As afternoon turned to evening on Wednesday, Washington state held its collective breath as the final tallies of absentee and provisional ballots flowed into Olympia and out over the web. In the end, after 2,882,435 valid ballots were counted, Dino Rossi (R-gladhanding stealth wingnut) led Christine Gregoire (D-triangulating equivocator) by some 261 votes. That's a difference of 0.009% of the total.
I ran a series of day-by-day diary entries as the post-Election Day counting continued, discussing the results of each day's newly-tallied ballots and updating my projections of the final outcome based on each new set of ballots. Here are links to the entire collection:
Monday, November 8
Tuesday, November 9 (morning)
Tuesday, November 9 (evening)
Wednesday, November 10
Thursday, November 11
Friday, November 12
Monday, November 15
Tuesday, November 16 (early evening)
Tuesday, November 16 (night)
But, until now, I didn't write a report on the Wednesday results. Before moving on to the upcoming recount, I wanted to close the books on the first count of ballots.
Going into the final day, Rossi led by only 19 votes. Based on each county's estimated number of outstanding ballots, coupled with each county's to-date vote percentages for Gregoire, Rossi, and Libertarian Ruth Bennett, I projected that Rossi would eventually receive 519 more votes than Gregoire. That projection assumed that the estimate of 6577 remaining ballots to be tallied was accurate. Note that between Tuesday afternoon and Wednesday morning, King County's estimate rose from 1000 to 1450, explaining the discrepancy between the projection I made on Tuesday (a 599-vote margin) and what I present here. Going into the final day, there were estimated to be approximately the same number of uncounted ballots in counties where Gregoire led and those with Rossi leads -- the largest numbers were in King (1450, Gregoire), Grays Harbor (1221, Gregoire), Benton (750, Rossi), and Stevens (600, Rossi).
As the count progressed on Wednesday, Rossi's lead fluctuated until King County came in in mid-afternoon. At that point, Gregoire took a small lead, which was erased by some pro-Rossi counties. When the next-to-last county, Grays Harbor, reported its tally, Gregoire was once again in the lead, by 13 votes. Benton County, a heavily Republican area, came in last, reversing the lead once again.
For those of us looking at the day-to-day changes, it was very exciting. There was good news when, say, Columbia County turned in far fewer new votes than had been estimated -- getting 33% of 160 votes is a lot better than 33% of 285, though that report still increased Gregoire's deficit by 48 votes. Another seemingly-positive sign was that Gregoire outpolled Rossi for the day in Pierce and Spokane, and that both sent in more votes than they'd estimated; she picked up 24 votes there. She also came within one vote of Rossi's total in Clark County.
Of the "big four" mentioned earlier, Stevens County was the first to report -- 677 ballots rather than 600, and an even stronger Rossi margin (64.6% to 31.8%) than in previous tallies, netting 202 votes in his favor. Before the next of the four reported, there was a surprise from Skagit County, which reported 645 new votes, far above its estimate of 200 and also somewhat stronger for Rossi than in earlier results ... add 51 net votes to his edge. King County gave Gregoire a much higher percentage than she'd been receiving there -- 64.7% of the day's votes (she'd been at 57.7%), but turned in only 1446 new ballots. The net Gregoire margin for the day in King County was 446 votes.
Grays Harbor was, without a doubt, the best news of the day for Gregoire. Going into the day, her overall lead in the county was quite small -- 49.3% to 48.7%. On Wednesday, however, she got 51.3% in the county, to Rossi's 46.1%. Furthermore, Grays Harbor reported nearly twice as many new ballots as they had estimated -- 2060 in total, giving Gregoire a 105-vote net advantage.
Benton County finished it off with a flourish. Not only did the county report exactly the same 750 new ballots as they had estimated, they went for Rossi even more overwhelmingly than they had been doing all along. Rossi came into the day with a 68.2% to 30.1% lead in Benton, and then won the final day's tally with a crushing 69.2% to 28.0% outcome. Ugh. All in all, that netted him a 274-vote margin in the county, changing that 13-vote Gregoire lead into the eventual 261-vote margin for Rossi.
Because the difference exceeds 150 votes, the mandatory recount will be automated rather than manual. So the ballots will be fed through the same scanners once again. This one should be completed before Thanksgiving. After that, the Secretary of State will certify the result on December 2, at which point either campaign can request another recount. This time, however, the campaign would have to put down a deposit to pay for the recount -- $0.15 per ballot for another automated recount, $0.25 per ballot for a manual count. If the result is reversed, they'll be reimbursed. I'm fairly certain that any requested recount must be on the entire state, not county-by-county.
Several factors remain uncertain, at least in my mind. I think it's possible that Grays Harbor County, which discovered irregularities in its tally on Monday and then started over from scratch, may have missed some ballots. The county's turnout was only 77.1%, some 5% lower than the overall state figure, and also much lower than any of the other seven counties that favored Gregoire over Rossi. In general, turnout in the Gregoire counties was 1.5% higher than in Rossi counties, so it seems unusual that Grays Harbor was low. A more ubiquitous question is whether ballots arriving at county elections offices after Wednesday will be counted during the recount. Similarly, I wonder whether rejected provisional and absentee ballots will be re-examined ... if the voter added a missing signature to the outer envelope on Thursday, would that ballot then become valid?
I believe there will also be a mandatory automated recount in one State Representative race. In Position 2 of the 28th Legislative District (Pierce County, west of Tacoma), Dean Dozen Democrat Tami Green received 23,697 votes (50.249%), and incumbent Republican Bob Lawrence got 23,462 (49.751%). That 235-vote difference translates to a margin of 0.498%, just inside the 0.5% threshold triggering a mandatory recount. I feel duty-bound to report that had Green received one more vote, her margin would have been 0.503%.
When the gubernatorial recount gets started, I'll continue updating the dKos community on the race. And when this thing finally ends -- it will, eventually -- I hope to write a report on the bigger picture of the 2004 election cycle in Washington. I'm putting together a county-by-county voting database of the results in the presidential, senatorial, statewide, and state legislative races. I'm interested in (among other things) the county-specific differences in percentages garnered by Kerry, Murray, Gregoire, and other candidates; the effect of the Libertarians and other fringe candidates; the relative proportions of ballots left blank and/or with write-ins.
Much more to come...