Blog

tl;dr– Almost every argument you’ve heard for/against the electoral college is a myth. The main problem is not giving small states more weight; the problem is swing states, which arise from states awarding their electoral votes as an all-or-nothing bloc. The motivations behind the EC are not as noble as you think, and no, New York and California would not sweep the election without it. It could easily be fixed if states allocated their electoral votes proportionally to the vote within each state. But this actually tracks National Popular Vote pretty closely, so we should just use that instead.

The electoral college is defended as being purposefully designed to protect the voice of rural areas. Without it, the argument goes, candidates would campaign only in the “few, major metropolitan areas” for an easy majority, completely disregarding the rest of the country.

This defense is complete bullshit.

While the few largest cities do indeed make up a formidable portion of the total U.S. population, it’s nowhere near a majority. And the next largest cities taper off in size quite quickly.

But yes, if you won every single vote with perfect turnout in greater New York, Los Angeles, Chicago, Dallas, Philadelphia, Houston, Washington, Miami, Ft. Lauderdale, Atlanta, Boston, San Francisco, Detroit, Phoenix, Seattle, Minneapolis, San Diego, St. Louis, Tampa, Baltimore, Denver, Pittsburgh, Portland, Charlotte, Sacramento, San Antonio, Orlando, Cincinnati, Cleveland, Kansas City, Las Vegas, Columbus, Indianapolis, San Jose, Austin, Norfolk, Nashville, Providence, Milwaukee, and Jacksonville, you would indeed sweep the election.

However, even in the most die-hard city a single candidate would never win 100% of the vote. With a still-landslide 80%, we’d have to throw in* Memphis, Oklahoma City, Louisville, Hartford, Richmond, New Orleans, Buffalo, Raleigh, Birmingham, Salt Lake City, Rochester, Grand Rapids, Tucson, Honolulu, Tulsa, Fresno, Worcester, Bridgeport, Albuquerque, Albany, Omaha, New Haven, Bakersfield, Knoxville, Greenville, Allentown, El Paso, Baton Rouge, Dayton, McAllen (who?), Columbia, Greensboro, Akron, Sarasota, Little Rock, Stockton, Charleston, Syracuse, Toledo, Colorado Springs, and Winston-Salem. A walk in the park, really.

* at the same time, it’s unrealistic to claim you’d get zero votes anywhere else, so you wouldn’t need all or even most of these extra cities to reach a majority. But the narrative does attempt to portray cities as these islands of liberalism with no support out there in “real America”, so I still think it’s a valid comparison.

It’s a pretty dubious claim to begin with, that rural areas deserve more representation than urban areas. And disconcerting how it’s so often taken at face value, given how at odds it is with the basic democratic principle “one man, one vote”. There are plenty of minority constituencies that we don’t carve out extra voting power for. Why should rural areas be special?

I often wonder if the pervasiveness of this meme is subconsciously seeded by how poorly we present electoral results in map form. Those vast, uninterrupted swaths of red… they’re so big; they must be important, population density be damned. It’s like some kind of neo-manifest destiny.

Really the only supporting argument seems to come down to “the Founding Fathers wanted it this way”. But, they didn’t. At the time they were busy founding, the entire country was overwhelmingly rural; it wasn’t even a concern.

History Lesson

The true origins of the electoral college are much less charitable:

1. Preserving the voice of small states vs. big ones.

   At the time, state identity was much stronger than national identity, and there were concerns that the largest states would look out only for themselves. There is even an arcane rule that an elector cannot choose both a President and Vice President from the same state as their own, lest people always vote for the ‘home team’. Yes, that’s actually in there.

   Even now it’s plain to see that the smallest states (which get the biggest relative electoral boost) are a mix of rural red states (Alaska, Montana, the Dakotas, Wyoming, Idaho), and urbanized or at least coastie blue states (Rhode Island, Delaware, Hawaii, Vermont, Maine, D.C.).

2. Avoiding direct democracy.

   Rather than giving the vote directly to the uninformed and gullible hordes, said hordes would vote for respected citizen electors, who would choose a qualified candidate.

