Paperless Ballots, Election Legitimacy and Voter Confidence:
Problems and Solutions
Prof. Nicholas Johnson
Direct Recording Electronic (DRE) voting machines
are used by less than 40% of voters in the
Problems with voting machines are not a new
phenomenon. In 2000, outdated technology
led to the embarrassing
The Help Americans Vote Act (HAVA), passed in 2002, afforded states the opportunity to replace old voting equipment with more modern technology. HAVA was intended to correct disparities in voting technology that gave rise to the Equal Protection issues raised in Bush v. Gore. Many states rushed to purchase electronic voting equipment in order to take advantage of HAVA money. Use of DREs tripled between 2000 and 2006.  However, as with all new technology, many DRE machines were not capable of perfect performance. Because of the importance of an accurate vote, DRE failures have garnered much publicity and raised the ire of voters and politicians. Just a few short years after the machines were hailed as the solution to outdated voting equipment, some states began junking their DRE machines because of a political firestorm about the accuracy of the machines and the motives of the companies who make them.
This paper will examine some of the legal issues surrounding the use of DREs.
Part I will briefly outline the role of technology in the voting process, the benefits and pitfalls of the most commonly used voting equipment, and importance of voter confidence in election results.
Part II will examine the equal protection implications of voting machine errors, and the evolution of voting rights in the wake of Bush v. Gore.
Part III will briefly examine the impact of DREs on the ability to verify and challenge election results, and the special problems created by the use of computer technology in elections.
Part IV will discuss the implementation of HAVA, and its relation to the growth of DREs.
Part V will discuss certain intellectual property issues regarding DREs, and the public policy and legal issues resulting from the fact that for-profit companies are manufacturing the equipment.
Part VI will advocate new standards for electronic voting equipment, including open source code, VVPAT, and enhanced certification procedures.
I. The Role of Technology in a Well-Functioning Democracy.
· the right to a cast a vote;
· the right to an equally weighted vote; and
· the right to have one’s vote accurately counted.
These rights and the decision in Bush v. Gore will be explored in greater detail in Part III.
It is clear from recent elections and subsequent litigation is that the right to vote does not include the right to utilize a perfectly functional balloting mechanism. Despite technological advances and congressional legislation like HAVA, American elections are still plagued by troubled voting equipment. The rights of individual voters are to a degree dependant on the type of equipment used in their precincts.
A. Current voting systems.
There are five types of voting technology
currently employed in
Paper ballots are the original voting technology, whereby voters are given a piece of paper listing all the candidates. Voters simply place a mark next to the name of their preferred candidates, and the ballots are tabulated by hand after the polls close. Paper ballots have benefit of providing a record of each individual vote which can be helpful in auditing elections, but obviously the speed at which the ballots can be counted is insufficient in the modern world. Additionally, paper ballots are extremely susceptible to fraud. Stories of ballot box stuffing are legendary and occasionally celebrated pieces of American folklore. Lyndon Johnson won election to the Senate in 1948 by a mere 87 votes, helped along to victory by 202 dubious ballots that arrived extremely late and happened to be cast in alphabetical order. Beyond the rampant possibilities for fraud, errors can occur if ballots are unclearly marked, or if mistakes are made by ballot counters. Paper ballots have largely been abandoned in favor of mechanical means that offer increased counting speed. In 2006, less than 1% of voters cast ballots using paper ballots.
Though they are still used by a little over
11% of voters,
lever machines are also dwindling in
use. Debuting in the 1890s, lever
machines were the first use of machinery in voting equipment.  Voters
enter a lever machine booth and close a privacy curtain. Voters select their candidates by flipping small
levers next to the name of their chosen candidate. Voters may change their votes by de-selecting
levers. After selecting their candidates, voters pull a large lever which opens
the curtain and records the vote on a tally sheet behind the levers, which is
unseen by voters. Lever machines are less
susceptible to fraud than paper systems, but do not produce a document record
of individual ballots. Additionally, the mechanical nature of lever
machines makes their moving parts subject to wear and tear, meaning that the
more a machine is used, the more likely it is to fail. Lever machines are no
longer manufactured, making it difficult to replace parts and faltering
worst case scenario for lever machines was borne out in
card systems also played a big role in
Since the passage of HAVA, the most popular
type of voting equipment are optical
Optical scan systems are similar to the systems used to grade standardized
tests such as the SAT or LSAT. Voters mark a paper ballot, which is then
scanned by a machine which senses and records the marks. Optical scan is currently the most common
voting system in the
Fully electronic voting machines debuted in the 1970s. Called “Direct Record Electronic” systems because they directly record a vote into computer software rather than on paper, DREs are the second most common type of voting equipment. DRE usage increased substantially after the 2000 election. 38% of voters now live in precincts that utilize DREs, up from 12% in 2000. Two types of DRE systems are currently in use. After checking in at their polling place, voters are given an electronic card which they insert into the DRE machine. The machine then displays the ballot to the voter. Voters make their selection by touching the screen, or by pressing buttons on older DREs. Most DRE machines produce an internal paper log of votes which can be used for audits.
Touch-screen DREs are versatile and user-friendly. On many systems, a “summary screen” listing all of a voter’s choices allows a voter to confirm his ballot before it is officially recorded. After viewing the summary screen, the voter may either cast his ballot, or go back and make corrections. On some DRE machines, an audio/headphone system informs sight-impaired voters of the ballot choices and the contents of the electronic summary screen, allowing them to verify and cast their ballots without assistance and in complete privacy.  Voter experience with DREs has generally been positive; voters have found the machines easy to use, likened to the ease of using an ATM. Other potential benefits for voters include online links to a voter information packet, photos of the candidates, and balloting in multiple languages.
