« New Map & List | Main | Pennsylvania: Additional Thoughts »

Saturday, October 25, 2003

Electoral College Reform

Posted by DavidNYC

It's a tiny irony: I think we should abolish the electoral college, but without it, this site wouldn't exist. Not to worry - I'm sure I'd find something else to write about. In any case, we've been discussing the unusual electoral arrangements in Maine and Nebraska recently, and they've prompted me to think a little bit about electoral college reform, rather than abolition.

Here's how it works in ME and NE: If you win the popular vote in a given Congressional district, you get one EV (which in theory represents each CD's congressman). If you win the overall statewide vote, you get two EVs (which in theory represent the state's senators). So in Maine, which has four total EVs, you could wind up with them getting split 4-0 or 3-1 (but never 2-2). In fact, in 2000, Bush nearly won Maine's 2nd CD. Had Bush won it, Gore still would have netted 3 EVs because Gore's statewide totals would almost certainly have exceeded Bush's. (Nebraska could go 5-0 or 4-1, but 3-2 would be pretty unlikely.)

To get rid of the electoral college, you'd need a Constitutional amendment, which would likely never pass because it would diminish the power of small states. But states could adopt the ME/NE system very easily - I imagine all it takes is some legislation, or perhaps a state constitutional amendment. (Not that I think this wholesale change is likely to happen, though.) On the face of it, the CD-splitting system seems to be more representational than the electoral college. At the very least, it's one step closer to counting the popular vote, I think. I'm not sure if it would truly be an improvement over what we have now, but it might be.

So I have some questions about this potential arrangement:

1) Had this system existed in every state in 2000, what would the EV count have been? (My instincts tell me it would have hurt the Dems, quite possibly very badly. But since this is just a theoretical exercise, this is beside the point, for now.) I can't seem to find any listing which shows every American CD in one place, but if anyone wants to look into this, I'd be very interested to see the results.

2) How would candidates campaign if this system existed nationwide? Obviously, candidates would ignore CDs that lean heavily one way or the other. But they would start visiting states that they otherwise skip entirely now because they'd want to hit populous "swing CDs" in otherwise solid states like California or Texas. This still means they'd skip most major cities (apart from fundraising jaunts) because they lean heavily Dem, and they'd obviously avoid rural areas just because those are both hard to campaign in (physically) and are usually reliably Republican.

The conclusion then, I think, is that candidates would campaign heavily in suburban swing districts. I'm not sure how much this differs from present practices - witness the devotion to "soccer moms" and "office park dads", both of whom are decidedly suburban figures. But at least candidates would hit up the `burbs in NY and GA, rather than just PA and OH.

On balance, I'd say this would be a qualified improvement. As an urba-phile (that's not really a word, is it?), I'm naturally reluctant to hand over any more clout to suburbia, but it's probably already too late to worry about that. And I do think this would get us closer to a true popular vote count. So maybe one of these days I'll be producing the "Swing CD Project". Nah - that sounds a little too Big Bad Voodoo Daddy to me.

P.S. A thought just occurred to me, which I think should have been obvious at the start: Could a state pass a law appointing electors simply in proportion to the total popular vote won? I don't see why note. Article II § 1 of the US Constitution says: "Each state shall appoint, in such manner as the Legislature thereof may direct, a number of electors, equal to the whole number of Senators and Representatives to which the State may be entitled in the Congress." Seems like the states have free reign here.

Posted at 12:42 AM in General | Technorati


Your insticts are correct. Doing it by CD's would kill Democrats.

Dems win states by getting huge voter turnout in the cities, which downplays the Republican rural voters. Overall, there are more rural CDs across the nation than city CDs. It would be a disaster for Democrats if the nation followed the Maine/Nebraska method.

You'll never see Electoral College reform. Democrats want to abolish it in favor of a popular vote (which favors them, because they can just campaign heavily in the cities and spend tons on GOTV efforts), while Republicans want to keep it because it forces Democrats to campaign in the midwest, the mid-atlantic, and the ocassionally the south (when we don't abandon it).

Actually, the Maine/Nebraska method would probably what Republicans consider Electoral College reform.

The bottom line debate is between whether we should use the final vote total from across the country (in effect, making every vote equal, though voters in the middle of the country would be ignored because there are less people living in that area), or should we weight the values of certain states to make sure all voters (or rather, states) have a "say"?

It's an intricate issue. Democrats lost the presidency the two times that the winner of the popular vote lost in the Electoral college. It's kind of shitty.

But there you have it.

Posted by: AEDem04 at October 25, 2003 01:12 AM | Permalink | Edit Comment | Delete Comment

Problem: the House district setup is severely gerrymandered in favour of Republicans, which only got worse after the 2002 redistricting. There's an excellent article on this, 'The United States House of Unrepresentatives: What Went Wrong in the Latest Round of Congressional Redistricting', in Election Law Journal, vol. 2 no. 2, 2003, pp. 179-216. I base my comments on data in that article and Dave Leip's election atlas.

