would be nearly impossibly to tell who is voting. If something like a Social Security number or other code was required, it would be all too easy to sit at home or at a public computer and randomly type in various combinations of numbers until a number that has not yet been voted from is found, thus enabling one person to cast many votes. People in other countries might be able to cast votes in a local US municipal election, throwing a wrench in the vote either maliciously or for fun.
With the requirement of a form of identification over the internet comes the concern that this information – one’s Social Security number, IP address, etc. – would be recorded along with his vote, associating his identification information with his vote, eliminating the secret ballot. With the possibility of the loss of anonymity, many voters could be scared away, knowing they can no longer vote in confidence.
Another concern is the “digital divide;” that is, the inequality of internet access across socioeconomic lines, as well as the concern of moving civic duties such as voting out of public places. According to R. Michael Alvarez and Thad E. Hall, showing up to vote is the most public civic activity many Americans engage in, and it is often their only overt participation in the democratic process.” Moving people away from public isolates them from their fellow citizens as well as from taking an active role in the democratic processes.
One of the big issues with in-person electronic voting is the lack of a paper trail, and certainly that is a big flaw with internet voting as well. The connection sending one’s vote to election headquarters is supposedly encrypted, but if anything goes wrong, there is no way to verify what votes were actually cast.
There are, of course, many arguments for electronic voting, but many of them, under scrutiny, do not always hold water. For instance, speed and ease of voting are two of the biggest arguments proponents have. However, a fast, efficient, user-friendly interface is in the eye of the beholder. It is difficult to truly determine what is “user-friendly.” What is user-friendly and easy to some is sure to be confusing and unfriendly to another. For instance, in the 2000 election in Florida, a lot of the voting problems with traditional paper ballots came in counties with high percentages of elderly voters. They had trouble with a system that had been in use for many years. To a lot of them, who probably do not use computers or automatic teller machines on a regular basis, these voting machines will be new and probably a bit intimidating and confusing. For internet voting as well, many elderly, as well as people from lower socioeconomic classes, may not have had much experience with the internet, so they would likely feel alienated by the system.
Convenience is another advantage being touted by e-voting supporters, with the hopes that, for internet voting purposes, people with busy jobs will be able to vote on their coffee breaks, people who are disabled will not have to leave home, and people who are on the move or do not have in-home internet access could use public kiosks set up at convenient locales. Convenience, however, does not always make a technology better. With voting being so widespread, the chance for hacking and tampering is greater. Even if the votes are sent through encrypted connections, it is still possible to tamper with the elections. PayPal, an online payment programs, was supposedly one of the safest ways to buy things online, but it was hacked and credit card information was endangered. This serves as just one example of how even the most stringent security measures can fail.
In-person voting machines are also not fool-proof, despite claims of reliability from proponents. Supporters say that programs written for the sole purpose of logging votes are less likely to be bug-ridden, because they are focused specifically on one task. However, testing to ensure this is highly flawed. Harvard researcher Rebecca Mercuri explains to Paul Boutin that “certification tests look for logic errors and vote-counting mistakes, not security holes. Much of the testing is automated, and layers beneath the voting applications – compilers, OSs, firmware on the machines’ chips – are not examined” (Boutin 124). Mercuri states outright that “the certification process is a joke” (Boutin 124). Thus, we cannot trust the failsafe reputation proponents speak of.
Of course, proponents have objections to the arguments made by opponents. Proponents feel that the three major manufacturers of voting machines will not have a near monopoly, and that it is really up to the counties to decide from whom they purchase their machines. However, while there are many options out there, the bottom line is what drives a lot of choices, particularly in our current economic state. With each state allocated only a specific amount to update their voting systems, they will have to buy what they can afford, and the larger three companies will likely offer the better deals. Thus, it is nearly certain that they will continue to dominate the market.
