Power to the People: Three Election Changes to Pull Us Together
For those of us who are concerned with the ability of our democracy to continue to be a going concern regardless of political parties and individuals, there is some hope of progress these days. In fact, three voting changes could be made to our electoral system that would help pull us back together as a country. We might even possibly see more compromise, more across-the-aisle camaraderie, and more satisfaction with government. None of these voting ideas are new. You have likely heard of all three of them. But to me, the key is putting them together to make a cohesive whole.
Sherman, set the way way back machine…
Back in my college days, I took a course on Constitutional Design. For the semester, the professor only handed out a single grade for the final paper you were to turn in. While that may sound horrible, I can guarantee you every person in that class loved it. In this paper, we were to take a fictionalized set of Pacific Islands that was setting out to create its own constitution and based on the demographics and culture of the various islands, write a constitution for them and justify why it made sense for that people. The only thing that would virtually guarantee a failing grade was to simply plop down the U.S. Constitution as an answer. I say “virtually” because he made it very clear that unless you had the most convincing arguments justifying it that have ever been made, you would get an F.
The class was a wonderful semester of looking at the way an electoral system works as a whole. It forced us to think about competing priorities, competing cultures, and shared values. The professor forced us to really dig in and identify how a constitution guides the direction day-to-day politics will actually take a country. He made us figure out what forces each piece of a framework would push people towards, and what it would do as a whole when put together.
With that, let’s get to the election changes that, when taken together, will make our system much more rock solid. Again, these are not new ideas in and of themselves. The secret is in the blend.
Right Round
First off, from a self-governance point of view one thing should be crystal clear: political leaders should not be deciding their own constituents. All you get are districts drawn to maximize the power of one group and minimize the power of another. That is anti-democratic. With today’s technology, we should be relying on independent, non-partisan commissions to draw districts that have clear non-partisan definitions on how they are defined. Legislators should not get a say. Governors should not get a say. State constitutions should have a definition for how districts shall be drawn in a non-partisan way, and any redistricting should follow that definition. I’ll say it again. Legislators should not be picking their own voters.
1 2 3 4 5…6 7 8 9 10…11 12!
What is the most common refrain you hear from discouraged voters?
- “It’s a choice between the lesser of two evils?”
- “Is this really the best we have to choose from?”
- “We really need a third party to step up!”
On that last point, let’s be clear. Third parties try. But our system of “first past the post” (i.e., whoever wins a plurality wins, or top X of the vote-share move onto a runoff) virtually guarantees that no third party will be a serious contender. This was actually highly debated in the aforementioned Constitutional Design class. But everyone finally came to the same conclusion: the math pushes our system to a “top two” format. Anything that splits the vote enough that the loser isn’t close to 50% will eventually have consolidation. A large minority will coalesce to try to counter whatever the majority is. “The enemy of my enemy is my friend” and all that jazz.
There is nothing in our Constitution or system of laws that say we have to have two parties. But the system as designed pushes us to that conclusion. In a perfect world, the two parties are nipping at each other’s centrist members, trying to get just enough people to switch sides to swing an election. Unfortunately, as we have seen, the primary system we have in place encourages each party to nomination the most extreme, yet still electable candidate. In places where one party dominates, the party primary is the de facto general election. Thankfully, Ranked Choice Voting (sometimes known as Instant Runoff Elections) is one of the answers to this problem
Ranked choice voting, or RCV, is getting a lot of attention these days. New York City is using it for its mayoral election right now. Alaska has put it in place for its Senate election next year. If you haven’t encountered it before, here is the quick rundown:
A voter ranks all candidates on a ballot, 1 to X, based on their own preference. Let’s say George Washington, Abraham Lincoln, Ronald Reagan, and Franklin Roosevelt are all running for office. I rank Roosevelt#1, Washington #2, Lincoln #3, and Reagan #4. After election day, Lincoln has 35% of the total vote, Washington has 30% of the vote, Reagan has 20% of the vote, and Roosevelt has 15% of the vote. Under our current system, depending on the office, Lincoln would win with 35% of the vote, or Lincoln and Washington would go to a runoff that almost certainly will have fewer voters. A lot of people who voted for Reagan and Roosevelt just won’t bother showing up, as well as just some portion of all voters regardless of preference. Runoffs are not the marquee, and they don’t get as much press. Therefore, lower turnout and results that reflect less of the will of the people than otherwise.