The article says Gore "would have carried fewer than 200 of the 435 newly redrawn districts." Let's say a round 200 of the districts as they stood in 2000. With those districts (200), 20 states (40), and DC (3), he would've run at least 20 EVs behind his actual total of 267.

Doing it by CDs is a bad idea until the redistricting process is fixed, because it adds yet another incentive to gerrymander.

As for perfectly proportional allocation, I don't have my spreadsheets on the 2000 election to hand so I can't do the numbers, but I'm inclined to suspect it hurts Gore, as his states tended to be closer than the Bush states, most of which W. won by 55% or better. The rule in these sort of things is: the bigger the number of warm bodies, the more proportional the allocation. So Bush would be piling up 2-1 and 3-1 splits in the 3- and 4-EV states, while Gore would be getting 10-8 splits in Michigan and 4-3 splits in Oregon.

Thing is, the electoral college actually distorted the popular vote percentages very little in 2000. Gore got 48.38% of the popular vote and 49.62% of the EV; Bush, 47.87% and 50.37% respectively. This is the least distortion of any election other than 1876.

That's the other thing -- proportional allocation would likely add a small but potentially significant number of third-party electors into the mix. Nader would be due somewhere on the high side of ten, going by his national vote. And proportional elector allocation in 1992 (and most likely 1996) would've thrown the election into the House of Representatives.

I'll do the detailed numbers on state-by-state proportional allocation and get back to you.

Posted by: schwa at October 25, 2003 06:05 AM | Permalink | Edit Comment | Delete Comment

I should also point out that Florida and Pennsylvania -- two big, important swing states -- have two of the worst gerrymanders in the country, right up there with that national embarassment in Georgia.

Posted by: schwa at October 25, 2003 06:08 AM | Permalink | Edit Comment | Delete Comment

I'm not sure about other states, but in CA it would be especially disaterous to allocate by CD. The democratic power of the state is entirely concentrated in the heavily populated coastal areas. According to the numbers, if we used the idea that each CD gets 1 elector and then the overall vote gets 2, then the results for the 2000 election would be

Gore - 35 Electoral votes
Bush - 19 Electoral votes

This would be awful for the dems.

Source - http://www.ss.ca.gov/elections/sov/2000_general/ssov/cong_dis.pdf

Posted by: WestCoastPat at October 25, 2003 02:47 PM | Permalink | Edit Comment | Delete Comment

Schwa, thanks for your great comments. A few thoughts in response:

1) My hope with electoral reform would not primarily be to make popular vote totals match more closely with EV totals, though this would be a good outcome. Rather, I'd like to see a system where candidates aren't incentivized to skip out on campaigning in large swaths of the country. I'm very generously describing 40% of the states as swing states, but I don't think the candidates will even seriously hit that many states.

2) I ran the numbers on strict proportional EV distribution. If you allow fractional electors (obviously you can't in real life), then you get this:

Bush: 259
Gore: 258 (ouch!)
Nader: 15
Buchanan: 2
Browne (libertarian): 2

The math doesn't work out properly if you round EVs:

Bush: 259
Gore: 258
Nader: 7
Others: 0

It seems to me that the only reasonable way to make this system would work be with instant-runoff voting. That way, a candidate who doesn't meet the threshold for getting at least one EV (which would start at a high of 33% in 3-EV states, and decrease all the way down to something quite low in California) would have his votes transferred to other candidates.

One interesting side-note is that this approach would not have hurt Gore (though that one-EV loss, if there were no IRV, would have been brutal). It's because for every MI which would yield a tight split that would have hurt Gore, there's a TN on the other side which would help him. In fact, in 2000, there were some 78 EVs in the Gore swing states but far more - 130 - in the Bush swing states.

Posted by: DavidNYC at October 25, 2003 02:54 PM | Permalink | Edit Comment | Delete Comment


I did a similar calculation as you did, but employed a slightly different methodology (and I think a similar methodology by which House seats are apportioned based on the census). I took the total number of votes cast in the state and divided it by the number of electoral votes. For example, in Michigan, which has 18 electoral votes and in which 4232501 votes were cast, each electoral vote can be considered to "cost" 235139 popular votes. Bush, with 1953139 votes, has enough votes to "buy" 8 electoral votes, and still has 72027 popular votes in "change". Gore, at 2170418 popular votges, "buys" 9 electoral votes with 54168 popular votes left. None of the other candidates has enough for one electoral vote, but Nader had a total of 84165 popular votes. Now, only 17 electoral votes have been apportioned, but Michigan gets to apportion 18. The extra EV goes to the candidate with the highest number of popular votes that have not been used to "buy" an electoral vote. In Michigan, this vote would go to Nader, whose 84165 total votes is more than either Bush or Gore have left after they have already "spent" the bulk of their votes on the first 17 electoral votes. Thus, Michigan gets apportioned 8 for Bush, 9 for Gore, 1 for Nader.

