Steve Landburg, guest blogging at Marginal Revolution, writes:

Amid all the scaremongering about a nailbitingly close election with a disputed outcome, it is worth observing that if you really believe in democracy, and if the election is close, then it doesn't much matter who wins. The theory of democracy (stripped down to bare essentials, and omitting all sorts of caveats that I could list but won't) is that the guy who gets more votes is the better guy. Surely, then, it follows that the guy who gets only slightly more votes is only the slightly better guy. And if one guy's only slightly better than the other, then a miscount is no great tragedy.

Bullshit.

There are two things that make democracy the best form of government. First, democracy is a system under which ambitious men and women can compete for power without spilling blood. To depose a king, you must kill him; to depose a president in a America, you only need to get enough people in enough states to vote for his opponent. The aftermath of the 2000 presidential election was pretty ugly, and the aftermath of the upcoming election may be as well; but no one was killed over the 2000 election, and it’s fairly safe to say that no one will be killed over this one.

The second great thing about democracy, as Matthew Yglesias has pointed out, democracy has salutory effects on the behavior of office holders who will be seeking reelection in the future:

Democracy, they say, is the worst form of government except for all the others. But why would that be? Not, certainly, because of the superior wisdom of the voting public who, if you read any of the public opinion literature you'll swiftly see, have almost no grasp of substantive policy issues and only a very vague familiarity with what the different candidates stand for. And yet, it seems to work pretty well. This is, I think, primarily because the voters have a habit of kicking incumbents out of office when thinks don't seem to be going well, and reelecting them when things are going well.

This is often not a very sound analytic approach. Candidates get blamed for economic problems that are not really their fault (see, e.g., Jimmy Carter in 1980) or get praise beyond what they deserve for improvements in living conditions (see, e.g., Rudy Giuliani in 1997). Nevertheless, this crude approach has certain merits. In particular, it encourages officeholders to try and make things better. If an incumbent mayor knows that whether the crime rate rises or falls will seriously impact his electoral fortunes, he has reason to try and make the crime rate fall. If an incumbent president knows that a solid macroeconomic situation will benefit him on Election Day, he'll spend at least some time trying to make it come about.

Knowing what we do about the American electorate, it would seem highly dubious that the elections are in any way a reliable selection mechanism for selecting the best candidate, by whatever objective standards you might appeal to, and even more dubious that a close election would indicate that the two candidates are equally good. If an election happens to select the best candidate, it’s mostly by chance.

And one certainly doesn't have to hold to Landsburg's naive "theory of democracy" in order to "really believe in democracy."

UPDATE: Mark Kleiman makes a similar point:

Just remember: Watching democracy in action is pretty depressing if you think of democracy as a noble project of collective self-government. But it doesn't look nearly so bad if you think of elections as an alternative to civil war.