For the government, DREs make administering elections more efficient by eliminating paper and speeding the counting process. DREs also provide assurance of certain voting rights by providing better ballot access for disabled voters, since they do not require voters to be able to hold a pencil, or in some cases, require voters to be able to see. On the other hand, the use of DREs has been marred by software malfunctions and questions of reliability. Voters using DREs have reported machines switching votes between candidates, irregular totals and general machine failure in a variety of jurisdictions. (These problems are discussed in more detail in part IV.) Additionally, the use of DREs requires poll workers who are familiar and adept at dealing with the technology. DRE (and optical scan) machines must be “calibrated,” meaning that the software running the machine must be coordinated with the specific ballot presented to the voter. Small errors in calibration can lead to huge errors in voting. While machine vendors often are able to assist in calibration, very few polling places have a technology expert on-site on election day. As a result, nearly any software glitch can shut down a polling place, or worse, impair the integrity of the count.
DREs are not the only voting system plagued by errors. No voting machine is capable of perfect performance; each system has a residual vote rate, which is essentially the percentage of votes that are cast but not counted due to machine error. These rates vary between companies and vary from election to election. A Cal Tech/MIT study determined that in presidential elections, voting systems have the following error rates:
· paper ballots: 1.8%
· lever machines: 1.5%
· punch cards: 2.5%
· optical scan: 1.5%
· DREs: 2.3%.
In practice, this means that 2.3% of votes cast on DREs are not counted, through no fault of the voter. While it may be tempting to conclude that optical scan machines are more reliable, other data suggests that the residual vote rate of optical scan machines can be much higher due to variations in the way in which voters mark their ballots. As alarming as these numbers may be, progress has been made in reducing the number of residual uncounted votes. In 2000, nearly two million votes (2%) were not counted. That number fell to just over one million votes (1.07%) in the 2004 election. Much of that progress can be attributed to the replacement of old machines with new technology.
In a perfect world, voting machines would accurately record and count every vote. However, due to imperfect technology and imperfect human beings, it is likely impossible to create a perfect system. Furthermore, while accuracy in a voting system is obviously a necessity, reliability is not the only factor that jurisdictions must consider in choosing a voting system. Decisions related to the purchase of voting equipment are often made under budgetary constraints. One of the oft-reported facts in the wake of the 2000 election was that in some low-income precincts, voters utilized old lever machines which were generally more susceptible to error. HAVA made it possible to replace some of those machines with new equipment. However, new voting machines are purchased once in a generation, if even that frequently. While DREs seemed like good solution to the problem of voting technology in 2003, some DRE converts are finding that the machines they purchased in 2003 are flawed in important technological ways, and perhaps, more importantly give voters reason to question the manner in which elections are conducted.
B. The importance of voter confidence.
While there are many concerns that influence a jurisdiction’s choice of voting mechanism, from a national and constitutional perspective, voter confidence must be of paramount importance. American Democracy is celebrated for its long tradition of peaceful transitions of power. The survival of democracy depends upon all parties trusting in the integrity of election results. While the losing side may not be happy with the results, as long as there is a belief that the will of the people was expressed accurately, the legitimacy of the election is not doubted. This idea was clearly expressed by computer science Professor David Dill, founder of the voter advocacy group VerifiedVoting.org. Testifying before a commission headed by former Presidents Carter and Ford on the issue of federal election reform, Dill said:
“[I]t is not sufficient that election results be accurate. The public must also know the results are accurate, which can only be achieved if conduct of the election is sufficiently transparent that candidates, the press, and the general public can satisfy themselves that no errors or cheating have occurred. Unfortunately, the advent of paperless electronic voting is moving us away from election transparency.”
With the increasing prevalence of DRE machines in American elections, there has been an increasing number of skeptics who doubt the validity of elections conducted using electronic voting. More than one-quarter of Americans believe that the vote count in the 2004 Presidential election was unfair. A report by the Congressional office of Rep. John Conyers (D-MI) analyzed the 2004 election results in Ohio, and concluded that voting irregularities “raise grave doubts” about the validity of the reported results.
It is not surprising that with the increased use of new technology, voter confidence has diminished. While all elections require a measure of faith in the process, DREs require an additional level of trust. The only documentation of a DRE ballot is in an internal paper record, which the voter never sees. Voter advocacy groups like VerifiedVoting.org contend that these records could easily be manipulated because there is no way to verify their accuracy. Additionally, with computer viruses, internet worms, and malicious hackers in the news on a regular basis, there is concern that electronic voting machines are vulnerable to security breaches in which a malicious programmer could alter elections results.
Concern over the impact of DREs on election legitimacy is not limited to academics or defeated candidates. Almost immediately after the passage of HAVA, Congressman Rush Holt (D-NJ) posted a question on his website: “On Election Day 2004, how will you know your vote is properly counted?”  Holt’s answer? “You won’t.”  Holt’s posting followed his introduction of The Voter Confidence and Increased Accessibility Act of 2003 (VCIAA). VCAIA would amend HAVA by imposing stricter standards on electronic voting machines. It was offered in response to the growing distrust of the DRE machines being used in many states.