In some cases, more than one extra vote would be left over after all the candidates have "bought" all the electoral votes that they could. In California, after Bush, Gore and Nader have "bought" 22, 28 and 2 electoral votes respectively, only 52 of the 54 electoral votes would have been apportioned. At this point, we rank the candidates based on how many votes each candidate now has in "change", and the order is Gore, Bush, Browne, Buchanan, Phillips, Nader, Hagelin. Thus, one electoral vote each goes to Gore and Bush. The total apportionment for California is 23 for Bush, 29 for Gore, 2 for Nader.

By applying this methodology to the entire country, I have determined the following results:

AL 5 4 0
AK 2 1 0
AZ 4 4 0
AR 3 3 0
CA 23 29 2
CO 4 3 1
CT 3 5 0
DE 1 2 0
DC 0 3 0
FL 12 12 1
GA 7 6 0
HI 2 2 0
ID 3 1 0
IL 9 12 1
IN 7 5 0
IA 3 4 0
KS 4 2 0
KY 5 3 0
LA 5 4 0
ME 2 2 0
MD 4 6 0
MA 4 7 1
MI 8 9 1
MN 5 5 0
MS 4 3 0
MO 6 5 0
MT 2 1 0
NE 3 2 0
NV 2 2 0
NH 2 2 0
NJ 6 8 1
NM 2 3 0
NY 12 20 1
NC 8 6 0
ND 2 1 0
OH 10 10 1
OK 5 3 0
OR 3 3 1
PA 11 12 0
RI 1 3 0
SC 5 3 0
SD 2 1 0
TN 6 5 0
TX 19 12 1
UT 4 1 0
VT 1 2 0
VA 7 6 0
WA 5 6 0
WV 3 2 0
WI 5 5 1
WY 2 1 0
TOTAL 263 262 13

None of the candidates makes it to 270 EV for a majority. So one of two things happens: the election gets thrown to the House (where, presumably, Bush wins) or Nader's electors become faithless electors and throw their support to Gore (what are the odds of that?).

Anyway, hope you enjoy.


Posted by: DaveOinSF at October 25, 2003 04:28 PM | Permalink | Edit Comment | Delete Comment

One thing, as I reflect on my results, is that it becomes tough to accurately represent the opinion held by voters in small states. Small states with an EVEN number of electoral votes, get screwed. Look at Hawaii and Maine, states that went pretty solidly for Gore. Well, they both end up 2 EV for Bush, 2 for Gore. They look like tossups, like New Hampshire or Nevada. For a candidate to come out with an advantage in these states, they's need to get supermajorities in those states. A 60/40 popular vote split is still more fairly represented by a 2:2 distribution in electoral votes than it would be by a 3:1 distribution.

Then look at New Mexico. Tight, narrow victory for Gore by a few hundred votes. But because it is a small state with an ODD number of electoral votes, the EV distribution has to be 3:2 in Gore's favor, looking like a safe Democratic bastion, which it was not.

These problems disappear with larger states. But changing this system might lead to some interesting dynamics: Republicans spending a lot of time campaigning in Idaho to try to get the popular supermajority to get a 3:1 electoral vote advantage, spending less time in large swing states like Ohio or Florida because victory is not any better than a narrow loss.

Posted by: DaveOinSF at October 25, 2003 04:50 PM | Permalink | Edit Comment | Delete Comment

Your numbers don't add up on the fractional-elector scenario, either: the college has 538 members and 259+258+15+2+2 is 536.

That aside, it doesn't actually represent a one-vote loss. Constitutionally, you must have a majority of the electors -- 270, in a college of 538. So what you've got is either (a) a situation where Nader tells his electors to vote for Gore, and Gore wins with 273 EVs, or the election gets dumped into the House of Representatives.

There's actually a better way to do this.

A variety of formulae have been developed to handle distributing seats proportionately in situations like this -- mostly in Europe, where they tend to use proportional representation in electing their legislatures. I'll illustrate with the most proportional, what's known as the Hare quota. (If you're interested in the mathematics behind this, the classic textbooks on the subject are Electoral Systems and Party Systems by Arent Lijphart and Seats & Votes by Rein Taagepera and Matthew Shugart.) I'll illustrate with Wisconsin's results.

You divide the number of votes cast in the election by the number of electors to be allocated. This result is the quota.

2,598,607 votes / 11 EVs = quota of 236,237.

Now divide each candidate's votes by the quota.

