Grab your popcorn, turn on MSNBC, and let’s watch Steve Kornacki map out tonight’s primary returns.
Primaries are being held in Virginia, South Carolina, North Dakota,Nevada, and Maine.
First results are expected:
Virginia: 7:11 p.m. Eastern
South Carolina: 7:30 p.m. Eastern
North Dakota: 8:05 p.m. Eastern
Nevada: 11:24 p.m. Eastern.
Maine – maybe in a week or two.
A week or two? Why is that?
The Bangor Daily News reports:
With seven Democrats and four Republicans running for governor, a first-round majority is unlikely. Clerks will report first-round votes to us on election night as they have in the past, but further rounds of ranked-choice voting will be run centrally by the secretary of state’s office.
The secretary of state has said that this will happen during the week after the election. State-hired private couriers must collect ballot information and bring it to Augusta for ranked-choice tallies. That’s a long period of uncertainty about crucial elections.
Consider this. Massachusetts has 351 cities and towns, and each municipality runs its own election.
Under Ranked Choice Voting, if someone doesn’t get a majority in the first ballot, you eliminate the bottom candidate and redistribute their second choice votes. This could set the stage for a long, statewide nightmare. Let’s set up a hypothetical statewide election with five candidates.
Candidate A gets 350,481 votes
Candidate B gets 225,218 votes
Candidate C gets 223,971 votes
Candidate D gets 151,012 votes
Candidate E gets 150,989 votes
Town clerks and election officials count the first round ballots, but then must wait for the statewide count before they proceed to the next step. Under the rules, Candidate E gets eliminated and the votes are redistributed. However, Candidate E says, “Wait a minute. Your count has me losing by 23 votes. That’s not right. I want a recount.” Whether the law permits a recount or not, the next round of counts is delayed until the state declares a loser and orders the ballots redistributed. Candidates D and E draw from a similar base, so Candidate D gets most of Candidate E’s vote ends up in second place, and now Candidate B and Candidate C gets into an argument about who should be eliminated.
And you thought the hanging chads in Florida were a problem.
Why would candidates B and C argue over who should be eliminated? It’s whoever has fewer votes at the end of counting the second round. Plus it seems in the computer age we should be able to calculate multiple rounds pretty quickly.
Not if we have 351 separate governmental units running elections.
But we already have to add up the results of 351 communities to elect a Governor, so now they would just have to send in a few more numbers and the state would centrally run the numbers.
It’s not a matter of sending a few more numbers. You can only do the first round of counting on election night. You can’t eliminate the bottom candidate until you determine who it is, which means receiving and verifying the first round counts from 351 municipalities, and you get a go ahead from the state.
If you have a multi-candidate race, the bottom candidate in Arlington is probably different from the bottom candidate in Dracut, so you can’t run the second round until we all agree who is at the bottom of the vote count.
Towns don’t need the go ahead from the state as the state itself will be counting subsequent rounds. At least the way I see this working the towns send their counts to the state just once, on election night. The state does the rest.
Didn’t you say this was a hypothetical statewide race? So what does it matter? The bottom candidate is the one with the least votes statewide. End of story.
You’re trying to say geography means something when you’ve already waved away concerns of geography.
And, if a candidate has particular concern about the vote count in a particular subsection of the geography… how is that any different than the present ‘first-past-the post’ system?
Yipes! Where are we ever going to get 351 computers!
We have ranked choice voting in Cambridge, and now get the results overnight. (Admittedly only a tenth the population of Maine.) The great advantage for us is that we don’t need to have districts. The problem with recounts is that outcomes may differ depending on the order in which ballots are counted.
Just FYI: that the results can theoretically depend on the order of the ballots is only because Cambridge uses RCV for multi-winner contents, and even then it’s a relic of the days when Cambridge hand-counted their ballots. Single-winner counts never depend on the ballot order, nor do modern multi-winner counts.
The difference is that Cambridge is just one jurisdiction, and can quickly tell who is in line to be eliminated from consideration. When you have 351 different municipalities, Cambridge voters may have a very different order of finish than the statewide count, and there would certainly be differences in the order of finish across municipalities.
This is why a statewide coordination of the count, starting with the second round of Ranked Choice voting, is essential for any contest that crosses municipal boundaries.
