UPDATE: Reliable sources tell me that King County took in 114,000 ballots this morning, now has about 300,000 unreported.
*****
Long-, long-, long-time Kossacks will recall that I was the go-to guy for discussion and analysis of the protracted recounts in the 2004 Washington gubernatorial race. In many ways, that Gregoire-Rossi contest was a tryout for the Franken-Coleman battle in 2008. I played the role of WineRev in the preview.
As you are surely aware, we have another squeaker on our hands in this year's Senate race, featuring (again) that smarmy, sleazy real estate speculator ... this time against feisty, fightin', liberal Patty Murray. Because it currently looks really tight, I'm prepared to help dKos take another venture through Washington vote-tallying.
I embark on this effort, however, in the anticipation that this won't be a particularly lengthy series. Patty's going to be declared the winner, and that declaration will happen in November 2010.
Washington elections are unlike any other state's.
I'm not referring to our ridiculous "top two" primary system. Its only relevance to the current situation is that it barred all potential candidates aside from Dino and Patty from appearing on the general election ballot. Washington joins neighboring Oregon as an (almost) entirely vote-by-mail state. I say "almost" because, for reasons that escape me, Pierce County still has polling places. According to the Pierce County Auditor, only about 29,000 poll votes were cast yesterday, a mere 7 percent of the county's 410,081 registered voters and less than one-sixth of the votes tallied (thus far) in Pierce.
The key phrase in the above paragraph is thus far. Washington differs significantly from Oregon in that it requires mail ballots to be postmarked, not received, by Election Day. Ballots arriving at county elections offices today will be counted. Ballots arriving tomorrow will be counted. Ballots arriving Friday will be counted. Due to the vagaries of the postal system, valid ballots will continue to arrive right up until, and maybe past, the counties' deadline for final counts, November 23.
A second point to consider is the immense range of characteristics among Washington's 39 counties. As many counties (eight) are under 8000 registered voters as are over 115,000 voters. King County's voting population is nearly 700 times Garfield County's 1537 voters.
Between the complexities of handling huge numbers of ballots and the attention it drew from the whisper-close recounts in 2004, King County has a very time-consuming procedure for verifying ballots. It generally takes a couple of days after receipt until a King County ballot can be tallied. Procedures in the other large counties -- Pierce (south of King ... Tacoma), Snohomish (north of King ... Everett), Spokane (eastern WA), Clark (bordering Portland OR ... Vancouver), Thurston (south of Pierce ... Olympia), Kitsap (west of King ... Bremerton), Whatcom (Canadian border ... Bellingham) -- slow down their counts as well. Except for Spokane, all of the large counties are on or very near the I-5 corrider. Except for Spokane and Clark, all of the large counties are on Puget Sound. Not surprisingly, the large counties (particularly those on the Sound) trend bluer than the small counties.
The upshot of all this is twofold:
- Democratic counts, especially King County's, are generally reported slower and later than Republican counts.
- What is shown on the very useful Secretary of State election website doesn't fully represent the status of the actual tally in large counties.
I'll get back to the latter point in a moment. First, though, a report on the way I'm looking at the results. Last night, I downloaded the couty-specific numbers compiled by the SoS. At that point, the tally was Murray over Rossi, 722,396-708,391 (50.5%-49.5%). It was also reported that the counties had 528,225 not-yet-tallied ballots.
My initial projection from this dataset involved applying the county-specific percentages from the tallied ballots to the reported number of untallied ballots, adding those results to the actual tally, and summing. To demonstrate the procedure, consider Adams County (first in the alphabetical list and in the middle of nowhere eastern Washington). It reported 758 votes for Patty and 2096 for Dino, with 800 more ballots awaiting processing. Murray got 26.6% of the tallied vote, so I would project her to gain 212 more votes, and Rossi to take 588, from the uncounted ones. Therefore, the Adams County projection is Murray 970, Rossi 2684.
Do the same thing with all 39 counties for last night's snapshot of the SoS tabulation, and the result is my projection. Which, as it turns out, comes out even closer than the actual tallies from last night ... 50.3% to 49.7%. That's mainly because Spokane County (56.3% for Rossi) reported 100,000 untallied ballots, almost 19% of the total. For comparison, the untallied count in King County (62.0% for Murray) was 137,358 (26.0% of all untallied), even though King County's voter-count is four times Spokane's.
That brings me back to the second point above. The SoS says King County had counted 368,921 ballots and had 137,358 waiting to be counted. That sums to 506,279 ballots. However, earlier last night, King County reported that it had already received 543,348 ballots. While a small proportion of those ballots were probably invalid, the great majority of them simply hadn't gotten through the county's processing far enough to qualify for the SoS's count. Note also that King reports a very precise number, whereas Spokane (and most other counties) gives what is clearly an estimate based on ... well, who knows what it's based on?
On top of that, King County will probably receive almost as many additional ballots today as came in yesterday (over 133,000), ballots that won't make it into King's totals until Friday at the earliest. Yes, the same sort of thing happens all over the state, but the raw numbers in King County dwarf the others. In addition, if we sum the counted and waiting ballots, calling that total the expected turnout, Spokane is at 66.5% whereas King comes out with only 47.3%. The SoS predicted 66% statewide, and King County Elections anticipated 68% turnout (and now says that might be a lowball estimate). Therefore, deep blue King County has a whole lot more votes not-yet-counted than it appears from the Secretary of State's conservative counts.
I plan to repeat this procedure tonight, and to simultaneously a) refine my projection, b) compare the vote distributions of newly-arrived ballots to early-arriving ballots, and c) otherwise examine the outcome. I'm very confident that Patty Murray has won this election, and that her margin will end up larger than it currently appears to be. Goldy called it last night, immediately after the initial ballot drop from King County, and I was nodding affirmatively as he posted that entry on HorsesAss.org.
Our experience in 2004, when Gregoire trailed Rossi on Tuesday night and throughout the initial count, the machine recount, and nearly all of the hand recount, is instructive. As is 2000, when Maria Cantwell only needed one recount to slip ahead of Slade Gorton for her Senate seat.