It's understandable, but wrong, to lay the blame for "swing states" on the Electoral College system; the possibility of a swing state is just as real in any proposed alternative which tries to mimic a direct national vote, because concentrated regional blocs of voters can easily be numerous enough to swing such an election. The only difference is that the "swing state" has to have a large population, or has to become a "swing region" made up of several states (so that, instead of saying the election hinges on Ohio, we'd say it hinges on the Upper Midwest, for example). It's also understandable -- but also wrong -- to dismiss the Electoral College, and the winner-take-all system most states use to apportion their electoral votes, as flawed. Perhaps the best explanation of why that's wrong is an analogy coined by mathematician Alan Natapoff (profiled in "Math Against Tyranny", Discover Magazine, November 1, 1996; a must-read article for critics of the Electoral College). Think of another quintessentially American institution: baseball. The champion of each major-league season is decided, ultimately, by the World Series, a best-of-seven contest. We understand on a pretty fundamental level that this is a pretty good way to pick a winner: by requiring the eventual champion to perform consistently well over multiple games, we're doing away with flukes introduced by isolated, one-sided blowouts. The analogy to various direct-vote emulation schemes would be a series which always goes seven games, and where the team with the most total runs over those games wins. But this opens the door to, say, a "champion" who had a 7-0 victory in Game 1, followed by six straight 1-0 losses. Any baseball fan would immediately say that such a team is no "champion", and that the "losers" who won six out of seven games should be the ones getting the ticker-tape parade. The Electoral College imposes a similar requirement of consistency on a potential President: by essentially creating many separate Presidental elections, the College requires the the eventual winner perform well across a broader spectrum. If not for this, we wouldn't be able to talk about electing the "President of the United States" with a straight face; most of the time, a few populous regions would elect a President, and the rest of us would be along for the ride. We'd have the "President of California, New York and Florida", who'd incidentally end up as chief executive of all forty-seven other states. And that would be nothing short of disastrous: the only response possible from the more sparsely-populated states (who would naturally want to ensure their interests were represented by the office of the President) would be to band together into a massive voting bloc, and we'd end up with an even more polarized electorate. What's worse, it would largely be polarized along geographic lines: the large population centers are mostly coastal, while the sparser areas lie in the nation's interior. The last time we were divided that sharply along a roughly geographic boundary, we fought a four-year civil war and got a lasting legacy of regional enmity thrown into the bargain. Keep the Electoral College, on the other hand, and the larger population centers lose some of their clout, the less-populated states gain a bit of power and a successful candidate is placed in a situation of needing to appeal to both. An analogy can be drawn to the structure of the Federal legislative branch: the bicameral Congress, with representation apportioned equally to all states in one house and apportioned by population in the other, makes the same trade-off in a far more explicit fashion, and in doing so provides an effective check against tryanny by the majority. The Electoral College was not, so far as I'm aware, conceived with any such goal in mind, but nonetheless accomplishes much the same thing for the executive branch, ensuring -- for all its perceived faults and antiquated notions about representative democracy -- that a candidate for the Presidency cannot glide in on a single blowout any more than a team in the World Series can ride a single game's run tally through to the end.

Whew !!!

You've made a very clever argument against a national popular vote for the Presidency.

Now try to apply that same argument aginst apportionment of the various state's electors based on the popular vote within each state.

Why is it beneficial to the citizens of a state to apportion all of their electors to one candidate after a 51% to 49% vote occurs?

How about a compromise? Keep the EC, but require apportioned appointment of each state's electors.

Go back to the baseball analogy: why do we say one team wins a game with a final score of 11-10, when the other team came so close? The answer is that -- if we value consistency -- we have to have some level where you either "win" or "lose" instead of being awarded a percentage. It's no different with voting for President: if the goal is to require the winner to have broad appeal across the country, we have to have some level of election where you either "win" or "lose" that entire level, instead of being awarded a percentage. Making electoral vote apportionment depend directly on popular vote apportionment would remove that and weaken the system beyond the point of effectiveness. Of course, this is predicated on the assumption that our goal is to make sure the President is broadly popular and accepted; I've explained why I think that's necessary.

I just disagree with the concept that the President needs to be given an artificial mandate by awarding a landslide victory to a candidate who might have won the election by only 1% of the vote.

I see it as blatant dishonesty and an affront to realism. Even if you can make an argument that this deception might improve governance, I don't think that is a fair trade-off for putting citizens in many states in the position of knowing that their vote for the highest political office in the land is meaningless.

Many different soccer competitions in Europe and abroad use a two-game aggregate scoring system to see who won (though admittedly, not for the final game) in certain tournaments. If the score is tied then the team that scored the most away goals wins, and if that is a tie then they do a shoot-out. The Europeans seem to be satisfied with this situation, and as a soccer fan I can tell you I enjoy it immensely.

1. The baseball analogy is inapposite. The rules of sporting events are designed not to foster political and economic justice, but to ensure excitement, drama, and entertainment. If that means prolonging the contest beyond a single game, so be it. For every sports analogy that might appear to be on point in some way, there's another that "proves" the opposite. For example, football is inarguably America's favorite sport today. The NFL championship is decided in a single, winner-take-all event. Is there a reason that the World Series should be deemed to provide better guidance for presidential elections than the Super Bowl? No; in fact, neither sporting event provides guidance because both are completely irrelevant.

2. The "interests" of "sparsely populated states" are already overrepresented in Congress, because each state is guaranteed two senators. That overrepresentation is more than enough to ensure them a voice in national policymaking. There is no need for this oligarchical tendency to permanently influence the executive branch as well.

3. The notion that the vast interior of the nation might end up being utterly unrepresented by a president elected through a national popular vote is unsupported. It stems from the improper application of Electoral College-style thinking to such an election, i.e., that a candidate who does not obtain a majority of a state's votes "loses" the entire state. In reality, a president elected through a national popular vote probably would enjoy at least 40-45% support even in the most hostile states. He or she might receive 49% of the votes in others that were "lost." The resulting feelings of enfranchisement and disenfranchisement among voters probably wouldn't be much different from the way they are today, thanks to the two-party system.

4. If the nation's population is disproportionately located on the coasts, why shouldn't those population centers enjoy an equivalent influence on election results? The president should represent all citizens equally, and should not attempt to represent states or, even worse, vaguely defined regions or "ways of life."

JMR, I am inclined to agree with you. However, I also believe in the Law of Unitended Consequences, which makes me skeptical of that inclination.

"The "interests" of "sparsely populated states" are already overrepresented in Congress, because each state is guaranteed two senators. That overrepresentation is more than enough to ensure them a voice in national policymaking."

1) Yes, and that holds to some extent to the House with gerrymandering thrown into the mix as well.

"In reality, a president elected through a national popular vote probably would enjoy at least 40-45% support even in the most hostile states. He or she might receive 49% of the votes in others that were "lost.""

2) I don't know about that. It seems that this contention is based experiences under the current system rather than what would prevail under another system. The current system is set up against a candidate seeking to trounce a competitor in any given state. Under the electoral college, if a candidate thinks he has almost 60 percent support in a state, he or she will reduce the amount of resources (and promises) that he or she commits to that state. However, under the popular vote, a candidate has incentives to continue to throw resources (and promises) at the state in hope of running up the vote total. To support my point, I would note that the vote totals by county of gubernatorial elections can be extremely lopsided.

(1) and (2) bring me to my point: Is it wise to have a President elected on a total different basis (and with a different base of support) than Congress? It seems like a recipe for gridlock (at best) or (much worse) power conflicts between the President and Congress as each argues who represents the "people."

President of New York? California?

George W. Bush lost these states but still won the popular vote. How do you explain that?

The fact is the Electoral College forces candidates to campaign in some places more vigorously than others whereas a direct election would force them to campaign everywhere. The country is not red or blue, it's purple.

A flaw with the baseball analogy is the game is not played the way we elect the president. Change the World Series to this format:

Game 1: 9 innings

Game 2: 3.5 innings

Game 3: 14 inning

Game 4: 18 innings

Game 5: 8 innings

Game 6: 6.5 innings

Game 7: 4 innings

and the game would be played in an entirely different way. The best pitchers would only be played on the short games, and eventually, both teams would probably agree to get through the long games, each trading off a win.

"(1) and (2) bring me to my point: Is it wise to have a President elected on a total different basis (and with a different base of support) than Congress? It seems like a recipe for gridlock (at best) or (much worse) power conflicts between the President and Congress as each argues who represents the "people."

Would it necessarily be that different that what we have now? The winner-take-all system used by most states for their electoral college votes produces a different state-by-state power balance than in Congress, where individual electoral districts have the freedom to disagree with the majority of their state. And, on an individual level, the direct election of each congressman varies dramatically from the electoral college system. A direct election of Presidents would actually bring them closer together, not farther apart.

Have electors chosen in the same manner as congress. Most of them represent their home districts, and are authorised to vote for the candidate who won the most votes in their district. Two are chosen state-wide, and authorized to vote for the candidate who won the most votes overall. Or, to compensate for the simultaneous choice of the state-wide electors (Senators are seldom, if ever, chosen simultaneously), have one state-wide elector chosen by majority vote, and the other chosen by vote of the state legislature, with the Governor getting a tie-breaker vote in the latter case.

This gives you the needed broad-based support for Presidential election, but makes it much harder for a single official to swing the entire election; even the state Governor would only effect a single electoral vote.

1.) The claim that the possibility of a swing state is just as real in any proposed alternative which tries to mimic a direct national vote, because concentrated regional blocs of voters can easily be numerous enough to swing such an election.

The concept of a swing state is predicated on the notion that a small change in the proportion of the vote within the state (say a 51-49 Pennsylvania instead of a 49-51 Pennsylvania) can produce a disproportionate result. Without that dynamic the concept of a swing state (or swing region) is meaningless, as it does not matter who wins this or that region, it only matters who wins the overall most votes. This would be a recipe for LESS regional polarization, not more.

2.) The Article "Math Against Tyranny" does not refute any of the criticisms of the Electoral College as it now stands- its marginalization of non-swing states, its anti-democratic nature, its anachronistic character, and it's arbitrariness. It is predicated solely on the idea that one's likelihood of influencing the outcome of the election by your vote alone is increased from a vanishingly nonexistent amount to a slightly less vanishingly nonexistent amount. In other words, an interesting theoretical exercise, but no more, and certainly does not lead one to the conclusion of the present College.

3.) The claim that the Electoral College is like baseball. The game of baseball, as has been pointed out above, is an exercise in recreation in which the most important unit of victory is the "game". It is not designed to measure how good a team's athletes are to the optimal level of perfection. If one team won a 7 game series by 4-3 but the losing team had numerous injuries and won their 3 games by huge margins while those injuries cost them the decisive 1 game, who would the money-changers rank as the best game if the teams were to face each other again, or in the next season without any trades? Even in baseball, those who are seriously studying the statistics do not just look at games won by RBIs, ERAs, etc. etc.

Furthermore, the analogy obscures that all states, unlike all games, are not equal. In fact, 80% of states are foregone conclusions. What if there was a rule that said that a team could only field its best players in game 1, while in games 2-5 it could only field two or three bad pitchers and its worst hitters? Such would be a game that was a foregone conclusion, and no baseball fan would stand for such rules.

4.) The claim that the College requires the the eventual winner perform well across a broader spectrum.

On the contrary, the College requires the eventual winner to perform across a NARROWER spectrum, geographically. He or she can get demolished in Nebraska, Montana, Texas, and Utah, but still win a victory by eking out tiny margins in Ohio and Florida. In fact, only the Electoral College allows the winner to win with less than a majority of the vote. By definition, it leaves more room for winners who leave more American voters behind than the popular vote.

5.) The claim that we would not be able to talk about the "President of the United States" but we'd have the "President of California, New York and Florida"

Does the author of this objection really think George W. Bush, the current President, really has 'wide geographical support' encompassing all areas of the country? Hardly. The President elected by the Electoral College has been one of the most polarizing Presidents in recent history. To many, he's the President of "Texas, Texas, and Texas". He's made absolutely no effort to reach out to the people or areas that did not support him, as evidenced by his win in 2004 again losing almost exactly the same states.

6.) The claim that the electorate would largely be polarized along geographic lines: the large population centers are mostly coastal, while the sparser areas lie in the nation's interior.

Again, the electorate is currently polarized (geographically) along roughly those lines. The author is merely describing the status quo. The only difference is that the author (apparently) imagines that this time it would be the coastal areas in the majority. This assumption is wrong-- as the Slate writer pointed out, Kerry, who had more support among coastal areas, would have been put at a disadvantage in a national popular vote situation.

7.) The claim that the Electoral College empowers small states while disempowering large ones.

Simply incorrect. The Electoral College empowers SWING states while disempowering non-swing states. Hence Florida, the 4th-largest out of 50 states (and hence in the top 8% percentile) decisively swung the 2000 elections due to the electoral college, while the much smaller states of Oregon and New Mexico were unable to do so. The small states like Montana and the Dakotas are DOUBLY punished by the Electoral College: not only do they have less people but they are not swing states. Hence, they receive even LESS attention than they would under a national popular vote.

8.) The notion that the Electoral College goes against the 'tyranny of the majority' like bicameral Congress, with representation apportioned equally to all states in one house and apportioned by population in the other, makes the same trade-off in a far more explicit fashion, and in doing so provides an effective check against tryanny by the majority.

This is a fundamental misunderstanding of how Congress's structure protects against tyranny of the majority. It does so by dividing powers among different parties, such as the two Houses of Congress; by staggering elections in the Senate, and by giving certain protections to the minority, regardless of which minority that is.

The Electoral College on the other hand, serves only to introduce an arbitrary element of bias into the system. It does not divide the Presidential power; whomever is elected President still has all the same unified powers. That makes the bicameral and staggering elements of Congressional diffusion irrelevant. Secondly, it does not give certain protections to the minority regardless of who is in the minority. It only protects the minority under certain, arbitrary cases. When the 'tyranny of the majority' benefits from the Electoral College (which is most of the time), it actually artificially enhances the stature of a smaller win by making it seem larger than it was.

In short, the claims made above are invalid; it is not impossible to defend the Electoral College, but they cannot be made under the grounds given by ubernostrum above.

The atomic unit of a democracy should be the individual voter. Seen that way, a popular vote victory is never akin to winning one landslide game in a seven game series, it's like winning one single game with a combined score in the tens of millions, which nearly any statistician (except apparently Napatoff) would easily endorse as the more sound sampling method than
the imaginary groupings created by state boundaries.

What good reason do we have for not valuing each voter equally, regardless of state?

Feline74, while I believe this suggestion might improve upon the current EC, if nothing else it would simply reinforce the absuridity of electing a national post, with a national electorate, through congressional districting. If this should be a desirable outcome, why not just dispense with presidential elections and let a special joint session of congress take care of it?

"It is predicated solely on the idea that one's likelihood of influencing the outcome of the election by your vote alone is increased from a vanishingly nonexistent amount to a slightly less vanishingly nonexistent amount. In other words, an interesting theoretical exercise, but no more, and certainly does not lead one to the conclusion of the present College."

While I agree that the article doesn't provide persuasive support for the college, I think you give it unduly short shrift. The statistical findings are interesting and relevant, even if the actual probabilities discussed are infintesimal (this brings up the whole question of the irrationality of voting in modern mass democracies). The point is simply that the premise of the article is that voting power is more important than equality of power, based on the straw-man argument of equal zero power in a despotism. The EC, like the majoritarian electoral systems used in the US for electing representatives to congress (contrast European proportional election system) maximises the possiblity of being a swing voter. The upshot is disempowerment of other voters. The article provides no reason to believe that this is a fair trade-off.