Gore: 1,242,987 votes = 5.26 quotas.
Bush: 1,237,279 votes = 5.24 quotas.
Nader: 94,070 votes = 0.40 quotas.

For each quota, its holder gets one seat - or one elector. Then, if there are any electors left over to allocate, you look at the largest partial quotas.

Gore: 5 full quotas = 5 electors.
Bush: 5 full quotas = 5 electors.

1 elector left to allocate.
Gore: 0.26 remainder.
Bush: 0.24 remainder.
Nader: 0.40 remainder = 1 elector.

Final breakdown, then, is 5 Gore, 5 Bush, 1 Nader. If you do this over the whole country, you get:

Gore: 262
Bush: 263
Nader: 13

...and, having hit Preview, I can stop my example here, because what DaveOinSF has done is essnentially this in another form. (You can bend this allocation to favour larger parties by shrinking the quota, making it EV+1 or EV+2, but I can't be bothered running another set of numbers.) I dissent from his conclusions, though; I don't think any of the problems he highlights are problems. I suspect that a proportional allocation would actually produce the opposite result: you would see a lot more money and time in big states. Suddenly it makes sense for the Democrat to campaign in Texas, or the Republican in New York. The Republicans get massive majorities in the small Mountain West states without even trying (how much time did Bush spend in Utah in 2000?) and the Democrat would have to move from about a 27-65 split to a 45-55 split to get an extra elector; it's simply not economical.

However you do it, though, one problem remains: any system which allows third-party electors a reasonably chance is likely to result in an election decided in the House. The last President to win heavily enough to make that unlikely was Bush Sr in 1998.

Posted by: schwa at October 25, 2003 05:17 PM | Permalink | Edit Comment | Delete Comment

Interesting analysis. I don't think the "vote buying" system would work simply because a) it's too complicated (and by "too complicated" I mean, "too complicated for the talking heads on CNN ever to figure out) and b) it could take ages to actually get results, because you'd need to count every last vote before you could determine the "cost" of each EV.

Would instant-runoff voting solve this problem? I'm not sure. You'd still wind up having to round electors for the top vote-getters.

I'm starting to think that all this tinkering might not leave us with a better system than we currently have. And as unlikely as an amendment to the US Constitution would be, the very fact that only two states presently use alternate systems suggests that we are unlikely to see much change on the state level here.

Posted by: DavidNYC at October 25, 2003 05:26 PM | Permalink | Edit Comment | Delete Comment

you'd need to count every last vote before you could determine the "cost" of each EV.

Granted the US has a blanket exception from all "most other democracies" generalisations because it enjoys being uncooperative, but most other democracies would consider this a matter of course. Besides, you could get a rough and ready approximation from exit polling. It'd be much more a case of "where do the last couple of electors go?".

The underlying problem is that most of the small states are solidly one-party anyway. Delaware, Maine, New Hampshire, New Mexico and Nevada are about the only states with 5 or less EVs that are ever seriously contested unless one party's nominee is a favourite son. With that said, nothing systemic can be done to make it worthwhile for a Republican to campaign in Vermont or a Democrat in Wyoming. It's a matter of demographics.

Electoral College abolition would actually promote the interests of the minority in a locked-up state, because then "Bush is going to walk away with Montana, there's no point in my contesting it" wouldn't be a legitimate argument. But there's a partisan advantage for the Republicans in retaining the college in winner-takes-all mode, so you'll never get reform, any more than you'll see DC with two Senators in the near future.

Posted by: schwa at October 25, 2003 05:40 PM | Permalink | Edit Comment | Delete Comment

Schwa, I hadn't seen your post by the time I was writing my last comment, so let me respond here.

Your numbers don't add up on the fractional-elector scenario, either: the college has 538 members and 259+258+15+2+2 is 536.

Darnit - I forgot to count the tiny fractional EVs for the ultra-obscure candidates. And the "one-vote loss" was just a total brain-freeze on my part. Maybe I've been looking at these numbers too long.

As for the effects of a switch like this... well, if Schwa is right, then this would be an awful idea (candidates spending MORE time in small states?). If you are right, however, then this change would achieve the goals of reform, which is to encourage candidates to campaign in more of the country, and to stop ignoring more populous areas.

And what happens in the Hare system if, instead of using partial quotas for the final EV or two, you use IRV? Does this get rid of the problem of having a chunk of third-party electors?

Posted by: DavidNYC at October 25, 2003 05:45 PM | Permalink | Edit Comment | Delete Comment

As long as all we're doing is tinkering with the means by which the Electoral votes are apportioned, not Constitutional amendment would be needed. States could just do it through legislation but, ME and NE notwithstanding, I can't see how it is in any state's interest to do it differently unless the entire country adopts the same system.

