Wednesday, November 27, 2019

Compromise Point System

Meeting halfway between winner-take-all and popular vote.

Winner-take-all, for multiple reasons, is the worst thing about the presidential election system. Someone could lose the popular vote by 10%, and still win the electoral vote.

Trump took the presidency while losing the popular vote by 2.9 million, but winning 30 states overcame that vote deficit. In 2004, John Kerry could have won by flipping Ohio, also with a 2.9 million vote deficit, with 20 states and D.C.

In winner-take-all, people who don't vote for the state winner are not only ignored, but their voting power is stolen away and given to a candidate they voted against. It doesn't matter how many millions of Trump voters turn out in California, because their state will disregard them and support the democrat. California's Trump voters could represent 5% of Americans, ignored, and that's just the tip of the ignored iceberg.

A compromise system, between state votes and human votes, between winner-take-all and opposition voters having any say at all, is necessary.

The new system must retain the advantage that exists for smaller states, because small states will be needed to approve a constitutional amendment. This advantage tremendously helps the states with 3 electors. States with 4 electors have, on average, double the population of the 3-elector states, but only 33% more voting power. This made a lot more sense when the most populated states had nine to twelve electoral votes, rather than 55. Most states are in the same boat as Texas and California (cheated by Vermont and Wyoming). We're stuck with the disproportionality, but we can certainly fix winner-take-all.

I have developed a new system that achieves the following:
- Provides a fair compromise between winner-take-all and popular vote.
- Will not ignore half of a state's voters.
- Uses a point system that does not require a national popular vote count.
- Eliminates electors, who are just a useless liability anyway.
- States will still be able to declare a winner without counting every single vote.
- Has a low-population-state advantage that is very similar to the current system.
- Reduces the odds of a large popular victory being a point-system defeat.
- Is much more believable as a way to measure public support.

(There are some conditions that I would add to this or any presidential election system, to most reliably find the people's favorite:
- An open, unified primary to narrow down the field to the four best candidates.
- An approval vote, which allows people to vote for as many candidates as they like.
- A second section of the final ballot, consisting of all the possible pairs of candidates (4 candidates make 6 pairs), to guarantee that the more popular of the top two is winning the championship (like the super bowl, two teams only).
But even without these conditions, this compromise point system is far superior to the current electoral system, which provides only a rough approximation of how Americans are actually voting.)

FORMULAS

The current formula for electoral votes, as required by the constitution, is:

 (electors of a state) = (number of representatives) + 2

 (And D.C. has the same number as the lowest-populated state)

Number of representatives of a state is somewhat proportional to population.

The "+2" is what gives low-population states their edge. Raising 1 to 3 is huge.

It is debatable whether winner-take-all actually helps anyone, but that rule is from state laws.

My new system applies two different formulas, to give the highest point-getter, the winner of the state, an edge.

STATE WINNER = (number of representatives + 1) x (winner's % of the votes) x 2

Notice the "+1" and the "x2," advantages that the losing candidates won't have. That's also the statehood bonus.

STATE NON-WINNERS = (number of representatives) x (candidate's % of the votes)

The percentages will be used as their nominal values, so 20.637% means multiply by 20.637. Technically, it will be:
100 x (candidate's votes) / (all votes in state)

Fractions of a point will be rounded to the nearest whole point in each state.

If a candidate has a small amount of voters, maybe 300, maybe 1200, it will sometimes not be enough votes to round up to one point, so they would receive zero. A handful of votes will usually not be enough to make a difference, so the future of the country won't depend on litigating over every single vote in every state. At the same time, rounding to the nearest point is tremendously more precise than rounding to the nearest entire state.

How does this work for small-state advantage? In a 1-representative state, the points will work out to be:

Winner = 2 x 2 x (winner's % of the votes)

Non-winner = (candidate's % of the votes)

That's a ratio of 4:1, if they receive a very similar number of votes.

In California, the calculations are:

Winner = 54 x 2 x (winner's % of the votes)

Non-winner = 53 x (candidate's % of the votes)

A ratio of 2.04:1 for evenly-matched candidates.

So in smaller states, the state winner gets a much larger proportion of the state's points. This should please those who believe smaller states need winner-take-all.

A large state's non-winners getting a significant portion of a state's points should please those who want the national count to resemble the actual popular vote.

One more ratio is important to note. In a 2-way contest, the influence that each state will have is the DIFFERENCE between the state's points for the winner and points for the loser.