   This scheme, however noble, was circumvented almost immediately. States have complete discretion in choosing their electors (popular vote is not even required), and states realized they could get an edge by forming a voting bloc of electors all pledged to the same candidate. Of course other states had to follow suit to keep their own edge, leading to our current winner-take-all system. This happened practically before the ink was dry.

3. If you elect by popular vote, you don’t get to vote on behalf of your slaves.

   Slaves counted as ⅗th’s population when allocating members of congress. By tying the number of electors to the size of a state’s delegation, you get to include your slaves in your clout on the national stage, without pesky complications like actually having to let them vote.

So the respected citizen was made moot, the slaves were freed and got the vote, and national identity is much stronger these days than state identity*. For all these reasons (as well as just plain simplicity) I support abolishing the electoral college and instituting National Popular Vote.

* you could argue it died when congress got the power to tax citizens directly – if states are no longer considered separate entities when funding the government, why should they still be separate when deciding who gets to run it?

However, that could be politically untenable, and I have at least some respect for the idea that electing the President is a time to reflect that we are a nation of states as well as of people, so perhaps some consideration of states as entities should persist in the electoral process.

The Trouble with Swing States

The electoral college itself is not the problem. What needs to die is the ‘swing state’.

States allocating their electoral votes winner-take-all is a disaster. It creates a small number of very high-stakes local elections, ripe for fraud or other interference, while writing off the vast majority of the country. For all the talk of National Popular Vote ignoring the preferences of rural voters, the current system ignores rural and urban alike, save for the lucky few that live in a state where by chance the race is close.

Should the millions of red votes in California (more than the population of most red states), and every blue vote beyond 50.00001% count for absolutely nothing? Should the votes of every safe red or blue state be treated as a foregone conclusion, the concerns of the voters behind them being completely irrelevant?

Indignation about the electoral college typically focuses on how a voter in Wyoming has 3.6x the weight of a voter in California. But this is not an accurate reflection of the system and this anger is misplaced. We should define a vote’s ‘weight’ as the relative impact it has on the results of the election. With a popular vote system, each vote increments the final count by one; the impact is clear and tangible. But in an electoral college the final ‘result’ is the count of electoral votes, so a vote is only worth it’s potential to move that needle. Most votes are swallowed unless they happen to be the magic one that tips how their state allocates its electoral votes. The more safe the state is, the more indistinguishable it is whether you actually voted at all. In this sense, a voter in New Hampshire or North Carolina or Florida has hundreds to thousands of times more weight – as in real ability to change the final tally of electoral votes – than a voter in either Wyoming or California.

We can quantify this voter power as a function of how close the race is (how likely it will end up in the magic 50/50 scenario), the state’s voting population* (how likely you are to be the one to cast the deciding vote), and the number of electoral votes you would flip. Obviously these metrics are very sensitive to current polling† but the general principle – and the order-of-magnitude difference between certain states – still holds.

* for simplicity we’re not actually considering how many people in each state are eligible to vote. The proportion varies by state (from ~64–80%), which would affect the relative rankings below, but only slightly.

† and particularly to how well that state has been polled– more and better polls reduces margin of error, which has a huge impact on the likelihood of an even split, especially for a not-so-close race

Using FiveThirtyEight’s final published polling data, we get the following figures as of 2016:

	Naive power per vote (μEV) EVs ÷ population	Actual power per vote (μEV) swing state dynamics	vs. strongest vote
New Hampshire	3.03	26.48	=
Nevada	2.21	24.83	1/1.07
Alaska	4.16	22.60	1/1.17
North Carolina	1.57	21.19	1/1.25
Iowa	1.96	20.86	1/1.27
Florida	1.53	20.31	1/1.30
Arizona	1.72	19.76	1/1.34
Maine*	3.00	18.76	1/1.41
Ohio	1.56	18.70	1/1.42
New Mexico	2.42	18.08	1/1.46
Georgia	1.64	17.71	1/1.49
Colorado	1.78	17.48	1/1.51
Pennsylvania	1.57	17.13	1/1.55
Michigan	1.61	16.06	1/1.65
Wisconsin	1.75	15.39	1/1.72
Minnesota	1.88	15.35	1/1.73
Virginia	1.62	13.92	1/1.90
South Carolina	1.94	13.28	1/1.99
Delaware	3.33	11.99	1/2.21
Rhode Island	3.79	10.77	1/2.46
South Dakota	3.66	9.52	1/2.78
Oregon	1.82	8.76	1/3.02
Texas	1.50	7.58	1/3.49
Montana	3.02	7.19	1/3.68
Missouri	1.66	6.09	1/4.35
Kansas	2.10	5.47	1/4.84
Connecticut	1.95	4.55	1/5.82
Indiana	1.69	4.21	1/6.29
New Jersey	1.59	3.93	1/6.73
Tennessee	1.73	3.89	1/6.81
Nebraska*	2.73	3.53	1/7.50
North Dakota	4.44	3.47	1/7.62
Mississippi	2.01	3.15	1/8.39
Washington	1.78	2.97	1/8.90
Illinois	1.55	2.34	1/11.3
Vermont	4.76	2.18	1/12.1
Idaho	2.54	1.31	1/20.2
Hawaii	2.93	1.21	1/21.8
Wyoming	5.28 (highest)	0.73	1/36.1
Louisiana	1.76	0.67	1/39.7
Kentucky	1.84	0.43	1/62.0
Arkansas	2.05	0.30	1/89.2
New York	1.49	0.17	1/155
West Virginia	2.69	0.10	1/253
Massachusetts	1.68	0.07	1/377
Alabama	1.87	0.04	1/617
California	1.47 (lowest)	0.02	1/1,710
Oklahoma	1.86	0.01	1/2,410
Maryland	1.73	< 0.01	1/5,440
Utah†	2.17	< 0.01	1/97,900
D.C.	4.96	0	1/∞

* assuming state used winner-take-all rather than district method

† ignoring McMullin spoiler effect and counting his votes toward Republican nominee

Disenfranchising a single vote in North Carolina has the same electoral impact as erasing an entire small town in California. If that’s not the antithesis of democracy, I don’t know what is.

The D.C. situation is also interesting in light of their struggle to get representation in national elections. Collectively they now matter, but the district is so safe that any individual has literally zero impact.

It’s hard to sympathize with hand-wringing over voter turnout when so many votes are demonstrably worth very little.

This swing state mess we’ve created for ourselves is the singular biggest problem with the electoral college.

Alternatives

States don’t have to use winner-take-all.

Two states don’t, in fact. Both allocate electoral votes by congressional district, with the two ‘senator’ votes going to the statewide winner.

No, this is not the solution. In fact, this is a terrible idea. It’s saying “hey, remember how great gerrymandering was? Let’s get more of that!” If you want your electoral votes to correspond one-to-one with votes for members of congress, you should just skip the foreplay and have congress directly appoint the president.

But, it turns out if states were forced (and they’d have to be forced, because prisoner’s dilemma) to allocate their electoral votes proportionally to the popular vote in that state, everything would work out pretty well. Minority votes in all states would be heard. There’d be less incentive for voter suppression, having neutralized the greatly amplified voting power of swing states. Small states would still get their originally intended bump by having more electors per capita.

Let’s see how it plays out:

		Clinton		Trump		Johnson		Stein		McMullin
Alabama	9	3.09	(34%)	5.59	(62%)	0.19	(2%)	0.04	(0%)	—
Alaska	3	1.10	(37%)	1.54	(51%)	0.18	(6%)	0.05	(2%)	—
Arizona	11	5.00	(45%)	5.39	(49%)	0.46	(4%)	0.15	(1%)	—
Arkansas	6	2.02	(34%)	3.63	(61%)	0.16	(3%)	0.05	(1%)	0.07	(1%)
California	55	34.25	(62%)	17.53	(32%)	1.87	(3%)	1.09	(2%)	—
Colorado	9	4.33	(48%)	3.89	(43%)	0.47	(5%)	0.12	(1%)	0.09	(1%)
Connecticut	7	3.82	(55%)	2.86	(41%)	0.21	(3%)	0.10	(1%)	0.01	(0%)
Delaware	3	1.60	(53%)	1.26	(42%)	0.10	(3%)	0.04	(1%)	—
D.C.	3	2.71	(90%)	0.12	(4%)	0.05	(2%)	0.04	(1%)	—
Florida	29	13.87	(48%)	14.22	(49%)	0.64	(2%)	0.20	(1%)	—
Georgia	16	7.34	(46%)	8.17	(51%)	0.49	(3%)	—		—
Hawaii	4	2.44	(61%)	1.18	(29%)	0.15	(4%)	0.12	(3%)	—
Idaho	4	1.10	(27%)	2.37	(59%)	0.16	(4%)	0.05	(1%)	0.27	(7%)
Illinois	20	11.17	(56%)	7.75	(39%)	0.76	(4%)	0.28	(1%)	0.04	(0%)
Indiana	11	4.17	(38%)	6.25	(57%)	0.54	(5%)	0.03	(0%)	—
Iowa	6	2.50	(42%)	3.07	(51%)	0.23	(4%)	0.04	(1%)	0.05	(1%)
Kansas	6	2.16	(36%)	3.40	(57%)	0.28	(5%)	0.12	(2%)	0.03	(1%)
Kentucky	8	2.61	(33%)	5.00	(63%)	0.22	(3%)	0.06	(1%)	0.09	(1%)
Louisiana	8	3.08	(38%)	4.65	(58%)	0.15	(2%)	0.06	(1%)	0.03	(0%)
Maine	4	1.91	(48%)	1.81	(45%)	0.20	(5%)	0.08	(2%)	—
Maryland	10	6.03	(60%)	3.39	(34%)	0.29	(3%)	0.13	(1%)	0.03	(0%)
Massachusetts	11	6.70	(61%)	3.66	(33%)	0.46	(4%)	0.16	(1%)	0.01	(0%)
Michigan	16	7.56	(47%)	7.60	(47%)	0.57	(4%)	0.17	(1%)	0.03	(0%)
Minnesota	10	4.64	(46%)	4.49	(45%)	0.38	(4%)	0.13	(1%)	0.18	(2%)
Mississippi	6	2.39	(40%)	3.51	(59%)	0.07	(1%)	0.02	(0%)	—
Missouri	10	3.82	(38%)	5.74	(57%)	0.35	(3%)	0.09	(1%)	0.00	(0%)
Montana	3	1.08	(36%)	1.69	(56%)	0.17	(6%)	0.05	(2%)	—
Nebraska	5	1.72	(34%)	3.00	(60%)	0.23	(5%)	0.05	(1%)	—
Nevada	6	2.88	(48%)	2.73	(46%)	0.20	(3%)	—		—
New Hampshire	4	1.88	(47%)	1.86	(47%)	0.17	(4%)	0.04	(1%)	0.01	(0%)
New Jersey	14	7.68	(55%)	5.89	(42%)	0.28	(2%)	0.14	(1%)	—
New Mexico	5	2.41	(48%)	2.00	(40%)	0.47	(9%)	0.06	(1%)	0.04	(1%)
New York	29	17.06	(59%)	10.86	(37%)	0.67	(2%)	0.41	(1%)	—
North Carolina	15	6.93	(46%)	7.47	(50%)	0.41	(3%)	0.04	(0%)	—
North Dakota	3	0.82	(27%)	1.89	(63%)	0.19	(6%)	0.03	(1%)	—
Ohio	18	7.88	(44%)	9.35	(52%)	0.57	(3%)	0.15	(1%)	0.04	(0%)
Oklahoma	7	2.03	(29%)	4.57	(65%)	0.40	(6%)	—		—
Oregon	7	3.64	(52%)	2.84	(41%)	0.34	(5%)	0.18	(3%)	—
Pennsylvania	20	9.60	(48%)	9.75	(49%)	0.48	(2%)	0.16	(1%)	—
Rhode Island	4	2.19	(55%)	1.61	(40%)	0.14	(4%)	0.06	(1%)	—
South Carolina	9	3.66	(41%)	4.94	(55%)	0.21	(2%)	0.06	(1%)	0.09	(1%)
South Dakota	3	0.96	(32%)	1.87	(62%)	0.17	(6%)	—		—
Tennessee	11	3.86	(35%)	6.76	(61%)	0.31	(3%)	0.07	(1%)	—
Texas	38	16.43	(43%)	19.85	(52%)	1.20	(3%)	0.30	(1%)	0.18	(0%)
Utah	6	1.67	(28%)	2.76	(46%)	0.21	(4%)	0.05	(1%)	1.31	(22%)
Vermont	3	1.67	(56%)	0.89	(30%)	0.09	(3%)	0.06	(2%)	0.01	(0%)
Virginia	13	6.47	(50%)	5.78	(44%)	0.39	(3%)	0.09	(1%)	0.18	(1%)
Washington	12	6.57	(55%)	4.60	(38%)	0.61	(5%)	0.22	(2%)	—
West Virginia	5	1.33	(27%)	3.45	(69%)	0.16	(3%)	0.06	(1%)	—
Wisconsin	10	4.73	(47%)	4.80	(48%)	0.36	(4%)	0.11	(1%)	—
Wyoming	3	0.66	(22%)	2.05	(68%)	0.16	(5%)	0.03	(1%)	—
Total	538	257.2 (47.81%)		251.4 (46.72%)		18.2 (3.39%)		5.8 (1.08%)		2.8 (0.52%)