With RCV, votes for candidates who receive the least amount of votes are reallocated to their 2nd, 3rd, 4th, etc…preference until one candidate has 50%+1 of the vote. In this fictional general election, my vote for Roosevelt would not be wasted. I wouldn’t have to take time off work to go vote again in a runoff. My preference has already been recorded. If Roosevelt cannot win, I’d prefer Washington. And so my vote would be reallocated to Washington. Let’s say Roosevelt’s split was 10% to Washington, 3% to Reagan, and 2% to Lincoln. So now the split is Washington 40%, Lincoln 37%, Reagan 23%. Now Reagan’s votes would be reallocated. Let’s say the split for Lincoln voters (including 3rd choices if they were originally a Roosevelt voter) was 12% Washington and 10% Lincoln. After being reallocated, Washington now wins with 53% of the vote compared to Lincoln’s 47%. Everyone has their say. This gives broadly acceptable candidates a chance to win. It gives voters a chance to vote for a candidate they are passionate about without the guilt of being a “spoiler,” which in effect will give those candidates a greater chance of winning (or at least show a more accurate picture of their vote share). Ranked choice voting pushes election results away from candidates who are extreme and have just enough devoted following to keep them high up the rankings, who then go on to win because they split the votes of the majority of voters who don’t like them.
Welcome to the Jungle
And the final piece of the puzzle: Jungle Primaries! You’re probably thinking “What?! No!! Those are crazy!!” In a vacuum, I agree. But hear this out. This is where the electoral ingredients come together to bake a delicious cake.
In a jungle primary, top two (or three or four) finishers make it to the general election regardless of party. This is used in places (most notably California) with localities where the top two finishers are almost always Democrats or Republicans, shutting the other party out of the general election. You might be questioning how that will help us get to a system where moderation and compromise are the natural outcome. Well, two wrongs may not make a right, but three left turns do.
If you have compact, non-partisan districts, you are overall going to represent the will of the local population better. While there will still be cases where compact districts overwhelmingly favor one party over another, that will be because the area is naturally that way. Overall, you would see fewer “safe” seats in Congress. This is especially true with state level legislatures where districts are generally sliced into smaller numbers of voters than Congress.
Then, add in jungle primaries and general elections where ranked choice voting is used. For instance, in a jungle primary where the top four finishers make it to the general, you are likely to have at least one Democrat or Republican running in the general election, and possibly even one or two third party candidates. This will motivate supporters of both major parties and the third parties. This will give those third parties a larger voice. Then in the general, people have a choice. A real choice. And can vote with their hearts without worrying about “throwing away” their voting on a candidate who “can never win.” Sometimes those candidates will actually win, because there won’t be a penalty for voting for them. But even if they don’t, those voters can then fall back to the more acceptable candidate of their choice among the other candidates.
Either way, you end up with elected officials that will trend more toward the center because they don’t want to alienate supporters of other candidates who may agree with them others. If you are following the New York mayoral race, you will have seen that this weekend when two opponents, Andrew Yang and Kathryn Garcia actually campaigned together at one point. The intention was to show the voters of the other candidate that they were worth supporting as second choices. More of that will come. More appeals to a larger base of voters vs trying to turn off voters from voting. Less animosity on the campaign trail. More willingness for actual compromise because you win when you actually appeal to a larger set of voters.
Wouldn’t it be nice…
Down here in Mississippi our state laws define five U.S. House districts. Based on U.S. Census Data, we got redrawn to 4 by the federal courts, because the state refused to follow federal law when our population shrank. This is now causing a state politics problem because the state supreme court has tossed out the voter initiative process on the grounds we can’t have voter initiative that collect equal numbers of signatures in the four federal congressional districts when state law requires collecting equal numbers of signatures in the five state mandated congressional districts. This has been an issue for over a decade and successive governors and legislatures have failed to fix it. Be careful what you ask for.
