NPV passes in Oregon


#1

Today is a dark day for voting system reform, and the culprit is not what you would expect. National Popular Vote passed in Oregon yesterday.

As currently written, NPV is fundamentally incompatible with alternative voting methods. I’m 100% for reforming the electoral collage, but as the NPV interstate compact is written, states who don’t use our archaic “Choose One Only” Plurality voting actually would not have their votes count towards the popular vote.

This is a disconnect that could have been fixed, but after over 2 years of reaching out a satisfactory solution has not been proposed. Will NPV leadership wake up, or will they essentially entrench vote splitting and the spoiler effect into our Presidential elections for posterity? Or is it too late?

We remain committed to not only fighting for better voting, but also working to make the election reform movement as a whole more cohesive. There are a lot of problems out there, and solving them all is possible. We need to think big picture.

"Chapter 6 - Every Vote Equal p. 263 starts to answer you question about what would happen if a state elected the president by anything other than a popular vote:" - Eileen Reavy

"Because the purpose of the compact is to achieve a nationwide popular vote for President and Vice President, the popular vote counts from all 50 states and the District of Columbia are included in the “national popular vote total” regardless of whether the jurisdiction is a member of the compact. That is, the compact counts the popular votes from member states on an equal footing with those from non-member states. Votes from all states and the District of Columbia are treated equally in calcu- lating the “national popular vote total.”

Popular votes can, however, only be counted from non-member states if there are popular votes available to count. As previously mentioned, Article II of the compact guarantees that each member state will produce a popular vote count because it re- quires member states to permit their voters to vote for President and Vice President in a “statewide popular election.” Even though all states have permitted their voters to vote for presidential electors in a “statewide popular election” since the 1880 election, non-member states are, of course, not bound by the compact. In the unlikely event that a non-member state were to take the presidential vote away from its own people, there would be no popular vote count available from such a state."


#2

" Addressing Concern 1: I think this is what Jacob & your colleague [Clay Shentrup] at Unrig [2018] were discussing. My understanding is the same as yours. If a state decided to select their electors in a manner that did not create a popular vote total they could not be counted in the national popular vote total. State legislatures get to decide how to select electors in the electoral college and any state looking to use an alternative voting method for this would have to weigh that decision against the possibility of their votes not counting towards a national total if the NPVIC goes into effect. Hopefully good governance would mean they’d do what you suggest and draft the new election law in a way that still created an output for a national popular vote, otherwise they’re adopting a way of choosing electors in the electoral college that only focuses on the votes cast in their state and is no better than what we currently have." -Eileen Reavy


#3

QUESTION: How can score vote totals be summed with plurality vote totals? If it can be done, then each states STAR runoff can counted be between the national top 2, which would prevent a states votes from ever being the spoiler in a presidential election.


#5

I’ve written on using a variant of STAR voting in an interstate compact before, though the compact I was considering was designed with alternative voting methods in mind. For the NPVIC specifically, you might be able to use the same process for scores, normalizing them to be between 0 and 1. It’s just that adding those scores with plurality votes is a lot more awkward than doing so with approval votes.


#6

I’m pretty sure that approval is compatible with the NPV, and I believe that STAR is.

With approval, you just report numbers of approvals for each candidate. That is a valid “popular vote” and it makes sense to add it with (presumably largely strategic) choose-one totals.

With STAR, you report “number of voters who rated the given candidate above zero and equal or above both of the state’s frontrunners”. By definition, at least one of the state’s frontrunners will get an approval from each ballot, unless it rated both frontrunners at zero. This is a kind of declared-strategy approval count. It will almost always mean that the statewide STAR winner has the highest statewide popular vote; though there are possible exceptions, they are quite unlikely to occur in practice, and are arguably an improvement over the STAR winner.

If all states used STAR, this would not mean that the national STAR winner would necessarily win, because different states could have different frontrunners. But the chances that this leads to a change are not too high, because the true national STAR winner is likely to be among the frontrunners in most or all states. And anyway, even with this flaw, this national voting method would still be dominant over national choose-one and over the current electoral college mostly-winner-take-all method.

(Note that this same idea could be used with 3-2-1, and is even less likely to change the winner with 3-2-1 than with STAR.)


#7

Re Jameson Paragraph 1-2: Wouldn’t summing approval votes with plurality mean that Approval voters would potentially have a stronger weighted vote due to the fact that you can approve multiple people?
Re Paragraph 3: I’m not following that. Can you explain it with numbers and not through the lens of Approval?