So I suppose an amendment would be needed to FORCE all the states to adopt the same syste, but all that's really needed it a nationwide campaign to get all the states the change their own laws. (However, I can totally see the situation where all the states move to proportional EV allocation, and then Tom Delay gets Texas to switch back to the winner-take-all system).

Posted by: DaveOinSF at October 25, 2003 05:54 PM | Permalink | Edit Comment | Delete Comment

And what happens in the Hare system if, instead of using partial quotas for the final EV or two, you use IRV?

I'm not sure how you could fit IRV into a system like that... well, I can think of ways, but if Hare's complicated without remainder transfer, it would only get worse with it. What you could very easily do, however, is set a threshold to be allowed to get any electors at all: either a requirement to attain a full quota, or some arbitrary percentage of the vote (say 5%). Either would wipe out all but about four Nader electors, and curb any third-party challenge but the most serious, like Perot.

I suspect that all that's needed to actually initiate a reform like this is for one big state - California, Texas, New York, Florida or Pennsylvania - to bring it in. The other party would immediately shower money on that state, and the other biggies would likely bring in proportional allocations of their own in order to recapture some of the attention. Over a couple of elections, you'd get a cascade effect. My preferred option is still abolition, though.

As a side note, now I've had time to do my numbers, I can point out why I would consider any 'reform' which increased the likelihood of throwing the election into the House to be a disaster -- voting by state delegations, as the House does when picking a President, there are 31 Republican delegations, 3 evenly split, and only 16 Democrats, counting Bernie Sanders; 32-3-15 if Texas' redistricting has its anticipated effect. Not only would that make it substantially harder for a Democrat to become President (a Democrat has to win 270 EVs; all the Republican has to do is hold the Dem to 269 or less), after the first couple of times a popular-vote-winning Democrat was stripped of the presidency, you'd be looking at an ugly anti-system backlash.

Posted by: schwa at October 25, 2003 06:09 PM | Permalink | Edit Comment | Delete Comment

Well, I agree that just about any system makes it silly for Democrats to campaign in Wyoming or Republicans in Vermont...It's hard to imagine any result other than a 2:1 EV split in favor of the majority party.

But then there's DC. All a Repub would need to do is win about 17% of the vote to get DC to go 2:1 rather than 3:0.

And then the next tier of states have 4 EV. A 3:1 advantage is a LOT better than a 2:2 tie. To make the formula simple, let's assume just 2 parties. If the popular vote split is 66:34, then the EVs go into the books as a 2:2 tie. But if it's 67:33, then it's 3:1. And there are a couple of states - Idaho, Rhode Island, Hawaii, Maine - where there's a definite majority party, but it would be in the minority party's interest to campaign hard to get to 34% of the vote to produce a net swing of 2 EVs.

Posted by: DaveOinSF at October 25, 2003 06:11 PM | Permalink | Edit Comment | Delete Comment

Wait, brainfart. I see what you're aiming at with IRV. Um. It could be done, I think; again, it would get rid of all but about four or five Nader electors, to Gore's benefit. I think it would add a layer of gratuitous complexity, though.

Posted by: schwa at October 25, 2003 06:16 PM | Permalink | Edit Comment | Delete Comment

The thing is, if you look at your numbers, most of the states with 4 to 8 EVs are almost evenly split anyway. If we look just at the 4, 6 and 8 EV states, leaving out the 5 and 7s which are highly likely to split 3-2 and 4-3 (I'll leave in Oregon, which split 3-3-1)

AZ 4 4 0
AR 3 3 0
CO 4 3 1
CT 3 5 0
HI 2 2 0
ID 3 1 0
KS 4 2 0
KY 5 3 0
ME 2 2 0
NV 2 2 0
NH 2 2 0
OK 5 3 0
OR 3 3 1
RI 1 3 0
SC 5 3 0

Take out the ones which are already split evenly:

CO 4 3 1
CT 3 5 0
ID 3 1 0
KS 4 2 0
KY 5 3 0
OK 5 3 0
OR 3 3 1
RI 1 3 0
SC 5 3 0

That leaves CT, OR and RI for the Republicans and CO, ID, KS, KY, OK, and SC for the Democrats where an extra EV might be gainable at the margin. The two obvious points here are (1) this would involve much more work for the Democrat than it would for the Republican, and (2) these are some pretty hard-core right-wing states. Competing in them would skew the (D) nominee even further right than he already has to run. What's the value in gaining 1 EV in South Carolina if fifty thousand left-wing New Yorkers stay home?

Then, looking at the split states:

AZ 4 4 0
AR 3 3 0
HI 2 2 0
ME 2 2 0
NV 2 2 0
NH 2 2 0

All of these except for Hawaii are swing states where the Democrat already spent a hell of a lot of time. I have enormous difficulty seeing anyone moving ME, NH, or NV from 2-2 to 3-1 either way.