Otherwise I’m all in on these ideas. Sadly, with the tossing of the citizen initiative process down here, I don’t have a way to make it so.Report
Every method to draw districts are partisan in the current U.S. Small Geographic districts would favor the Republicans since Democrats live in large cities that would allow packing Democrats into fewer districts. Drawing competitive districts would create large spread out districts that would pie slice large cities and close in urban counties. Drawing districts that ensure election of members of communities of interest would help the Democrats since all of those districts would ensure the election of Democrats while diluting non-hispanic white voters.Report
I’m all for the non-partisan redistricting commissions, as long as they have clear rules to follow and can not be easily captured by a party.Report
I agree with the principle that the foxes should not guard the hen house.
But or should I say ‘And’ settling upon a framework/algorithm won’t be as easy as we think… and honestly I’m skeptical that we won’t come up with 50 different models based on, ‘well you have to understand…’
Which is a slightly more agnostic way to say: who’s really ‘independent’ if we can’t agree on a universal district theory that’s applied without regard to ‘independent’ configurators changing the ‘universal district theory’ for … reasons.
At this point I usually post the interactive 538 Redistricting toolReport
I took the liberty of correcting your URL reference.Report
much obliged… what did I do wrong this time?Report
You had “href:” rather than “href=”.Report
Yes, but mine is aesthetically better.Report
To paraphrase from a very old computer-science story, “But yours doesn’t work! If it doesn’t have to work, I can write stuff that’s aesthetically gorgeous.”Report
Take it up with WHATWG. Alternatively, encourage enough people to make that mistake that browsers will support it whether it’s in the standard or not.Report
Also… can you tell me how sites are telling me I’ve ‘used up my free articles’ when I open them in a non-standard browser that I keep perfectly pristine with absolutely no ID information? Is it IP-Address? Mac address? I’m kinda baffled because it just started happening.Report
One possibility…
Almost all news sites now load an actual HTML page only for the front page. For articles, they download a bunch of JavaScript that looks around, and if the code is satisfied by the cookies and other data it finds, downloads and displays the article. Said JavaScript used to consider a lack of cookies to mean you’ve read zero articles. Since many people now run add-ons that delete cookies anytime you leave a page, more and more of said JavaScript is treating a lack of cookies as if you’ve reached your article limit.
My (non-systematic) observation is that Gannett is pushing this strategy out to all of its newspapers.Report
Bastards. This wouldn’t happen if we coded for aesthetics.Report
Like the Arizona frauditors are now rejecting ballots that were filled in too neatly.Report
There is no perfect way to redistrict. But as long as it gets close that seems fine. I don’t want Wisc level insanity where 60 +% for one party but the other party gets 60+% of the seats. Seats should be reliably close to the vote totals.Report
Sure. One vote for “Proportionally Partisan” that’s reasonable.
What about other ‘goods’ like compactness, county borders, COI/majority minority districts? Also, do we have no interest at all in maximizing competitiveness?
But, to be truly Proportionally Partisan, we’d have to allocate the districts *after* the election or we wouldn’t know what the partisan proportionality *should* be, only what it was the last time the districts were drawn… which might create unbalanced districts over the usual 10-yr period.Report
Well we redo every 10 to correct for changes. Counties are lame. Compactness, meh. I’m more concerned about bizarre rorschach shaped districts.
It’s complicated like most things
Except for counties being silly.Report
Compactness is the counter to bizarre shaped districtsReport
Depends on the area. Some rural areas will have giant but sensible districtsReport
You misunderstand, compactness isn’t about size, it’s about, for lack of an easy way to describe it, simplicity or uniformity of shape.
A square is more compact than a U-shape, even if the U-shape takes up less area. You want the border to be as convex as possible (minimizing concavity).Report
Time for some Schilling bait.
The square is convex, the U-shape is not. Whether either is compact is a completely different sort of question to a mathematician.
When politicians say compact, they mean something along the lines of minimize the sum of the differences in area of each district and its convex hull.Report
Compact means that every open cover has a finite subcover.Report
You two just stop!Report
NO MORE COUNTIES.