Here’s what I was proposing, but I’m not sure the math works out. I’m not sure if step 2 below is needed or if it does the trick. STAR Voting runoff votes can be summed easily with plurality, but what I’m looking for is the simplest way to sum a 0-5 score ballot with plurality votes. Once that’s done we would need to do the STAR runoff using Oregon ballots, but between the national top 2. That would ensure that

IDEA:
Step 1: Divide all candidates scores in the 1st round by 5 (ie. normalize ballots to 0-1)
Step 2: Re-weight all Oregon votes to ensure that Oregon vote totals equal the number of voters and that all votes are equally weighted. Then add to NPV vote totals for each candidate.
Step 3: Determine national top 2 candidates
Step 4: Conduct the runoff of Oregon STAR ballots, but between the national top two candidates, rather than the local top 2. (This would ensure that Oregon wouldn’t be a spoiler, and that the election wouldn’t get thrown to the House if candidates failed to get 270 electoral collage votes.)
Step 5 with the National Popular Vote Interstate Compact: Divide Oregon’s number of voters proportionally between the top 2, (using the runoff preference between those 2).
Step 5 with the Electoral Collage as it is now with winner take all: Oregon’s votes would all go to the national front-runner who was preferred in the Oregon Runoff between those 2.
Step 5 with Equal Citizens proposal for proportionally distributed electors: Divide Oregon’s electors proportionally between the top 2, (using the runoff preference between those 2).


#8

I’ve often wondered if there’s a way for one or more red/blue flyover state pairs to make a pact with each other to break the winner-takes-all arms race (making candidates start to care about them again) but without giving one party an immediate advantage over the other (or else it would never pass). I’m not sure how it would work, though.


#9

Does my idea above work? This is a time sensitive question as NPV will be signed into law by our governor any minute now and I’m meeting with their rep and need to know if there is a proposal possible.

spephomancy- the proposal you described is a plan that Lawrence Lessigs group Equal Citizens is working on. They are challenging winner-take-all in court, in pairs of red and blue states, using the equally weighted vote ruling. If they succeed the legal president could strongly support making FPTP illegal for the same reason and will mandate a switch to a voting system that passes the Equal Vote criteria (STAR, Score, Approval, etc.)


#11

That’s good, but I mean a voluntary pact between small numbers of states. NPVIC Lite.

Almost every state adopted winner takes all in an arms race with each other, and won’t relinquish it if it gives their party a disadvantage, but this means that safe states get no attention/money from candidates. Is there a way that safe states can enter into a pact to regain relevance without obviously disadvantaging their party?


#12

I have no words for this. Really. I must have written 3 to 5 posts with various “solutions” or complaints only to delete them immediately after.


#13

Re Jameson Paragraph 1-2: Wouldn’t summing approval votes with plurality mean that Approval voters would potentially have a stronger weighted vote due to the fact that you can approve multiple people?

Strategically optimal approval voters would have the same voting power as strategically optimal plurality voters. It’s just that it’s easier to be strategically optimal under approval. Which IMO isn’t really unfair, it’s just life; a reason to switch to approval.

Re Paragraph 3: I’m not following that. Can you explain it with numbers and not through the lens of Approval?

Say I vote 54332100 on ABCDEFGH. If the frontrunners are A and anyone else, my ballot counts as 1 vote for A and none for anyone else. If the frontrunners are C and F, my ballot counts as 1 vote for ABCD and 0 for EFGH. If the frontrunners are C and D, same; that’s the only way this ballot can end up giving 1 point to both frontrunners. If the frontrunners are G and H, my ballot counts as 1 vote for ABCDEF and 0 for G and H; that’s the only way this ballot can end up giving no points to either frontrunner.

Here’s what I was proposing, but I’m not sure the math works out. I’m not sure if step 2 below is needed or if it does the trick. STAR Voting runoff votes can be summed easily with plurality, but what I’m looking for is the simplest way to sum a 0-5 score ballot with plurality votes. Once that’s done we would need to do the STAR runoff using Oregon ballots, but between the national top 2. That would ensure that

IDEA:
Step 1: Divide all candidates scores in the 1st round by 5 (ie. normalize ballots to 0-1)
Step 2: Re-weight all Oregon votes to ensure that Oregon vote totals equal the number of voters and that all votes are equally weighted. Then add to NPV vote totals for each candidate.
Step 3: Determine national top 2 candidates
Step 4: Conduct the runoff of Oregon STAR ballots, but between the national top two candidates, rather than the local top 2. (This would ensure that Oregon wouldn’t be a spoiler, and that the election wouldn’t get thrown to the House if candidates failed to get 270 electoral collage votes.)
Step 5 with the National Popular Vote Interstate Compact: Divide Oregon’s number of voters proportionally between the top 2, (using the runoff preference between those 2).
Step 5 with the Electoral Collage as it is now with winner take all: Oregon’s votes would all go to the national front-runner who was preferred in the Oregon Runoff between those 2.
Step 5 with Equal Citizens proposal for proportionally distributed electors: Divide Oregon’s electors proportionally between the top 2, (using the runoff preference between those 2).

