Get to Know Your County Clerk

In the past several years, I've become very skeptical about the integrity of our democratic process.  Increasingly, blogs, followed by traditional media, have revealed instances of voting machine glitches, security holes in almost every voting technology on the market, and also good old-fashioned voter suppression and alleged election fraud in some cases.

Election integrity organizations and activists like Black Box Voting, Verified Voting, and Open Voting Consortium have had a crucial role in educating voters like me about the problems with existing voting technology and what can be done to improve it.  I've been following this issue closely, but the multitude of problems nationwide is daunting and downright scary.  What's a voter to do?

It's become clear to me that educated, active voters are the key to ensuring the integrity of our elections.  Spend some time looking at the web sites of groups like Black Box Voting, Verified Voting, and Open Voting Consortium.  Read blogs about election issues.  Find out what problems have been reported with various voting technologies, then find out what is used in your area.  If you educate yourself about potential security flaws in elections, you may be able to spot potential problems in the elections in your own area.

After you become educated on this issue, the next step is to make use of that education when you go to the polls.  When you vote, pay attention to what kind of machine you're using and what the process is.  You may want to ask some questions, or do further research when you get home. 

If you ask questions of poll workers or election officials, remember that it's their job to ensure a clean and smooth election, and most of them want to do so.  It pays to be polite and respectful to them and not to interfere with their jobs.  They, in turn, should respect your right as a voter to be satisfied that votes are being counted accurately.  In most cases, they'll be happy to answer your questions.  If they don't, it may be because they are feeling busy or harried, or it may be that they don't know the answers and are uncomfortable being asked, or it may be a cause for suspicion.  In any case, don't be afraid to ask the questions.  It's your right to know.

I recently voted in the Michigan primary in Washtenaw County, Michigan.  I'm a new arrival in the area, so the system there was new to me.  I had just moved there from Santa Cruz County, California, where I was lucky enough to have confidence that my vote was counted correctly.  Part of my confidence in that county's election system came from the machines they used at that time (before HAVA was implemented), part of it was because I was not quite as skeptical as I am now, but part of it had come only after getting to know the County Clerk, at least enough to get some questions answered. 

My experience as a new voter in Washtenaw County raised some questions, and I was able to get some answers.  My activism as a voter here in the past weeks was not perfect, but it was educational.  Maybe I didn't always ask the perfect questions, enough questions, or the right questions, but I did get information that helped me evaluate my level of confidence in the system here, and have some ideas of where it could be improved.  That's a good place to start.

My new voter experience went like this:

I went to the poll, wrote up my application for a ballot (a new thing for me here in Michigan), and received a ballot from the poll worker.  The ballot application required me to write my address and sign for the ballot.  It seems that would help ensure against unauthorized votes, so I guess that's good.  Or is it?  Does this amount to a mini Jim Crow literacy test of the kind that was used, and then later outlawed in the South?  Does this mean people who can't write down their addresses can't vote?  Shouldn't they be able to, if otherwise eligible?  If this concern sounds ridiculous, consider that 1% of Americans are illiterate, according to the CIA. In some recent elections, that has been equal to the margin of victory. Even people who cannot write are able to listen to candidate's speeches and probably pattern match the names of the candidates to names they have seen before.  I would guess that the illiterate are among the poorest in our nation.  Aren't they entitled to a vote?  Nevertheless, though this a good question for debate, it was not my primary concern.

I then went to the booth provided and marked my choices on the optical scan ballot. The poll worker then showed me where to feed my ballot into the Diebold optical scan ballot counting machine.  I took note of the type of machine.  The name on it, Diebold, had me worried.  Their name is not synonymous with election integrity, in my opinion. 

The poll worker stood behind the machine while I did this, so if I hadn't been very careful to keep my ballot in the folder provided until the machine sucked it in, she could have read my ballot, at least if she could read upside down.  If this sounds like an unusual talent to you, it doesn't to me.  I can read upside down. 

She also may or may not have operated something on the back of the machine.  I couldn't tell.  I'm sure it was all perfectly innocent, but with all the stories about voting machine flaws and hacks floating around, I want the poll worker's hands where I can see them.  Know what I mean?