I guess the thing about proportionality is two-fold: If we build it to be proportionally accurate on the year ending in 0, what happens
1. If the districts change demographically over 9-yrs to deliver non-proportional results. (e.g. Narrow Blue district flips red which doesn’t reflect true proportionality)
2. Or conversely if the districts deliver non-proportional results because the demographics changed un-evenly but at very accelerated rates in one or two districts… (e.g. Blue district goes Double-Dog-Blue by vote totals, changing the overall state proportionality, but not the seat count).
The point here is that if we agree that the only (or most primary thing) is pure proportionality we’re going to run into issues that aren’t mitigated by other things like, say, COI or Compactness, or Competitiveness.
I think there’s maybe a hybrid way out where you have COI/Boundaries, calibrate for some competitiveness, then maybe allocate some ‘at large’ seats to balance proportionality.
But that doesn’t work for any state with fewer than, say, 10 seats.
But true proportionality could only be 100% at large.Report
There was a proposed test for SCOTUS, which had more math, but could be approximated as “if the seat distribution isn’t within roughly 10% of the popular vote, there’s something hinky”, which is a fairly good rule of thumb.
Sure the actual method was somewhat better (there are plenty of methods to calculate that), but frankly this is a case where the mark one eyeball is probably sufficient to “know” if districting is gerrymandered.
The end results of, say, the House elections of State X should be approximately (give or take 5 or 10%, because real life) if you’d simply summed up all the House seat votes across the state and allocated them proportionally.Report
Colorado’s new commission is going to do Congressional redistricting for the first time this year. The considerations, in the order they now appear in the state constitution are: equal population; comply with the 1965 federal Voting Rights Act; as much as possible preserve communities of interest; as much as possible keep counties and cities intact; compactness; and explicitly last, politically competitive. There’s a lengthy definition of community of interest. Are those clear rules, in your opinion?
For the first time, there are eight districts. I assert that, broadly, there’s an obvious way to do this. We’ll see what the commission thinks. If they follow my reasoning, the division between the Front Range and rural Colorado is going to become much more pronounced.Report
I think that ‘community of interest’ is a problem, as it is necessarily a fluid thing. I would make it an aspirational goal, rather than a hard target. Political boundaries are more reasonable, but even those have to include some measure of give and take (which I think it does). Compactness (and existing political boundaries) can, and often will, I suspect, run counter to ‘communities of interest’, which is another reason COI should be aspirational (if you can do it, great, but don’t twist things into knots for it).Report
For the last 40 years in Colorado, the three communities of interest that mattered for Congressional districts were Denver, the vast empty Eastern Plains, and the equally vast and empty Western Slope. This year El Paso County (Colorado Springs) may get added to that list. If the commission adopts my approach, those fall out almost automatically.
The 35-seat state senate and 65-seat state house are on a different scale, and COI can be complicated.Report
Regarding #2 on RCV:
I’ve shifted my political discussions with all the Anti-Trump republicans around me to just this… it is in ‘our’ interest to change the voting regime (at the non-Constitutional level) so that we can exert proportional influence in areas we care about and have the freedom signal support that runs contrary to either of the two parties.
I think you could augment your position by pointing out that even if your 2nd or 3rd choice is elected in Election #1… the actual # of votes and proportional support would influence Election #2 and potentially create some fluidity across the party lines as coalitions can shop their positions to candidates looking to either build a bigger influence coalition for Election #3 or to influence policy direction in Election #2.
Which is to say… the very first election after RCV will probably follow fairly predictable lines, but it might change the iterative aspect over time. Depending upon actual voting preferences, of course… not by magic.
I also support #1 on redistricting… and note some of my observations/concerns above.
If we get RCV, I would not support #3 (Jungle Primaries) as I want no primaries and actual parties to select their candidates, and #3 would be redundant or counter productive if you have RCV.Report
I cosign all of thisReport
Follow up:
Unless we abolish House districts and just elect House members by proportional vote which is probably betterReport
1. Non-partisan redistricting commissions are a no brainer to me and they seem to work. A lot of Democratic states have them. The issue is getting red or reddish states to adopt them seems impossible and GOPers in the Senate are deadset to block them with an assist from Manchin and Sinema.