Basically what we've got here is an unequally heavy load on the Democratic nominee. Meanwhile, the Republican, with safer small states and more money, is spending it in New York and California and picking up EVs by threes and fours. If you're the Democrat, what do you do?-- chase a single electoral vote in Kansas or haul your ass back to San Francisco to shore up your biggest state?

Posted by: schwa at October 25, 2003 06:35 PM | Permalink | Edit Comment | Delete Comment

To get the basics of what you were saying about the district plan the way Nebraska and Maine do things here is how the past few elections would have turned out. After each is the +- in E.V. from the acctual total.

Bush-271 +-0
Gore-267 +-0
No change in EV

Clinton-345 -34
Dole-193 +34

Bush-214 +46
Clinton-324 -46

All of the other elections since 1956 have no major effect except for the following

Kennedy-245 -58
Nixon-278 +59
Byrd-14 -1
Nixon defeats Kennedy

Ford-269 +29
Carter-296 -28
Others-0 -1
Ford and Carter Tie ends up in House of Reps.

So it appears that Republicans would benefit from the District plan.
I have the results from 1956 to 2000 and the two above are the only ones that switched the final results. If you want to know the others not listed, just ask me and Ill give them to you.

Posted by: Josh at October 25, 2003 07:09 PM | Permalink | Edit Comment | Delete Comment

In the case we switch to proportional EV distribution, the race becomes all about winning the extra EVs here and there, wherever you can find them. Thus, one of the things that will result is that odd EV states become a lot more important than even EV states. Who cares whether you win MN by 50000 votes or lose it by 25000? EV splits 5:5 either way. But a 500 vote win in NM is the difference between being up 3:2 or down 2:3.

It would actually be to the detriment of a state like NM to gain an extra EV with increased population. The odds of winning with a large enough margin to switch the ratio from 3:3 to 4:2 is small enough that the major candidates would spend less time in the state.

And I still think small states become proportionately more important. Each EV in HI only takes 91000 votes, while in NJ it's more like 215000. So, even if you have no chance of actually winning the state, shifting an EV here and there by just changing the margin of victory and/or defeat becomes easier to do because the numbers of voters you need to switch are fewer.

Gore won NJ 56/40, but still got only a 8:6 advantage over Bush, and without Nader, probably only would have had a 9:6 advantage. In HI, the 56/37 vote split resulted in a 2:2 tie. Adding some resources to HI to change about 45000 votes would be a lot cheaper than to do the same to NJ, where you'd need to change about 108000 votes.

Of course a lot of this changes if a 3rd party candidate is even a little bit viable.

Posted by: DaveOinSF at October 25, 2003 08:10 PM | Permalink | Edit Comment | Delete Comment

Josh, what was your methodology/what sources did you use to compile that data?

Posted by: DavidNYC at October 26, 2003 01:04 PM | Permalink | Edit Comment | Delete Comment

I obtained the information from 'The Road to the White House 2000, The Politics of Presidential Elections. The Post Election Edition', 1st ed., by Stephen Wayne.

Additionally this information can be double checked in the 'Alamanac of American Politics' published every two years, and it breaks down Presidential votes by congressional district as well. It may not be available in most libraries, but its worth a shot to look for it.

Posted by: Josh at October 26, 2003 03:30 PM | Permalink | Edit Comment | Delete Comment

Wow, this is all so complicated, it makes me wish for a direct election system. I agree that any system where the election is thrown to the House is bad, and not just because we lose. It's even less democratic than the current electoral college system!

What do you think of the direct election alternative? It is simple to understand and seems fair. It also makes competing in non-swing states worthwhile, to prop up your total vote count.

Posted by: Luke FranclLuke Francl at October 28, 2003 11:07 AM | Permalink | Edit Comment | Delete Comment

In strictly partisan terms, this kind of "electoral college reform" is a terrible idea for Dems. Here's the bottom line. The electoral college is "undemocratic" in two different ways: (1) because of the "winner takes all" system that prevails in most states; and (2) because of the overrepresentation of very small states. Roughly speaking, the first undemocratic element favors the Dems, while the second favors the GOP. Adopting the Maine/Nebraska formula would "remedy" the anti-democratic feature that favors the Dems while leaving the other, arguably more powerful anti-democratic feature intact. The scary part is that the first can actually be accomplished without a constitutional amendment while the second cannot. That's why I have been worried since about November 2000 that the GOP will actually pick up this ball and run with it. As long as CA, NY and MI lean Dem, any partisan Democrat (and I mean that as a term of endearment) should fight any such "reform" tooth and nail.

