Posted: September 18, 2007
What Congress Should Be Talking About
The U.S. House of Representatives may soon resume consideration of H.R. 811, a bill that would require voting machines to generate a "voter-verified" paper record. This bill is the outgrowth of a bitter and protracted debate over the security of direct record electronic (DRE), which has consumed considerable time and attention across the country over the past few years. Even among the fiercest critics of DRE technology, there is persistent disagreement over whether the current version of the bill will really improve electronic voting security.
More distressingly, the current Congress' preoccupation with electronic voting continues to serve as a distraction from much more serious problems with the administration of elections, including problems in how voter registration is being handled in the states. Rather than squandering further time and energy on the electronic voting debate, Congress should turn its attention to improving registration, enhancing participation, and collecting better information if it really wants to ensure a fair electoral process in 2008. In this comment, I briefly canvass the current debate over electronic voting, and then turn to the more pressing election administration issues that Congress ought to be considering.
The Electronic Voting Debate
Prompted by Florida's 2000 election and the Help America Vote Act of 2002 (HAVA), new voting equipment has been implemented throughout the country in the last few years. Few dispute that changes were sorely needed, given the high number of votes lost with punch card voting equipment that was prevalent before HAVA. There has, however, been enormous controversy over the new technology, especially DRE machines. This equipment has the advantage of reducing the lost votes which resulted from the use of punch cards and other paper-based systems. By providing voters with notice and the opportunity to correct errors, DREs reduce the number of unintentional undervotes and eliminate overvotes. This equipment also makes it easier to accommodate non-English proficient voters, and can allow many people with disabilities to vote independently.
Electronic voting has nevertheless become the bete noire of some activists, mostly those on the left side of the political spectrum. DRE critics argue that the machines are insufficiently secure, leaving open the possibility that software could be manipulated to "steal" an election. There have also been well-publicized problems with the implementation of DRE machines in some jurisdictions. The most notable was in Sarasota County, Florida, where there were over 18,000 undervotes in the 2006 election for Florida’s 13th Congressional District.
According to some activists, the solution is to require that all voting equipment generate a "voter-verified" paper record. For electronic voting machines, the most common configuration is a reel-to-reel printer attached to the touchscreen voting unit, on which the voter’s choices are contemporaneously printed as the voter votes. In theory, these printouts could be verified by the voter and later audited to ensure the accuracy of the electronic voting totals. In practice, it is questionable whether voters will actually check the paper record -- hence the quotation marks around the term "voter verified." There are also serious issues of ballot integrity associated with the printed paper records, as well as privacy concerns recently reported here.
All of these issues make contemporaneous paper records a dubious solution to the security concerns associated with DREs. In fact, some DRE critics -- most notably the N.Y. Times editorial page -- have shifted away from the oft-repeated argument that a mandatory paper trail is the solution (see here and here) and now call for a total ban on touchscreen DREs.
This is not the place to engage in an extensive discussion of the paper trail debate, as I have previously done in this article and in prior weekly comments (see here and here). It should be noted, however, that electronic voting is not the greatest problem facing our democracy. In fact, the transition to new voting technology is one aspect of election reform that has really worked. According to one study, one million lost votes were avoided in 2004 due to improvements in voting technology and administrative procedures. That same study found that counties moving from punch cards to DREs saw the greatest improvement in lowering the number of uncouted votes.
None of this is to deny that there are legitimate concerns associated with new voting technology. But at this point, the best solution is far from clear, and very unlikely to be achieved through federal legislation. County governments are strongly opposed to the current bill, given the massive changes in voting equipment that it would require. Many e-voting critics are also skeptical, given that one of the more promising means by which to address their concerns -- a strong requirement that source code be disclosed for testing -- is absent from the current bill. As demonstrated by the recent experiences of California and Ohio, which were among the first to pass paper-trail laws, this requirement will not resolve the debate over voting technology.
With election officials strongly opposed to the current federal bill and voting activists divided, it appears unlikely that we will see productive voting technology reform coming from Congress anytime soon. As a practical matter, it's not realistic to require election officials to procure new technology for next year's elections. And even if it were possible to make major technological changes in time for 2008, it is far from clear what direction that reform should take. At least for now, these issues are best left to the states.
The Debates That Aren't Happening
Congress' preoccupation with electronic voting would be less costly, were there not other pressing issues of election administration desperately in need of attention. If left unaddressed, these problems could result in systematic inequalities that impede many voters from participating in the 2008 election. Here are three issues to which the current Congress should be devoting attention right now, if it's serious about improving election administration in 2008 and beyond.
1. Statewide Registration Databases
One of the most important changes mandated by HAVA was that each state have in place a computerized statewide voter registration list. A few states, including Michigan, had statewide voter registration lists in place before HAVA that worked well. In other states, voter registration lists were maintained at the local level -- and often had serious problems. According to an influential report by the Caltech/MIT Voting Technology Project, registration glitches were probably a bigger source of lost votes than bad voting equipment in the 2000 election. The idea behind HAVA's statewide voter registration list was to improve the quality of lists and to allow for address changes to be more easily tracked when voters moved across county or municipal lines. HAVA also included a requirement that voters' information be "matched" against motor vehicle and social security records.
