Posted: November 14, 2006
Local Blackouts on the Electoral GridThe early returns declared that the nation’s new electoral system to have operated effectively—“relatively problem-free,” in the words of one Wednesday morning headline. Doug Chapin, who directs the deservedly prestigious electionline.org, was quoted as saying “fender benders but no major tie-ups.” On Sunday, the Washington Post editorialized: “The System Worked.” My assessment is a bit different. To be sure, there were many success stories. Maryland did not have the meltdown that occurred during its September primary. Likewise, Cuyahoga County, Ohio, did not repeat the severe problems that crippled its own primary, in May. Indeed, one important lesson from November 7, 2006, is that local election administrators usually are able to avoid the problems that caused a catastrophe in their own immediately preceding election. The administrative challenge, however, is to get localities that have not recently experienced system failure to implement or maintain effective safety mechanisms to prevent their jurisdictions from becoming the next ones to break down. The most significant successes occurred, of course, in Virginia and Montana, where extremely close results in their U.S. Senate races were quickly accepted by the losing candidates, thereby avoiding recounts that would have prevented a determination of which party controlled the Senate itself. When Senator Allen conceded defeat, he was reportedly behind by only about 8000 votes, or approximately one-third of a percent. Similarly, Senator Burns stepped aside trailing by only about 3000 votes, which in his smaller state amounted to three-quarters of one percent. In both cases, the lack of a basis on which to challenge an outcome this close means that the state’s voting system easily passed my Election Validity Test. Moreover, it may still turn out that most of the prominent races that still remain undecided end up passing this Election Validity Test as well. As I explained last week, this test gives local officials ample time to thoroughly and carefully count—and if necessary recount—all the votes. Only one week after the casting of ballots is complete, it is still early on in this official counting process, as distinguished from the unofficial returns of election night and the next morning, which are useful only in races that turn out not so tight. Thus, for example, there is no necessary reason to think that the race for Washington’s 8th congressional district will flunk the Election Validity Test, just because it is taking a long time to count ballots there as a result of flooding and other difficulties. As long as both candidates ultimately agree, after the counting is complete, that the result is fair and accurate, then the election will have been a success despite the delay. (The allegation that some bags of absentee ballots there burst open because they were stuffed too full is a bit troublesome, but apparently all ballots remained in a secure chain of custody, in which case there would be no basis for attacking these ballots as corrupted.) The same point applies to Wyoming’s single congressional district and Georgia’s 12th. Even though the leading candidates in both races have margins of about half a percent, the counting of provisional and absentee ballots is proceeding apace, and there is no reason to think that the losing candidate will contest the result. Even in Connecticut's second congressional district, where the margin stood for a while at a mere 60-some votes, or about one-tenth of one-quarter of one-percent (0.024%), there does not yet appear to be any basis for challenging the ultimate result as invalid. (The margin climbed back up to 109 by the end of yesterday.) And if the final certified margin of victory dips below one hundredth of one percent 0.01%), or 25 votes out of a total of 250,000, then my Election Validity Test would deem the race a statistical tie and no longer expect the state’s electoral system to produce an outcome that both candidates accept as accurate. As of this writing, however, there appears to be one congressional race that will be unable to satisfy the Election Validity Test: Florida’s 13th. Whatever the cause of the 18,000 undervoted ballots in Sarasota County, whether bad ballot design or machine failure, it would seem difficult for supporters of the losing candidate to accept the outcome as the product of a fair process, especially if a statistical analysis strongly shows that the losing candidate likely would have prevailed had the machines and/or ballot design employed in other counties also been used in Sarasota. This point does not mean necessarily that a court should invalidate the election’s result. Rather, even if the result must be accepted, the conclusion may remain that the system failed to permit voters to register a choice between the competing candidates in such a way that the result is a fair and accurate reflection of their choice. There are two other unsettled congressional races that conceivably could end up failing the Election Validity Test, although in each case it appears unlikely at this point. Both of these races, New Mexico’s 1st and Ohio’s 15th, have involved preliminary courtroom squabbling over the procedures for counting provisional ballots. Therefore, it is possible that partisan disputes down the road over the eligibility of particular provisional ballots will become determinative of which candidate wins, with the consequence that the losing candidate refuses to accept the legitimacy of the opponent’s victory. But the number of provisional ballots in these two races is relatively small in comparison with current margins by which the leading candidates are ahead, and therefore this end-game scenario seems only a remote possibility right now. But even if it turns out that Florida’s 13th congressional district is the only federal or statewide election to fail the