By Christopher W. Larimer, University of Northern Iowa Professor of Political Science
In addition to candidate stump speeches, one of the things Iowans have become accustomed to hearing every four years are the criticisms of outsiders regarding the state’s “first-in-the-nation” status. For example, last Tuesday, the New York Times ran an article titled, “Why Almost Nobody Will Defend the Iowa Caucuses.” (The Times also ran an article earlier this summer titled, “Take the Iowa Caucuses. Please.”)
One of the arguments against Iowa’s status is the relatively low participation rate, and indeed this is justified. According to the United States Election Project, just 15 percent of eligible voters participated in the 2016 Iowa Caucuses, compared to 52 percent of eligible voters in the New Hampshire Primary held just 8 days later. (On average, about one-fourth to one-third of party members show up on caucus night.)
But, does low participation mean Iowans are caucusing for “bad” (unelectable) candidates? Put another way, do the relatively few Iowans who show up on caucus night hold unique candidate preferences? Or, as Christopher Hull put it in his book, Grassroots Rules, is Iowa a “cess poll”?
That depends on how you look at the data.
To get at this question, I followed Hull’s analysis of looking at the vote share candidates running for president received in the Iowa Caucuses and the New Hampshire Primary for each presidential cycle going back to 1976. While Hull’s data goes through 2004 (see Chapter 3 and Appendix B of his book), I’ve extended it through 2016.
On the Democratic side, excluding 1992 because of Tom Harkin’s candidacy, the predictive value of what happens in Iowa is quite high. About 73 percent of the variation in candidate support observed in New Hampshire can be predicted by what happens in Iowa. On the Republican side, however, that number is cut in half to just 35 percent. (Even if you exclude John McCain, and only look at the years 1980-2016, it is still just 39 percent.) As careful observers of the caucuses know, the “social conservative” vote in Iowa is much stronger than it is in New Hampshire. If South Carolina were to follow Iowa, rather than New Hampshire, there is a good chance the predictive value on the Republican side would increase.
The critics are right, sequencing matters, but that also means that wherever you start the nominating process, there will likely be criticism.