It will take longer to get results and more expense to count the votes. Is that the best you can do to criticize Rank Choice Voting? On the other side of the argument we have a system that requires a candidate to get a majority of the vote and not a plurality. Maine is attempting to stop the repeat of the last two election in the state, where the Governor was elected each time with 40% of the vote. Waiting a bot longer and spending a bit more money seems like a small price to pay for electing someone who has the support of a majority of people in the state.
Think California. Run a “top two” primary in June. Two names advance to the November ballot. Winner will get a majority of the votes.
If you’re going to have a non-partisan primary, why go through the expense and turnout cost of two separate elections when you can have all the information on one ballot?
By your own numbers, you are claiming that “a week or two” (it’s not) is much worse than 5 months!
Also, with an entrenched polarized two party system, the “top two” primary makes vote splitting even worse than our primary system. CA Dems got lucky this time to barely squeeze into all the November run-offs. Even so, the dynamic hurt the progressive candidates because voters had to strategize.
“Top-two” might work starting from scratch, but RCV provides a smooth transition from a highly polarized political environment to a somewhat less polarized future.
In that system I don’t like the possibility that one party might get shut out of the general.
Some parties are more party than others.
In the Fourth Middlesex special election last year, here is the vote from the Democratic primary:
In the general election, the results:
Friedman (D) 7876
Jackson (Green-Rainbow) 832
Current rules, you get a blowout of the Green-Rainbow candidate.
Top two: Friedman and Garballey emerge as the top two candidates, and voters have a legitimate choice in the general election.
Friedman/Garbelley is what the primary is for. I’d prefer that GR had a candidate for the general, with IRV. Without it I prefer a GR candidate contest the Dem primary.
You could have that same result with a non partisan IRV with the GR getting eliminated first round and saving the time and expense of a second contest. Even if you kept partisan elections, the IRV contest for the Democratic nomination would ensure the voters who selected the third place finisher would not be wasting their votes and would force all three candidates to cooperate in order to compete for the other candidates votes. This totally eliminates the spoiler effect and helps make a multiparty democracy a reality. This will make the main parties stronger in the long run while allowing third parties and outsider candidates within major parties greater influence to move the needle on issue they care about.
Lastly I could easily see a jungle primary situation on the Cape, South Shore, or even the CD-3 specia where the smaller GOP would have outsize influence as the Democratic side fragmented.
First, why is that an emergency? The world won’t end and the sky won’t fall in a week. The state needs to wait for all the overseas and absentee deadline before the official results can be certified anyway. The general election in Maine isn’t for 5 months, so you can calm down.
Moreover, the delay isn’t inherent to RCV, but due to the Secretary of State’s interpretation of a Maine-specific provision, which he believes requires all the ballot data be physically transported to Augusta. If done statewide in Massachusetts, we would know the unofficial results on election night, just like plurality elections. The country of Australia is more than 3 times our size and has run nationwide RCV elections for more than a century, without the problems you envision.
We know you’re on the record in favor of top-two, Pablo, as superior to RCV. I imagine the chaos that has erupted from top-two in California and the widespread dislike you’ve encountered for that reform is at least partially to blame for this petulant post. Some of your posts are quite good, but I guess you just shot from the hip this time and missed wildly.
I will echo Greg’s specific critiques and also that the entire Maine political establishment (though thankfully none of the Democratic primary candidates) was against this proposal and dragged their feet on implementation, particularly Gov. LePage who is threatening to delay this process even further. Also important to note that voters overwhelmingly decided 54-46 to keep this system in a second referendum. Support for the system is strong across all voter groups-Democratic, independent, and Republican.
A benefit of this system as we look at crowded primaries in our own state in the CD-3, and a potentially ugly primary in the CD-7, is that it encourages candidates to be cooperative toward one another. Can you imagine this kind of political ad becoming the norm? Can you imagine how much better the last primary would have been had Bernie and Hillary cut a similar ad?
There is also no way a candidate can win the nomination of a major party without a plurality. That means no LePage in Maine, it also means no Donald Trump. He only won 43% of the national Republican popular vote. This means all those Kasich, Rubio, Cruz, and Jeb voters would not have cancelled one another out.
Finally, it means no third party spoilers in the general. You can support third parties without endangering the outcome. The slow count is due to old voting systems in Maine. Update the systems, and RCV can be tabulated instantaneously as it is in Cambridge.
