State legislatures are given broad Consitutional authority to appoint electors:

Each state shall appoint, in such manner as the Legislature thereof may direct, a number of electors. . .

Most states employ a "winner-take-all" strategy where the candidate who wins the state gets all the electors. But a few (Maine, maybe others?) assign electors based on proportional voting. Legislatures don't even have to hold a popular vote -- they can assign electors themselves. So on the surface, I can't see any Constitutional prohibition against the idea of assigning them based on the national vote.

In terms of substance, I don't think it solves the problem. When the 2000 election was close in the electoral college, it resulted in litigation. If a future election is close in the popular vote, it will result in litigation.

What would really help is some true bipartisanship around voting integrity and protecting ALL votes that are cast, regardless of whether they are from the inner-city poor or suburban dads or gun-owning country folks. There is no excuse for the US not to be leading the world in terms of a voting system that gets it right the first time, every time.

So on the surface, I can't see any Constitutional prohibition against the idea of assigning them based on the national vote.

Swell, but don't you think that somebody's going to come up with a colorable court challenge? And that it's going to be stuck somewhere in the judicial system during the next election?

The last thing we need is one more cloud over the elections process. I'm with Miles -- much more useful than some clever "end run" around the existing rules is serious improvement in basic reliability of registration and vote counting.

It might be helpful to have a better understanding of why our elections are so closely contested. Clues about this can probably be found in the demographics of who votes and who abstains, in the way campaigns are funded and run, and in the perception that both major parties are much the same. But close votes make fraud much easier, and I would venture to guess that no voting system devised and run by humans can get it right the first time, at least for some of those times when there is no statistically meaningful margin of victory.

"But Oregon's status could change under a pending bill in the Legislature that would award the state's seven electoral votes to the candidate who wins the popular vote nationally, regardless of who wins the state."

What. The. %$*&?

In an age where undetectable wholesale election fraud is possible, that's a really, really bad idea. It would place our electoral votes, our only representation in the choice of President, completely at the mercy of 49 other state electoral processes that we don't control.

The current system sucks to be sure, but at least there's some compartmentalization to it. Electoral fraud in just one state won't necessarily swing the whole contest.

(If we're going to change the system, it would be better to split our electoral votes proportionally to the popular vote in our own state.)

Alan has an excellen point. Oregon's vote-by-mail system may be one of the few left that has some shred of integrity to it. Throwing our little bucket of extracted voter juice into the national pool would dilute it beyond recognition.

I would venture to guess that no voting system devised and run by humans can get it right the first time

What I struggle with is this: How many online stock trades are made every day? How many transactions does Fred Meyer's process every day? How many checks does Bank of America process every day (Google says over 30 million)? What is the margin of error for these processes? It's very, very small.

Now, why can't some of this technology and these systems be applied to the 100 million or so votes cast every four years?

"It's very, very small."

Yes, but it's not zero. That's the problem with really close elections.

Allan,

OK, so the private firms regularly tabulate and collate data much more accurately than do counties, states, etc. But, if they don't do it perfectly, then we shouldn't try to adopt better means and methods?

Or have you already judged this matter?

Yes, but it's not zero.

Oh, well then; so private firms and institutions regularly tabulate and collate data much more accurately than do counties, states, etc. But, if they don't do it perfectly, then we shouldn't try to adopt better means and methods because we'll never reduce the error rate to zero?

Or have you already judged this matter?

do you hear an echo?

sorry.

Errr, rr, while putting words in someone else's mouth is a long and distinguished tradition on teh intarnets, I don't think that's what Allan was trying to say.

He appears to me to have been saying that when you're talking about elections with millions of votes cast, even very small nonzero error rates may not be close enough. (According to Wikipedia, Gregiore won Washington in 2004 by 0.0045%; if the actual error rate is 0.005% then the result has no meaning whatsoever.) However, I do not imagine him to be saying that trying for better error rates is futile.