Now, fractional votes are liable to make most people’s head explode. Electors are supposed to be actual people, too, which would present some logistical challenges. (Can an elector now be just 30% faithless?) So, we must round each state’s votes to whole numbers.

Properly rounding a set larger than two can be a bit tricky, but there are various fair schemes. One such scheme results in:

	Clinton	Trump	Johnson	Stein	McMullin
Alabama	3	6	—	—	—
Alaska	1	2	—	—	—
Arizona	5	6	—	—	—
Arkansas	2	4	—	—	—
California	34	18	2	1	—
Colorado	5	4	—	—	—
Connecticut	4	3	—	—	—
Delaware	2	1	—	—	—
D.C.	3	—	—	—	—
Florida	14	14	1	—	—
Georgia	7	8	1	—	—
Hawaii	3	1	—	—	—
Idaho	1	3	—	—	—
Illinois	11	8	1	—	—
Indiana	4	6	1	—	—
Iowa	3	3	—	—	—
Kansas	2	4	—	—	—
Kentucky	3	5	—	—	—
Louisiana	3	5	—	—	—
Maine	2	2	—	—	—
Maryland	6	4	—	—	—
Massachusetts	7	4	—	—	—
Michigan	7	8	1	—	—
Minnesota	5	5	—	—	—
Mississippi	2	4	—	—	—
Missouri	4	6	—	—	—
Montana	1	2	—	—	—
Nebraska	2	3	—	—	—
Nevada	3	3	—	—	—
New Hampshire	2	2	—	—	—
New Jersey	8	6	—	—	—
New Mexico	3	2	—	—	—
New York	17	11	1	—	—
North Carolina	7	8	—	—	—
North Dakota	1	2	—	—	—
Ohio	8	9	1	—	—
Oklahoma	2	5	—	—	—
Oregon	4	3	—	—	—
Pennsylvania	10	10	—	—	—
Rhode Island	2	2	—	—	—
South Carolina	4	5	—	—	—
South Dakota	1	2	—	—	—
Tennessee	4	7	—	—	—
Texas	17	20	1	—	—
Utah	2	3	—	—	1
Vermont	2	1	—	—	—
Virginia	7	6	—	—	—
Washington	7	4	1	—	—
West Virginia	1	4	—	—	—
Wisconsin	5	5	—	—	—
Wyoming	1	2	—	—	—
Total	264 (49.07%)	261 (48.51%)	11 (2.04%)	1 (0.19%)	1 (0.19%)

Which presents another wrinkle. You need an absolute majority of electoral votes (270 presently) to win the presidency, otherwise the House of Representatives decides via arm wrestling or something. With proportional allocation, third parties now get a small residue of electoral votes that can much more easily push the major party candidates below the 50% threshold.

