I was always puzzled by why McCain did so badly in Iowa in 2000. The finish order was Bush=1, Forbes=2, Keyes=3, Bauer=4 and McCain was way down there with only 5% of the votes.
Now this seemed crazy. McCain had nationwide popularity in polls, exceeding both Bush and Gore. He was a highly successful politician with a long record, whereas Forbes & Keyes lost every election they ever ran in up to and including the present day (late 2007).
An 8 September 2000 Portrait of America US-wide poll
asked: "Suppose George W. Bush loses the presidential election this year. In
2004, would you vote for Al Gore or John McCain for president?"
Results:
8 September poll 3 March poll
Al Gore 34% Al Gore 37%
John McCain 45 John McCain 48
not sure 21 not sure 16
which is a very wide margin for McCain.
The same pollsters had asked essentially the same question back on 3 March (just not
mentioning Bush,
and about McCain vs Gore in the 2000 election)
and had got essentially the same answers (righthand table).
Since Gore beat Bush in the actual (popular vote)
November election, one presumes by "transitivity" that McCain would have beaten Bush also.
(That conclusion can only be avoided by postulating the presence of a remarkably strong
Condorcet Cycle. We don't have any poll data about a
McCain-Bush head-to-head matchup except among Republicans only, and among them, Bush
defeated McCain both in polls and in the actual primary. McCain's large presumed edge over Bush
would only happen if all voters, not just those in one party, were polled.)
A possible explanation (but which still seems unsatisfactory) is this. McCain did not campaign in Iowa. Then the Iowans figured therefore he had "no chance" to "win Iowa" and hence strategically they felt forced to vote for the top perceived two, namely Bush & Forbes. (Forbes "vastly overperformed the polls" in his own words, presumably exactly because Iowans saw those polls, then voted more strategically in the real election than in those polls, since they now had more information. In plurality voting, voting for anybody besides the top two in the polls, is usually strategically foolish since it just wastes your vote.) OK, sure – but still, only 5% for McCain seems still hard to believe, even with the strategy-distortion.
The true explanation might not only be that, but also the effect of the "20% cutoff rule" in Iowa caucuses. Suppose McCain in a typical district would have got 16% (on average) votes, with a standard deviation of 5. Result: every time McCain gets 19%, that district automatically cuts him down to 0. So McCain's 5% really came from about 25% of the districts giving him 20%, and 75% of the districts being nulled out. So he got 5% instead of the 16% he deserved, in this model. (That – 16% – seems much more believable.)
But from the rest of the country's point of view, all we knew was "McCain got totally hammered, probably it is Bush v Forbes now."
So we thus see that Iowa's 20% cutoff rule can introduce a huge distortion as well as making everything vulnerable to gerrymandering.
Of course I do not know for sure this was the explanation. But it seems plausible that some combination of cutoff-effect and/or strategic-must-vote-for-top-two thinking played some role.
Another distortion caused by Iowa's caucuses is the fact that absentee voting is impossible and forbidden. Thus, Iowans in the army in Afghanistan are unable to vote in 2008.
Was this all a bad thing? Probably. It seems plausible to us that McCain would have been a better choice than Bush for both the Republican party and the USA. What do we mean by that?
It is not our purpose to say that McCain was necessarily better than Bush. We are merely pointing out that it is a plausible possibility. If so, then it might well be that the world and USA suffered tremendous damage due to Bush's election rather than McCain's. Whether or not that is true, our point is that the American people were not given the chance to decide that for themselves, largely because of the huge distortions in Iowa's voting system which made McCain look far worse than he actually was, and knocked him out of the Republican contest in spite of, as we said, having more support in nationwide polls than either Bush or Gore. That isn't democracy, or at least is less-good democracy than one would desire.
First of all, there would not have been any Iowan thinking of the form "since McCain did not campaign here, he has no chance, so I cannot afford to waste my vote on him and must vote Bush or Forbes." Instead, the thinking would have been "I can give McCain whatever score he deserves – including maximal, if McCain is my favorite – and that will in no way hurt Bush in his battle versus Forbes. So I can be honest about my true feelings for McCain."
Second of all, the ultra-distortionary 20% cutoff rule would be gone. Result: true, undistorted, honest results. Almost certainly McCain would then have done better than one or both of the never-won-an-election "candidates" Forbes or Keyes, for sure he would have gotten far more than 5% of the votes, and perhaps even beaten Bush. Then history might have been different, and certainly it would have been less distorted.
Third, if the US national elections had employed range voting, then McCain (or Bush) would not have felt forced to drop out of the race merely because he lost the GOP nomination. In our present ridiculous system in which third-party candidates can never win, and in which a third-party candidacy would merely have split the vote with Bush so they'd both have lost, McCain was forced to drop out. With range voting, "vote splitting" does not exist, and McCain could have run third party if he so desired. Then the US public would have had more than two candidates to choose from, and in fact would have a candidate (McCain) with more popularity in polls than either. Why is it good for the USA to be deprived of having that choice? To us it doesn't seem "good," it seems "insane."