The Umpteenth Voter List Maintenance Post

I've written this post so many times that my long time readers can just skip this - though I'm sure they're the only ones who will actually read and understand it.

But when it comes to voter registration statistics, people keep making the same incorrect assumptions over and over. So with another news article, it's time for me to once again dust off my 18 years of election office experience.

Funny thin is, this latest article, which ran last Friday in the Register and is not recirculating is actually one of the better ones, once you get past the doom and gloom headline Iowa Democratic voter numbers down sharply.

This conjures up a vision of a disillusioned voter who's mad as hell and won't take it anymore.

This happens. I helped Mike Thayer change his affiliation last January when he made a special trip. But it happens so rarely that it's noteworthy when it does. I've seen it happen maybe... a dozen? times in my career, and between work and politics I have probably registered tens of thousands of voters in my life.

Before I get deeper into party changes, I want to go over how registration numbers rise and fall.

Registration numbers rise most sharply, of course, just before general elections. Depending on the local election or primary, they may rise significantly or not at all. There's also a constant steady trickle of registrations, many from drivers licenses, but a few from diligent types who just moved to town.

There's also un-registrations. Deaths are steady, and moves away are cyclical, lagging a little behind the wave of new registrations at big elections. When people register in a new place, the new place sends notice to the old place - IF the voter remembered to list it. (Within the state, counties just grab voters from each other on the statewide voter registration database, iVoters.)

But the way most people get un-registered is long and slow and happened in two stages. I'm going to crib from the page I wrote at work on the subject:

Until 1994, Iowa voter registrations were cancelled after four years with no activity.  But since the National Voter Registration Act, better known as "Motor Voter," took effect in 1995, no one's registration is cancelled simply for not voting. Instead, the cancellation process depends on whether or not you can get mail at your address.

There are two main mailings used for list maintenance, and in Iowa both usually come out late winter or early spring. The National Change Of Address (NCOA) mailing goes to everyone who has reported an address change to the post office but has not updated their registration yet. (Read the work page for all the details.)

The other mailing is the Four Year No Activity mailing. Instead of cancelling your registration after four years without voting, we just send you more or less a reminder.

The important thing that happens with these mailings is: many of them come back as undeliverable. Sing it, Elvis:



If I have to write this post again, then you have to watch Elvis again.

Voters whose cards get Elvised are placed on "inactive" status. No campaign staffer anywhere seems to get this concept. "Inactive" does not mean what staffers call "a weak voting D"who will vote if we just knock their door one... more... time. "Inactive" means the auditor has documentation that the person has probably moved away.

Now, back to how this effects party totals.

In May 2011, Iowa Democrats held an edge in voter registration with 645,899 voters, while Republicans had 609,365 voters and there were 696,696 no-party voters.

In May 2015, Iowa Democrats' voter registration rolls have dropped to 585,178, a loss of more than 60,000 voters that represents a decline of 9.4 percent. Meanwhile, the number of registered Republicans is holding firm at 609,042, while the number of no-party voters has increased to 703,208.

Give the Register credit - they used the active numbers. Way too often, stats stories cite those wildly misleading "total" numbers which includes inactive voters. And they also address some of the reasons Democrats dropped more, though they do it in the usual Objective Journalist he said, she said format.

Even the shallowest analyst of politics and demographics knows that younger voters skew Democratic (and no party) and older voters skew Republican. Younger voters are also much more mobile and much more likely to get Elvised into inactive status. That more than outweighs the fact that older voters are more likely to die.

But that headline and the content seems to emphasize changes. So let's look at changes.

Set aside that rare Mad As Hell March To The Courthouse scene. Most voters change party for a more mundane reason: to participate in a caucus or primary, Many don't even know that voting in a primary can change your affiliation.

You know what I've seen many more times than a special trip to the office to change party? A person mad that their party "got" changed, who then sees the slip with their signature on it from the primary. "Oh, I though I was just voting for Millard Fillmore." Yep, and to do that you changed registration to the Know Nothing Party. (The dead parties are safer examples, and Know Nothing is inherently funny. Unless you're an immigrant.)

The percentage pattern of party affiliation is highly predictable. A slow, steady trend to no party, mostly from drivers licenses, as people aren't in "partisan" mode at the DOT and since there isn't a contest at stake. Then a big shift at the primary or caucus, sometimes up or, if the other party is more interesting, sometimes down. Followed by a very small counter trend as a handful of people reverse their changes. Finally, a big no party trend right before the general election as the less partisan voters tune in.

So let's look at the last few cycles.

The last major, statewide, hotly contested race within the Democratic party was the 2008 caucus. For 7 1/2 years, most people have not had a compelling - as in, you HAVE to register as a Dem to vote for this person -  have not had a compelling reason to register as a Democrat.

It turns out party unity, counter intuitively, is bad for your registration numbers.

Since the May 2011 benchmark in the article, we've had a wide open and very close - too close - 2012 Republican caucus, while Democrats had just a poorly organized Uncommitted effort against Obama. Indeed, quite a few anti-war Democrats crossed over to the GOP caucuses, because they felt like a vote for Ron Paul was better than a vote for literally nobody.

Then we had a presidential election, where lots of people checked the No Party box when they registered, and the Obama box when they voted.

We also had the 2014 primary. Republicans had an epic US Senate race and serious primaries in three of the four congressional districts. Democrats had only one congressional primary and, unfortunately, no primary sparring partner to get Bruce Braley (D-Colorado) in shape.

Lots of "compelling" reasons to be a Republican For A Day... and a lot of people whose intentions are "for a day" forget, or don't fully realize that it's an affiliation change.

The clincher, for me, that the Democratic statistical decline is not a real decline in Democratic identity is the local evidence.

In those 2012 and 2014 cycles, Democrats saw counter-trend Democratic registration INCREASES. That's because we had compelling local contests for courthouse offices.

The year Johnson County had a Republican trend, 2010, Democrats had no local contests, only a lame US Senate primary. Does Roxanne Conlin beat Tom Fiegen and Bob Krause by 40 points or 50?

Republicans, though, had the REAL ball game. Vander Plaats vs. Branstad felt almost like a general election for governor, because we all know Chet was in big trouble. So people became Republicans either because they were so scared of Vander Plaats, or because they couldn't wait to vote against Branstad. Same thing happened in 1994, when Republican registration in Johnson County climbed from 20% to 24% in ONE DAY thanks to Fred Grandy.



They're like the Viet Cong.

Look, folks. If you're looking to argue that Iowa Democrats are weaker than we were in 2011, there's enough REAL arguments to make, starting with last year's loss of a Senate seat and two congressional seats. There's no need to use the weak indicator of registration statistics to "prove" things which they don't prove at all.