On A New System (Sort Of), Or, Referendum 71 And Mail-In Voting

We are now about two weeks away from the November election in Washington State, and one item on the ballot that has national attention is Referendum 71, the so-called "everything but marriage" proposal that would give same-sex couples more rights and protections than they have today.

There has been a lot of conversation about whether it will or won't pass--and a lot of conversation about whether it should pass.

I hope it does, and if you live here I encourage you to vote "yes" November 3rd.

But that said, you may not be aware that Washington has an electoral system in transition, and that as a result of the transition Washington has some idiosyncrasies that will make forecasting the results a bit tougher, and determining the results a bit slower.

We'll talk about that today, and by the time we're done you should have an appreciation of the odd way in which things can work out--and that, absent a landslide, we aren't likely to know the results on Election Day.

These Are Not Normal Times

We have the strangest weather here: it is not quite 50 degrees F. as I write this, in midafternoon; but by tonight it’s expected to get warmer as the rain moves in.

In normal times, this is the kind of thing experts would be considering as they tried to estimate what turnout might be in the upcoming election—but these are not normal times. After the November ’08 election, Washington, following Oregon’s lead, became the second “vote-by-mail” state, and now the question has become not whether weather will impact the turnout…but if it will matter at all.

“Democracy is only an experiment in government, and it has the obvious disadvantage of merely counting votes instead of weighing them.”

--Dean William Ralph Inge, “Possible Recovery?”

The first unusual thing about Election Day in Washington is that there no longer is an Election Day. Voting now begins when the ballots begin to arrive in voters’ homes (20 days before Election Day), and as of Sunday, October 25th, King County Elections (Washington’s largest county; the county that includes Seattle and almost 1/3 of the State’s population) reports that 8.59% of the ballots are already in. All ballots with a postmark before November 4th will be counted, which means there will be new ballots arriving for several days after the “polls close”.

(As you may have guessed, each county operates their own elections office. All elections in the State are regulated by the Washington Secretary of State, which is also the office that handles paperwork for State-level candidates, initiatives, and referenda.)

This is driving the professional political community nuts, because it means every day there is a smaller pool of voters to influence, even though the cost of advertising time isn’t going down. Additionally, it is at the moment unclear exactly who has voted and how; over time, I think we’ll begin to see patterns emerge.

For example, in King County in this election cycle, the locations most likely to have already voted are, for the most part, the wealthiest regions of the county. A group of six communities clustered around Bill Gates’ house all have “in” rates above 10.5%, including three above 13%. The Town of Beaux Arts Village is at the top of that pack, running almost double the countywide rate at 16.74%.

The other communities most likely to have already voted are among the most rural in the County. Skykomish has 16.31% in, Enumclaw 12%. Unincorporated rural King County, however, is only running 8.49%, suggesting that the trend to vote early among the wealthy is more predictable than that same trend among the rural voters.

Among the many communities with average “in rates”, however, are clusters of low- and upper-income housing—and that’s where it is impossible to determine precisely who’s voted already and who is left to influence. With polling reports on Election Day you can track by precinct (and that type of tracking will be available after November 3rd), but for now an effective method of tracking has not emerged.

We assume that over time we’ll see the development of some form of “exit polling” of those who have already voted…but this is the first significant election since all-mail voting began, and prediction tools are as of yet untested.

“Message, We Have A Problem”

All of this is affecting advertising—after all, if you don’t know what portion of the electorate has already voted, how do you target your message to the remaining voters? When we get a week out, if we have 20% or more of the ballots in, this question will begin to loom very large as campaigns have to decide whether they have spent enough campaign dollars to buy airtime…or not…and whether the target audience they seek to influence is actually responding to the message…or not.

This all becomes even tougher to figure out because it’s a series of state and local races that are being contested in this election; as a result there is no daily tracking poll data available from which we might draw some near real-time conclusions.

Speaking of polling data: here’s some. A Survey USA poll conducted October 3rd and released October 6th of 548 likely voters suggests R-71 was winning 45%-42%. Women were both more likely to vote for the measure and more unsure as to how they would vote, relative to men (48% yes, 36% no, 16% unsure for females; 42% yes, 46% no, 12% unsure for males).

Voters 35-49 were simultaneously the least supportive of the measure and the most unsure as to how they’ll vote (35% approve, 49% reject, with 20% unsure). Voters over 65, the group most likely to vote, were supporting the measure (44%-40%, 16% unsure) as of October 6th.

The poll has a 4% margin of error, and some of these results are within that range, so as of October 6th this was still a race that’s very much up for grabs.

There are no Federal or State offices being contested in this election, and the only other statewide ballot issue, Initiative 1033, seeks to limit the growth of State income. The presence of the two ballot measures is likely to increase voting by 3% to 8%. It is suggested that a lower turnout will help the anti-71 crowd, a higher turnout, the pro-71 crowd.

All of this has had a major impact on “get out the vote” efforts as well—for example, no one volunteers to drive voters to polling places anymore…because there aren’t any polling places left. (There are a few exceptions for the disabled.) Instead, the effort here is to make sure those ballots get in mailboxes before Election Day.

It is possible to construct ads that attempt to “close the deal”: suggesting, in the last 20 days, that voters vote right now for or against the candidate or issue, but I haven’t seen ads of that type yet.

Finally, a few words about the “after Election Day” action. If this election is close, the number of votes that are in the mail in the days following the close of voting (and where they’re from) will be critical—and in the ‘08 cycle 50% of the total votes cast were in that “in the mail” category.

(Washington has been moving to voting by mail for some time, and in the 2008 cycle more than 90% of the votes cast were mail-in ballots. At that time 37 of the State’s 39 counties were voting entirely by mail.)

The bad news: it could take anywhere from several days to several weeks before we absolutely know the results. This process may include “reevaluation” of votes after Election Day and efforts by either party to disallow votes based on what they think they can get away with, and the result could be litigation.

The good news: there are no electronic voting machines in this system, and every ballot is a paper ballot. This means we can determine, eventually, exactly how the votes were cast—and if it takes a few recounts before we know the results, well, that’s what it will take.

So as of right now, that’s where we’re at: it’s the first major election since mail-in voting was adopted statewide, we are not sure of exactly how the impact of early voting is being felt, even though we know that almost 10% of the votes are in, professionals are still not exactly sure of what’s going on, and there should be a higher turnout due to the fact that we have two questions on the ballot for the entire voting public to consider.

Don’t expect a final result on Election Night, and if we do have to go to a recount, there won’t be any electronic voting machines to screw things up. Instead, every vote will be on a paper ballot. Most importantly of all: this ain’t Florida, we’ve been through recent close elections and recounts before—and we were able to work things out just fine.