Predicting the outcome of the Iowa caucuses is challenging enough. Six different Republican candidates — Michele Bachmann, Herman Cain, Newt Gingrich, Ron Paul, Rick Perry and Mitt Romney — have led at least one poll of the state at some point in this cycle. A seventh, Rick Santorum, is closing fast in the polls and has a realistic chance to win on Tuesday night.
What may be even more challenging is predicting how the results of the caucuses will reverberate throughout New Hampshire and the other states. As the political scientist Jonathan Bernstein notes, and as my research has found, performance relative to expectations can matter almost as much as the order of finish. The Iowa caucuses are a two-step process: first comes the voting, then comes the spinning.
Critics of the Iowa caucuses sometimes take the argument one step too far, arguing that they are entirely a media-driven obsession with little practical importance. It is true that no delegates will be chosen on Tuesday night, and it is true that the Iowa caucuses have far from a perfect predictive track record, especially on the Republican side.
The caucuses provide important evidence, however, as to how voters behave in the flesh and blood after exposure to the campaigns — rather than in the abstraction of a poll. It is not coincidental, for instance, that Newt Gingrich is expected to perform much worse in Iowa than implied by his national polling numbers. Iowans have seen the commercials that target Mr. Gingrich for his ties to Freddie Mac and for his various deviations from the conservative orthodoxy. Voters in the rest of the country (unless they are political junkies) have not. But they will see those ads once the campaign comes around to their states, and if they have proved to be effective in Iowa, they might be expected to have similar effects.
Still, the post-caucus spin plays a crucial function in mediating the relationship between a political party and its voters. The equation for determining spin is complex. Come-from-behind victories, like John Kerry’s in 2004, can produce more momentum than those that had seemed preordained all along. Second-place finishes, like Gary Hart’s in 1984, can sometimes produce a bigger impact than finishing first if the candidate performed better than his polling had projected.
The critical thing to keep in mind is that a wide range of outcomes remain possible in Iowa even at this late hour. Our forecasts, which are derived from the polling but account for its historical accuracy or lack thereof, project for example that Mr. Romney is most likely to receive about 22 percent of the vote. However, the margin of error on Mr. Romney’s forecast (enough to cover 90 percent of all possible outcomes) is about plus or minus 10 points, meaning that Mr. Romney could plausibly finish with as much as 32 percent of the vote or as little as 12 percent.
The campaigns and the news media may be used to looking at general election polling, which is much more reliable, and will usually be taken by surprise at these outcomes, even though they occur with some frequency. Iowa (and other early-voting states like New Hampshire) can therefore be much more consequential when things don’t go to plan than when they do.
The following, therefore, represents an attempt to consider how each candidate’s range of possible outcomes might play out in terms of the post-caucus narrative — and then in the subsequent early-voting states. The figures associated with each candidate represent the 90 percent confidence intervals from the FiveThirtyEight forecast model.
538 forecast (most likely outcome): 22 percent
High end of forecast range: 32 percent
Low end of forecast range: 12 percent
By far the most consequential outcome in Iowa would be if Mr. Romney performed much worse than expectations. In practice, this would probably mean if Mr. Romney finished with about 15 percent of the vote or worse, which would most likely leave him no better than a distant third place, and would risk his finishing behind supposedly flailing candidates like Mr. Gingrich.
Mr. Romney’s campaign had once done a very good job of managing expectations in Iowa. But that has been less true in recent days, following various boastful statements that his campaign has made on and off the record.
Meanwhile, the most high-profile poll in the state, that conducted by The Des Moines Register, had Mr. Romney ahead. That candidates like Mr. Paul have led in other recent polls of the state is of consequence for our forecasts, but may not matter as far as expectations go, since The Des Moines Register poll has received much more news media attention.
None of that will matter if Mr. Romney lives up to expectations. But if he unambiguously fails to achieve them — and candidates with his profile sometimes have — it may be taken as a sign that voters have vetoed the conclusion of the “invisible primary,” the phase of the campaign when the candidates seek to position themselves for institutional and financial support. That could turn the race into a free-for-all.
At the other end of the scale, if Mr. Romney romps to an easy victory — say he receives almost 30 percent of the vote — the conclusion may be that he has all but locked up the nomination, particularly given his strength in New Hampshire. True, the voters in the other 49 states will still have something to say about that. But if Mr. Romney has performed impressively in Iowa, which has among the most conservative Republican electorates, it would augur well for how he should finish elsewhere in the country.
Then there are the in-between cases, where the top two or three candidates finish close to one another as is currently projected by the polling. I am not quite ready to say that a close second-place finish for Mr. Romney would be exactly the same as a close first-place finish — it will affect the tenor of the coverage. But either would leave Mr. Romney in a position where he was poised to win New Hampshire. It might become more important how the rest of the candidates performed, particularly insofar as it positioned them for South Carolina.
Finally, there is the case in which Mr. Romney performs slightly below expectations but not disastrously so — say, he finishes in third place but with a reasonably healthy total like 18 percent of the vote. In this case, it would be harder to dismiss the importance of New Hampshire. But it might not be enough to make Mr. Romney truly vulnerable there. This is the perversity of the expectations game: by failing to achieve expectations in one state, a candidate may lower them in another.
538 forecast (most likely outcome): 21 percent
High end of forecast range: 31 percent
Low end of forecast range: 11 percent
The actual voting total achieved by Mr. Paul may be less important than that of most other candidates for purposes of narrative and spin. It is not clear that many of Mr. Paul’s voters are open to persuasion (especially from the mainstream media). Nor are very many voters who are not currently supporting Mr. Paul likely to come around to him.
