Why did Dean do so "unexpectedly" badly in the 2004 Iowa Caucus elections?

Vermont Governor Howard Dean appeared from pre-election polls to be the overdog. He had the most popularity, the most money in his kitty, and the most endorsements among all the Democratic contenders. But Dean came in far behind both Kerry (1st) and Edwards (2nd) in the 2004 Iowa Caucuses, and was never able to recover in the race for the Democratic presidential nomination. Kerry and Edwards finished nationwide in the same order as in Iowa and then ran as a President/VP team, losing to the Republican Bush/Cheney team.

Why did Dean finish so far below predictions?

The highly-distorted Iowa caucus voting system undoubtably played a large role. In each Iowa caucus meeting, each candidate with below 20% support is automatically zapped down to zero support, and his supporters then must decide among the remaining candidates. Only candidates with over 20% final support in that meeting have their votes transmitted to Central Headquarters. Actually, the "20%" figure is not fixed and varies from caucus to caucus. I.e. for some it might be 15%. I am simplifying.

This 20% cutoff rule tremendously reduces the vote counts for weak candidates, below what the polls would have predicted. For example, a candidate with 20% support (really) could get "0%" support (final Iowa results).

That is a tremendous distortion. But it is worse than that. Suppose there are 10 candidates. If all were about equal, then they would each have 10% support. That would mean they all would be eliminated with zero votes coming out of that meeting. Well. That is unacceptable. So in practice, caucusers then start arguing and making deals to try to build over-20% coalitions. (And obviously that process can be highly manipulated by a skilled agenda setter, etc.) So eventually a few candidates survive this process at each meeting. But the result can be more due to arm-twisting than representative of what the voters actually want. This is more a joke than a democracy. (Also, everybody's votes are non-secret, a violation of what, in the view of many, is a core democratic principle.)

Another huge 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 the 2008 caucuses. Oh, and what if you are the kond of person who doesn't have hours to waste and wants to vote in only a minute like in other states? Sorry. This caucus system is an absolute outrage.

But it is worse than that. The 20% cutoff rule makes Iowa vulnerable to gerrymandering. Suppose Iowa's gerrymandering of election districts "packs and cracks" the Dean supporters into Dean-rich and Dean-poor districts. In that case, Dean could easily have, say, 30% support but solely in the form of 19% districts (nulled out) and 74% districts (in a 4:1 ratio). The result would be a huge distortion due to a combination of the gerrymandering and the 20% cutoff rule – this particular scenario would give Dean under 15%, not 30%, in the final count.

I am not saying I am confident this was the cause of Dean's troubles. There certainly were other causes. I am saying, these distortionary effects must have played a role, and the only question is how large a role.

Is Iowa gerrymandered? Supposedly Iowa has a non-partisan district-drawing process that improves on most of the Nation. But the fact is that the jigsaw-like map of congressional districts looks gerrymandered, although it is not as blatantly obvious as in, say. Arizona. (Arizona's maps were also drawn by a "nonpartisan" commission. Ha! These commissions often are not "non" but rather "bi"partisan and thus draw districts intended to keep everybody from both sides "safe." And indeed the Iowa legislature rejected the first district drawing attempt by the commission, and then after accepting the second try enjoyed a 98% re-election rate.) But when you examine maps of state-house districts gerrymandering seems clearer: there are numerous fractal-shaped districts with no explanation that I can imagine for their weird shapes, other than gerrymandering. So I believe there is some gerrymandering going on, although my eyeballs say Iowa seems comparatively good in that respect.

Return to main page