2. I’m not enamoured of jungle primaries. I think sometimes part of the problem is that we think paritsanship is the problem. Primaries should be on the party level. Plus in a lot of states you will just see one party cut out. California Republicans largely are cut out off races because of the Jungle primary. This has not helped to moderate them. In fact, they seem to just let the freak flag fly more because of it.
3. Rank-choice voting is good as well.
I would also add:
1. National vote by mail.
2. Election Day is a federal holiday
3. Election voting lasts two weeks
4. No off year elections, all elections for all offices are held on even years. This is local, state, and national. The only exceptions are special elections to fill an empty seat.
3.Report
Co-Sign. Again.Report
Just to be pedantic, you mean something other then “federal holiday.” We have many federal holidays — non-essential federal offices are closed. States make their own decisions about whether to have a corresponding state holiday. Businesses make their own decisions. Juneteenth is now a federal holiday. It is unlikely that very many businesses will observe it, as they are disinclined to add another paid day off to their work calendars.
You mean some sort of never-before-in-the-US holiday, where the federal government mandates that non-essential federal and state offices will be closed, as well as all non-essential businesses.Report
Yep. FWIW I have seen some companies announce plans to make Juneteenth a day off.Report
Odd numbered year elections are off-off year elections and are meant to decrease turnout (See Virginia). However, governor races seem to be helped by not occurring at the same time as presidential elections.
If every states goes to vote by mail, then all of the states need to be in-hand states instead of postmark. If everyone votes by mail, then there is no point is a federal holiday since the proposal is meant to help Democrats.Report
Utah would like to disagree with the assertion at the end of the sentence. Colorado too.
That aside – yes if every state went vote by mail there would be no need of a holiday.Report
Arizona’s >80% vote by mail system and Montana’s >75% system were both installed by Republican state governments.
It was entertaining to watch Mike Lee on one of the talking heads shows running down a list of “bad” things in HR1. I believe 5 of the 6 were things Utah was already doing, some only recently adopted.Report
There would be no reason for Colorado to have a holiday on election day (day off for goverment workers) if 95% or more have voted by mail. Colorado also is an inhand state that gets it ballots counted quickly.
Considered that the Republicans are no longer relevant for state wide office in Colorado, then the measure was meant to help the Democrats since the turnout of non-whites has gone up. The all mail ballot could also be blamed for the election of Boebert to Congress since the turnout in the Republican Primary went up and Boebert won over a more main stream incumbent.Report
Damn, somehow clicked “Report”. Didn’t mean that.
Colorado votes mostly by mail. There are in-person voting centers to handle edge cases — eg, people who move to the state too late to register and get a ballot by mail, or people who need assistance and have no relatives or friends to do that. The only pure vote-by-mail state is/was Oregon. At least at one point you had to live there and be registered far enough ahead of the election to get your mail ballot. I remember reading that people who need assistance in voting in Oregon call a number and a qualified volunteer comes to their house. I believe even Oregon is changing that to handle last-minute registration.Report
Oregon is the pure case. The Democratic Primary is the only relevant election and the Democrats operate a closed primary. That means that the winning candidates are picked by the most active Democrats and the general election is moot due to the turnout of not so active Democrats in the Portland area.Report
I’m constitutionally (small “c”) opposed to Jungle Primaries, ever since David Duke used that voting system to get in a runoff with a criminal. I don’t think a multiplicity of candidates encourages moderation. The polls showed that either under a traditional party primary system or ranked-choice voting, Duke would not have had a chance. I’m not sure of the underlying dynamic, but it seems like the two step process encourages a certain type of voter in the first round (attracted to otherwise marginal candidates) and a different in the second round (less politically engaged, wait and see the meaningful choices).