Right now, each city or town is responsible for their own votes.
To do a statewide IRV without stopping after each round, we would need to network everybody into a statewide feed for the scanned ballots. Our current system is not tied into the Internet, and thus is more resistant to hacking.
A statewide, networked ballot count would make Vladimir Putin very happy.
Every city and town clerk’s office is already connected to the Secretary of State’s office via a non-internet-connected, dedicated T1 line, for the transmission of sensitive files. That could easily be used to transmit the cast vote records on election night from the cities and towns to the Secretary for the unofficial election night tally. The official count, once all the absentee ballots are in, could either be transmitted the same way, or if we don’t trust even that, rely on a hand-delivery of the ballot data. Either way, you get unofficial results on election night, just as with plurality elections.
There was much drama and angst in California over “top two,” but Chicken Little is very much alive, the sky hasn’t fallen, and the players have figured out how to operate within the game theory of that structure.
We could add a provision in which a state convention would authorize someone to be listed as a Democrat (or Republican) on the ballot, either as a designee or as authorized to be listed as a party candidate. There are other tweaks we could use to make it work in our political culture.
Let me note that our primaries are in September, which leaves a winning candidate a short window to get into general election mode, raise some money, and win the general election. This late primary has hurt Democratic candidates for Governor.
Adding a week or two to determining the winner of the primary will give a general election opponent a head start in a short campaign window.
A week may seem like a short period of time, but in a two month (or less) run from the primary to the general election, a week is an eternity.
Then move the primary. I might add, under IRV in the last party primary the anti-Coakley vote would not have split both ways and you may have ended up with a stronger general election candidate who would have been required to have the support of the majority of the party. The third parties you have blamed on her loss in the past would also have been a non factor with their votes eventually transferring to the major party candidates. An IRV 2014 race would have likely resulted in a different nominee and possibly a different general election outcome. We also would not have seen the GOP nominate Donald Trump.
We should also move the primary so it is no later than June 15, regardless of voting system.
Let’s remember Maine is no longer Massachusetts (as of 1820), and the political realities are very different.
Maine is a two-plus party state, with a tradition of three-way contests that lead to the election of someone like Paul LePage.
Massachusetts is a one and a half party system. We have a weak Republican party whose nomination is virtually worthless in a majority of legislative districts. In that context, we have RINOs running on the Democratic line who can fend off primary challenges and face little or no opposition in November.
A “top two” system guarantees that, if there are at least two candidates in the primary, voters will have a choice of two candidates in November. Charlie Baker might have missed the November, 2014 ballot, with the possibility of Coakley and Grossman being the two choices. In any case, Falchuk, Lively, and McCormick wouldn’t have been on the ballot, and Baker wouldn’t have won with 48.4% of the vote.
Now all of a sudden we’re talking about Massachusetts? The fact of the matter is many people want more than a binary choice. Maine’s history of strong third parties is more likely to spread, given the near universal distain for the two party system. Without a system like RCV, you’ll have many cases where candidates will win with 40% or less of the vote.
You can’t get to a majority without some structure to narrow the field.
Besides, multi-candidate elections are an oddity in Massachusetts. In most cases, there is no choice, especially when the Democratic (or Republican) primary is tantamount to election.
Are you saying Democratic party primaries are not elections?
Besides, all but one statewide general in 2014 had 3 or more candidates.
Primaries have a restricted list of participants, and lower voter turnout.
You write, “You can’t get to a majority without some structure to narrow the field”, and you are correct. And that is another advantage of RCV…you create as many instant runoffs needed for one candidate to achieve 50% plus 1.
Steve Consilvio says
Deleting the lowest voted for candidate is the wrong thing to do. They could have the maximum number of #2 votes, and therefore meet the 50% plus one vote necessary to win. I first came across this idea years ago, and somebody pointed me towards Australia. Unfortunately, I couldn’t find the link or graphics that I saved explaining it. It is very possible that everyone’s second choice is the best choice, and nobody’s first choice.
Trickle up says
Thanks, Pablo, for the reminder that the purpose of voting is to provide material for instant analysis from Steve Kornacki.
No, but if Steve Kornacki isn’t interested, chances are the campaigns find us to be equally uninteresting. I want my vote to matter without the need to move to New Hampshire.
