This is waaaaaay oversold.

The Iowa Democratic Party is very good at getting caucus results out early in the evening.

Also, the networks will be reporting that their entrance polls are not good reflectors of the final results.

We had the exact same speculation about entrance polls in '04, the entrance polls were reported, and then we had the final results early enough in the evening that they were all that mattered. Dean screamed about the results, not the entrance polls.

All that matters is the results. There's a reason John Kerry didn't take the oath of office in January 2005.

We couldn't even get the electoral college scrapped after the mess of the 2000 general election. The American political system is tremendously resistant to change. Even if Ezra's scenario played it, it wouldn't result in any changes in the caucus system itself. At most it would discourage the networks from doing the entrance polls next time.

"Then, Bill Richardson fails to make the 15% threshhold for viability and releases his caucusgoers to Clinton."

I've been wondering for a while why this scenario hasn't been getting more play.

It seems overwhelmingly likely, and quite important.

The only real question is whether or not Richardson caucus-goers will follow the instructions.

For clarification of your Richardson scenario, the 15% threshold is on a precinct-by-precinct basis, right?

Also, in re. the Edwards part of the scenario, no one could logically damn the caucus system without simultaneously damning the electoral college. I'm sure that's not an unpopular opinion around these parts, nor do I think political coverage tends to come with a lot of logic attached, but it's the same principle to a lesser degree of distortion.

This is weird ... last time, they conducted entrance polls, but agreed not to release them until after the networks had projected a winner.

"This is weird ... last time, they conducted entrance polls, but agreed not to release them until after the networks had projected a winner."

Not true.

The entrance polls were reported early in the evening. Everyone understood they might not correlate particularly well with the final results, and so no one treated them particularly seriously.

Ezra, over at Slate Mickey Kaus has linked to the two definitive pieces from the Slate archives about how this exact same situation broke down spectacularly in 1988. It explains how this can go so wrong as to actually report the wrong winner -- resulting that the worng person gets the bounce into NH! Please go read it...

"This is weird ... last time, they conducted entrance polls, but agreed not to release them until after the networks had projected a winner."

And there is no way to "project a winner" to the caucuses.

You just wait until the Iowa Democratic Party tells you who won.

And again, that comes quite early in the evening.

"Mickey Kaus has linked to the two definitive pieces from the Slate archives about how this exact same situation broke down spectacularly in 1988."

But, of course, it wasn't anything close to the "exact same situation" in 1988.

The NES was trying to something utterly different in '88 than what the entrance polls of '04 and '08 are trying to do.

If you've bothered to read Salentan's piece, you shouldn't need it explained to you.

The 11:57 comment is anonymous no more.

Will the networks announce the overall results, or just talk about certain demographics? If the former, then it probably will have some influence over the result. But if the latter, it seems similar to how exit polls are reported in general elections. If I recall correctly, networks do not report on overall exit poll results before the polls close (I remember relying on blogs and Drudge), but they will report exit polls showing a certain candidate doing better than expected with a certain demographic. Without any context about what percent of the electorate that demographic is, it's hard to predict who's actually winning or losing. This gives the networks something to talk about but diminishes the influence over the results. It also minimizes network humiliation, since such information will be as good as any available on demographic breakdown and can't be contradicted by actual results.

Remind me again of how Richardson is supposed to release his caucusgoers to Clinton, in this hypothetical?

Remember, this isn't happening on a statewide or even a countywide basis. It's happening in each of 3,562 precincts, according to Politico.

So this can't happen in real time, that is, Richardson's operatives can't be wheeling and dealing with the reps of other campaigns on some "I'll release my people in Precinct X to you, if you release your people in Precinct Y to me" sort of deal, and have any guarantee that it'll work out.

Second, they're not his people; they're just Iowa residents who came out to caucus, and opted to support Richardson.

Richardson's precinct captains are Richardson people, but how many precincts will he have precinct captains in? Are there enough Dem political volunteers in Iowa for the fourth-place candidate to have 3,500 precinct captains, which would imply that the top three also have that many apiece? That's 14,000 precinct captains, in a caucus were 125,000 is a big turnout.

So my hypothesis is that in many (most?) precincts where Richardson's under 15%, there's no Richardson rep in the room - just a bunch of Iowa Dems who decide on the spot who their second choice is.

Plus, if I'm right about that, Richardson's people can't even deal like that in advance, except in a minority of the precincts: if there's no one there to say "Richardson said we should support X" then how's it gonna happen?

Finally, even if Richardson's people make some deal in advance, will the caucusgoers who made him their first choice listen to his local reps? Why should they do that, rather than make their own decisions on who to throw their support to, if Richardson gets less than 15% in that room?

It just seems to me that the logistical obstacles to such wheeling and dealing are pretty steep.