Still, the caucuses will be an important test of Mr. Paul’s organizational strength — and his ability to expand the electorate to groups like young voters and independents who are more inclined to support him.
A truly impressive finish — say, if Mr. Paul finished with close to 30 percent of the vote — might create some momentum for him to the extent that it mobilized new voters. That might give him some chance of winning New Hampshire, an idiosyncratic state where there are a large number of independents. But his campaign would survive a poor finish as well and could continue to play a spoiler role.
538 forecast (most likely outcome): 19 percent
High end of forecast range: 29 percent
Low end of forecast range: 10 percent
Expectations are formed not just in the days immediately before the caucuses but also in the weeks and months that precede them. Since Mr. Santorum has already made a tremendous amount of progress relative to early expectations, even something like a third-place finish in which he got 15 percent of the vote might yield its share of favorable headlines. More important is that Mr. Santorum finishes ahead of other conservative candidates like Mr. Gingrich and Mr. Perry.
Still, the stronger Mr. Santorum’s performance, the more credible his claim to being the one and only “anti-Romney” candidate. Moreover, Mr. Santorum could use the earned media from a strong finish more than most of his opponents, since his campaign has little financial or organizational strength beyond Iowa. Positive momentum from the caucuses, particularly in the event of a clear first-place finish, would bide Mr. Santorum time, giving him the chance to build a more robust campaign operation before South Carolina and Florida. His near-term objective would be finishing second or a strong third in New Hampshire, which might require as little as 15 percent of the vote there.
538 forecast (most likely outcome): 15 percent
High end of forecast range: 24 percent
Low end of forecast range: 7 percent
It is perhaps hardest to anticipate how the news media will react to Mr. Gingrich’s performance. Given that Mr. Gingrich led the Iowa and national polls mere weeks ago, the post-mortems for his campaign have literally already been written, with the news media depicting him as an Icarus-like figure who fell back to earth on the basis of a flimsy foundation.
Would, say, a fourth-place finish in Iowa be regarded as the denouement to the story? Or are expectations for Mr. Gingrich already so low that it might be regarded as redeemable?
One factor that might help Mr. Gingrich is that he can claim to be viable in South Carolina and Florida, where his polling was once extremely strong. This claim may be dubious since few polls of these states have been conducted recently. But if Mr. Santorum’s finish is underwhelming, there may be doubts as to which candidate can challenge Mr. Romney in the South, which could help Mr. Gingrich to continue along.
What is clearer is that if Mr. Gingrich unambiguously beats expectations by finishing in the top three, he might have some real upside. If the deluge of advertising against him in Iowa was not enough to stop him, he might earn a reputation for resilience. Such a finish might also call into question the power of the party establishment, which has gone to great lengths to stop him.
Mr. Gingrich will be done absolutely no favors by either the party establishment or the news media, however, with a fifth-place or worse finish.
538 forecast (most likely outcome): 10 percent
High end of forecast range: 18 percent
Low end of forecast range: 4 percent
Mr. Perry would have fewer excuses than Mr. Gingrich for a poor finish in Iowa; his campaign has spent a great deal of money there, and Republican caucus-goers have historically been kind to Southerners.
If Mr. Perry’s position were stronger in the other states, he might be able to excuse a poor finish in Iowa because of his late start, arguing that he was among the only nationally viable alternatives to Mr. Romney. However, that claim will now appear questionable given that Mr. Perry’s polling is weak even in South Carolina, and that (like Mr. Gingrich) he was sufficiently disorganized that he failed to qualify for the Virginia ballot.
At an absolute minimum, Mr. Perry will need to beat his polling average and finish ahead of Mr. Gingrich to get any favorable spin out of Iowa. A finish in the mid-to-high teens, meanwhile, which remains mathematically possible although not likely, could revitalize his campaign.
But if he finishes at or below his polling average, Mr. Perry is perhaps the most likely candidate to drop out after Iowa. This is particularly so given his opportunity cost — Mr. Perry has a very good job down in Austin.
538 forecast (most likely outcome): 8 percent
High end of forecast range: 15 percent
Low end of forecast range: 2 percent
Although Mrs. Bachmann’s polling has sometimes been decent in South Carolina, Iowa has long been perceived as her strongest state, particularly given that she won the Ames straw poll and that she once led the polling there. I don’t know whether or not she will decide to continue in the race, but unless she finishes toward the highest end of her forecast range, she is unlikely to exert much influence upon it.
Jon M. Huntsman Jr.
538 forecast (most likely outcome): 4 percent
High end of forecast range: 8 percent
Low end of forecast range: 0 percent
Mr. Huntsman abandoned Iowa long ago and has had little presence in the state. There is an outside possibility that he could finish ahead of a candidate like Mrs. Bachmann, which could garner him a favorable headline or two. But mostly Mr. Huntsman has a rooting interest in seeing how the other candidates perform.
In particular, Mr. Huntsman might be hoping for a highly ambiguous finish, especially an effective three-way tie as is projected by the current polling, and which would leave no candidate with demonstrable momentum. That would free up news bandwidth for him in New Hampshire, where his polling is stronger but where he will have to compete with several other candidates for attention. The less news coming out of Iowa, the more time the news media will have to speculate about whether it is finally Mr. Huntsman’s turn to surge.