Electoral College Abolition Amendment and IRV Bill

scoobrs writes "Two bills, H.J.R. 109 and H.R. 5293, were introduced in the US House by Congressman Jesse Jackson, Jr. (D-IL). The first is a constitutional amendment abolishing the electoral college. The latter is a bill providing for instant runoff voting in all federal elections by 2008."

Comments

Why IRV?

[Gadzinka]

I don't know why modern political-reformists are so fixated on IRV. Of all the technical criteria of "fair voting" IRV fulfills NONE. In this respect it's worse even than "majority vote".

I mean, why would you want to go with a voting scheme, that makes possible situation that adding votes for a candidate causes him to lose, and converselly, removing votes for a candidate causes him to win?

Why not go directly with "aproval" or even "condorcet"?

Robert

PS Go, read the above link to find out what's exactly wrong with IRV.

Re:Why IRV?

[dschuetz]

If 20% of the population prefer Nader to Kerry and Kerry to Bush, 35% of the population prefer Kerry to Nader and Nader to Bush, and 45% of the population prefer Bush to Kerry and Kerry to Nader, then 80% prefer Kerry to Nader and 55% prefer Kerry to Bush, so Kerry wins by every pair-wise comparison and hence would win the election.

Ah, that seems much easier than what I just wrote. Lemme try and clean it up:

20% vote: Nader, Kerry, Bush
35% vote: Kerry, Nader, Bush
45% vote: Bush, Kerry, Nader

You end up with three two-way elections: Kerry/Bush, Kerry/Nader, Bush/Nader.

Kerry beats Bush in 55% of ballots.
Kerry beats Nader in 80% of ballots.
Nader beats Bush in 55% of ballots.

No other candidate beats *both* of the other two in more than 50% of the ballots cast, so Kerry wins.

Re:Why IRV?

[dschuetz]

Thats fine for a 2-5 person race, but imagine a 15 seat city council election, with over 25 people running, you would have a 10 page ballot.

No, because the "A > B, B > C, C > D" comparisons are inferred from the ranking. You only need to rank them all at once.

So a 25-person race would just have 25 names listed, and you put a "1" next to the person you like best, "2" next to your second choice, etc.

That said, I'm not sure how Condercet works for a multi-seat election like a county council. I guess it's just a question of ranking the final results.

Re:Why IRV?

[Minna Kirai]

That said, I'm not sure how Condercet works for a multi-seat election like a county council. I guess it's just a question of ranking the final results.

Pick the winner as per normal.
Delete him from all ballots.
Repeat until there are no more slots to fill.

(The same repetitive approach can work with concordet, plurality, or IRV)

Re:Why IRV?

[the morgawr]

Can anyone who's had more than 20 minutes with Condercet comment on this?

Not the way I'd explain it but it is pass-able. Personally I prefer simplified examples.

Why I'm NEVER going to support IRV in a National Election:

We used to use run-off voting in our Fraternity Elections before we swapped to Condorcet. What run-off voting does is eleminate compromise candidates early on. In a national election this will favor the more extream candidates over the moderate ones.

Example:

We have three candidates X,Y,Z. Let's say the voting goes like this:

40% Like X the most.
40% Like Z the most.
20% Like Y the most.

However, 30% of the voters for X, would rather see Y win than Z and 30% of the voters for Z would rather see Y win than X. The remaining 10% only want their candiate to win. So the break-down of the ballots looks like this:

10% - X --- They only like X, Y & Z are equally bad.
30% - X,Y --- They prefer X but like Y more than Z
10% - Y,X --- They like Y but lean toward X
10% - Y,Z --- Y with leaning toward Z
30% - Z,Y --- Like Z; like Y less than Z but more than X
10% - Z --- Only like Z, Y & Z are equally bad.

Now, this is a democracy so our voting should try to make the maximum number of people happy (alternatively, we could define fairness as minimizing the number of unhappy people, more in a moment). Ideally Y should win, because the most people support him, the fewest oppose him, and he would win one-on-on against both other candidates.

However with plurality or IRV we end up with a tie between X and Z (because the "compromise" candidate is eliminated in the early round).

Condorcet solves this problem by breaking each election up, into a bunch of one-on-one elections and figuring out a winner in all of these simplified cases.

Condorcet has certain interesting properties by design: it is essentially stratagy free (being dishonest with your vote does not get you any ground), and it will find an Ideal democratic winner if one exists. However it does have several practical limitations (that are mostly irrelevent to smaller groups but could cause problems if used in something as big as a Presidential election): Because it is ranked, adding one vote can swing a very close election in unexpected (but technically correct) ways, as such, you cannot break the counting up as you can with plurality or approval the counting must all be done on all ballots simultaniously (this is no problem for a small group but for a large election, it would require computer-systems to count up the vote), and finally, some people claim that Condorcet implicitly compromises on behalf of the voter.