There are a few solutions:

• “works as intended”

• in the same constitutional amendment you needed to force all states to adopt a common allocation method, lower the threshold to win to 40% or so, and require only a plurality (easiest)

• allow electors multiple rounds of voting to converge on a majority winner, not unlike delegates at the conventions (in the original spirit of things, but not a chance)

• acknowledge the inherent two-party hegemony built into our system and treat third parties like the annoying flies to be swatted at every four years that they are

If we limit proportional allocation to just the top two winners in each state, we get:

	Clinton	Trump
Alabama	3	6
Alaska	1	2
Arizona	5	6
Arkansas	2	4
California	36	19
Colorado	5	4
Connecticut	4	3
Delaware	2	1
D.C.	3	—
Florida	14	15
Georgia	8	8
Hawaii	3	1
Idaho	1	3
Illinois	12	8
Indiana	4	7
Iowa	3	3
Kansas	2	4
Kentucky	3	5
Louisiana	3	5
Maine	2	2
Maryland	6	4
Massachusetts	7	4
Michigan	8	8
Minnesota	5	5
Mississippi	2	4
Missouri	4	6
Montana	1	2
Nebraska	2	3
Nevada	3	3
New Hampshire	2	2
New Jersey	8	6
New Mexico	3	2
New York	18	11
North Carolina	7	8
North Dakota	1	2
Ohio	8	10
Oklahoma	2	5
Oregon	4	3
Pennsylvania	10	10
Rhode Island	2	2
South Carolina	4	5
South Dakota	1	2
Tennessee	4	7
Texas	17	21
Utah	2*	4
Vermont	2	1
Virginia	7	6
Washington	7	5
West Virginia	1	4
Wisconsin	5	5
Wyoming	1	2
Total	270 (50.19%)	268 (49.81%)

* McMullin was within striking distance of finishing 2^nd place in Utah, snatching Hillary’s two electoral votes and leaving no one with a majority

To summarize:

	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Clinton	48.31%*	232 (43.12%)	257.2* (47.81%)	264* (49.07%)	270 (50.19%)
Trump	46.45%	306 (56.88%)	251.4 (46.72%)	261 (48.51%)	268 (49.81%)
Johnson	3.31%	—	18.2 (3.39%)	11 (2.04%)	—
Stein	1.07%	—	5.8 (1.08%)	1 (0.19%)	—
McMullin	0.45%	—	2.8 (0.52%)	1 (0.19%)	—

* plurality, but not a majority

† ignoring faithless electors

So it looks like everything behaves pretty well. Final results are much more in line with the popular vote, but the margin is narrower, which is the desired effect from boosting certain segments of voters.

This isn’t just sour grapes that all the alternate scenarios result in a Hillary win (albeit some prett-aay narrow ones). Standard caveats apply that campaigns are run to optimize for the system currently in place, and if a different one had been, then the entire election could have played out completely differently. The point here is to highlight the undemocratic and chaotic effect that winner-take-all and swings states exact on the election outcome – in this case that 0.06% of the votes flipped an election with a popular vote margin of almost 2%.

I stand by these results despite such caveats, since although swing state dynamics depress voter turnout in safe states, I see no reason why they would materially alter the proportion of votes currently cast within each state. It’s actually impressive these alternate vote counts track the popular vote as well as they do.

Historical Results

And what would have happened in elections past (…and future)?

2020	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Biden	51.32%	306 (56.88%)	273.3 (50.81%)	275 (51.12%)	276 (51.30%)
Trump	46.86%	232 (43.12%)	254.6 (47.32%)	262 (48.70%)	262 (48.70%)
Jorgensen	1.18%	—	6.5 (1.21%)	1 (0.19%)	—

2012	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Obama	51.06%	332 (61.71%)	272.2 (50.60%)	275 (51.12%)	276 (51.30%)
Romney	47.20%	206 (38.29%)	256.1 (47.61%)	262 (48.70%)	262 (48.70%)
Johnson	0.99%	—	5.5 (1.03%)	1 (0.19%)	—

2008	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Obama	52.93%	365 (67.84%)	283.4 (52.68%)	289 (53.72%)	289 (53.72%)
McCain	45.65%	173 (32.16%)	246.7 (45.86%)	249 (46.28%)	249 (46.28%)

2004	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Bush	50.73%	286 (53.16%)	274.6 (51.03%)	280 (52.04%)	280 (52.04%)
Kerry	48.27%	252 (46.84%)	257.8 (47.93%)	258 (47.96%)	258 (47.96%)

2000	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Gore	48.38%*	267 (49.63%)	258.3 (48.01%)	266* (49.44%)	269 (50.00%)
Bush	47.87%	271 (50.37%)	259.2* (48.18%)	265 (49.26%)	269 (50.00%)
Nader	2.74%	—	14.9 (2.77%)	7 (1.30%)	—

This election was very close no matter how you slice it.

It also forces us to confront a supposed ‘benefit’ of winner-take-all: that in close races recounts can be localized to just the state that was close. Otherwise, a small change in any state enough to eke out one more electoral vote could make or break the election. So rather than just Florida, a close race like this may have triggered recounts in most if not all states.

I am not very sympathetic to this argument.

Saying you can restrict the recount to just one problem state exposes the lie that you’re not really running a national election in the first place – just an election in several key states. And what is so bad if we have recounts in many states– if we have the resources to run a national election in the first place, we certainly can afford to perform a national recount. States should be auditing their results already as a matter of course, not only where there is a problem! Furthermore, if many state recounts had been happening simultaneously, the Florida recount itself may not have been such a shitshow without the pressure of knowing that everything depends on these few-hundred hanging chads.

Speaking of chads, a nationwide recount would also do well to expose the absolute insanity of not having consistent national standards for how polls are run.

To me this argument exposes an extremely disturbing lack of concern for the rigor with which we run elections.

1996	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Clinton	49.23%*	379 (70.45%)	263.6* (49.00%)	267* (49.63%)	294 (54.65%)
Dole	40.72%	159 (29.55%)	220.1 (40.92%)	225 (41.82%)	244 (45.35%)
Perot	8.40%	—	45.2 (8.41%)	45 (8.36%)	—
Nader	0.71%	—	3.9 (0.73%)	1 (0.19%)	—

1992	Popular Vote	Actual EVs†	Proportional EVs	Proportional EVs, rounded	Top-2 EVs
Clinton	43.01%*	370 (68.77%)	230.5* (42.85%)	237* (44.05%)	287 (53.35%)
Bush	37.45%	168 (31.23%)	202.3 (37.61%)	197 (36.62%)	247 (45.91%)
Perot	18.91%	—	101.5 (18.87%)	104 (19.33%)	4‡ (0.74%)

‡ Perot placed 2^nd in two states (one won by Clinton and one by Bush), so he still gets some electoral votes even with the top-two method

So no huge upsets. Just the end of an archaic and entrenched system that only still exists because whoever makes the first move to fix it loses.

The margins with proportional electoral votes are actually surprisingly close to the popular vote. So close, in fact, that it makes this electoral college business seem really not worth the trouble. Whatever theoretical arguments you can make in support of the electoral college, if it still tracks the popular vote to within ~1%, what’s the point, really? National Popular Vote would also challenge the assumption that each state deserves a fixed stake in the choice of President, regardless of turnout in that state or how many of their own voters they disenfranchise.

But I can understand the resistance to dismantling the electoral college for both emotional and logistical reasons.

Nonetheless, it doesn’t have to be either National Popular Vote or Exactly What We Currently Have. I just thought I’d take a moment to advocate for a third option that I rarely see discussed– one which leaves nearly all the machinery of our current system in place, yet solves nearly all its problems.

Honestly, it would be great if we could just use a preference-based voting system like Condorcet or Single Transferable Vote, but that is firmly in the realm of fantasy.