Column: Ann Arbor Election Autopsy

Only a few minutes after voting ended at 8 p.m. on the evening of Nov. 2, CNN used exit polling conducted throughout the day to call the Michigan governor’s race in favor of Republican Rick Snyder. Even before polls opened, the only real question for most analysts was the margin of Snyder’s expected victory.

Margin of victory was also the main interest offered in local races, but with expectations for the identity of the victorious party reversed from the gubernatorial contest. Ann Arbor voters returned Democratic incumbents to five city council seats and the mayorship. For Steve Bean, who mounted an independent campaign for mayor, and for city council challengers Republican John Floyd (Ward 5), independent Newcombe Clark (Ward 5) and Libertarian Emily Salvette (Ward 2), the final raw tally did not offer many bright spots.

Bean managed about 18% of the vote in the mayor’s race. Floyd and Clark drew 22% and 9%, respectively, in the Ward 5 city council race, and Salvette received 21% in the Ward 2 council contest. Unless they are robots, it’s hard to imagine that any of their egos escaped completely unscathed. And despite the fact that Newcombe Clark’s door hangers depict a very cheerful robot with an NC insignia, I do not believe that Clark himself is a robot. So at some level, given their sheer humanity, the results must feel at least a little bit like a personal rejection by the electorate.

On the flip side, it’s hard to imagine that an incumbent like mayor John Hieftje, or Carsten Hohnke (Ward 5) or Tony Derezinski (Ward 2) – perhaps even more so Sandi Smith (Ward 1), Christopher Taylor (Ward 3) and Margie Teall (Ward 4), who were all elected unopposed – could interpret the results as anything less than an overwhelming endorsement of their job performance.

Challengers and incumbents alike would be wrong in those interpretations, I think.

But as far as local races go, far more interesting to me than performing a postmortem on the council and mayor’s campaigns would be to take a look at the race for the library board, where there was little campaigning by the candidates. The outcome was not completely clear until the votes from outside the city and all absentee ballots from the city of Ann Arbor had been counted. That came at around 4 a.m. – almost eight hours after CNN had already called the governor’s race.

Vivienne Armentrout would have been a winning choice of city of Ann Arbor voters who voted in person at the polls. But once absentee ballots and votes from outside the city were included, she narrowly missed joining the board. Instead, incumbents Barbara Murphy, Edward Surovell, and Jan Barney Newman retained their seats.

Why The Local Democratic Landslide?

If challengers are wrong to interpret the local Democratic landslide as a personal rejection, and incumbents are wrong to interpret it as an endorsement of their job performance, what is a reasonable way to look at it? One alternative within easy arm’s reach might be to blame it on straight-ticket voters.

Why the Democratic Landslide: Straight Ticket?

Taking Ward 5 as an example, 38.4% of ballots were voted straight-ticket Democratic, while 7.25%  were voted straight-ticket Republican. That’s a total of 45.6% of all ballots that were voted straight ticket, leaving 54.4% of voters who voted for each race individually.

It’s worth noting that individual race votes will override a straight-ticket selection. By way of example, in Ward 5 a voter could have voted straight-ticket Republican, but chosen to vote for Carsten Hohnke, the Democrat. In that case, the specific Ward 5 selection for Hohnke would have overridden the Republican straight-ticket vote. It’s reasonable to assume, however, that there are not prodigious numbers of voters who make their choices in this way.

If the 54.4% of non-straight-ticket voters had been distributed randomly among the three candidates, then the expected percentage of non-straight-ticket voters for each candidate would have been 18%.

Let’s assume the same straight-ticket behavior that was recorded at the polls, but with a random distribution of non-straight-ticket votes. And let’s further assume that a straight-ticket vote actually resulted in a tally for the Ward 5 candidate from the straight-ticket party. That would give an expectation of the following vote totals: 56.4% for Hohnke, 25.3% for Floyd, and 18% for Clark. Variance from that expected outcome based on random voting can be taken as some indication about attitudes of voters who made a specific selection in the Ward 5 race:

WARD 5 EXPECTED OUTCOME
ON RANDOM VOTE
W/ STRAIGHT TICKET

CANDIDATE  RND STRGHT  EXPT  ACTUAL  VAR
Hohnke     18%  38.4%  56.4%  69%   +12.6
Floyd      18%   7.3%  25.3%  22%   - 3.3
Clark      18%   0.0%  18.0%   9%   - 9.0

-
So even peeling away the straight-ticket voters in this way, Hohnke did 12.6 points better than you’d expect from a random distribution of non-straight-ticket voting. Floyd did slightly worse than you’d expect if non-straight-ticket voting were random – by 3.3 points – and Clark did about half as well as he could have expected.