The following conversation with the poll workers is one that I wrote down after I got home, so I've captured the general flow, not the exact words.

After the ballot was sucked in, the poll worker indicated that I was done.  I said "Don't I get a receipt or something?  Is there some way I can verify that my vote was counted correctly?"  I shouldn't have asked for a receipt.  It actually wouldn't be good if voting machines handed out receipts showing how you voted, because then you could prove it to someone who wanted to buy your vote.  Vote buying would become big business.  I know this.  It was a dumb question.  I asked it because I was surprised to have my ballot sucked into a machine without the machine giving me any confirmation that it had counted my vote.  At least in Santa Cruz, I used to be able to keep the stub of my ballot.  Come to think of it, that didn't prove the machine counted my vote correctly either.  I was so naive then.

The first poll worker said "You get a sticker." This was one of those "I voted" stickers that many polling places hand out. I bet I could buy a roll of 5,000 of them on the internet, but that wouldn't prove I voted 5,000 times.

Then, a second poll worker stepped in and said "It was counted."

I said, "but not in a way that I can verify as a voter."

He said, "Yes, it verifies your vote on the front of the machine."  He then showed me the front of the machine, which had a display reading "105."  I guess I was to understand that I was the 105th person to vote and that before I voted, it probably said "104."  If I had been aware of that display before I voted, I might have checked the number before my ballot was sucked in.  The display was small and inconspicuous, however, so I suspect that few people notice it.

I said, "but that doesn't verify who I voted for."

He said, "but it's on the paper ballot."

I said, "but how do I know the machine counted the paper ballot correctly?"

He indicated it was on a tape in the machine. I asked whether, in the event of a recount, the paper ballot or the machine count would be used. At first he said the machine count, but later changed his mind. I'm not sure whether he misunderstood the question at first, didn't know, or simply wanted to tell me what he thought I wanted to hear.

I told him I wanted the machine to show me how I was voting before I confirmed the vote. He said it was impossible.  I told him it was impossible with that machine, but not with some others I had seen in California.  He told me "I'm not sure you'd want it to."

I said "I wouldn't want to be able to verify how my vote is counted?"

"But if you could read it, we could read it."

"Not if it scrolls up into the machine after printing."

"But we could open the machine and read it."

"But you wouldn't do that after each voter."

He said, "But we can't change how you voted on the card."

I said, "But the machine could.  These Diebold machines are easy to hack."

Then the first poll worker chimed in and said they had a seal.  I've heard about similar seals before.  In some places, they've been reported to be easy to peel off and replace.

I wasn't getting anywhere with the poll workers.  I wasn't sure what good it would do to discuss it with them anyway.  They were just volunteers.  I brought it up mostly to make them aware of concerns that probably they should have too. 

Later, I googled "Diebold optical scan voting machine."  I found an article on Black Box Voting's web site that said that the Diebold optical scan machines make a call to a program on a removable memory card.  It also said the programs on the memory card could be changed.  In other words, it wouldn't be too difficult to tamper with these machines and make them count the votes incorrectly. In fact, Black Box Voting said it could be done by someone with no more knowledge than an advanced TV repairman.

The fact that I had marked my vote on a paper ballot didn't give me as much confidence as it once would have.  It's great that the paper ballot is there in case of a recount, but if no recount is done, the machine count is all there is.  Some localities don't do recounts by hand anyway, and I wasn't sure of the procedure in Washtenaw County.  It was time to ask some questions of my local county clerk.

I found the County Clerk's name, Larry Kestenbaum, in the next day's edition of the Ann Arbor News, in an article about the election.  An easy search on Google turned up the web page for his office. 

I e-mailed Mr. Kestenbaum asking if he was aware of Black Box Voting's findings about Diebold's optical scan machines.  I was careful to tell him that I was asking the question, not because I had any reason to think something had gone wrong with the election, but because I was concerned about the vulnerability of the machines.  I followed my initial question with several other questions:

"What is the recount procedure here? If an election were called into question, what do you check? Is there a hand count of the paper ballots? Is the machine count checked? How is the final decision made about the vote tally? Is there an audit process to make sure the machines are counting correctly? What triggers it?"