I realize that the French Presidential election system isn’t technically a Jungle Primary because parties and their primaries are retained, but they do have a centrist, albeit pretty unpopular President who appears to be aiming for a runoff election with Le Pen in which he will have to be the choice as lesser of two evils. Some on the Left are signaling that they will sit such an election out. Maybe they will or maybe they won’t.Report
I was in graduate school home in Louisiana for that election and had a bumpersticker that said “Vote for the crook – its important.”
good times . . . .Report
Ranked choice voting sounds good on paper but it has several problematic elements. Firstly it’s relatively confusing for voters but far more importantly as a second point it takes a LONG time to count the votes and the vote counting process is necessarily byzantine. That’s a serious problem in our current voting environment.Report
Yeah, there’s something a little odd about ranking every candidate… like I wouldn’t even want to put a ’13’ beside some candidates.
Maybe a way to ease into it might be to require 50% + 1 to win, with a RCV of 3 candidates… if no candidate gains 50% in RCV-3… then actual Run-off between #1 and #2.
RCV doesn’t take long in any automated system; it takes a long time in systems that don’t want to be automated, though.Report
Last year Colorado had 20-some candidates for President on the ballot. I think California was even worse.Report
What about it makes it take a long time to count? It looks like a pretty simple algorithm.Report
It’s a simple algorithm if you’re counting it electronically but that isn’t, as I understand it, how votes are counted by and large nationally.Report
Ranked choice shouldn’t take any longer to tabulate.
Report the results? Yes. You wouldn’t want to report partial results, or if you did you’d have to do a lot of work (“Candidate X is the first choice of X%, second of Y, third of Z”, repeat per candidate) which would probably be confusing.
You’d wait until you had enough votes tabulated to be certain someone either was certain to win or certain to be eliminated and report that. (Which might vary from “We’ve tabulated roughly 70% of the votes, and Bob’s gonna win even if he’s last choice for everyone else” or “We won’t know until we’ve got the last ballot in, it’s real close”).Report
Slow counting/slow reporting, layers of results. I think it’s a problem considering how the country is accustomed to having a pretty good idea of the outcome. I mean it’s academic since there’s no way in Hades that the existing diad would endorse a massive change that’d allow third parties to have a shot at winning in second round picking.Report
Eh, that’s not really “political parties” so much as it is “inevitable outcome of a system in which you vote on things, and the majority wins”.
It forces everything down into two parties of roughly equal size — anything else is an aberration that is worked out of the system. This is true no matter how your government is set up (parliamentary systems might have a million parties, but they have long-standing stable coalitions — the difference between the US and the UK is simply that in one the platform is worked out before the election, and the other after).
If you need 50%+1 votes in your Congress-equivalent to pass legislation, you are incentivized to put together a stable coalition or party that can muster 50%+1 votes. Which means you’ll end up with two groups with a fairly equitable split. (And the whole ‘life goes on, things happen’ bit means eventually those coalitions will fragment and reform, during which you might see larger majorities or brief third groups until it’s all worked out).
If you grow too large past 50%+1, your coalition is stressed — the other group starts making more and more serious overtures, or you get purity tests (“why do we need those sellouts”) towards one end or the other, and you’ll fragment.
But it trends towards two stable groups, because in the end — you need 50%+1 votes to pass legislation. or stop it.Report
How is picking 1,2,3 confusing? First past the post can create situations where a minority of the voters picks the winner especially in a crowded primary.Report
It strikes me as simpler to assign points to the candidates. If you list up to five, the your first choice wins 5 points, your second choice wins 4, etc. It seems that this could be toted up immediately. So to take three hypothetical sets of choices: (1) Adams [5], Garcia [4], Wiley [3], Stringer [2], Donovan [1]; (2) Stringer [5], Garcia [4], Donovan [3], Wiley [2], Adams [1]; (3) Wiley [5], Morales [4], McGuire [3], Garcia [2], Adams [1]. You end up with: Garcia 10, Wiley 10, Adams 7, Stringer 6, Donovan 5, Morales 4, McGuire 3. (I chose these numbers at random and see that Garcia and Wiley are tied. I suspect that with more voters this becomes less likely. If it happens anyway, there could be a head-to-head runoff.) Is there some reason it isn’t done this way? I suspect that there is a technical reason I don’t understand.Report
I had this same thought, but it doesn’t work. The first step is to see if one candidate has a majority of 1s. It’s game over at that point, if it happens.Report