Posted by: JohnL at October 28, 2003 01:05 PM | Permalink | Edit Comment | Delete Comment

Perhaps the answer to the Electoral College favoring the Republicans is for the Democrats to make some type of long term effort to win back the rural, more conservative states. After all, Truman and Johnson each carried every western state except one. Even Clinton picked off a handful of western states (with Perot's help).

Posted by: Dave at October 30, 2003 09:34 PM | Permalink | Edit Comment | Delete Comment

There's another method I haven't seen proposed:

States could have an even-as-possible split
for percentage splits close to 50/50, a
winner-take-all for popular votes above a
certain percentage (60% ?), with provisions
to the effect that any candidate above 40%
gets at least one EV.

Small states still have their EV's, so they
won't have an overwhelming desire to veto
it. Large states will usually have a
significant number of EV's in play, so they
will get attention from candidates. The
40% threshold freezes out minor candidates
(good or bad depending on your viewpoint),
so the two major parties would not be overly

When I ran the numbers using a scheme like
this in Dec 2000 (I don't remember the exact
formula anymore), I got a tie! (Of course,
in reality, the campaigns would have been
different, so the results would have been

This could be done state-by-state or via
Consitutional amendment.

Posted by: James McDonald at June 1, 2004 09:20 AM | Permalink | Edit Comment | Delete Comment

I don't understand why any legislature, or any ballot measure which would win enough votes to win, would indulge in such complex systems as have been suggested here. To my mind they seem to be non-starters.

What we're essentially talking about with Electoral College reform is taking a step towards a more democratic and more representative system. It wouldn't be as democratic and representative as junking the EC entirely and going to direct national popular election, but, as David said, that ain't going to happen. State-by-state proportionate electoral vote assignment also doesn't require a consitutitional amendment, and can be instituted over time, one state at a time.

So, there's no real requirement that the method used to divide the vote be absolutely proportionate to the popular vote in the state, as long as it's *more* proportionate than the winner-take-all system.

So why not something like: take each candidate's percentage of the popular vote, to one decimal point of accuracy, and multiply it by the number of electoral votes in the state. Drop *all* fractional votes, and add up the votes thus assigned. Subtract this number from the total number, and assign those few votes left over to the candidate who received the most in the popular vote, as a bonus for coming in first.

Wouldn't a system such as that work OK? True, it's still not the whole hog, but it's better than what we got.

And, really, the issue here is what's best for us as a *democracy*, not what's best for the Democratic party. If the party cannot compete under a fairer and more proportionate system, then the party had better change itself so that it can. Similarly, the "third party problem" is only a problem because we've never done things like that before. It's much more likely that third parties would do a deal with the major party closest to them in philosophy than it is that they'd stick to their principles and allow the determination to go into the House, where they'd get nothing at all for their electoral votes. Typically, the third party would want some guarantee of access, and some of their people appointed to high positions. We'd end up with a watered down version of coalition government, but without the structural ability for small parties to bring down the government, as they can in parlimentary systems, since withdrawing their support from the coalition would not bring down the government (but would have severe ramifications for the next election's coalition-building).

Posted by: Ed Fitzgerald (unfutz) at June 15, 2004 10:39 PM | Permalink | Edit Comment | Delete Comment

FYI -- in a vote splitting system such as in ME and NE, Bush would have won 7 of Kentucky's EV's with Gore winning 1.

Posted by: JN at July 29, 2004 08:24 PM | Permalink | Edit Comment | Delete Comment

There is a way to elect Presidents by the nationwide popular vote, if the 11 largest states will sign on.

The U.S. Constitution says:
Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors, equal to the whole Number of Senators and Representatives to which the State may be entitled in the Congress....

In other words, the Constitution specifies how many electors a state gets, and the legislature of that state can then make whatever state laws it wants as to how it will appoint its electors to the Electoral College.

The California state legislature could pass a law saying that all of its 54 electoral votes will go to the candidate who wins the nationwide popular vote. Texas, New York, Florida, etc. could pass similar state laws. If the 11 largest states pass these laws, then 271 electoral votes will automatically go to the candidate who wins the nationwide popular vote.

Since 271 is a majority of the total 538 electoral votes, that would be enough to give the Presidency to the candidate winning the nationwide popular vote.

For instance, suppose the legislatures of the 11 largest states had passed these laws before the 2000 election. Since Al Gore won the popular vote by 50,999,897 to 50,456,002, he would have automatically won 271 electoral votes (plus whatever the smaller state gave him), enough electoral votes to give him the Presidency.

Of course, while the state legislature are passing these laws, other states might be tempted to keep their current system in order to gain an advantage. This could be overcome by including a trigger in the new state laws which would cause them to take effect only when enough legislatures have passed it to secure 270 electoral votes.

The current system favors the smaller states, because each state gets a minimum of three electoral votes, even if their population is extremely small (e.g. Wyoming). That is why the larger states have an incentive to cooperate in overcoming it.