I closed the letter with a suggestion that he might want to talk to Alan Dechert of Open Voting Consortium about developing a more reliable voting system for Washtenaw County.  Open Voting Consortium has what I consider to be a prototype of a system that would be pretty close to ideal, so I described it in brief:

"In a nutshell, a more secure system should allow the voter to verify that what the machine is reading in is the same as what was marked on the paper, should use the paper as the final word in any recounts, should be regularly audited, and should ideally use open source, not proprietary, software for transparency."

The very next day I received an e-mail from Larry Kestenbaum that I feel exemplifies the responsiveness that citizens wish they could receive from public officials every day.  In his detailed reply to my questions, Larry instilled some degree of confidence in me as a voter that he is aware of potential flaws in Washtenaw County's system, and is doing the best he can to remedy them within the political realities of budget constraints and state law. Some of the e-mail is quoted below (in italics), with his gracious permission and my comments:

"All elections in Michigan are conducted using optical scan paper ballots.  The ballots are an ideal voter-verified paper trail.  This is better than punch cards, better than old-style paper ballots, and enormously better than unrecountable methods such as mechanical voting machines or touch screen computers with internal counters."

All very true, except that I don't agree that it's an ideal voter verified paper trail.  The paper trail itself will only be used in the event of a recount.  If there is no recount, we have to trust that the machine counted the votes correctly.  If the machine count were way off from the actual vote count, but there was a clear winner in a given race, probably no recount would be requested.

"All of Washtenaw County uses the Diebold optical scan devices.  The choice of vendor was made by my predecessor as county clerk, and the equipment was provided by the state using federal HAVA funds."

The fact that the choice of vendor was made by his predecessor was filed in my cynical partisan brain under "of course," when I found out from another voter that his predecessor was a Republican.  It is well known among election integrity activists that Diebold's former president, Wally O'Dell, promised to do everything he could to deliver Ohio to Bush in 2004. 

Perhaps I wasn't being fair.  Many county clerks across the nation have chosen Diebold voting machines, and some of them are not Republicans.  There are many honest public servants across the nation, and difficult as I sometimes find it to believe, probably a few of them are even Republicans.  I had no legitimate reason to believe that the previous Washtenaw County Clerk had chosen Diebold for nefarious reasons, but I'm afraid that until our nation's electoral system is cleaned up and becomes reliable, such partisan suspicion is inevitable. 

In any case, any nefarious reasons the previous County Clerk may or may not have had did not work out to the Republican Party's advantage.  These days, Washtenaw County is pretty firmly in Democratic control.  This may give Republicans reason, justified or not, for partisan suspicion of their own.  All of this illustrates why election integrity is one issue that odd to bring us together in agreement as voters.  If we can't trust that our election system works, we will always suspect that the other guy is cheating.

Mr. Kestenbaum went on to further address my concerns:

"My elections director, Derrick Jackson, and I are very much aware of the Leon County test, and we have taken measures to increase the security of the memory cards which go into the tabulators.  Unlike Leon County, we do not allow clerks to leave unsecured memory cards out in public areas. The memory cards are locked into the devices with a numbered seal, so that any physical tampering or changing of the memory card will be evident."

This should theoretically bring me some peace of mind, but I've seen complaints that seals used in other localities for the same purpose have been nothing more than tape that could easily be peeled off and replaced, undetected.  In a follow-up e-mail, Mr. Kestenbaum invited me to the County Clerk's office to examine the seals.  As a concerned voter, it's now my responsibility to follow up on this.

One thing in Larry Kestenbaum's e-mail that I was very glad to hear and that really does improve my confidence is that Washtenaw County does real recounts using the paper ballots.

"When a candidate petitions for a recount, they specify which precincts they wish to have recounted.  Very often, the answer is "all precincts". The other candidate also has the option of counter-filing to make sure that precincts neglected in the first filing are included.

In each precinct subject to recount, for the contested race, we do a complete hand count of the paper ballots.  This is our understanding of the meaning of the term "recount".  The hand count is done with two election workers of different political parties looking at the ballot, and challengers from both sides watching them."