"Will the networks announce the overall results, or just talk about certain demographics?"

In '04, it was the former. I expect it would be the former in '08 as well.

"If the former, then it probably will have some influence over the result."

How?

So much for 'one person, one vote', if the votes don't count equally across the rural/urban spectrum.

Ezra: wholesale reassessment of whether this caucus system makes any sense at all?

Shorter Ezra: Iowa Caucus's Really Suck.

(and yes, so does the Electoral College).

Iowans are supposed to be level-headed, sane, salt-of-the earth folks. How can they keep this crazy system?

[I admit that if my preferred outcome is attained in IA, I'll be less harsh and shrill, and I'll even take a shower to wash off the DFH image - if that's enough to achieve results.]

"Remind me again of how Richardson is supposed to release his caucusgoers to Clinton"

The same way Kucinich and Edwards made an alliance in '04, by making an announcement on the morning of caucus day.

"Iowans are supposed to be level-headed, sane, salt-of-the earth folks. How can they keep this crazy system?"

The Iowa caucus rules are fucking brilliant on multiple levels.

Most of the folks who don't like the Iowa caucus rules don't understand them.

And I'll feel the same way whether my candidate finishes first or third.

Petey: The Iowa caucus rules are fucking brilliant on multiple levels.

Please explain how FB these rules are to us dimwits who think it is FUBAR.

Here's a take (by me) on a better primary system, at Brian Beutler's place.

Ezra:

I understand that entrance poll results will very likely not accurately predict the outcome of the Iowa caucuses, but so what? The networks have been fielding primary polls all year NONE of which are predictive of the primary results, for the very simple reason that polls taken over the summer are too far from the primaries to be predictive. Yet they still get fielded.

Networks don't field polls to accurately predict the primary outcome. They field polls because poll results garner viewers, and viewers are what they're after.

And when the day comes and the networks DO release the results of the Iowa entrance polls, I'll bet dollars to donuts that you, and the folks at Tapped, and Yglesias, and Kos and Atrios and all the rest will post them. And you'll all put a caveat at the very beginning saying that these polls don't mean anything, and then go on to write an entire post about them as if they did mean something. Because that's what everyone's been doing all year long.

"Please explain how FB these rules are to us dimwits who think it is FUBAR."

The quickie version (and I'm not joking - the non-quickie version would be dissertation length):

1) It greatly increases the megaphone given Democratic activists in multiple ways, thus providing a counterbalance to media, money, and celebrity.

2) It deals well with preventing large multi-candidate fields from distorting the results.

3) It forces consensus candidates to the front, which is why it is so influential in terms of the nomination compared to the Republican caucuses, which don't have the same FB rules.

4) It gives extra strength to regions in a similar manner that the US Constitution does, thus putting the Party in a better general election position.

If you were to dream up the perfect rules from scratch for a first in the nation nomination contest, you'd come up with something astoundingly close to the Iowa Democratic caucus rules.

"2) It deals well with preventing large multi-candidate fields from distorting the results."

For example:

I believe Bill Richardson is explicitly working for the Clinton campaign, and has been from the start. I believe the strategy worked out for him from the beginning was for him to get to the left of the entire field on Iraq, so he'd draw voters who were already lost to Clinton.

But the viability requirements in the FB Iowa caucus rules stop this strategy dead in its tracks.

Sure the Richardson campaign can tell its supporters to caucus for Clinton in the second round after he doesn't hit viability, but many of them likely won't follow such instructions.

To piggyback on something Nate W. said: it's struck me for a long time that the power of the Iowa caucus (and the NH primary) is not merely that it's first, but that, because it's first, it's polled into infinity. The effect is that, even though nothing has "counted", we've had dozens of virtual-IA caucuses already, with Edwards and Clinton and Obama variously reported out as "winners"; same with NH primaries. (And pundits speak of these results not as snapshots, but as facts -- "Obama's now winning in Iowa")

Historically, nothing remotely similar has been given to later states -- though in the current insatiable cycle we've at least had a greater number of polls in the later states (insignificant places like NY, CA, MI, IL) if not anything like the attention given to an IA or NH number.

Does this distort the process, too? The Dem situation in IA is still at least within the ballpark of where it began, but have polls showing Hillary in first place led at least a few Edwards supporters to switch tactically to Obama? And given the free-floating chaos of the GOP scrimmage, you certainly have to wonder how each rise and fall (often based on polling samples or fluke reaction to minor events) is impacting what happens on voting day.

"Remind me again of how Richardson is supposed to release his caucusgoers to Clinton"
The same way Kucinich and Edwards made an alliance in '04, by making an announcement on the morning of caucus day.

Could you elaborate? That means zilch to me. How well did it work?