Ireland uses this system and has results in less than a week for an entire legislative body. It’s not the system, it’s the implementation.
The phrase “losing by X votes” has no clear meaning in ranked choice because all candidates (theoretically) may receive some form of a vote. (after all, are they X first place votes? Second? Third? What’s the ranking of the differentials?) What matters are what ranking of votes you did get and the absolute value of Candidate E’s vote count in relation to Candidate any-other is essentially meaningless.
Actually on this point Pablo is describing the system correctly. “votes” here pretty clearly means first-choice votes and in his hypothetical example the exact number does matter because the last place candidate will get eliminated first. The second-to-last in first-choice votes could still possibly win. That’s unlikely in his example, but it all depends on how the support breaks down.
Pablo is describe the system correctly. Pablo is NOT describing the candidates possible options correctly: “Losing by X votes” is meaningless in contesting the count which is what Pablo described. First-choice votes are the only meaning if the candidate gets greater than 50% of the vote: not having achieved that goal the hierarchy of voting preferences are taken into account and the meaning is changed.
It’s one thing to look at it from the point of view of the candidate and say whether or no he/she ‘got’ X votes… and contest the absolute number of votes. But ranked choice allows for a different, in fact less adversarial, perspective.
Trickle up says
If Joe Smith is knocked out of contention in the second round with x votes, to a field of candidates who received y through z votes, his margin of loss is equal to y minus x.
Suppose that there is a third round (i.e., that no one achieved a majority in Round 2).
It is perfectly intelligible, and accurate, to say that Smith lost by (y-x votes) +1, since had he gotten (y-x votes) +1, he would have remained an contender in the next round.
It would also be possible to characterize his loss in terms of the smallest number of votes Smith would have needed to win outright in round 2. Majority of votes cast minus x.
Finally, if Jill Jones is the absolute winner in Round 2 with j votes, it would be accurate to say that Smith lost to Jones by a specific number of votes, j minus x plus one.
So there are 3 different ways to characterize margin of loss depending on circumstances. They are all meaningful and might all be contestable.
Another side benefit of RCV is the elimination of special elections. During a vacancy the retiring incumbents votes are redistributed according to the intention of the ballot and the second preference candidate is selected. Has worked so far in Cambridge for almost 80 years.
So, if a progressive Democrat wins by a significant margin, and leaves office, the distant second place finisher, a Trump loving far right Republican, would automatically ascend to that office?
Trickle up says
NO, Pablo. The votes of the retiring incumbent get distributed according to the next-ranked choice to select the replacement.
There is no “second place finisher” until the process selects one. You are conflating the person who had the second greatest number of votes in the voting round that elected the incumbent with the actual second finisher elected.
Distant Trump-loving etc. only wins if he or she gets the majority of votes; is that likely?
JConway’s observation is spot on. The cost savings could pay for same-day registration and early voting on a much greater scale.
I’d rather have the special election, which would give me the opportunity to vote for someone who was a supporter of the departing incumbent, who wouldn’t have been on the ballot against a friend or ally.
Where there are only two candidates, there is a second place finisher. Remember, the problem in Massachusetts isn’t usually the oversupply of candidates, it is usually a total lack of opponents.
Interesting. The original objection you stated in your post was the time and cost it takes to learn the winner. But now you advocating a system that would require separate elections…what is the timeline and cost of that process?
Pablo, this need not take weeks.
Let’s say for the sake of argument that there are 10 candidates. That means that every single IRV ballot can be described with less than 24 bits of data (10! = 3628800 = 1101110101111100000000). No need to count locally. You send the results of every single ballot to a central computer (several, for redundancy). You’re going to want to go ahead and use error correcting code, and of course your headers will be beefier for security, plus encryption.
But 24 bits is 3 bytes. Even if each ballot requires transmission of 30 bytes due to security, it’s no biggie. If you had 50 million votes, It totals 150 megabytes per race. Even if I’m off by a factor of 100 due to encryption, and even if there’s no ability for compression [there is!], it’s really not a big deal because each location is only sending in a tiny fraction of that total. And Maine is a hell of a lot smaller than California.
If Maine is going to take weeks to figure this out, it’s because it’s new and the process hasn’t yet been streamlined. They’ll improve the process.