Larry invited me to observe the recounts of last Tuesday's election that they expect to have.  Again, the ball is now in my court as a concerned voter.

He went on to provide further interesting information:

In last Tuesday's election, we did a pilot project of a hand count audit, doing a different race in each of six precincts.  The results of the hand count were very close to exactly the same as the machine tabulated count.  The state Bureau of Elections has proposed legislation to mandate random precinct hand count audits statewide.  Based on our

consultation with the Bureau, we will not be doing more random hand counts until the legislation is in place.  The bill should have bipartisan support.

The Bureau of Election's goal of requiring random precinct hand counts is a good one.  It's important, given the level of justified distrust in the voting machines, to audit every election.  How effective a random hand count will be depends largely on the size of the sample and whether or not it is really random.  In some localities that have claimed to do "random" hand counts, the counts are not random at all, and in some cases, the precincts that will be counted know that ahead of time. I hope that Michigan implements a truly random hand count.

I also wonder what he means by "very close."  Even when they haven't been tampered with, optical scan technology is usually off by about 1.6% (PDF link).  This is good enough in most races, where the margin of difference between the candidates is significantly greater, but is not good enough to inspire confidence in very close elections.

Depending on how much time I have to devote to this in the near future, the specific provisions of the random hand count bill, as well as further questions to Larry Kestenbaum about what "very close" means, would be good things for me to follow up on.

Here's another section of his e-mail that I should follow up on to get more information:

Prior to the primary election, all of the tabulators in Washtenaw County were modified and upgraded by Diebold technicians.  A complaint has been filed with the Washtenaw County Election Commission as to the implication of those physical changes to the equipment for the integrity

of elections, and the Election Commission is gathering information to determine whether an investigation is needed.

Oh really?  Was the software they were upgraded with certified?  Is this the subject of the potential investigation?  I'm going to have to get more information about the complaint on the web.  Without details, I can't tell how worried about this to be, but I do know that there have been questions about modifications and upgrades by Diebold technicians to their machines in other places before.  At the very least, upgrades just prior to election time by people who work for a company run by partisans is troubling.

I very much agree that open source technology would be more trustworthy than proprietary software, but changing to a different system is not feasible.  First, we are only permitted to use voting technology which has been certified by the state.  Second, there would need to be some

funding source to pay for the cost of the transition, including training and support as well as equipment and software costs.

Note, too, that the actual tabulators, though chosen by the county clerk, are the property of the individual cities and townships in each county.  Though the current tabulators were paid for by the state, I expect that the cost of any future tabulators would be charged to municipalities, as it was in the past.  And voting equipment is expensive.  Given the financial squeeze on all Michigan local governments, that means there would be fierce political resistance to any attempt to replace the current equipment using local dollars.

As I told him in response to this e-mail, recounts cost money too, and there will be more and more of them as people get more information about the threats to the integrity of our elections.  It would probably save money in the long run to invest in reliable machines, even if the machines still need work to be developed and certified.  I also think there will come a time when voters will demand other machines. 

I do think he's right, however, that the time might not be right now, and that the political will might not be available to make changes to systems that local governments bought only recently in response to HAVA.  That's too bad, because HAVA, despite its noble intent of election equality for everyone including disabled voters, came at a time when local governments were forced to choose between a whole range of bad options.

I was encouraged by his statement that he thought open source would be more trustworthy.  It helps when public elections officials like him are aware of this.  A change to more reliable voting machines will come when enough people demand them.

If you are concerned, as you should be regardless of party, about the integrity of our elections, I encourage you to get in touch with your local election officials, ask them questions, share your opinions, and find out whatever you can.  You may find out, like I did, that although the system is far from perfect, at least some steps are being taken to help correct it.  Above all, demand the right to vote and have your vote accurately counted in every election.  Get out there and vote, just in case some of the votes on your side somehow get missed, for whatever reason.  Observe the counting and the recounts whenever you can.  The only solution to saving our democracy is active participation.

P.S. I'm working on an article about Alan Dechert's work with Open Voting Consortium. Watch for it.