So when all candidates have 50%:

Small state (winner points) - (loser points) = 150

California (winner points) - (loser points) = 2750

* The ratio between the influence of California vs the small state is 2750/150 = 18.33.

* The current ratio between electoral votes of California vs D.C. is 55/3 = 18.33.

The same ratio holds up whenever the top two candidates are evenly matched. In a 4-way race that has them at 25% each, CA vs DC will still be 18.33:1.

When the winner receives many more votes than the loser, the ratio increases a bit, but not enough to make a big difference. If, in each state, the winner has 1.5 times the votes of the loser (60/40), the ratio becomes 21.8:1, about 55/2.5. If it's 70/30, the ratio is 55/2.3.

The population ratio of California to Wyoming is about 67:1.

The small-state advantage has not had much effect on elections since 2000. D.C., Vermont, Delaware, New Hampshire, Rhode Island, Hawaii, and Maine usually support democrats, so they balance out the smallest republican states.

USING THE NEW SYSTEM WITH ACTUAL VOTE COUNTS OF RECENT ELECTIONS

Always remember that if the rules of the election are different, everyone will behave differently, and the vote counts and results could be quite different. For example, votes for small party and longshot candidates will be more common if the people know that protest votes can reduce the points the state winner will take.

I created spreadsheets to test four recent elections, and the results are:
2016, Trump wins by a smaller margin.
2012, Obama still wins by a lot.
2004, Bush wins by a strong, and now secure, margin.
2000, Bush wins, but just barely.

2016
Electoral system should have produced a result of 306-232, that’s Trump winning with 56.9% of electoral votes. Although he lost the popular vote by 2.9 million, Trump would still win the compromise system by 49.81 - 46.78%, a more than 3% advantage.
It is understandable that popular-vote supporters would want Trump to lose, but it seems his 2016 turnout would win any reasonable system that provides a statehood bonus.
It's hard to lose when you win 30 states, most of the medium-sized states, Ohio, Texas, Florida, Pennsylvania, and 4.5 million votes in California.
HYPOTHETICAL: If Clinton took 113,000 more votes to win Florida, Trump would still win the current system, by 3% of electoral votes. But the compromise system would give Clinton the win with a 0.9% point lead, and a 3 million vote advantage.
HYPOTHETICAL 2: Clinton needed to win all three, Wisconsin, Pennsylvania, and Michigan, to win the electoral system. Flipping Pennsylvania would make it possible for Clinton to win the new system, but it would have to be a huge flip, taking away 5.5% of Pennsylvania's vote from Trump.
However, she could win the compromise system by barely flipping the two smaller states, winning 23,000 more votes in Wisconsin, and 11,000 more votes in Michigan.

2012
Obama won 26 states and DC, 51% of the popular vote, and a 5-million vote majority.
Obama won 61.7% of electoral votes.
New system, Obama wins with 55% of electoral points, to Romney's 43%.

2008
Obama landslide, no need to run the numbers.

2004
Bush won 31 states, 3 million more votes than Kerry, and took 53.1% of electoral votes. Bush would win the compromise system by 5%.
HYPOTHETICAL:
If Kerry had done a little better in Ohio, the electoral system would have given him the White House, despite losing the popular vote by 2.9 million. Bush wins 30 states by higher margins, for naught!
The new system, if Ohio flips to Kerry, Bush still wins by more than 2.1%.
To win the new system, Kerry would need a ridiculous 78% landslide in Ohio, with 800,000 more Ohioans voting. In that case, he would be winning the national popular vote by 700,000. That many votes are necessary to overcome his lack of states.

2000
“Winner-take-all,” coupled with Florida's incompetence, made this the election from hell. Evenly-divided states should not have as much influence as the electoral system gives them.
The supreme court ruled that Bush won the presidency, with 50.3% of the electoral votes. It was eventually certified that Bush won Florida by a little over 500, and New Hampshire by about 7200. Iowa, New Mexico, Oregon, and Wisconsin were also very close, in Gore's favor.
Gore won the nationwide popular vote by over half a million.
Bush wins the compromise system by 0.21%. Flipping Florida, or any state to Gore would still flip the election.
OK, I admit that my system doesn't fix the 2000 situation, but people would more readily accept the result of an election that doesn't throw out half the country's votes.
And again, if the rules were different, the campaigns and the vote would be different.
HYPOTHETICAL:
Gore could win the new system by flipping 200,000 votes in California, or 120,000 votes in Texas. It was a very close election.
The best remedy for Gore would be for Nader to not act as a spoiler! And the best way to do that is to collect more information from voters, for example, allowing people to vote for more than one candidate.
Winner-take-all distorts voter turnout. Maybe 59% of Texas preferred Bush to Gore, maybe it was 55%, maybe 75%. Maybe 35% of Americans wanted Nader, but we'll never know. A lot of potential voters stay home, because winner-take-all discourages people from actually voting when the state's outcome seems certain.