The problem with a plan like this is you need to look at the NPV vote total to figure out the first round, then change it in the second round. You can’t guarantee other states would let you do that, or that it would even be compatible with the voting method they’re all using.

I think it’s safer to have a system that looks at the state ballots only, and produces an approval total for each candidate. Each ballot should count as 1 or 0 for each candidate, and this should be semi-honest and not obviously strategically stupid. My suggestion meets these criteria. It’s not perfect if the frontrunners differ in different states, but it’s strictly better than the current system or than plurality NPV.


#14

All right, the compact needs to be scrapped and replaced with the following:

As soon as the compact reaches 270+ votes within 365 days of the (first?) presidential election, the compact awards its votes to the popular vote winner of all and only states within the compact. This means that once it gets to 270, every state will be forced to join the compact. (The 365 days give every state time to actually ratify the compact.)
All states within the compact must use the same voting system (or pick from a list of compatible ones). The system (and compatible systems) can be changed by a 60% vote of the governors, again at least 365 days before the first state’s election.

This seems to be the only way to allow voting system experiments without weird compatibility rules that IMO destroy the essence of the voting system they claim to emulate.

IRV can be included in a Score/STAR election

For Score, or STAR’s first round: Run the IRV system. The top two finishers get their vote counts in the final round. All other candidates get as many votes as they got the round they were eliminated. Alternatively: In the final two, eliminate the second finisher and transfer as normal. This means that ranking someone last is NOT the same as leaving them blank!!
For the runoff: An IRV voter’s full vote goes to the candidate you ranked higher.

(Basically, your ballot is converted to an Approval ballot for all the candidates your vote ever touched.)

This is as simple as IRV is and makes sense from the perspective of someone familiar with IRV. It is IMO far better than WDS’s suggestion.

Of course, this does nothing to resolve the precinct-summability problem (other than that you do it within each IRV state unless they agree to combine), and it makes batch elimination impossible. But richie is clearly not going to listen to those concerns, so we might as well compromise.

EDIT: Well, of course, the IRV propagandists may not like that the overall vote is done via Score/STAR, and it is IRV that has to bend. Too bad. The other way around leads to problems with vote equality.


#15

Can you explain this more please?

Step 4: Conduct the runoff of Oregon STAR ballots, but between the national top two candidates, rather than the local top 2. (This would ensure that Oregon wouldn’t be a spoiler, and that the election wouldn’t get thrown to the House if candidates failed to get 270 electoral collage votes.)

Why can’t whatever method a state uses be fully totalled at the State Level to produce the Winner of the state election and then provide the number of votes for that winner up to the other Chief Election Officers of the other states.
That is how the Plurality winner of each state is determined.


#16

How do you propose to force states to join a voluntary compact?


#17

OMG this is just one more of the 726 “Rube Goldberg” election schemes being proposed here. (“Goldberg is best known for his popular cartoons depicting complicated gadgets performing simple tasks in indirect, convoluted ways.”)

All I ever asked for was simple score voting – Voters get to grant from (1) to (10) votes to each of as many candidates as they desire, and then all the votes are hand counted. And they can exercise any strategy they prefer. Freedom at last. Very fast, simple, safe, and easy. What more could we want? Whoever gets the most votes wins. Period.

I question whether people even begin to understand this electoral college issue at all. Wyoming gets one Electoral College vote per 195,000 people. California gets just one per 712,000 people. Note that Wyoming, with 577,737 population gets the same number of senators as California, with 39,557,045 population. =/ 39,557,045/577,737 = ~68.47 /=! So each Wyomingite get over 68 times as much representation in the US Senate! This is simply because the perennially fuzzy concept of sovereignty attached vastly more tightly to the several states during the era of the constitutional convention. That has obviously changed almost completely.

The thing to understand is that the “electoral college elector vote” is in fact just a sort of proxy for a “state electoral vote”. The constitution says that US senators are pretty much chosen by states – not so much by individual voters at all (essentially by indirect democracy). There was never any absolute need for any body of “electoral college electors” at all. The term of the incumbent house of representatives could easily be extended for a few days to certify the raw “state electoral votes”. How many here actually know the names of the “electoral college electors” whom they voted for – show of hands?

