A few interesting items in the blogosphere today.
First, Mickey Kaus argues that the recent parity of the Democrats and Republicans is no coincidence, but instead results from the parties' moving towards a competitive equilibrium:

Imagine that we have a two party system, and each party is a collection of status-seeking individuals looking for power by winning a greater "market share" of the vote. Imagine that they each have their ideological principles --one is more to the left, one more to the right -- but these principles are quite flexible in the face of imminent or repeated failure at the polls. Over time, as each party crafts its message to maximize its appeal -- and adjusts its message after each election to regain any lost share of the votes -- wouldn't one expect the system to reach a roughly 50-50 equilibrium, in which every election was a cliffhanger?

Jacob Levy expands on Kaus' theory. (Levy's specific link is broken; scroll down to the post that begins "Mickey Kaus argues".) He expands on the ramifications of the dead-locked country at great length.
It appears obvious that we're going to see more Floridas, with so many races being so close and control of Congress at stake.
I thought of one other ramification. In the wake of the substitution of Frank Lautenberg for Bob Torricelli in New Jersey, a number of commentators argued that the move was unlikely to have much precedential value. In the words of Josh Marshall:

Can anyone who makes this argument have ever spent any time around elected politicians? Not a chance. Especially these days with weak parties there's really no institutional force capable of knocking a candidate out of a race. And people who run for office just don't have egos that work that way. To put it mildly.

All true. But that argument doesn't fully account for two aspects of the "Kaus theorem" (not to be confused with the Coase theorem):
1) With the high stakes of each race, the parties now have added incentive to try harder than ever to force hopeless candidates out of the race. While it is true that such withdrawals have been historically rare, I'm not sure the numbers wouldn't rise if parties tried really hard.
While parties may have been stronger in the past, they may have been less willing to pressure candidates to withdraw, because...
2) Campaigning and predicting elections were much less scientific in the past. By contrast, today's polling & focus-grouping - maligned by many - actually does its job fairly well. Kaus' strongest point is his argument that the obsessive use of polling and focus groups are a major factor in creating today's political equilibrium.
In the past, a Bob Torricelli may have been able to convince himself and party leaders into thinking he could overcome the hole he'd dug for himself. (The name "Harry Truman" would probably come up - and as that example shows, it may even be true.) With present polling technology, that scenario seems less applicable. With the better information provided by today's polls, it is arguable that even an elected politician's confidence would be dented, and it seems likely that party leaders would be more confident in telling such a candidate that he was in an unwinnable position. Couple that with the added stakes of each race in an equilibrium, and I think you may see more Torricellis in the years to come. (God help us all.)