Election Validity Test this year, there remains cause for concern. While the Election Validity Test measures a system’s ability to determine which candidate is the rightful winner—thus identifying arguably the most important attribute by which an election system can be judged—this test does not attempt to determine the degree of voter disenfranchisement that may occur in any given election. It is possible for there to exist widespread disenfranchisement of eligible voters in a particular election, and yet for the electoral system still to be able to identify the rightful winner. The margin of victory just needs to exceed the number of disenfranchised citizens. That’s what occurred, in my judgment, in the presidential election of 2004 in Ohio. A large number of voters in Franklin County and elsewhere in the state were effectively turned away at the polls because of excessively long lines in 2004, but I do not believe that the number of voters who went to the polls attempting to cast an eligible ballot yet were unable to do so exceeded President Bush’s 118,601-vote margin of victory over Senator Kerry. But to disenfranchise thousands of eligible voters is no minor matter, even if doing so does not prevent the system from identifying the rightful winner. For this reason, most observers were outraged by what occurred in Ohio in 2004, whether or not they went so far as to think the malfeasance of the state’s electoral system deprived Kerry of a victory that was rightfully his. Yet anyone outraged by Ohio 2004 should be equally outraged by Colorado 2006, from what has appeared so far. Just as excessively long lines prevented untold number of voters from casting their ballots in Ohio two years ago, so too widespread lines of over two hours in length—sometime lasting over 4 hours—prevented an unknown number of voters in Colorado from casting their ballots this year. While the problem was most acute in Denver last Tuesday, just as the problem was most prevalent in Columbus in 2004, excessively long lines were reported elsewhere in Colorado in the same way that they were reported elsewhere in Ohio. The only reason not to care about voter disenfranchisement in Colorado this year, in contrast to the voter disenfranchisement in Ohio in 2004, is that this year’s disenfranchisement did not have the potential to affect the outcome of a presidential election. But the issue right now is not the Electoral Validity Test, rather only the magnitude of the harm to individual voters caused by their inability to cast a ballot when they went to the polls. In this respect, the harms in Ohio 2004 and Colorado 2006 seem exactly the same, assuming that the number of voters affected is equal (which is difficult to judge from this vantage point so soon after this year’s vote). Moreover, there was a gubernatorial election on the ballot in Colorado this year (it just wasn’t that close), and the disenfranchisement in Denver and elsewhere may still end up affecting the outcome of the Secretary of State election. Looking ahead, as the Washington Post did yesterday, Colorado may turn out to be a swing state in 2008. Thus, even if one wants to focus solely on how a state’s election machinery may affect the outcome of a presidential election, without paying any attention to the civil rights perspective of disenfranchised citizens, there is currently just as much a reason to worry about Colorado’s electoral infrastructure as there is about Ohio’s. For this reason, I do not believe that there is such a night-and-day contrast in the quality of election administration between 2004 and 2006 as the headlines, editorials, and other commentary during this past week would suggest. The prevailing wisdom of the past week is that this year the operation of the nation’s election machinery should be judged essentially a success, whereas 2004 was essentially a failure—although we were lucky then to dodge the proverbial bullet. As I see it, however, the two years are much more similar than that. If Colorado’s calamity ends up being irrelevant to the outcome of any race there, then that state would appear to have dodged a bullet in the same way that Ohio did in 2004. Using the total number of voters disenfranchised as a result of election administration as a yardstick, I do not have sufficient information at hand to know whether 2004 or 2006 was worse, either when considering solely Ohio 2004 and Colorado 2006 as a side-by-side comparison, or looking at the nation as a whole in both years. There was disenfranchisement of voters as a result of long lines in states other than Ohio in 2004, but there was also disenfranchisement of voters as a result of long lines in states other than Colorado this year. It would be interesting—and perhaps sobering—to see the result of a scientific study that attempted to measure as precisely as possible the number of eligible voters nationwide who were deprived of their right to cast a countable ballot as a result of election administration failures. I would be pleased to see that a comparison of 2004 and 2006 according to this metric showed the nation moving in the right direction, by reducing the number of voters disenfranchised as a result of administrative failures. But from what I have been able to ascertain so far about the overall operation of the electoral process last Tuesday, I would not be altogether surprised if the evidence went the other way.
Edward B. Foley is Director of the Election Law @ Moritz program. His primary area of current research concerns the resolution of disputed elections. Having published several law journal articles on this topic, he is currently writing a book on the history of disputed elections in the United States. He is also serving as Reporter for the American Law Institute's new Election Law project. Professor Foley's "Free & Fair" is a collection of his writings that he has penned for Election Law @ Moritz. View Complete Profile