Note that Condorcet is not perfect; it is however the closest to perfect that exists.

There is another method that has most of the good points of Condorcet but trades away some of them to get a few practical benefits: Approval. Approval voting asks the voter to mark all candidates who he approves of for office (the goal being to minimize disatisfaction).

Unlike Condorcet, under Approval voting, adding one more vote does not cause an unexpected outcome (but the outcome might not be technically correct); Approval voting also allows the vote counting to proceed in smaller groups and have the result total up and make sense. Additionally it forces the voter to decide what they are (or arn't) willing to compromise on.

However approval voting is subject to some voting stratagy and a successful implementation relies on explaining the strategic aspects to the voters. Because of the strategy element, Approval voting is not as accurate as Condorcet for small groups and groups that don't follow polls.

An example of a good stratagy is to watch the polls and vote for everyone you prefer to the front runner, then vote for the front runner if you prefer him to the second-place candidate.

Re:Thanks!

[Tyndmyr]

Ditto for that!

The electoral college does need to remove winner take all...but this aint gonna solve that.

And why, oh why, did they choose IRV? Possibly one of the worst systems they could have chosen. Alright, you could make an arguement that it might be better than the current system, but its vastly inferior to concordent(which is unfortunately complex) and my personal favorite, Approval Voting.

On the bright side, Im glad people are taking note of this, though I fear this will be used as a reason to ignore other pushes for election reform.

Re:Ya know...

[alwayslurking]

This article suggests approval;

The Mathematical Association of America and the American Statistical Association each elect their committees by a new method called approval voting.

Re:Thanks!

[Canthros]

I suppose I'm a hardliner; I favor leaving the broader system as is.

I think the electoral college works fine, and the state-level winner-take-all approach forces candidates to appeal to a broader base of voters in most states (New York and California being anomalies in which very large urban areas completely dominate the whole state).

Likewise, I see nothing wrong with the present voting system. It's simple, and it works. While I don't disagree that this can limit national support for third party candidates in marginal situations, I am also fairly convinced that the existing style of voting works plenty well provided that there is broad enough support for the third party in the first place. Which is to say, if a third party candidate were to provide a platform that was interesting to a broad enough number of Americans, I am pretty sure that they could win the Presidency. Especially if they can cough up the funds to campaign effectively.

Re:Thanks!

[aqkiva]

I was for eliminating the electoral college until I read this: Math Against Tyranny. It also makes the analogy to baseball runs vs. games. Alan Natapoff has mathematically shown that voters have more power with the current system where power is defined as the ability to tip an election in any one direction. Basically, if it was purely a popular vote, the only way your vote would matter is if the rest of the voters split exactly down the middle. Given the size of the US population, the probability of this is extremely low. Especially given that people tend to lean towards one candidate or the other, the chance of deadlock is essentially nil under a popular vote. That means each voter has no power to tip an election and thus politicians have no reason to listen to them. Dividing into smaller groups means that each group is more likely to deadlock and so each voter has more power. Thus, what happened in Florida in 2000 was a good thing. In fact, the best thing to do is to re-divide up the nation into groups such that every single group would be very likely to deadlock. The winner would then take-all from each group, making it so that all politicians would have to work to win votes in every single group.

Re:Colorado will become irrelevant if they pass th

[michael_cain]

What would happen if 95% of all Americans lived in cities? Would the 5% of rural voters still get 50% of the representation?

The obvious question to follow up with is "Which cities?" If 95% of all Americans live in Chicago, the West Coast cities, and the stretch from Boston to Washington, D.C., (call it 12 states) then they will be under-represented. Very badly in the Senate, where they would have 24 out of 100 senators, least badly in the House where they would have a large majority of the representatives but still not 95%, and somewhere in between in presidential elections.

Speaking as someone from a large western state with relatively few people, great scenic beauty, and rich in natural resources, let me say that replacing the current system with one that was based solely on population would be terrifying. I can easily envision the 95% who live in the 12 states (in this example) passing federal laws that do a variety of things: requiring that we strip-mine the resources; requiring that we operate massive land-fills in the non-scenic areas to dispose of waste from the urban states; requiring that we ban all development in scenic areas (even though the large majority of that 95% will never visit them); requiring energy-efficiency standards that make sense in an urban setting but are simply not practical in my state.

One of the key issues that the Founders wrestled with in writing the Constitution was how to make it difficult for a small group of states with large populations to impose their will on the other states. I would be happy to entertain systems other than the current one. Can you suggest one that guarantees my state's ability to have a meaningful say in governing the nation that doesn't give me "over representation" relative to our population?