Hey, remember when Gore couldn't be president, because he was too much of a liar? Hahahaha. Ha. We are so doomed.

Anyway, the people's actual preferences should mean something, but winner-take-all prevents that.

Monday, November 18, 2019

A Real Instant Runoff

An infinitely better presidential election

Change the spring election to a unified open primary, with all candidates on one ballot, and the favorite four will advance. Each voter may choose two from the primary field. Parties are limited to having one candidate advance in the top four.

The fall ballot will have two separate parts:
1. Voters may choose as many of the four as they wish. (Approval vote.)
2. Voters are also presented with the six possible head-to-head pairings, and may choose their preferred candidate of each pair.

The primary determines the top four.
Part 1 in the fall narrows the field to two.
Part 2 is a real runoff between the top two. (Only one of the six pairings comes into play.)

The statehood bonus, and state-by-state tabulation could still be used, with a point system instead of electors. One option would be 300 points minimum (instead of 3 electors) per state for national tallies for the top four and top two, and 300 minimum for the final, with the winner of each state taking 100 points, and the remainder allocated proportionally to the top two per state.

Instant Runoff?

The term "Instant Runoff" is usually used interchangeably, and erroneously, with "Ranked Choice," a voting system that many people are suddenly excited about.

I have been aware of ranked choice since the 2000 election. I have done a lot of looking at it, and I can confidently say that the concept is very flawed. Ranked choice creates the illusion of a majority, when the actual majority might prefer another candidate. The 3rd choices of fringe voters are considered, while the 2nd choices of many mainstream voters are ignored; this is unequal voting power, in favor of weirdos. The complexity of shifting votes between candidates requires computer tabulation, and hand recounts have been called off due to difficulty. Ranked choice is not worth using. There are multiple superior alternatives.

1st and 2nd choices are apples and oranges. Ranked choice only counts some people's oranges as being worth apples. Approval voting makes it clear to all that all votes will count the same.

Ranked choice is not even an accurate "instant runoff", it is only approximate. While some people will vote strategically in a ranked choice, the only possible strategy in a top-two is to vote for the better one.

Top 4, Top 2, Head-to-Head

This system takes the primary out of the polarizing hands of the parties. Candidates who advance can be selected by a majority of all of the voters, not just a majority of the 1/3 who are party members.

The top four concept should result in a redistribution of power to at least four major parties, which would provide more options, and would make it less believable for one partisan to say "You have to support me, because the other three are all the devil!"

Limiting the field to four candidates also makes possible a real, "instant," head-to-head runoff, as a safeguard against possible strategic voting or spoiler effect. (Six pairs is doable, but six or seven candidates would make 15 or 21 pairs, which would dazzle voters.)

The fall election produces the actual top two with an approval vote that decides which two are acceptable to the most voters. Approval voting allows people to not have to choose between their favorite and one "who can win;" they can choose both. Or negative voters can choose the three opponents of the one they hate most, instead of giving the lesser of two evils disproportionate support just to keep the other one out.

The pairs section allows ALL voters to give input as to which of the top two is better. This is like insurance against a weird situation such as one liberal vs three conservatives who might split some votes, or a virtual 4-way tie. Although no one voting will know who the top two will be, it should still work for people to vote on each of the six possible pairs, or on as many pairs as they care to.

Why use Approval instead of only head-to-head pairs? It is important to have two sections because this allows voters to precisely answer two different questions. The first question is "Which of these would you want to win vs the others," and the second is "Which in each pair is better." If the fall ballot was head-to-head only, voting for a poor candidate over a terrible one could cause the poor one to beat the best one! With two separate sections, we are free to choose one per pair, without fear of putting a poor one over the best one.

[UPDATE 1/22/2020: I did another  blog post featuring this plan, complete with PICTURES!
https://americarepair.home.blog/2020/01/17/head-to-head-matches-make-a-better-instant-runoff/ ]