That the US is the longest-standing democracy in existence today doesn’t prevent it from having massive reliability problems.

Iceland might differ about the priority of the American upstarts… their Althing is not only the oldest democratic government, but the oldest continuously-operating human organization, beating out the Catholic Church by some centuries.

(It does help that they’re an island with a fairly homogeneous population!)

OTOH, the counterexamples being proposed, Kenya and Mexico, are not generally considered part of the developed world, and neither has a long tradition of genuinely competitive elections.

First, they get run by politicians within the party system. Florida:2000 is an obvious example. Most of these people may be honest, but it only take a few to spoil things.

Second, many of the most debatable features of the Electoral College system are outside the Constitution. The Electoral College could still work, if people were electing an individual representative, rather than the whole state’s EC representation going for whichever candidate won the majority of the votes.

As it is, the value of winning a large state is so high that almost any skulduggery can be justified, and the election is being run by a person who can further their political career by putting their thumb on the scales.

That is a surprising statement to make given that the procedural problems that haunt US elections are pretty much unique in the developed world. That the US is the longest-standing democracy in existence today doesn’t prevent it from having massive reliability problems. I would even argue that these problems are due to the fact that the US constitutional system is so old, anachronistic in parts, and has never been overhauled. As has been pointed out by others, the electoral college system does not tend to produce close results. Rather, it usually lets the victory appear much clearer than it really is and it supports a media narrative in which the President “has won” this and that state, as if millions of dissenting voters didn’t exist.

One of the main problems of this system is the distortion that it produces by making the votes of some people more valuable than others. It should normally be an embarassment in a democracy to not count each vote the same. Yet Americans in general seem to be totally oblivious to this concern.

The Democratic primary, on the other hand, is not a very good process for determing a clean succession. But it has a different goal. The primary as currently constructed gives the challengers a much better opportunity to make their case and develop support. Clinton would have been a lock this year if not for the early January states. Obama was able to establish that he was a viable candidate, which greatly helped him moving forward. Also, it provided him with funding momentum and more time to campaign in later states. In many states that Obama later won, polling in January had Clinton well ahead. Since we don’t want the primary to be a name-recognition process, it is structured in the fashion that it is. Occassionally this will give rise to the current scenario. But actually, the chance that the political party will be split 50-50 on two candidates is pretty low compared to the chance that the nation as a whole will be roughly 50-50 between the two major parties. So normally this structure would not be much of a problem.

If the US Elections were to be decided on the basis of the National popular vote alone, there would be no need for the Electoral College, nor the States which represent it. Once this happens, the question then becomes, “If the President shall be decided by National popular vote, then why not Congress too?”. This is where abb1’s idea immediately inserts itself.

The vote differences between candidates percentage-wise are tiny. Very rarely do candidates get more than 60% of the vote. Politicians will always hug the center—that is the central concept of a democracy. Winning 52% of the popular vote (as Obama probably will in the primaries) means nearly half the country didn’t vote for you.

Furthermore, as we are seeing in the democratic primary, political divisions often reflect demographic divisions. However, demographics is destiny: friendship circles, where one lives, and the media one pays attention to all correlate. The like minded cluster.

The typical voter may be surrounded entirely by supporters of one candidate. The impression generated by such unanimity can be so strong that even when the voter’s candidate loses, that loss is perceived as illegitimate. People simply refuse to see themselves in the minority.

Or are you saying that his early smallscale victories were essential to his later success (by maintaining momentum)?

Momentum may have been important, but I think more important was Obama’s initial victory in Iowa, which reassured black voters that whites were willing to vote for him. Before that the black vote was more evenly split between Obama and Clinton.

But the 2006 Mexican election is still a strong counter to Quiggin’s conjecture. Mexico had a much praised electoral commission. In 2000, Mexico was helped up as an example to how the US should manage its counting. In Mexico, part of the controversy was that the court only ordered recounts in certain areas. Indeed, one of the key problems with most electoral systems, including the US and Mexico, is that it is difficult to obtain a national recount. In the end, the actual number of ballots recounted was only 9% of the polling places in Mexico because specific legal complaints had to be filed for each polling place.

Quiggin is correct that if it were popular vote only and with national controls on the polls that the significance of narrow state margins would decline (remember the question about how many votes Kerry lost the election by: by the popular vote, 3 million, but he lost OH by only 118,000 and a victory there would have given him the presidency). But narrow votes nationally (less than 1% will still raise questions about how votes were tabulated).

Picture the U.S. with a combined primary/general presidential election where after a small amount of winnowing, the remaining dozen or so candidates from both parties combined were put on a solid Condorcet preferential ballot. This is ostensibly more complex, of course, but would have the advantage that the Condorcet principle directly and specifically addresses the question of close elections between any two candidates, and the wider field should reduce the odds of a near-tie.