While the idea behind the statewide registration database requirement is commendable, the transition has not gone smoothly everywhere. As summarized in this table, many states are not yet in full compliance with HAVA's requirements, as they were supposed to be by 2006. More disturbing is the possibility that state "matching" practices may result in some eligible voters being wrongfully stricken from the rolls. On Monday of this week, the Brennan Center for Justice and other voting rights advocates filed suit in Florida, arguing that the state's overly strict matching criteria may wind up excluding up to 30% of voters due to typos and other data entry errors. Latino voters who use both maternal and paternal surnames are especially at risk. If Congress is really intent on preventing the systematic exclusion of voters in 2008, it should pay close attention to how states are implementing HAVA's statewide registration database requirement.
2. Barriers to Registration and Participation
Since 2004, a number of states have moved to impose more stringent requirements as part of the registration or voting process. A voter-enacted initiative in Arizona requires proof of citizenship in order to register and vote, driven by the spectre of noncitizens fraudulently participating in elections. In Georgia, Indiana, and Missouri, state legislatures adopted laws requiring voters to provide government-issued photo ID in order to vote, despite arguments from voting rights advocates that such requirements would disproportionately burden poor, minority, disabled, elderly, and college-age voters. In Ohio, the legislature declined to adopt a photo ID requirements, but instead enacted a confusing identification law that leave voters and poll workers alike unsure as to exactly what must be done in order to have one's vote counted.
There's precious little evidence of voting fraud to justify these laws. It's extremely rare for voters to appear at the polls pretending to be someone they're not, the only form of fraud that an ID law would address. There is, on the other hand, a significant risk that such laws will disproportionately exclude some groups of eligible voters. By conditioning voting on identification that some voters do not have -- and that certain groups of voters are much more likely not to possess -- the composition of the electorate is likely to be affected. Even in states with less stringent requirements, the confusing patchwork of rules for registration and voting make it more difficult for all voters to participate in the electoral process.
The time has come for Congress to consider imposing clear and uniform requirements for participating in federal elections. Such requirements should be designed to preempt the unduly restrictive, exclusionary, and confusing ones adopted by the states noted above. In the long run, the best solution might be to couple uniform identification requirements with election-day registration (EDR), which has now been adopted in eight states with excellent results. This is a reform that could at once increase participation while enhancing the integrity of our electoral system.
3. Bad Information from the States.
For those seeking to understand the administration of American elections, one of the most serious problems is the absence of reliable data from the states. As part of HAVA, Congress created the U.S. Election Assistance Commission (EAC), which is supposed to serve as a clearinghouse of information from which better election policies might emerge. One of the things the EAC did in 2004 and 2006 was to survey states on various aspects of their election system, including voting technology, registration, provisional voting, and disability access. Unfortunately, the quality of information received from the states in 2004 was poor. Many states provided incomplete or inaccurate data. At present, neither the EAC nor any other agency of the federal government has the tools to make the states provide better information.
This is a serious problem for both researchers and reformers. Without accurate information, it is difficult if to know where the most serious election administration problems lie. In fact, it is hard to know whether state and local entities are complying with existing laws such as HAVA and the National Voter Registration Act of 1993 (NVRA). Without adequate information, it is impossible to diagnose problems, much less prescribe appropriate remedies.
To deal with the poor quality of election administration information, Thad Hall and I have suggested a money-for-data swap. States would receive needed federal funding and, in exchange, would be required to provide accurate and complete data on the administration of their elections. Under our proposed carrot-and-stick approach, those that provide this information would receive monetary benefits; those that fail to comply would face monetary sanctions. Most recently, the Midwest Democracy Network -- a coalition of political reform advocates -- issued a questionnaire to presidential candidates challenging them to support federal funding for election administration in exchange for better information from the states (Question IV-A) .*
It may be possible to come up with other means by which to improve the quality of information on election administration. What cannot be seriously questioned is that better data is a necessary precondition for future election reform. Now is the time for Congress to focus on what can be done to get better information on election administration in 2008.
These three items are not meant to be an exclusive list. There are undoubtedly other things that Congress could do to prepare the 2008 election season and pave the way for long-term improvements in the administration of elections. Instead, Congress remains mired in an electronic voting debate that is unlikely to result in any real improvements in our election system, but at best to provide a placebo for activists who view electronic voting as the problem and paper as the answer. This may score some members of Congress points with their base, but it will do nothing about the more serious barriers to a fair democratic process.
* Disclosure: I have attended meetings of the Midwest Democracy Network and consulted in the preparation of this portion of its questionnaire.
Dan Tokaji is an authority on election law and voting rights. He specializes in election reform, including such topics as voting technology, voter ID, provisional voting, and other subjects addressed by the Help America Vote Act of 2002. He also studies issues of fair representation, including redistricting and the Voting Rights Act of 1965. View Complete Profile