On the predictive properties of the Iowa & NH primaries, redux

This is an updating of a post from back in 2007.

Q. What is the importance of winning the Iowa and/or New Hampshire primaries?

Short answer: Winning both makes a candidate the overwhelming favorite for his/her party’s nomination. Winning neither means you’re a longshot. Winning one generally means you’re in the race, but you’ve still got to make something of it from there.

The longer version:

Iowa first joined the New Hampshire primary as a launching pad for Presidential hopes when Jimmy Carter won the Iowa caucuses in 1976, so we’ll look at 1976 through 2012. That gives us 10 nomination cycles for each party, or 20 altogether.

In five of those cycles, the incumbent President didn’t face significant opposition for the nomination. (Dems 1996, 2012; GOP 1984, 1992, 2004.) So that leaves us with 15 cycles. (You can make a case that Buchanan in 1992 constituted a significant challenge to GHW Bush, but it doesn’t really make a difference in how this analysis plays out.) So we’ll leave them out.

Here’s what it looks like (party year: Iowa winner, NH winner, nomination winner bolded):

R 1976: Ford won both
D 1976: Carter won both
D 1980: Carter won both
D 2000: Gore won both
D 2004: Kerry won both

R 1980: Bush, Reagan
D 1984: Mondale, Hart
R 1988: Dole, Bush
D 1988: Gephardt, Dukakis
D 1992: Harkin, Tsongas (Clinton finished second in NH)
R 1996: Dole, Buchanan
R 2000: Bush, McCain
R 2008: Huckabee, McCain
D 2008: Obama, Clinton
R 2012: Santorum*, Romney

The asterisk by Santorum is because Romney was declared the winner at the time, but Santorum eventually turned out to have won. One could argue that effectively Romney won both, and I can go either way on that one.

In every instance where one candidate won both primaries, that candidate ultimately won the nomination. And with the exception of 1992 on the Dem side, the ultimate nominee won at least one of these two primaries. (And 1992 was an unsurprising exception in that there really wasn’t a signifying Iowa primary that year: everyone skipped Iowa, letting favorite son Tom Harkin have an easy but worthless win there.)

The results don’t tell us much, though, about who will win when one candidate wins Iowa and another wins New Hampshire. If there’s any sort of pattern there, I sure don’t see it.

Makes sense. Two very different electorates. If you can appeal to both, you should be able to win anywhere.

The fact that they are very different primary electorates is what makes this interesting, especially on the GOP side. It’s also what’s made me realize I can’t just laugh off Jeb Bush’s potential candidacy. Because even if he is an also-ran in Iowa, he can win NH, and if he does, he’ll have enough money behind him to compete the rest of the way.

A Bush can never be laughed off, I just don’t see it as a win even if he wins. A third Bush Presidency, even if it was a successful Presidency, would mean 20 straight years of Bush Presidents for Republicans and 28 straight years of a Bush being on a winning ticket. That can’t be healthy for a party.

But Bush’s strength is that he appeals to all sides of the party. Even the Tea Partiers will prefer him to Romney or McCain. He actually has the most conservative record of any Bush from his time in Florida.

This. In addition it gives you the big ‘Mo’, and it cuts off the availability of money for the losers. A split doesn’t seem to mean anything.

It’s not always that easy though. It took the Republican electorate awhile to finally get behind Romney. Gingrich won in SC and was briefly the frontrunner.

Well, it’s not a ‘hey, I won these two primaries, the win is automatic’ situation; the IA/NH winner still has to go out and run the rest of the race. But given the normal motivation of a candidate to keep running hard when the nomination’s probably but not definitely his, this seems to be how things can be expected to work out.

Well, being one of the two candidates in that split seems to put you in a much better position than not being one of those two.

But no, it doesn’t tell you which of those two to put your money on. The specifics of a given race are all that can tell you that - if even then. Pat Buchanan’s NH win in 1996 didn’t derail Bob Dole in the least, but when [del]the Empire[/del] Hillary struck back in NH in 2008 after Obama had won Iowa, it was anybody’s game.

Sometimes. In 2004, for reasons that can only be explained by a misguided desire to beat Bush at any cost, the Democrats rallied behind a candidate they weren’t all that enthusiastic about just because he won in Iowa and NH.

How do you determine the “winner” of a caucus? Wasn’t there an argument over whether Clinton or Obama won the 2008 Nevada caucus (something along the lines of, Clinton got more “votes”, if you can call them that in a caucus, but Obama ended up with more delegates)?

Besides, in the end, the winner of the Iowa Republican Caucus was “The RNC Won’t Tell You” (i.e. Ron Paul), with 22 of the 28 delegates.

You’ve criticized Bush severely in your posts here; I assume in an attempt to appear even-handed. In that context, this post makes no sense at all.

I think a key factor is that Iowa and New Hampshire are both small states and their political contests are scheduled when there’s no other significant events going on. I think the theory is these states serve as a means of separating out the wheat from the chaff - if you and your organization can’t produce a good showing in these relatively small arenas then you’re not going to be able to run a full Presidential campaign. The political season goes into Iowa and New Hampshire with a few dozen potential candidates and only a handful emerge as “serious” candidates.