VCIAA would have required all voting machines to produce a voter-verified paper audit trail (VVPAT) that voters could use to verify the accuracy of their votes. Additionally, the measure would have banned the use of undisclosed software and wireless communications in voting systems. VCIAA did not come to a vote in the House, and the 2004 election went ahead without any of the safeguards Holt proposed. Upon proposing his bill, Holt stated that “current law does nothing to protect the integrity of our elections against computer malfunction, hackers, or any other potential irregularities.”
proven to be something of a prophet in 2004, as DRE glitches marred elections
Holt reintroduced his bill in 2005 and attracted a number of co-sponsors. However, Holt’s bill, formally known as H.R. 550, saw no action after it was introduced into the house. VCIAA is just one of several similar measures currently pending in Congress. In 2006, there were at least 17 different pieces of legislation proposed in Congress by members of both parties. Senators Hillary Clinton and Christopher Dodd, and Representative John Conyers have all introduced legislation calling for a VVPAT;  on the Republican side, Iowa Congressman Steve King has proposed the “Know Your Vote Counts Act,” which calls for a VVPAT in all voting equipment. Yet there was little agreement on the specific provisions of these bills, and all of the bills died in committee. In early 2006, it was apparent that the fall election would proceed without the safeguards advocated by the skeptics of electronic voting.
extremely close and bitterly partisan elections of 2000 and 2004, it is likely
that politics has much to do with concerns over voting machines. Democrats are far more likely to question the
accuracy of the vote count than Republicans,
which is hardly surprising given which party has ended up on the losing side of
the vote counts. Nevertheless,
complaints about DREs and calls for reform have come from across the political
spectrum. Indeed, the rift between “red
states” and “blue states” only serves to underscore the need for voting
mechanisms that not only provide an accurate a count as possible, but also
engender voter trust. In a closely
divided nation, it is more important than ever that voters believe that
election results have integrity, lest the democratic process be subverted by
hysterical allegations of cheating in the aftermath of a close vote. While
the 2000 Presidential election in
II. Voting Rights in the wake of Bush v. Gore.
Six years after the
A. The right to cast a ballot.
Bush v. Gore highlighted a peculiar aspect of voting rights: Voters do not have a constitutional right to vote for Presidential electors. The manner by which states choose representatives to the electoral college is determined by the laws of each particular state. A popular vote of the citizenry is the method for choosing electors in all 50 states, but legally any state is free to change its laws, and revert the power of choosing electors to the legislature. The peculiarity of the Electoral College highlights the enormous power granted to states on matters of voting rights.
The “Time, Place and Manner” clause creates
dual sovereignty over voting regimes, giving a state the authority to set its
own rules, but reserving to Congress the power to alter those regimes as
necessary. Congress has used its authority to implement
a national election day,
but states have retained the power to set polling hours and locations, absentee
balloting procedures, and the mechanisms for vote casting. Providing that no discrimination against
voters occurs in violation of the Federal Constitution, states may determine
their own specific voting procedures. The primary constitutional challenges to state
election procedures involve the Equal Protection clause, where voters are
denied the right to vote because of discriminatory state law. Bush v.
Gore put a new twist on Equal Protection issues by highlighting the
disparate treatment of voters in different
After the controversial election, a fierce national discussion over creating a procedures prompted congress to explore implementing national standard for voting procedures. However, given the importance of dual sovereignty in the federal system, a state’s authority to determine its own election procedure creates considerable obstacles to the implementation of uniform national voting standards. Subsequently, HAVA attempted to provide states with the ability to remedy voting disparities, but left the states considerable power to choose their form of compliance. The states in turn allow counties to determine which machines to purchase. As a result, jurisdictions within each state have set different priorities relating to their purchase of voting machines, and it is not uncommon for states to have two or more different types of voting systems in use across the state
B. Right to an Equally Weighted Vote.
The right to vote extends to the manner in
which the vote is weighted as well the grant of the franchise itself. The historic “one-man/one-vote” decisions of
once granted the right to vote on equal terms, the State may not, by later
arbitrary and disparate treatment, value one person's vote over that of
Bush v. Gore, 531
The Supreme Court found that the Florida
Supreme Court recount decision had mandated a disparate treatment among voters in
violation of their Equal Protection rights. During the chaos of the
New voting technology put into use after the 2002 election was supposed to eliminate this kind of problem. Computerized voting terminals theoretically capable of alerting voters to potential over/under votes, and additionally offer the potential of standardized recount procedures. However, given the technological glitches that have plagued some states, it may be that the very use of DREs in some counties but not others, creates an Equal Protection violation.
Within most states, counties remain free to choose which machines to purchase. As previously stated, no voting machine can function with 100% accuracy, and each of five voting systems in use has its own unique error rate. Consequently, voters in some counties will have their votes tallied more accurately than voters elsewhere. Given that optical scan machines tend to be more reliable than DREs, voters who cast ballots on DREs may have a claim that their vote is valued less than voters who vote in a jurisdiction with optical scan machines.
While Bush seems to leave the door open to this kind of Equal Protection claim in hundreds of jurisdictions, lower courts have not found disparate use of technology within a state to be a constitutional violation. Many lower courts have relied on Justice Souter’s dissent in Bush, in which he stated that local variation in voting technology did not violate Equal Protection, even if the technologies have different levels of effectiveness in recording voter’s intentions, so long as there is some rational basis for choosing the machine variety. Since rational basis review only requires some plausible explanation for a legislature’s decision, most lower courts have easily done away with local challenges to disparate election equipment, without considering the nature of the alleged voting errors. The issue has not reached the Supreme Court for consideration.