Then why did the “founding fathers” create the body of “electoral college electors” in the first place? Simple: They were convinced that the selection of the president needed to be kept as far from the dirt-encrusted hands of the voters as possible. They certainly did not believe in democracy, and they did their very best to set up a fake democracy. Ben Franklin and maybe George Washington were exceptions to this. People like Thomas Jefferson were psychopathic hypocrites who always did precisely the opposite of what they so eloquently and extravagantly proclaimed. Hamilton truly believed that the 1% were finer folks, so maybe he at least was sincere.

This “National Popular Vote Compact” is a pathetic joke. What if some members of this “compact” simply renege and decide to throw their votes in whatever direction they want? What then?

Everybody so loves complicated “solutions” to simple problems, at least until their tricky “solutions” turn into seething quagmires.

“KISS”: Keep It Simple Stupid!


#18

Hey, nobody’s saying that they want this. It’s a plan to attempt to mitigate a possible bad situation.

Also, I can’t help but find the rebranding of FPTP as “the popular vote” to be a little sickening.


#19

So, I looked up the OR 2016 election results as something to consider for this discussion.

Presidential general election, November 8, 2016[19]
Party	Candidate	Votes	%
Democratic	Hillary Clinton	1,002,106	50.07%
Republican	Donald Trump	782,403	39.09%
Libertarian	Gary Johnson	94,231	4.71%
Pacific Green	Jill Stein	50,002	2.50%
write-ins	72,594	3.63%
Total votes	2,001,336	100.00%

Let’s say that Approval Voting was used instead and enough Democrats also voted for Jill Stein to actually bring her up to 2nd place. Then the total number of votes would be higher than voters in Oregon so other states would not want to accept those numbers.

Let’s say that STAR was used instead and enough Democrats also scored Jill Stein high enough to actually bring her up to 2nd place. Because of the Automatic Runoff of the top 2, the results from OR would only be Hillary and Jill. Trump would not have any numbers. Also, no way to include the people who scored Hillary and Jill the same.
Now when the totals from all the states are added together, would this cause any problems? The Jill votes are not spoilers. However, not being able to provide the number of people who wanted Trump seems wrong. So, would there need to be a top 3 or top 4 runoff in OR to be able to provide numbers for all the leading (15%?) candidates?

Let’s say that IRV was used instead and all 4 leading candidates made it above 15% and only the write-ins were eliminated. Then by providing the numbers for those 4 candidates, it would end up being a Plurality national popular election.

Of course using the same Alternative method in all states for a combined alternative popular vote would be better but still think the NPVIC is better than the electoral college winner take all situation - and an alternative method could still be adopted nation wide.

If I am missing something please let me know.


#20

Re Chris Mason, Step 4. Yes, you can total the Oregon ballots all the way through the runoff and determine who won Oregon, but the problem with that is that if Oregon’s winner is not among the national top 2, then Oregon’s votes would become the spoiler. This is a real risk, not only because Oregon is so progressive, but also because the rest of the country would be using the FPTP 2 party system, while I believe that STAR could break 2 party domination. If STAR Voting in Oregon ever spoiled a national election it would be catastrophic for voting reform. It’s preventable, and I would like to see it prevented. Preventing a spoiler is the point of an automatic runoff, and there’s no good reason not to use it to do that nationally.

The hard part is how to sum the first round score votes. I still don’t understand Jameson’s assertion that an approval ballot can be summed with plurality ballots in a way that’s fair to the plurality voters. His suggestion also seems more complex from a tabulation standpoint.


#21

I think he’s saying that a voter maximizes their voting power by voting for one frontrunner and against the other(s), which they can do under both approval and plurality (though it comes at a higher cost in plurality, hence the incentive to switch to approval). Thus, neither method gives a voter more control over the outcome than the other. I’m not convinced this is perfectly fair, but I don’t think it’s any less fair than plurality voting itself.

I think the exact opposite is true. Your suggestion requires that two rounds of national totals be calculated under a compact which is set up to handle only one round of national totals, whereas his suggestion avoids that and thus a lot of complexity. I do understand your concerns about Oregon acting as a spoiler, but I think producing approval totals is the best that can be done to fix that.


#22

Thanks, BTerary Tau, that’s helpful. To clarify, my suggestion requires looking at the national sum to determine the top 2, then counting the local ballots and doing a final runoff of Oregon ballots. It’s still only one round from a national standpoint, the runoff is just of Oregon ballots. (And same for any other states that use STAR.)

There are really 2 distinct questions.

  1. How do you sum score votes and plurality votes.
  2. Should you have a runoff between the national top 2

Let’s focus on just answering #1 for starters.