Now that Republicans in Washington State have chosen not to appeal yesterday's decision in their contest of the gubernatorial election, many citizens there may share the sentiment expressed by the ebullient victor, Gov. Christine Gregoire: "I am happy for the state of Washington that we can finally move on here." Meanwhile, citizens of other states haven't paid much attention to either the two week trial before Judge John Bridges of Chelan County Superior Court or the months of pre-trial proceedings in this contest. Yesterday's decision did not make front-page news in the New York Times or Washington Post or even, perhaps surprisingly given its geographic proximity, the Los Angeles Times.
But Washington 's experience with this election, and particularly the litigation that followed it, raises many important questions for its 49 sister states. Its lessons are at least as significant as those from the 2004 presidential election in Ohio. After all, unlike Ohio, Washington went to court: the margin of victory, 129 votes out of more than 2.8 million cast, was well within the now-proverbial margin of litigation. Indeed, when Gregoire's victory was announced on the penultimate day of December - already essentially two months after Election Day - it was perceived by Republicans as well worth fighting over, especially because the Republican candidate had been the winner of the first two counts, conducted by machine, and Gregoire prevailed only in the third count, which was manual. More than five months later, and a full seven months after Election Day, with a judicial ruling against them, the Republicans have now abandoned their effort to invalidate the results of the manual recount. By declining to appeal, have the Republicans signaled that in hindsight the litigation shouldn't have occurred in the first place?
The results of the trial, although interesting in many ways, largely confirm what was known in advance of the trial or, indeed, in January when the contest was filed. Judge Bridges found that there were 1678 illegal votes cast in the election:
This finding is obviously significant: there were 13 times as many illegal votes as the margin of victory (again, 129). Indeed, there were almost twice as many provisional ballots that should not have been included in the final certified count, 252, as the margin of victory. (There is some confusion in Judge Bridges's oral decision concerning these improper provisional ballots: when stating his Conclusions of Law, he says that 96 were from King County, 79 from Pierce County, and an additional 77 from Pierce County; earlier, when setting forth his Findings of Fact, he more clearly explains that 96 from King County were improperly fed through machines without checking, 79 also from King County that lacked appropriate labels and were inadequately verified, and 77 from Pierce County. In any event, the total number of improperly included provisional ballots amounts to 252.)
The problem with these illegal votes, as has been apparent from the outset, is that it is unknown - and, given the secret ballot, inherently unknowable - for which candidate each illegal vote was cast (or, indeed, whether the illegal ballot did not mark a vote for Governor, confining itself to other races in the election). Except for five felons, who testified for whom they voted, there was no evidence directly linking a particular illegal ballot to a preferred candidate. Perhaps the court could have called all the felon voters to the stand, as well as all the double voters and the identifiably unregistered provisional voters, but asking individuals how they voted is unreliable. They have an incentive to dissemble. If a felon voted for Gregoire and want her to prevail in the contest, he'd be inclined to say that he voted for her Republican opponent, so that throwing out his illegal vote wouldn't make any difference to the result. As a felon, his testimony even under oath might not be so reliable and, given the irretrievable secrecy of the ballot after having been cast, could not be impeached by documentary evidence to the contrary. The same point would apply to any individual who voted twice, or one who voted without ever registering. And, of course, when there is no available evidence concerning who voted on behalf of a dead person, it is impossible to call that witness to the stand, as is also true when the identity of the individual who cast a provisional ballot cannot be ascertained.
Given this problem, the legal system has three basic options, as has been clear since this contest was filed. First, the judiciary could throw out the results of the election simply on the ground that the number of illegal votes exceeded the margin of victory, thereby indelibly tainting the result, since no one can be sure that these illegal votes didn't contribute to the victory. Second, the judiciary could dismiss the contest on the ground that, conversely, the contestants could not show that the illegal votes ultimately made the difference in the outcome of the race. Third, the judiciary could attempt, by some statistical method, to apportion the illegal votes among the candidates, with the consequence that the election would be invalidated if the margin of apportioned illegal votes exceeded the margin of victory (i.e., if this apportionment method assigned to Gregoire over 129 more illegal votes than it assigned to her Republican opponent).
Facing this choice, Judge Bridges squarely and emphatically selected the second option. He categorically rejected the idea that a court would invalidate an election simply on the showing that there were more illegal votes than the margin of victory. He equally firmly ruled out the notion that a court would attempt to use statistical methods to apportion illegal votes among the candidates, although he also added that that Gregoire still would have won even if he had done so. In light of this last finding, this case turned out to be unsuitable for testing on appeal the proposition that a court should adopt the apportionment approach when doing so would make a difference in the outcome. And in light of precedents from the Washington Supreme Court in previous election contests, it was highly unlikely that the high court would have, contrary to Judge Bridges, embraced the first option - that a contestant need show only more illegal votes than the margin of victory. Therefore, the Republicans wisely chose not to appeal.