C. Right to an accurately counted vote.
In his philosophical play “Jumpers,” playwright
Tom Stoppard wrote, “Its not the voting that's democracy; its the counting.” The Supreme Court has recognized the
fundamental right to an accurate count
of one’s vote, holding in Gray v. Sanders that “[e]very voter's
vote is entitled to be counted once. It must be correctly counted and
However, vote-counting decisions have
generally arisen out of classification-based denials. Prior to Bush,
matters related to vote counting were regarded primarily as the province of the
state, rising to the constitutional level only if specific Equal Protection
violations were asserted. For example, refusing to count the votes of
African-Americans would clearly be an Equal Protection violation, but the
disqualification of a vote based on an incorrectly marked ballot would not be a
extended the vote-counting right by applying Equal Protection rights more
broadly. The affected
While the court also stated its decision in
Bush was limited to “present
it at least established a clear rule that recount procedures within a state
must be consistent. It also created numerous questions about whether
initial counting procedures themselves must be consistent within a state. In the wake of the Bush, several Equal Protection challenges were filed in courts
around the country, charging that counting deficiencies created an Equal
Protection violation.  Notably, in
uncertainty and embarrassment following the 2000 election created a legislative
push to improve the manner in which Americans voted. The legislative response to the
III. HAVA: Helping Americans Vote?
Ensuring that votes are recorded accurately
and counted correctly is fundamental to the right to vote. The
· permit the voter to verify (in a private and independent manner) the votes selected by the voter before the ballot is cast and counted,
· Notify the voter of an over-vote prior to the ballot being cast,
· provide the voter with the opportunity (in a private and independent manner) to change the ballot or correct any error before the ballot is cast,
· produce a permanent paper record with a manual audit capacity for such system, and 
· provide the voter with an opportunity to change the ballot or correct any error before the permanent paper record is produced.
At first blush these requirements would seem to ban nearly all current systems. Punch card and lever machines would seemingly be in violation of the “second-chance” provisions giving voters the opportunity to correct an error. However, HAVA allows an exception for jurisdictions still using these machines, providing that those jurisdictions implement a voter education program, and offer instructions to voters on how to correct a mis-marked ballot. Consequently, HAVA’s requirements do not actually mandate that the voting machines themselves provide the opportunity to correct mistakes. Similarly, the paper-record and voter-verification requirements would seem to limit the use of DRE machines that do not have a VVPAT. However, the two provisions are not read together—it is not necessary for the voter to verify the permanent paper record in order for a voting system to be HAVA compliant. DRE equipment contains an internal paper record of the votes cast, which is sufficient for HAVA compliance, while the opportunity to change votes is preserved prior to the creation of the paper record.
Another key HAVA provision is aimed at increasing the accessibility of voting for disabled citizens. States must also ensure voting systems are accessible to voters with disabilities; with special emphasis placed on ensuring that visually impaired voters are afforded the same opportunity to vote.  States may satisfy this requirement through the use of at least one DRE voting system or other voting system equipped for individuals with disabilities at each polling place, and HAVA specifically mentions DREs as an acceptable method of disability compliance.
The disability requirement in HAVA led to a rush on the purchase of DREs. Not only were machines used to fulfill the disability requirement, but state also took the opportunity to replace outmoded systems with new technology. Thirty states took more than $300 million in HAVA money to replace punch card and lever machines.  In the 2000 election, just 12.5% of voting equipment in use nationwide was electronic, but by November 2005, that number had tripled to 38%.  DRE machines are in use in 42 states, with at least two states using DREs in every precinct. 
IV. DRE: Does it Record Everything?
States turned to DREs for a variety of reasons, but HAVA’s disability requirements were chief among them. Many jurisdictions chose to use DREs for all of their voting equipment in order deal exclusively with a single vendor while meeting HAVA disability specifics. However, in their rush to get a piece of the HAVA pie, some states purchased equipment that was not capable of providing maximum accuracy.
A. The Trouble with DREs.
Numerous problems have been reported with the current generation of DRE technology. The three primary manufactures of DRE equipment, Diebold, Sequoia, and Electronic Systems and Services (ES & S) have all been plagued by problems and questions about the use and validity of their machines. Since the 2004 Presidential election:
2006, a City Council Candidate in
S manufactured machines malfunctioned on Election Day 2005 in
in several precincts in
These issues built on problems with the use
of DREs in the 2004 Presidential election. A
At the moment, the use of DREs in elections undermines fundamental voting rights. While making it easier for voters to cast a ballot, the unreliability of the technology calls into question a voter’s ability to be assured their votes are accurately counted. Furthermore, the lack of a voter-verified paper trail actually makes it more difficult to guarantee an accurate record of each vote, and makes it nearly impossible to audit election results.
B. Just because you’re paranoid doesn’t mean they aren’t out to get you
DREs are unique among voting equipment in that they do not rely on paper. Optical scan machines, punch cards, lever machines and hand counted-ballots are all rooted in paper records; the voter physically marks a ballot, or else pulls a lever that creates a metal punch in recording paper. With DREs, the voter sees only a representation of the ballot on a computer screen. Votes are registered electronically, and stored in an electronic file.
There is legitimate concern about the
security of electronic voting. Independent
testing determined that Diebold voting machines in Leon County, Florida were
vulnerable to outside tampering, and that election workers could alter voting
tallies without detection. Election Supervisor Ion Santo was roundly
criticized for allowing the tests,
but was somewhat vindicated when a similar study in
These concerns have led to a bi-partisan
political backlash against DREs. In March 2006, Democratic New Mexico Governor
Bill Richardson enacted a bill that will transition his state entirely away
from DREs, in favor of optical scan voting equipment.