Posted by: Nathan Larson at October 8, 2004 02:23 PM | Permalink | Edit Comment | Delete Comment

Nathan, interesting idea! Very interesting indeed.

Posted by: DavidNYC at October 8, 2004 06:28 PM | Permalink | Edit Comment | Delete Comment

Interesting idea, but it'll never happen. Can you imagine the outrage if voters in state X overwhelming vote for one candidate, and yet all of their state's EVs go to the opposing candidate? I don't think that would play well in any of the big states. Of course CO might make things interesting this year, so there could be some type of change in the air. We shall see.

Posted by: pepe at October 8, 2004 07:57 PM | Permalink | Edit Comment | Delete Comment

Personally, I think we should abolish the entire Presidential system and have a Parliamentary system. Large Presidential democracies may get unwieldy. On the other hand, we have an existence proof of a Parliamentary Democracy that is 4 times larger than the US(India).

Posted by: erg at October 8, 2004 08:44 PM | Permalink | Edit Comment | Delete Comment

I like the idea of the Palementary system as well. It gives minor parties a better chance at being represented. But, it's not ever going to happen in the U.S. The one thing the U.S. could do is expand the House of Representatives. There's nothing in the Constitution that says it has to be 435 members. We could expand it to 1,000 members and provide more representation and more diverse points of view.

Posted by: Rock_nj at October 8, 2004 10:33 PM | Permalink | Edit Comment | Delete Comment

We need to keep the system that we have. The US has the longest lasting constitution in the world. Obviously our constitution is the right way. Changing it would be wrong. So many countries have tried to adopt constitutions like ours because they realize how good it is. Any other way would be bad.

Posted by: Zack Zulock at October 8, 2004 11:51 PM | Permalink | Edit Comment | Delete Comment

Very interesting discussion here... I wish I had seen when it started. Some random thoughts first�Ķ
Let's set aside the problem that gerrymandering in favor of one party over the other. (This is a major problem still.) Let���s just hope that congressional gerrymandering would become a much larger issue and there would be greater voter backlash for having done this.
The discussion on proportional representation has been interesting, including the comments on third party candidates. (Just remember, those third party candidates may not just be extremist Nader-like candidates. It also may encourage guys like McCain to run.) The possibility of no one getting 271 and having it decided in the House is pretty scary, too. Maybe some sort of Election Day ranking of candidates to redistribute electoral votes of minority candidates would fix this.
There is one idea that has been left out so far (I think)�Ķ
When the founders devised this system, their goal was to encourage candidates to gather broad support nationwide. If a candidate already has the support of a large portion of one group (however you want to build these groups), that candidate shouldn���t be encouraged to get more. Instead, they would be encouraged to gather support elsewhere. The founders placed more importance on making a system that encourages the right type of campaigns from the candidates even at the expense of making elections ���fair.���
At the time, and maybe other readers with a better feel for history will comment, building these groups by state fit this model relatively well. Now, this is the part of the Electoral College that is out of date. We have major variations in regional politics in virtually any medium to large sized state.
Today, there is a major benefit to either party to cater to their base; it just isn't the base nationwide. The base they cater to is their base in the individual swing states. I remember one day when I checked the CNN website and Bush was campaigning in Michigan's Upper Peninsula and Kerry was in Flint. Anyone familiar with Michigan politics would tell you that these voters have already made up their minds. I���m sure there are similar stories for PA, FL, CO, OH, etc�Ķ I firmly believe that if this system were in place during this election, we would not have watched Bush push the gay marriage ban or the debate on government funds on stem cell research.
Again, we shouldn���t be concerned about being ���fair,��� at least when talking about major changes to the way that we elect the next leader of the free world. We should be talking about how those electoral changes would effect the way campaigns are run, the policies that are encouraged, and the way the victors govern.

Posted by: Dan Hogan at October 14, 2004 04:08 PM | Permalink | Edit Comment | Delete Comment

Let's return to the original plan by the leading constitutional theorist of the Convention, James Wilson. This future member of the Supreme Court advocated direct election of the Senators by the people and direct election of the President. The 17th Amendment has taken care of half his program. Let's finish the other half.

The Electoral College was never intended to be a permanent feature of the Constitution. It was intended to serve a temporary function only. In a republic, all power comes from the people. No one spoke against the people having the power to choose the President. They only set forth certain conditions: communication in a large country, travel, information, education, etc. All these condititons were overcome long ago. In fact, members of the House were arguing that the conditions were extinguished in 1826.

After two centuries, it is high time to take the training wheels off this democracy and implement the plan of the leading speakers of the Convention for direct election of the President.

Gary Michael Coutin, Esquire

Posted by: Gary Michael Coutin at January 19, 2005 12:15 AM | Permalink | Edit Comment | Delete Comment