A recent post by Stratfor intelligently discusses the dynamics of US presidential elections — why they tend to be so even, why low voter turnout does not seem a good indicator of voter alienation, that big [60%+] wins (Harding 1920, FDR 1936, LBJ 1964, Nixon 1972) have not generally led to historically well-regarded presidencies.
But it misses out on why two-Party system tends to be so even in voter supporter–such that a 55-45 result is a “landslide”. It is a result of the dynamics of coalition building.
For, in any political system dominated by two Parties, even a unitary “Major Party” is, in reality, a coalition. Not only a coalition of various shades of political ideology among its activists (who may or may not form as overt factions) but a coalition of voters. The level and strength of support within the various elements of said coalition for a specific Major Party will vary and will be subject to different “triggers” of support or opposition. The weaker that support, the more “floating” considerations — such as apparent competence — will become dominant.
As long as enough of the electorate is open to persuasion, trying to build a majority coalition will be somewhat like herding cats. Just as you apparently have one in place, another will wander off. What appeals to one group may be a matter of indifference, or even hostility, to another. And your bids for support are subject to continual competition from your opponents.
Being in office is an advantage, in that one can “deliver”, but also something of a disadvantage, in that you are responsible for policy outcomes, so your chances of alienating people tend to accumulate over time. Hence a tendency for the electoral support of Governments to decline over time; for being in office to be a wasting political asset.
So, given a certain political competence and openness, there will be a constant tendency for support for each “side” of politics to converge on about 50% of the vote. This is related to the median voter theorem, but one should be wary of seeing the politics as being captured along a single dimension or axes.
Moreover, such convergence is what one wants. If politics is divided into a permanent majority and minority (as Ulster politics is, for example), the level of alienation among the permanently excluded is likely to be dangerously high. One wants as large a section of the electorate as practical feel that “their side” (someone they voted for) will be in office at least some of the time. That concerns of folk like them get a say.
Which is one of the advantages of the Australian system, in that the proportional representation Senate gives even supporters of Parties not up for gaining a House of Representative majority a vehicle for having some influence without causing political instability.
If one sees the Major Parties as competitors seeking to build a voter-majority across disparate social groups who nevertheless all feel that they are in the same political “game”, then the tendency of electoral politics in Two-Party systems to result in remarkably persistent even divisions of total votes is neither mysterious nor problematic.