Unfortunately, the decisions to replace DREs are motivated by political desire rather than sound policy or constitutional principles. Scrapping the DRE machines is not only a waste of public dollars, but a waste of an opportunity to improve election administration. In most cases, DREs have functioned perfectly. The errors that have occurred are likely attributable to voter error or error by poll workers. Problems reported with DREs are generally the types of error that occur when people are unfamiliar with new technology. While some voters equate using a DRE with an ATM, there is still a large segment of voters who have never used an ATM, or had the benefit of any kind of computer instruction. As computer usage becomes increasingly ubiquitous, voter familiarity with the machines will improve DRE performance.
This is not to suggest that the problems associated with current DRE usage should be tolerated. A “figure it out as we go” strategy may be appropriate when users are learning how to use a word processor or the Internet, but is clearly an insufficient way to conduct an election that determines who govern the nation. In the long run, DREs have clear advantages, but the price paid in the erosion of voter confidence and decreased result validity in the short term are too problematic to ignore. In the short term, a better solution than scrapping the machines altogether includes changing machine certification standards, improving poll worker training and voter education. Additionally, election machine manufacturers must recognize that confidence in the voting process is essential to democracy and make elections conducted on DREs substantially more transparent.
V. Conflict of Interest: profit motives versus public elections.
One reason why there is a lack of transparency involved with elections conducted on DREs is that development of election technology has been left entirely to the private sector. This creates a difficult problem surrounding the proprietary code with the voting machines. Given that states must purchase equipment from private companies, and that such equipment is used exclusively for the exercise of rights fundamental to democracy, how much interest should a private company retain in the operation of those machines, and how can states balance private company rights with the public’s right to a full accounting of an election? Vendor contracts can be a barrier to transparency. State election agencies alone can conduct testing and regulation of DREs, placing the awesome responsibility of providing for the security of the voting process in the hands of partisan officials who are given little or no guidance.
A. Proprietary Code.
The 2004 election is still being questioned
Diebold DREs use the company’s Global Election Management Software (GEMS) to create and log votes. Each machine contains an internal database file which include the raw data records — the votes — collected by each machine. Diebold claimed that though the information is public, the company owned “the structure of the database,” and that the files were proprietary information. Diebold originally told the Department of Elections that the public data could be released only after being transferred to a common format such as Microsoft Excel. The Alaska Democratic Party rejected this notion, maintaining that it is imperative to see the information in its original structure —the format in which the data was created, stored and reported.
From the public’s point of view, the database files in the DRE are akin to the old paper ballots —the most direct evidence of the votes, and the most accurate device used in auditing the results. On the other hand, Diebold claims the files contain not only the vote record, but also software engineering that is company property. Allowing public access to the files implicates Diebold’s property rights, and potentially leaves the company vulnerable to infringement. Because the raw vote data is intertwined with the company’s proprietary software, the debate over the release of the information creates a host of legal issues, pitting intellectual property and contract rights against voting rights, without any clear standard of which should take precedent.
Complicating Diebold’s proprietary claim is the fact that the structure of the database is publicly available on the Internet and has been for several years.  A website, complete with instructions on “how to build your own GEMS box” is maintained by a computer technician affiliated with an electronic voting rights activist group. The instructions were discovered on an unsecure site on the Diebold company server, and subsequently published on the web.  While the publishing of this information may constitute “fair use,” it nevertheless raises troubling issues about the security of Diebold’s code.
The Alaska Democratic Party seized upon this invasion of Diebold intellectual property rights to further their claims. Essentially, the ADP contended that since the information was publicly available, Diebold had no proprietary interest to protect, and therefore could not claim “company secret” to stop the release of the database files. While this claim is of dubious legal merit, it does hold considerable force in the court of public opinion. With the discovery of the public nature of the code that the company had claimed was secret, Diebold simply could not refuse to release the database files without appearing as though it were hiding something. Diebold told the Alaska Division of Elections that the company would waive its proprietary claims, and leave it up to the state to decide whether or not to release the files.
In Diebold’s letter to the state waiving its claims, Diebold advised the state that the information contained in the files included the user-names and passwords of all the system’s users, as well as the phone number used to modem-in the election results. Diebold advised the state’s election officials that “due to the nature of the sensitive information contained in the files” they should “carefully consider the ramifications” which the state would have to deal with upon release of the files.” Subsequently, the state determined “the release of the database, the database backup file, and the audit files … present a significant security risk to the confidentiality, integrity and availability of the Election System and information.” In the spring of 2006, the state had yet to release the files.
While anecdotal in nature, the
The only data available for an audit are the election returns themselves, which could be flawed if the programmer or a hacker altered the method by which the machine records and counts. At this point, auditing becomes a pointless endeavor, because there is no reliable information on which to base the audit—the only information available is contained in the machine whose accuracy is being questioned. Additionally, given the lapse in security at Diebold in protecting its code, it certainly seems possible for the workings of the machines to be compromised by a nefarious third party. Without a public knowledge of the code, such malfeasance is undetectable.
B. Privatizing elections?
Currently, the task of creating of voting machines
and software has been placed exclusively in the hands of private companies who
shroud their innovation in secrecy. This
nearly eradicates the transparency required for a reliable election. The "independent" testing
laboratories who review these machines are often chosen and paid by the
vendors, casting doubt on the reliability of their reports. While
all machines must be certified before use, certification can be subverted by
the need to install last-minute software “fixes.” While such fixes may be
necessary to ensure the machines function properly, they invariably raise
integrity questions by the losing side. In
the 2002 election, Diebold installed software patches in each of
Further complicating this is the limited
number of vendors who serve the market.