So the incumbent victory can’t be blamed on the straight-ticket vote – voters who made a specific selection still preferred Hohnke – by a lot. A similar exercise for the mayor’s and the other city council races yields similar conclusions.

Why the Democratic Landslide: Are You Listening?

In these races, I don’t think voters were rejecting challengers personally, or rejecting their ideas. Instead, voters did not perceive that the challengers would represent their interests any better than the incumbents. That perception was based mostly, I think, on the failure of any challenger to run a campaign that would convince voters they’d get better representation than what they’ve already got.

Bean’s campaign for mayor included one kind of message that could have succeeded: I’m here to listen and facilitate discussion. But that message needs to come at the very beginning of a campaign, and then evolve at some point to a message that goes something more like this: I have listened, and here’s what I’ve been hearing people say … and here’s how I can effectively represent that point of view.

I don’t think it’s actually necessary that you go knock on doors of people’s houses as a part of the “listening phase.” In fact, I’m astonished that candidates will actually knock on doors – you couldn’t pay me enough to do it. Selling newspaper subscriptions door-to-door when I was young left a certain mark. So it’s easy to understand why two years ago, when Carsten Hohnke contested the Democratic primary for Ward 5 – his first run for office – he tagged along with a seasoned political veteran, Chris Easthope, who was running for district court judge at the time.

But if not literal door knocking, then there needs to be some alternative, towards the beginning of the campaign, that translates into an engaged listening phase. Maybe it’s hanging door hangers on every door – no knocking required – with a handwritten note that invites people to a series of listening sessions you’ve set up. You’ll also encounter people while you’re out and about hanging stuff on doors. Inevitably, someone will ask you what you’re doing with that batch of door hangers in your hand, and then you’ve already turned the tables in your favor – they have interupted you, instead of the other way around.

Why do I think this early listening phase is a key? Because that’s what I hear local politicians who actually win elections talking about later in their campaigns. They’ll say things like, “When I’m out in the neighborhoods knocking on doors, what I’m hearing is …” The listening phase is important not just in its own right, but also so that you can talk about it later.

I did not discern any real listening phase in Bean’s campaign. He made himself available – but that’s not the same thing as investing the time and energy to seek out and elicit ideas from the people whose vote you want.

Floyd’s campaign included one message that I think could have been presented slightly differently, to convey that he was interested in listening to voters. Part of Floyd’s message was that he would ask probing questions. Fair enough. Close readers of The Chronicle’s meeting reports might reasonably conclude that Hohnke does a relatively poor job at this basic council function. So it was not crazy for Floyd to focus on a weakness of the incumbent.

Floyd’s message however, was presented in the form of a series of his own questions: Have you ever wondered …?  For example, Have you ever wondered who will pay to park 4-5 stories underground in the new underground parking garage? The easy smart-aleck answer is: No, I’ve never wondered that – because I can’t wait to park down there in that mine shaft myself. So Floyd would have needed to ask the quick followup: Okay, then, what have you wondered about? Something similar to Floyd’s message could have been presented early in the campaign, with voters’ concerns front and center: What are your questions … I can get them answered.

Part of Clark’s message actually tackled the listening issue head on. He proposed a mechanism for “open source government” where constituents would use an online interface to vote on the issues they wanted him to work on for just the two-year term he would have served. That might have been more effective, I think, if the interface had been set up very early in the campaign, and during later stages of the campaign, he had been able to talk about what voters’ priorities seemed to be, based on his open source government.

Why the Democratic Landslide: The Sky is Not Falling

I think it’s a mistake to try to sell Ann Arbor voters on the idea that our city’s condition is somehow dire, and that city councilmembers and the mayor are completely derelict in fulfilling their responsibilities as public officials. Even if it is dire, and even if the council is derelict in their duties, that’s probably not the basis of a winning campaign. Why don’t I think so?

Over the summer, several challengers in local Democratic primary races attempted to appeal to the East Stadium Boulevard bridges as a campaign issue. And it’s plausible to think that pointing at a giant gaping hole in a piece of major infrastructure could convince voters that maybe things have gone horribly, horribly wrong. It didn’t work. I don’t think it was a matter of the specific narrative that the challengers told about the bridges. It’s simply that a big hole in a bridge is not dramatic enough to convince Ann Arbor voters that things aren’t still pretty much okay.