No, I actually don’t like GWB. But picking your candidate for stupid reasons because you’re so deranged with hate for the man is foolish. “OMG! He’s a veteran! Pick the veteran! Bush can’t beat a veteran!”

The mistake with Kerry was in thinking that Republicans want what they say they want, and that we could thus pick off some votes by running a candidate that Republicans say they would like. You like war heroes? OK, we’ll run a war hero. Except it turns out that Republicans don’t like war heroes, after all, and in fact attack them for their heroism.

The problem with Kerry was that the Swift Boat Liars For Bush were successful in tearing down his character. If the “liberal media” had been diligent in fact checking and pointing out their lies, things might have been different.

What Republicans wanted, if they had to choose among the Democratic candidates, was Joe Lieberman. Some might have supported Wes Clark, but Clark got handled to death by Clinton pros, thus destroying the whole point of nominating an outsider candidate.

Kerry was one of the most liberal members of the Senate and never had any chance of being supported by very many Republicans.

The liberal media was diligent in fact checking, too much so, I thought. It was the liberal media that nitpicked some of their allegations by saying “they didn’t serve on Kerry’s boat”, when none of the Swifties ever claimed to have done so in the first place.

The liberal media did their job.

Naw, what you get are the talking heads wagging their fingers and saying things like “I think it’s terrible when either side makes up lies about the other.”

There’s some of that, but the media actually destroyed the Swifties, and despite the target rich environment, went even further than they had to, resorting to disingenuosness themselves.

In the end, the Swifties did not influence the election. They convinced no one who wasn’t already a Bush supporter.

George F*. Will has a column up today titled “The most important state in the 2016 primary,” which he claims is South Carolina. His logic:

Nice sinecure you’ve got there, George. You can publish any idiocy you want twice a week, and no matter how dumb it is, you’ll never get fired.

But since you raised the issue, let’s see what a great job the South Carolina primary has done in predicting the GOP nominee. ‘Every year since 1980, except 2012.’ Let’s see what we have here.

  1. Using 1980 as the starting point is fair: there was no South Carolina GOP primary in 1976. (Good thing for Will’s argument that that’s so: since Reagan won both NC and GA in 1976, he almost surely would have won SC as well. But we’ll let that go.)

  2. As with Iowa and New Hampshire, we’ll toss out 1984, 1992, and 2004, when the incumbent was a Republican and was either unopposed or ran against trivial opposition.

That leaves 6 data points: 1980, 1988, 1996, 2000, 2008, and 2012. So at best, SC has predicted the winner 5 times out of 6. That’s good, but hardly overwhelming. Flip 6 coins, and you’ll get 5 or more heads 11% of the time.

  1. My recollection is that before 2000, the South Carolina primary was just another primary, and was pretty incidental to the 1980, 1988, and 1996 races. It became a big deal in 2000 when Bush beat McCain there in a famously dirty primary, and reinforced that status when McCain beat Huckabee there in 2008, with an assist from Fred Thompson. But Newt won SC in 2012. So during the time that SC has been seriously contested, it’s picked 2 winners out of 3. Big whoop!

  2. Still, one can say that it could have predicted the winner in earlier years, even if nobody was paying much attention. In 1980, Reagan hadn’t yet nailed down the nomination when he won SC, so you can say the SC result predicted the nomination.

But 1988 and 1996 are less clear. In 1996 in particular, Bob Dole’s strongest opposition was from Pat Buchanan and Steve Forbes. Nobody needed the SC result to tell them who was going to win this race.

My recollection from 1988 was that it was pretty clear which way the wind was blowing after [del]the Empire[/del] Bush struck back and won NH after Dole had won Iowa, and the Wikipedia summary of that contest supports that:

But OTOH, Dole won primaries or caucuses in three states (and Bush won two) between NH and SC, but Bush was undefeated from SC on. So I’ll give Will the benefit of the doubt on 1988.

So SC has determined or predicted the winner in 4 of 5 instances. Flip 5 coins, and you’ll get 4 or more heads 18.75% of the time.

And of course, the failed prediction was just last time.

No doubt, the SC primary is important. In this cycle, if it’s still a 3-way contest going into Iowa, then whoever’s the odd man out after IA and NH will really need a win in SC to stay alive. But its record of picking winners is a lot less impressive than that of Iowa and New Hampshire.
*Stands for ‘fuckhead,’ I believe.

Seems like an apropos time to bump this thread, given that today’s primary day in New Hampshire.

On the Dem side, it looks like the pattern’s a near-certainty to hold, given that Hillary won Iowa, Bernie’s almost certain to win NH, and they’re the only two left.

And despite the reality that this year’s GOP race is much more unpredictable than usual, I really think the pattern will hold on that side as well. Cruz won Iowa, somebody will win NH tonight, and either Cruz or the NH winner will be the nominee.

If it turns out I’m wrong, feel free to bump this thread at the appropriate time and [del]taunt[/del] remind me. :slight_smile: