Bell Curve The Law Talking Guy Raised by Republicans U.S. West
Well, he's kind of had it in for me ever since I accidentally ran over his dog. Actually, replace "accidentally" with "repeatedly," and replace "dog" with "son."

Tuesday, March 10, 2009

The Recount of the Great Minnesota Recount, Recounted

A week after the November election, I wrote that the razor-close Senatorial contest in Minnesota between Coleman and Franken would now become, "The most important test of the integrity of the Federal elections process since Florida 2000." A few weeks after that, as the initial recount neared completion, I noted that much had been learned from Florida and this time, "Everything is being done out in the open, and standards are uniform and clear. This may be the best-managed recount in US history."

But it has now been more than four months since election day, and no one has been certified as the winner. It took two months for the canvassing board to complete its work, and of course immediately the result was challenged in court. The current hearings are expected to last another two weeks, and regardless of the outcome there will very likely be an appeal to the Minnesota State Supreme Court and probably to the U.S. Supreme Court as well. Thus while the recount was managed fairly well on a technical level, in the larger sense the Minnesota election has failed to achieve the primary purpose of all democratic elections: to make a choice, in a manner deemed fair (albeit grudgingly) by most of the people.

I assume I am not alone in my conviction that more reforms are needed. This time, the problem is not with the regular ballots. No doubt there could be some improvements with the regular balloting process--determining voter intent is not the cleanest standard that could have been applied--but with the exception of about 30 misplaced ballots in one Minnesota precinct (which would not have been dispositive) there were no major challenges to the integrity of that process, and the result of the initial recount was essentially the same as the original election result. This time the fault lies with the mail-in absentee ballot system.

There are on the order of 10,000 absentee ballots that were irregular to some degree in terms of the identifying information provided on the outside of the envelope. This is the most serious area of contention. In some cases, the voter name and address did not match those in county records. In other cases, the voter signature did not match or was omitted. Pretty much every possible thing that could go wrong with the mail-in process happened with hundreds of ballots, whether through fault of voter or registrar or both. From a purely engineering point of view, the problem is that the process of registration and voting is not sufficiently standardized and has too many points of failure.

But no system is immune from these problems, of course, and so once again the real issue is how to handle the disputes when they arise. In my opinion the Minnesota Supreme Court made a very stupid ruling regarding counting absentee ballots. Rather than devising a set of clear standards for the entire state, the Minnesota Supreme Court ordered each irregular absentee ballot to be decided on a case-by-case basis: each individual ballot would be counted only if the county in question and both campaigns agreed. The court's preoccupation with being Minnesota Nice has produced a hodgepodge of non-uniform non-standards which has now become the underlying basis for the seemingly endless legal appeal.

From this I draw three lessons. First, we really need to fix voter registration rolls in this country. This business of separate records for the Department of Motor Vehicles, the Social Security Office, the Registrar of Elections, etc. (not to mention the differing systems between states) has got to go. It is high time for strong, secure national ID cards. (You may commence the tossing of rotten eggs at me now.) Second, we need uniform federal standards for all federal elections. Senate business has been slowed down by the absence of Minnesota's Senator, and the peculiarities of Minnesota's process will likely make this a federal issue anyway before it is through. Finally, state elections officials need more money to hire more staff and have faster recounts when necessary.


Robert said...

As a resident of MN, I'm frustrated by the lack of action. But at the same time, the process has been incredibly orderly. Ultimately, I blame Coleman because he has kept this in the courts. He broke his promise that he made the week following the election.

Minnesota has a good system, but there are holes to fill. But ultimately there is always going to be an absentee ballot that is incorrect for a minor reason, so I don't think we can reasonably expect the future of elections to look much different than this.

Dr. Strangelove said...

I don't like Coleman, but I'm not sure it's fair for me to blame him really. If the situation were reversed, I'm sure Franken would be equally tenacious. In this case, as the irregular absentee ballots that were counted by the canvassing board broke strongly for Franken, it is understandable that Coleman is suspicious. That his campaign agreed to each and every vote that was counted is not sufficient, as it is easy for him to believe Franken prevented the counting of votes that would have helped Coleman. (Really the choice of which votes to be counted should never have had anything to do with the candidates--talk about conflict of interest!)

Minnesota has a good system, surely among the best in the nation. And you are right, Robert, that no matter how good the system is, there will be issues with some ballots. But I'm sure you'll agree that just because we cannot fix the problem completely does not mean we cannot make meaningful improvements. If other states can learn from Minnesota's pain, the efforts made in that state to resolve this dispute will have benefits for years to come.

Raised By Republicans said...

I'll resist the temptation to take Dr. S.'s bait about ID cards. But I will ask how he thinks that having such an ID card would fix the problems with the absentee ballots?

With regards to Dr. S's cultural stereotype as an explanation for the Minnesota Court decision: Dr. S, pointed out that the process of voting "has too many points of failure." In a situation where the types and sources of failure are many (I would hesitate to volunteer to count all the possible ways of failure), is a standardized process that could be applied to all individual ballots fairly possible? Or would it necessarily be so complex as to be impractical to impose such a scheme through the limits of a single judicial ruling under intense time pressure? Might not the "case by case" ruling really be a concession to the practical problems being faced rather than some mindless attempt to be "Minnesota Nice?"

Also, I agree with Bob. The problem isn't that Minnesota has a bad process it is that the process it does have is being milked for every last drop of redress by the (so far) losing candidate - Coleman.

In a tangentially related anecdote. I have actually met one of the absentee voters in question. He is the friend of a friend of mine. He is a Coleman voter who described his ballot to me. I'm a little foggy on the details but I remember that he had several errors that would defy interpretation absent sitting down with him and interviewing him about his ballot and intended vote. I've met this guy before and he has a noticeable mental disability (serious learning disability among other "issues"). Since the election he has also been convicted of assaulting his nephew because his nephew was teasing him. I have no idea what this says about this election or voting rights or anything other than that it confirms my general prejudices about Republican voters - mentally handicapped and violent.

Dr. Strangelove said...

Regarding ID cards: Many registration mismatch errors occur because when someone moves (or dies, for that matter), it can take quite a while for that information to percolate through the system--and moreover updating the information requires the citizen to notify many government bureaus separately. Simplifying the system to one major point of contact would help, and would make error-checking easier. (It would also reduce fraud in the voter rolls, another perpetual complaint from some quarters.)

The "case by case" ruling of the Minnesota Supreme Court is in my view not so much a concession to practical problems as an abdication to them. Categorizing the ballots and devising standards is actually quite feasible--the number of points of failure is manageable. We know this because for the several thousand ballots covered by the original Coleman complaint, the three-judge panel hearing the current case has indeed developed a scheme to classify the vast majority of the irregular absentee ballots into about twenty different types. Broadly speaking the court has decided that those classes of irregular ballots in which the voter was not at fault should be counted, but others should not.

The current panel has refused to widen its purview to consider the absentee ballots outside of the scope of the original Coleman complaint, which is probably the most legally defensible position for a lower oourt, but it also means the Coleman camp has a strong appeal going forward. Coleman wants a higher court to order all the irregular absentee ballots statewide to be counted again, this time under the same standards, rather by "standards" that varied county by county. (They cite Bush v. Gore, of course.)

Given that you are such a staunch advocate of the adversarial system, RbR, I am surprised you would blame Coleman for "milking" the system for "every last drop of redress" he can get. On the contrary, Coleman is doing exactly what a zealous advocate for his own interests should do. The problem lies not with Coleman but with the Minnesota system, because it currently leaves too wide of an opening for disputes. Reasonable judges must at least consider requests to count additional ballots when the standards by which those ballots were excluded are not clear, are not uniform, or were simply (as in this case) subject to the veto of the opposing campaign.

Dr. Strangelove said...

Just to be clear... I think Minnesota has a very good recount system, as these things go. Nevertheless, the election still has not produced a winner, so I believe there is room for improvement.

Also, I hope I did not offend anyone by invoking the phrase "Minnesota nice." I was just trying to be "California cute" but may just have ended up being "West Coast insensitive." My apologies. The Minnesota Supreme Court's attempt to get a consensus was obviously much more than just mindlessly trying to be nice.

Spotted Handfish said...

Living in a country that has a standardised voting system, it can work very well. I think the biggest advantage that you gain from it is trust: people know what to expect and trust the process and the results. The most immediate analogy that comes to mind is the current situation with the banks: simply the lack of trust kills the system. Our system isn't perfect: voter registration can be defrauded, and they have to have mechanisms for counting dubious ballots. But every election I have attended in the last twenty years has had the same structure to the point that the ballots are the same format with the same font. I can also walk into any office in the country and vote in local elections.

The system is run Federally. Of course we have compulsory voting which means that you have to fund it properly...

Raised By Republicans said...

RE: Minnesota's not offensive, it's just not a very good explanatory variable.

RE: The advesarial system...While I think Coleman's milking the system is the cause of the delay, I wouldn't say that means the system is fatally flawed. This is, after all, an extremely rare case where the vote count was virtually tied. We're talking about a vote difference of hundreds in an election where nearly 3 million votes were cast. Even a nationally standardized system would be challenged by that.

Raised By Republicans said...

Oh, I forgot, I still don't see how having an ID card will help a sloppy voter from screwing up their absentee ballot form.

Dr. Strangelove said...

Having a national ID card would not help a sloppy voter fill out their absentee ballot form.

The Law Talking Guy said...

The only problem I see with the Minnesota system is the requirement that all court challenges be resolved before seating someone. First, it would make more sense to issue a provisional certificate of election. Judicial review of electoral processes is generally too slow. Second, it would make more sense to have a very high standard of review (say, "clear and convincing evidence,") to overturn a certified electoral result after the (fair) formal recounting process. The judiciary should be brought in only for issues of fraud, not second-guessing the recount board.

The idea of a provisional certificate of election would not work well in a presidential race for other reasons - there we would need a system for swift recounting before a certain date.

The Law Talking Guy said...

I think the Minnesota Supreme Court's requirement that both sides agree to accept a ballot was dumb but effective in one narrow respect. Once Coleman okays a ballot, he is justly estopped from challenging it in the future. This was no doubt their intention, and why they are sticking to their guns. The problem is the assumption that Coleman is the aggrieved party. Is he?
Isn't it the voter? If so, then estoppel is a poor choice. On the other hand, it is important to have finality in elections, and for practical reasons, it may be best to put the onus for objecting or agreeing to a ballot on the candidate (who is more motivated than the voter, one would think), then use that to whittle down the field of potential challenges so that we can achieve finality.

The problem, as I said, is not so much the court challenge as the fact that the court challenge is permitted to last so long and interfere with the need for representation.

The Law Talking Guy said...

I'm puzzled by the suggestion of strong national ID cards. I don't believe that voter fraud is really at issue in Minnesota. The problem has to do with absentee voters not following the rules properly, or the rules not being properly explained to them, or interpretations of the rules not being uniform.

Dr. Strangelove said...

The suggestion of strong national ID cards was not about helping sloppy voters fill out the absentee ballot forms. It was not about helping absentee voters follow the rules, nor about explaining the rules to voters better, nor making the rules more uniform.

What I actually wrote was: "First, we really need to fix voter registration rolls in this country. This business of separate records for the Department of Motor Vehicles, the Social Security Office, the Registrar of Elections, etc. (not to mention the differing systems between states) has got to go. It is high time for strong, secure national ID cards."

Because errors in the registration were most of the problem. You can see the 16 categories (it was not 20) of rejected ballots here. In category C-2, the voter was mistakenly sent the wrong ballot. In category D-5, the witness address was an incomplete match to the records. In category D-6, the voter's proof of address did not match the registrar records. In category D-7, the voter was improperly listed as unregistered. In category D-8, an already-registered voter was mistakenly sent ballot materials for an unregistered voter (this was the single biggest reason for rejection). In category D-9, the witness was marked invalid even though the witness was in fact a registered voter.

All these could reasonably be improved with a better registration system.

The Law Talking Guy said...

I am more comfortable without strong national ID cards, and I don't think these problems here merit such a change. They speak to improving state registration systems. Or rather to improving the data entry skills of state employees.

Dr. Strangelove said...

Hopefully we can all at least agree that improving state registration systems is sensible.

Raised By Republicans said...

Dr. S. I agree that improving the registration system would be great. National ID cards would probably help with that.

But, as near as I can tell, the problem in Minnesota is caused by the problems inherent in identifying very small differences between very large counts. Mechanical and hand counting both have error rates and computers are very vulnerable to fraud. So when the real difference between the two candidates is so small that random fluctuation can change the result, there will be the potential for messy challenges.

I think that it is fairly clear that when our elections produce results where one candidate clearly wins we don't see these messy delays. I'd be interested in finding out if we've been seeing an unusual number of very close election results lately or if they are just getting a lot of press. I'd also be interested in hearing how Australian election officials (or officials from any other democracy with a majoritarian electoral system for that matter) deal with vote counts that are as close as the election we're talking about in Minnesota (a few hundred vote difference out of 3 million votes cast).

I think LTG is on the right track by focussing on how Minnesota law could allow for a probationary certification of some sort.

Dr. Strangelove said...

I know this may sound strange, but this is honestly not a problem of small numbers or random fluctuations. The mechanical error rate was actually extremely low and the subsequent hand-count essentially confirmed the original result. The standards for determining voter intent were remarkably clear, really, and the number of genuine disputes in the end was not significant to the outcome. (And fraud via computer has not been alleged even by Coleman.) The original recount went quite smoothly and efficiently, and would have been finished in time to seat the new Senator.

The real issue--the only real issue--has been with the absentee ballots that were not previously counted for various reasons. Some votes were rejected because voters were sloppy, others because the registration system was flawed. But the real problem here is that the standards for accepting or rejecting absentee ballot envelopes were neither clear nor uniform. Had the standards for accepting envelopes been as robust as those for counting the ballots within, we would not be in this situation.

While there is always going to be some minimum error rate we cannot reduce, we can certainly make them good enough to determine a margin of a few hundred votes out of a few million. With clearer standards, better registration rolls, and perhaps some simplified procedures, this mess could have been greatly curtailed.

I would focus on improving the process rather than kludging it with a probationary certification that could crown the wrong person in the eyes of public opinion

Raised By Republicans said...

OK, good point about the counting. But the absentee ballot issue is still just an issue of how to deal with a handful (a thousand or two) ballots out of 3 million.

I guess I'm asking how often this is a problem in our democracy.

Dr. Strangelove said...

I don't have the statistics, unfortunately. Even so, I am comfortable with the general sense that very close votes are relatively rare (hence newsworthy). I believe the current Minnesota race is a closer contest than any of the most recent elections for the other 99 senators currently serving, so we're about 1 in 100 here.

So these events are not common. But they are devastating. These events are like political earthquakes: rare but devastating. We make a huge effort to construct buildings that can weather the moderate earthquakes, even though we all know the most severe earthquakes cannot be helped. Similarly, we can and should build our electoral structures to withstand stronger "earthquakes" at the ballot box, even though there could always be an election so close that it cannot be helped.

We saw in Florida, 2000 how bad it can be. After that, many states made great efforts to modernize their voting systems and to provide clear, uniform, unambiguous rules to handle irregular ballots. Minnesota has shown us that there are still some outstanding issues that require attention. In my original post I wrote, "From this I draw three lessons. First, we really need to fix voter registration rolls in this country... Second, we need uniform federal standards for all federal elections... Finally, state elections officials need more money to hire more staff."

I still think these make good sense. If need be I would be willing to back off the "federal" requirement in most cases, with the exception of the Presidential contest.

The Law Talking Guy said...

I'm not sure why uniform federal rules are needed. The problem seems to be a need for predictable and workable rules, period, whether they vary at the state level or not. The presidential contest may not need federal rules either, although it will require earlier finality than other contests.

I hate to say it as a lawyer, but Minnesota had a great process until the judiciary got involved. Florida had a crappy process that the judiciary did not improve. Somewhere between allowing one side's campaign manager to act as secretary of state and unilaterally certify a vote (Florida) or letting legal challenges tie things up forever (Minnesota) there must be a better set of rules.

The other problem, as Dr.S notes, has to do with absentee balloting. That system needs to be seriously fixed. I can see uniform federal standards for absentee ballots being of use. Simply say the following.

1. Each absentee ballot must be postmarked on or before the day of the election.
2. The envelope for the absentee ballot must be signed by the voter.
3. There need be no verification of addresses against databases, but voters must submit a social security number with the ballot that is unique.
4. A witness is only necessary where the voter is incapable of filling out the form without help, in which case the witness must sign the envelope.
5. The ballot inside the envelope must be the same ballot that everyone else uses in that election.