Maybe that says something interesting about Ann Arbor voters, but hey, that’s us. So candidates for local office, take note: We Ann Arbor voters will not be persuaded we are in a crisis, as long as the sky is not on fire, and as along as when we flush our toilets, the shit still disappears. So if you think the sky is merely falling and not actually aflame, go ahead and run for office, but not on that message – keep it to yourself … unless that’s what you hear people telling you everywhere you go.

For incumbents who believe that an overwhelming victory in a local race means a personal endorsement by the electorate of your job performance, I would suggest that most voters still have no idea who you are or what you’ve done – they just know that the sky is not on fire. And that’s good enough. If there’s someone out there who can persuade the electorate that they will be better at listening to us than you are, they’ll be elected, whatever party they represent.

Absentee Voters

For one of the library board races, the voting patterns of absentee voters actually made a difference. But tabulation of absentee ballots for the city of Ann Arbor was not complete until the early morning hours of Nov. 3.

The group of election inspectors who were counting the absentee ballots cast by city of Ann Arbor voters were sequestered in the basement of the county administration building at Main and Ann streets for the whole day that polls were open and they worked through the day. So why did the absentee voter ballots take until 4 a.m. to get counted?

Several factors contributed to the delay. First, it’s worth noting that the ballots were not processed as one giant batch. That’s because not all paper ballots in the city of Ann Arbor looked the same on Tuesday. Some city residents live in state House District 53, while others live in District 52. Some city residents live in county District 8, while others live in 9, 10 or 11.

So for each ward, the precincts were collapsed into as few absentee voter count boards as possible, based on the form of the ballot. That resulted in 19 different batches of ballots to be processed.

The ballots themselves were physically long. Sending them through the mail required two folds, which meant that each ballot had four creases in the paper. Based on conversations with some of the election workers who performed the absentee ballot counts, the voting machines scanned the creased ballots without great difficulty. Where the creases caused problems was when the ballots dropped into the catchment bin after scanning. Instead of floating gracefully to the bottom, the creased paper tended to snag against the sides or fold back in on itself. A fair amount of time was apparently spent trying to get the ballots to drop cleanly to the bottom of the bin. I watched election workers smoothing out the ballots – one at a time – before they were inserted into the scanner.

Another issue that arose on at least one occasion was a mismatch between the number of paper ballots in a count board – as determined by a hand count – and the number that should have been in the batch, based on the list of names that had been checked off as the ballots had been returned through the mail. The count was off by one. It was a batch of over 1,000 ballots that needed to be recounted by hand – and the hand recount still did not resolve the mismatch. Much discussion ensued about how to handle it. Scott Munzel, a local attorney who was part of the team that processed that count board, wound up writing a memo to the board of canvassers, who would certify the election results the following day, clarifying exactly what had happened. To be clear no special legal skills are required in order to work the elections – Munzel is just coincidentally an attorney.

City clerk Jackie Beaudry and election inspector Jeff Micale, who had directed the day’s activities for counting absentee ballots, were still wrapping things up after the counts were finished and The Chronicle was departing the scene at 4 a.m.

Library Board Numbers

The early returns for the four-year term library board of trustees race showed Vivienne Armentrout narrowly ahead of incumbent Edward Surovell, who were both led by the other two incumbents, Barbara Murphy Jan Barney Newman, in a very tight race. The four were competing for three four-year terms on the board.

The other library board race was fairly clear-cut from earliest results, with Nancy Kaplan eventually winning a two-year term on the Ann Arbor District Library board. Kaplan took just over 55% of the votes in that three-way race against Lyn Powrie Davidge and incumbent Carola Stearns, who was appointed to that seat in 2008 following the resignation of Jean King.

Armentrout wound up finishing fourth in the four-way race, with 16,975 votes or 22.3% of total votes cast. Murphy got 20,404 votes, or 26.8%; Newman had 19,834 votes, or 26.07%; and Surovell got 18,415 votes, or 24.2%.

The earliest returns The Chronicle received were from the city of Ann Arbor polls where people had voted in person. Later, non-city precincts started to be reported to the county clerk’s office, and a lot later, the absentee ballots came in.

Looking at the final numbers, it’s clear that Armentrout did better among the in-person voters (22.65%) than absentee voters (20.91%). She also did better among city residents (23.19%) than non-city residents (20.39%). And among  in-person city of Ann Arbor voters Armentrout did well enough to finish third, outpointing Surovell by 23.75% to 23.13%. [Sheet 2 of this shared Google spreadsheet contains the library board elections results breakdown. It's identical to this downloadable MS Excel version.]

That still doesn’t put Armentrout on the library board, but it’s fun to know.

It makes up for the fun that CNN ruined on election night by calling the governor’s race so quickly.