Proponents will also argue that computer glitches are the exception, not the rule. They have a valid point in that it is difficult to say whether or not these glitches are in fact exaggerated in the news, because the technology is still so new. However, seeing as how the school board election referenced earlier had numerous votes that were not counted, another election had software malfunctions, and other elections had problems transmitting their results, out of the relatively few times the e-voting technology had been used, it seems that it is much more likely to be a big problem as the technology spreads across the country. There have been lawsuits already in India, where electronic voting is widespread, due to major problems with the system, such as software bugs and “deliberate backdoors for tampering,” and the US would likely be in for the same problems (Noronha). Also, we, as the voting public, cannot trust the programmers of electronic voting machines to create software the way they are supposed to; often, they do not follow the precise guidelines set out for them, and due to arrogance, ignorance, or just plain human error, they are likely to write flaws into their programs.
Additionally, security issues can be taken care of with encryption and firewalls, proponents says. These, however, as was mentioned earlier, are not fool-proof, and can be broken through in the right hands. Also, problems such as spoofing and DoS attacks, while they would likely be difficult to pull off against voting systems, they have been used against many other websites, and could possibly wreak havoc with internet voting sites as well.
Proponents also claim that fidelity checks will eliminate the need to worry about ballots not recording anonymously or by the right person, and will qualm fear about not having a voter verifiable paper trail. However, with identity theft rampant nowadays, anyone could forge the information needed to vote. Even if a national identification card were issued to ensure the identity of voters, they could be stolen or forged, just like all other forms of identification. Fourteen-year-olds have made headlines for registering illegally to vote under the current system, without being discovered for quite some time. If we cannot regulate who registers to vote, we certainly can not trust ourselves to regulate who is actually voting. No identification system is adequate for ensuring that not just anyone is casting a ballot.
Of course, if we do not implement an e-voting system, we are left with an outdated paper ballot system which is not, as we know, fool-proof either. To outright drop the idea of ever voting by computers or the internet would also be foolish. Instead of trying to implement the new system so quickly, it would be advisable to hold off for a few years in order to allow for ample testing of the machines and technology to ensure as few problems as possible. A bill, the Voter Confidence and Increased Accessibility Act of 2003, introduced in both the House of Representatives and in the Senate, aims to modify the HAVA to include a provision for paper audit trails. Bills such as these that can narrow down a specific set of requirements are good, and we should wait until more narrowing down can be done before implementing a flawed system. In the meantime, while e-voting undergoes further explorations, modifications could be made to the current paper ballot system, such as coming up with an easier to read, universal ballot that could be used nationwide until the new technology is truly ready.
E-voting, through both polling place machines and the internet, is a risky idea that at the moment is in no way ready to take the place of our current system. There are too many flaws and risks with the system to ensure that we would not have a fiasco like the one that occurred in the 2000 Presidential election. With modifications and time, it might be a wonderful system for the masses to use and ensure a fair election, but until we can be promised a fail-safe system, this is not the answer to our voting problems.
Works Cited:
107th Congress. “Help America Vote Act of 2002, Public Law 107-252.”
Alvarez, R. Michael and Thad E. Hall. “Point, Click, and Vote: The Future of Internet Voting.” The Brookings Institution. 2004. 5 June 2004.
Boutin, Paul. “Is E-Voting Safe?” PC World. June 2004: 121-125.
“Denial of Service Attacks: White Papers and Webcasts from leading IT vendors.” Bitpipe. 2004. 6 June 2004.
“Election Reform Briefing: Securing the Vote.” Election Reform Information Project. April
2004. Washington DC: The Pew Charitable Trusts. 1-6, 9.
Fitrakis, Bob. “Diebold, Electronic Voting and the Vast Right-Wing Conspiracy.” The Columbus Free Press. 2004. 27 May 2004.
“Frequently Asked Questions....” Georgia Counts. 2003. 27 May 2004.
Noronha, Frederick. “India’s electronic voting faces lawsuit over accountability.” Linux
Journal. 2004. 27 May 2004.
Oakley, Freddie and John Oakley. “Can America trust electronic voting? Much clout, no
Regulation for big firms.” The Bee. 2003. 27 May 2004.
“Site of electronic voting firm hacked.” CNN.com. 2003. 27 May 2004.