Diebold, ES & S and Sequoia command 89% of the voting machine
Since no hard evidence of any past election being compromised by DREs failures exists, it is preferable to believe that past elections have not been compromised by the use of DREs. Nevertheless, placing the intellectual property and profit interests of voting machine manufacturers over the security and sanctity of voting can potentially compromise the democratic process.
VI. Moving beyond fear: solutions to the dilemma of voting technology.
HAVA represented a positive step toward the goal of ensuring that the intent of each vote is correctly recorded and counted, to the maximum extent possible. Given the rapid rate at which technology changes, creating national voting machine standards requires flexibility with an eye toward the future. In the four years since HAVA passed, technology has advanced well beyond 2002 levels. VVPAT technology that was expensive and impractical in 2002 is now readily available. However, unlike personal computers which users often upgrade ever few years, voting machines are purchased by states with much less frequency. Many states have invested in new machines in recent years, making it unlikely that they will junk machines in favor of better technology, unless provided with financial incentives, significant political pressure, or federal mandate. Nevertheless, until accuracy and transparency can be increased, the use of DREs undermines voter confidence and election integrity. Several reforms can improve public confidence in DREs, and their reliability.
A. Open Source Code.
A potential solution to the proprietary code problems posed by the current regimes may lie in open source code. Open source code allows access to the precise instructions for how a program functions, meaning that a programmer can determine the methods the software program uses to achieve its functionality.  In an open-source model programmers forgo intellectual property protection, and distribute their software for public use and improvement.  The Linux operating system and Mozilla Firefox web browser are two widespread examples of open source software. This model could be adopted to DRE technology.
DREs vulnerability lies not in the machines themselves, but in the software that runs them. In an open source model for voting machine software, a cooperative group of programmers would collaborate on software design for the recording and counting of votes. The group could continually improve voting process, working together to look for errors, poorly written code, or security weaknesses. This process can improve the efficiency or design of the source code or add additional features. Somewhat counter-intuitively, open source software improves security. If the source code is public known and regularly scrutinized by a wide community, it becomes much harder for a nefarious programmer to insert malicious code without detection. 
Open source technology used in voting machines would satisfy the Equal Protection standards set forth in Bush v. Gore. Unfortunately, under current election practices, open source coded voting machines are contrary to public policy, and in some states, against the law. Escrow provisions and proprietary code keep secret nearly all source code used to operate a voting system. However, a HAVA amendment could prohibit the use of proprietary code in the counting and recording of votes.
Furthermore, the National Science
Foundation has recently endowed A Center for Correct, Usable, Reliable,
Auditable and Transparent Elections (ACCURATE), a multi-institution research
community headquartered at John’s
Allowing widespread access and educated scrutiny to the software that runs voting machines may go a long way toward increasing public confidence in DREs. A completely independent association charged with maintaining the integrity of voting systems would provide a uniform standard and an heretofore unseen level of accountability.
While manufacturers might be reluctant to turn proprietary code into open source, given the rate at which states are turning away from DREs alters the market considerably. The profit in providing voting machines derives not from the software, but in selling the machines themselves. Furthermore, the manufacturer’s proprietary interface (what the voter sees while using the machine) does not necessarily need to be open sourced. This system would allow manufacturers to specialize in user-friendly interfaces to distinguish them from competitors, while leaving the recording and counting functions open for public scrutiny and improvement. A move to open-source vote counting software would be a sea change in the industry, which would go a long way toward restoring voter confidence in the integrity of the balloting process.
A Voter Verified Paper Audit Trail is the most popular reform advocated in Congress, touted by Rep. Rush Holt and many others as a way to boost voter confidence and election integrity. Voters would be given a paper record of their vote as a way of assuring accuracy.
While HAVA requires that all machines contain a paper trail to be used for the purposes of an election audit, internal records currently satisfy this requirement. In 2002, the technology for a voter-verified paper audit trail was in its infancy and was thought too expensive to be required. . However, advances in technology and cryptography have made VVPAT an achievable goal.
VVPAT’s inherently present logistical problems of security and secrecy. A paper record of a voter’s ballot carried out of a polling place could easily be perverted by a vote-buying scheme, while the lack of a paper record leaves voters with their votes could be compromised by unreliable technology. Computer scientist David Chaum has proposed a solution for the creation of a VVPAT that solves the problems of secrecy and privacy in a novel way.
Chaum’s scheme involves a verifiable ballot that uses a unique form of cryptography that allows a guarantee of an accurate record of his or her ballot, while preserving ballot security and secrecy. Under Chaum’s proposal, after the voter selects their candidates, the DRE machine prints out a specially formatted version of the ballot on two transparencies. When the layers are placed on top of each other, they clearly show the voter a vote record. However, each of the two transparencies employs visual cryptography so that one alone does not reveal any information without the other. After confirming their vote, the vote gives one layer to a poll worker who destroys the ballot in a special shredder, while the voter keeps the other layer as an encrypted vote record. The DRE retains an electronic copy, which is identical to the encrypted layers presented to the voter. In this manner, both the secrecy ballot and integrity of the election are preserved. The system remains affordable, with only additional costs being the printer and transparencies.
Chaum’s proposal makes the VVPAT advocated by many a reality. While not the only method available to solve the problems of security and secrecy with regard to VVPATs, Chaum’s proposal has been lauded by computer scientists and voting advocates.