Anyway, counting 30 million personally-identifiable things every day for profit is a much different task than counting 150 million purposely anonymous things in one day, once every four years, at taxpayer expense. BoA's process gets a lot more testing every week than any election system gets in a year.

This method of reforming the electoral college is reminiscent of the way Oregon pioneered the way toward direct election of senators. This was a part of the "Oregon system" which included initiative, referendum, and recall.

Thank's to third-party state legislator William U'Ren, initiative and referendum were added to Oregon's constitution in 1902, primaries in 1904, and recall in 1908.

Prior to 1913, U.S. senators were elected by state legislatures. In 1907, U'Ren maneuvered the Oregon legislature to force it to always elect to the U.S. senate whoever got the popular vote (which until then was advisory if a popular vote was made at all). Some twenty-nine states followed Oregon's lead, and by 1913 Congress passed the 17th amendment requiring all states to directly elect senators.

The electoral college is archaic and should be replaced. This may be the way to do it.

Alan,

Thanks for sharing what "appears to be" your insight.

I'm sure that the words you, in your infinite wisdom, put in Allan's mouth are much more succinct and accurate than mine. Never mind the point of my comment.

Oh, and it's Gregoire...

Of course, details like that are unimportant - as long as the result is correct.

Don't make me post this comment twice!

rr: Hey, I'm all for tradition. :-) So what was your point?

Responding (on my own behalf, thankyouverymuch) to what Miles said, I think Oregon has adopted those systems where possile. Hand-marked, optically-scanned ballots seem pretty reliable as far as counting goes. I'm pretty sure those are now in use statewide.

The real killer problem that makes vote-counting harder than banking, though, is the requirent for anonymity: secret ballots are hard to track.

If you're a bank scanning checks, any mistakes can (and probably will) be caught by the other bank, the check issuer or the check recipient when they each balance their accounts. When a mistake becomes apparent, the paper trail will clearly lead back to the problem's origin, and the problem can be corrected if the banks happen to believe in the quaint old idea of customer service.

With voting, though, that can't happen. As we discussed here last year, ballots are de-identified after they arrive at an Oregon county election office, and before they are counted. That makes the sort of error-checking the banks do - giving the issuer feedback about the transaction's results - just about impossible.

So anyway. I think Oregon has nailed two aspects of a good voting system: encouraging high turnout, and quick accurate counts. (There's more work to be done, but it's in areas like registration processes where technological solutions are not necessarily helpful.) The last thing I want to see is our pretty good system drowned out by whatever noisy crap the other 49 states might be doing with their voting systems - systems that I have no say in.

I'l grant that this is a well-meaning proposal, but it's misguided and definitely not a win for Oregonians.

I'm amused by this proposal - mostly by the audacity and simplicity of it.

But imagine for a moment that the overall national vote total in 2000 was roughly as close as the one in Florida.

Officially, Bush beat Gore in Florida by 537 votes out of 5,963,110 cast - or 47.8468% to 48.8378%.

Extrapolated to the national vote total, that's 51,482,914 votes to 51,473,423 votes -- a margin of 9,491 votes.

We'd have a national recount -- which would be a horrifying, heart-wrenching, process... one in which every state would apply their own laws, their own standards, their own procedures.

If we move to a national vote, we need a national vote-counting law. Take the townships, counties, and states out of it. One law, one standard, one technology, one process.

The point I was trying to make is that we have a long ways to go on error rates before we're at a "reasonable" level, and I think the technology exists in the private sector to help. I'm always amazed by the number of votes that "switch" in a recount. How is that happening? Why is that happening?

Kari is right -- if the popular vote is ever that close, we're doomed. But a few hundred million from the feds could significantly improve the vote-counting systems we have in most states.

This isn't a national vote, this merely requires Oregonians to disenfranchise themselves and support whoever wins the popular vote total in the remaining 49 states.

We might as well just abstain and let the rest of the country pick the President, because we're too pissed off about Bush winning in 2000 to ever vote again.