Even so, the conclusion of the case remains somewhat unsettling. Suppose there had been 16,780 illegal votes, rather than 1678. Suppose further that it was fairly clear that these 16,780 illegal votes most likely affected the outcome: perhaps they largely were improper provisional ballots from urban counties that voted heavily Democratic. Or, conversely, in a different election in which a Republican candidate comes out on top by an extremely slender margin, suppose that a large number of improper provisional votes from heavily Republican rural counties appear to have made the difference. In these situations, do we really want the judiciary to say that there is no remedy for the illegal votes? In other words, should the state and its citizens be stuck with the tainted result even though no one can be sure that the illegal votes made the difference although it seems quite obvious that it did? In such circumstances, the apportionment approach might prove to be the least bad of three unpalatable options.
Perhaps, then, the main lesson from Washington's experience is that the 1678 illegal votes, while significant and regrettable, were not large enough to set aside the results, especially when it did not appear that these illegal votes were consequential. It is undeniably true, as it is often said, that there never will be a perfect election, and in the rare case in which the margin of victory is inside the margin of unavoidable error, perhaps the winner gets to take advantage of good luck. The losing candidate may be able to live more easily with the short end of the stick when, as Judge Bridges ruled, there is no evidence that the illegal votes were a consequence of deliberate fraud or misconduct by election officials or partisan groups.
I tentatively suggest, however, that the most significant lesson that emerges from this Washington contest may concern the timing of the litigation. Five months after Gov. Gregoire was inaugurated and had begun to exercise the responsibilities of her office, including the successful completion of an initial legislative agenda that by many accounts demonstrated impressive leadership, there was little stomach among citizens or other public officials for kicking her out of office. The fact that an exceedingly narrow victory was tainted by non-fraudulent and probably non-determinative errors of election administration seemed insufficient reason to unseat a Governor who already was well under way in implementing her policy agenda. Had the Secretary of State announced on December 30, however, that he was unable to certify any winner in the election because the 129-vote margin of the third count was compromised by 1678 illegal votes that could not be attributed definitively to any candidate, the public likely would have preferred a quick reelection, with an acting Governor in place temporarily, rather than proceeding with the inauguration of a candidate unable to claim a clear-cut victory.
This point indicates that, however a state chooses to handle the problem of indeterminate illegal votes that exceed the margin of victory in an election, it should endeavor mightily to resolve the matter before the winning candidate takes office. In hindsight, it would be preferable for a state to have in place a procedural mechanism, including whatever expedited judicial review it wishes to provide, to identify before Inauguration Day the number of illegal votes and to determine what to do about them. Ideally, of course, felons would not have voted in the first place; nor would there be mistakes in the processing of provisional ballots. (Indeed, ideally, there would be no need for provisional ballots to process.) But, again, recognizing the impossibility of perfection on Election Day, there ought to be a way to arrive at a conclusive audit of the election results before winner takes office, over two to three months later.
It is true that the possibility always exists that some further error will be unearthed after inauguration. Indeed, it is conceivable that more errors in connection with this 2004 governor's race in Washington will be uncovered subsequently, even after the completion of this exhaustive trial. We still don't know why so many votes, particularly absentee ballots, cannot be attributed to specific voters. Judge Bridges recognized the serious problem of discrepancy, but ruled that he could not conclude that these votes were illegal rather than just unaccounted for. If subsequent evidence emerges showing these absentee ballots to be duplicate or otherwise illegal votes, that would cast further doubt on the results of this election. But it would be too late, coming after the conclusion of this five-month contest.
The point, then, is that there is an inevitable need for finality and, given this need as well as the practical and legal significance of inauguration itself, it would be better if inauguration marked the concluding point for contesting an election based on the number of illegal ballots exceeding the margin of victory. To be sure, fraud is a separate matter, as Judge Bridges clearly recognized: the courthouse doors remain open even after Inauguration Day to potentially unseat a sitting Governor if the evidence convincingly demonstrates that the Governor obtained office only as a result of fraudulent voting. But the clear message of Judge Bridges's ruling yesterday is that, once a Governor has taken office and begun to govern on behalf of the people of the state, a court will not put an end to the Governor's tenure absent fraud, even when the number of illegal votes exceeds the margin of victory by more than 10-to-1. Since this kind of judicial restraint is to be expected, rather than attempting to rely on the illusory hope of post-inauguration litigation to address the problem of illegal votes, it would be preferable to develop a pre-inauguration procedure for deciding whether the number and character of illegal votes is sufficient to render an election irreparably defective and therefore require a return to the polls.