Hard-core opponents of DREs contend that even a VVPAT is unreliable, since hidden programming could cause the machine to print one ballot record while recording another. However, combined with open source code that reveals the inner workings of the voting machine, an encrypted VVPAT would provide much improved security and integrity to current DREs. Furthermore, voter confidence would substantially be improved by the simple act of a receipt, assuring the voter that his or her ballot was properly received.
While state retain the right to choose the voting machines that will be used within its borders, the standards for voting machine certification should be codified at a federal level. Stronger federal regulation of the voting machine industry in the form of stricter certification procedures would alleviate many concerns. HAVA’s disability requirement led many states to rush into purchases of DRE machine which were long on bells and whistles, but short on reliability. Now that states are decertifying certain machines, the money that was invested in DREs was essentially wasted. The optical scan machines many states are now selecting offer a VVPAT in the form of the ballot filled out by voters, but face software concerns similar to those in DREs. Additionally, optical scan machines are manufactured by the same vendors as DREs, creating the same issues of proprietary code. Federal requirements beyond the minimum HAVA standards are required to enhance voting security. New certification requirements should include:
Increasing the minimum federal standards for voting technology will reduce the burden on states to guarantee machine accuracy, while at the same time preserving state ability to choose its preferred type of system. By raising the bar required before vendors and machines can be certified, federal standards can improve voter confidence in the voting process and provide better functioning equipment.
An accurate vote count is essential to the survival of democracy. While technology can offer efficiency and standardization that can enhance the voting process, this efficiency must not come at the expense of the right to vote. All citizens are promised the right to vote and the right to an accurate vote count. The rush to embrace new technology should not compromise those rights. The potential for malfeasance and the appearance of illegitimacy present in the current generation of DREs can and must be remedied through better legislation and more focused use of technology.
Most importantly, proponents of DREs must
earn public confidence in the integrity of the process. While manufacturers are confident their
machines are reasonably accurate and fair, the only confidence that matters is
that of the voters using the machines to cast their ballot. For voters, issues
of proprietary technology and open-source code are less important than the
trust that must be placed in those administering the election. In writing about
“For many of us, the way e-voting works is as mysterious as the way images of SpongeBob SquarePants travel through the ether and onto our TV screens. The key difference is that we don't bank on SpongeBob to lead us out of war, balance the national budget or address our social ills.”
Merging technology with the fundamental right
of voting can only be successful if the public believes the process works. The public
will accept technology it does not understand as long as that technology is
proven reliable and verifiable by independent experts. In order to create voter trust, elections
must be transparent. Unless states
implement measures to make DREs more transparent and reliable, it is likely the
diminishing voter confidence will plague American elections for several
 Tom Stoppard, Jumpers, Act I, (Grove Press 1971).
 In 2004,
28.9% of the registered voters in the
 Eric Fisher, Voting Technologies in the United States, CRS Report for Congress, March 21, 2001, http://www.cnie.org/nle/crsreports/risk/rsk-55.cfm.
 Bush v Gore 591 U.S. 98 (2000).
 See “Election Reform Since November 2000,” supra note 1.
 Henry E. Brady et al., Counting All the Votes: The Performance of Voting Technology in the United States 10 (2001), http://ucdata.berkeley.edu/new_web/countingallthevotes.pdf.
 Marshal Camp, Bush vs. Gore, Mandate for Election Reform, 58 NYU Ann. Survey Am. L. 402 (2002).
 Fisher, supra note 2
 Robert Caro, Means of Ascent (Random House 1990).
 See “Election Reform Since November 2000,” supra note 1
 Fisher, supra note 2.
Post, “Uniform Voting Machines Protect the Principle of
 Election Reform Since November 2000,” supra note 1.
 Fisher, supra note 2.
 Election Reform Since November 2000,” supra note 1.
 Post, supra note 12.
 Fisher, supra note 2.
 Election Reform Since November 2000,” supra note 1.
 Coney , supra note 1.
 Some evidence has been offered to suggest the use of DREs can increase voter turnout, particular among young people who view the technology as “cool.” See Alternative Ballot Techniques: Hearing Before the House. Subcomm. on Elections of the Comm. on House Admin., 103rd Cong. (Sept. 22, 1994).
Report on the 2004 Presidential Election (Academy
 Caltech/MIT Voting Tech. Project, Residual Votes Attributable to Technology 4 (Mar. 2001), available at http://www.hss.caltech.edu/%7Evoting/CalTech_MIT_Report_Version2.pdf.
Jones, “Statement regarding the optical
mark-sense vote tabulators in
 Harvard University Civil Rights Project “Democracy spoiled,” (2004) available at http://www.civilrightsproject.harvard.edu/research/electoral_reform/ResidualBallot.pdf
 Charles Stewart, ‘Residual Vote in the 2004 Election,” CalTech/MIT Voting Technology Project, February 2005, available at http://www.vote.caltech.edu/media/documents/vtp_wp21v2.3.pdf.
 Jake Tapper, Down and Dirty : The Plot to Steal the Presidency, (Time Warner 2001).
De Tocqueville, Democracy in
 Tesitmony of Dr. David Dill, Carter Ford Commission, April 18th, 2005, available at http://www.verifiedvotingfoundation.org/article.php?id=5987.
 NBC News/Wall Street Journal Poll, Study #6050, p. 11 (question 8 ) , available at http://online.wsj.com/public/resources/documents/poll20041217.pdf (last visited June 13, 2005).
Report on the 2004 Presidential Election (Academy
 Fisher, supra note 2.
 Dill, supra at note 40.
Ornstein, Remarks to the Commission on Federal Election Reform,
Hansen, “Beyond the Margin of Litigation: Reforming
 Camp, supra note 7.
 Bush, supra, note 2, quoting McPherson v. Blacker, 146
 National Election day set as first Tuesday after the first Monday in November, codified in 2 U.S.C. 1 § 7.
 See Camp, supra, note 10.
 Bush, supra note 10.
 See e.g. Harper v. Virginia Bd of Elections, 383 US 663 (1966).
 See Camp, supra, note 10.
 Bush v. Gore, 531
 Id at 105.
 Id at 106.
 Id at 105.
 See e.g. Stewart v. Blackwell, 356 F. Supp. 2d 791.
 Tom Stoppard, Jumpers, Act I, (Grove Press 1971).
 Gray v. Sanders, 372
 Edwin Chemrisinsky, How we Should think about Bush vs. Gore 34 Loy. U. Chi. L.J. 1, 17 (2002).
 Bush supra note 10.
 Edmund S. Sauer, note, "Arbitrary and Disparate" Obstacles to Democracy: The Equal Protection Implications of Bush v. Gore on Election Administration” 19 J. L. & Politics 299 (2003).
 See E.g. Common Cause v. Jones, Black v. McGuffage.
 Wexler, et al v. Lepore, et al. 342 F. Supp 2d 1097 (S.D. Fla) (2004).
P. Tokaji, Law and Democracy: Early
Returns on Election Reform: Discretion, Disenfranchisement, and the Help
 Lillie Coney, note, Legislative Reform: A Call for Election Refor” Journal of Law & Social Challenges 7 J.L. & Soc. Challenges 183 (2005).
 42 U.S.C. 15481 (2005).
 Tokaji, supra, note 38.
 Conny B. McCormack, Los Angeles County Registrar-Recorder, Testimony before the California State Senate, January 18, 2006.
 42 U.S.C. 15481 (2005)
 In 2000, Punch Card machines accounted for 30.76% of the voting machines in use, while level machines accounted for 17%. By January 2006, those number had dropped to 4.75% and 11.20%, respectively. Election Online, “Election Reform Since November 2000, available at http://www.electionline.org/Portals/1/Publications/2006.annual.report.Final.pdf.
 Election Online, “Election Reform Since November 2000,” available at http://www.electionline.org/Portals/1/Publications/2006.annual.report.Final.pdf (2006)
breadth of DRE use varies from state to state.
 “Votes won’t count in Carlisle Precinct,” Middletown Journal, November 17, 2005, available at http://www.middletownjournal.com/news/content/news/stories/2005/11/17/mj1117carlisleelection.html.
“Incumbents Return in a
 Voters Unite, Report, Myth Breakers: Facts About Electronic Elections,
 Fisher, supra note 2.
 “Election Whistle Blower Sytmied by Vendors,” Washington Post, March 26, 2006, pA7.
Michael A. Carrier, Vote Counting,
Technology and Unintended Consequences,
 Open Letter from Governor Bill Richardson, http://billrichardson2006.com/reform/paperballots/index.html
 Election Whistle Blower Sytmied by Vendors,” Washington Post, March 26, 2006, pA7.
According to CNN Bush won
 “State Rebuffs Raw Vote Demand,” Anchorage Daily News, January 24, 2006 available at http://www.adn.com/news/alaska/story/7386582p-7298824c.html.
 Letter from Jake Metcalfe, Alaska Democratic Party, to Whitney Brewster, Commission of Alaska Division of Elections, Jan. 23, 2006 http://www.bradblog.com/Docs/AlaskaRecordsRequestResponse_012306.doc.
 The program and data files needed to "build your own GEMS box" (including GEMS itself, sample MDB/DBF data files and instructions on loading/testing them) can be viewed online at http://www.equalccw.com/dieboldtestnotes.html. This web site is owned by Jim" March, a computer technician and analyst with Black Box Voting, a non profit organization that monitors problems with electronic voting across the nation.
 Letter from Jake Metcalfe, Alaska Democratic Party, to Whitney Brewster, Commission of Alaska Division of Elections, Jan. 23, 2006 http://www.bradblog.com/Docs/AlaskaRecordsRequestResponse_012306.doc
 Fisher, supra, note 67.
Carrier, supra note 72.
 Fisher, supra, note 67.
 “Election Whistle Blower Sytmied by Vendors,” Washington Post, March 26, 2006, pA7.
 Massey, supra, note 60.
S. Evans, Bernard J. Reddy, note GOVERNMENT
PREFERENCES FOR PROMOTING OPEN-SOURCE SOFTWARE: A SOLUTION IN SEARCH OF A
 Massey, supra, note 59.
 Fisher, supra note 67.
 Fisher, supra note 67.
 Bush, supra, note 10.
 Massey, supra note 59.
 Massey supra note 59.
 Dacid Chaum, “Secret Ballot Receipts: True Voter-Verifiable Elections,” IEE Computer Socierty, 2004, http://crypto.csail.mit.edu/~rivest/voting/papers/Chaum-SecretBallotReceiptsTrueVoterVerifiableElections.pdf.
See e.g. Electronic Vote foundation report, available at http://www.electronic-vote.org/introduzione_en.php.
 Dick Rogers, “Retaining faith in our democracy” San Francisco Chronicle, December 9, 2004 available at http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2004/12/09/EDGSVA88PB1.DTL .