- Scott Ellington
Numbers are at the heart of electoral politics. From vote totals to money raised to doors knocked on by canvassers, numbers tell us much about what happens in the political world. "By the Numbers" is a new regular web-only feature with a goal of illuminating some of the key numbers in Arkansas politics. Have a comment or have a request for a future piece? Email Jay Barth at firstname.lastname@example.org.
Runoff elections have been around for over a century and are regular features in politics in the South. Most southern states adopted the runoff once one-party Democratic reign was established with the disenfranchisement of African Americans and the demolition of Populist threats to Democrats. Ensuring that the ultimate winners of Democratic primaries (then the only elections that mattered) would have a majority of the vote, Democratic parties in the South created the runoff as a bow to "majority rule" at the same time the majorities of southerners were denied the right to participate in those primaries.
Because runoff elections have now been around for decades, we now have a lot of data on them. Scholars who have studied runoffs find that the leader of the first primary ends up winning the runoff about two-thirds of the time. Moreover, if a candidate gains 40 percent of the first primary vote and has a 5 point lead in the first primary over the candidate they are to face in the runoff, the candidate rarely loses. That is, unless all of their first round opponents coalesce behind the second place finisher or the second place finisher has a decidedly better turnout operation, that "40 + 5" lead is enough to hold off the challenger.
Going into yesterday's runoff elections for the First and Fourth Congressional District's Democratic nominations, the first primary leaders — Prosecuting Attorney Scott Ellington in the First and State Senator Gene Jeffress in the Fourth — each were in strong shape according to the "40 + 5" rule of thumb. Yet, Ellington limped across the finish line yesterday barely defeating State Representative Clark Hall despite the fact he'd nearly avoided a runoff altogether three weeks previous. Jeffress, on the other hand, blew by his runoff opponent Q. Byrum Hurst on the way to an easy victory. Why the disparities in outcomes?
In the First District, analysis of the vote returns across the counties of the district, shows that very little changed in terms of the candidates' support across the runoff campaign except that Ellington nicely coalesced the support of third place finisher Gary Latanich across the district. To be more specific, a simple regression analysis shows that nearly 70 percent of the variance in Ellington's vote across the counties of the district yesterday is explained by only two factors: his share of the vote in a county in the May primary and the percentage of the vote received in a county by Latanich. This means that not much happened in terms of electoral dynamics over the past three weeks in that these results were a reiteration of what happened before.
The story of the First District runoff vote that made things close in the end was turnout. Ellington showed an extraordinary inability to get his voters back to the polls for the runoff. This was particularly true in the two large counties — Craighead and Greene — where he ran up huge majorities in the first primary (and large percentages in both elections). In his home county of Craighead, voter turnout dropped by over 70 percent, causing a reduction in Ellington's vote total of 2,944 votes from the first primary. The numbers were even worse in Greene County, where turnout went down by just under 80 percent and Ellington's vote total went down by 2,173. In contrast, while participation dropped in Hall's best counties, it was by much lower amounts. Most importantly, in his home county of Phillips, turnout decreased by only 30 percent costing him 703 votes. This basic pattern held across the district.
While I was bullish on Ellington's general election chances following his surprising showing in the first primary, his inability to put the race away should give Democrats pause. In addition to raising the resources (and getting help from the national party in DC) to present his message against incumbent Rep. Rick Crawford, Ellington has to also build a campaign infrastructure that can get his voters to the polls, particularly in the southern part of the district where he did poorly, in way that didn't happen in the runoff. The clock is ticking on both fundraising and building campaign apparatus.
In the Fourth District, on the other hand, Jeffress strolled to an easy victory. Again, the pattern in terms of county-level returns between the first primary and the runoff were markedly stable showing that very little shook up the race over the last three weeks. Moreover, somewhat surprisingly, Jeffress nicely coalesced the vote gained by third-place finisher DC Morrison who endorsed Jeffress after the primary. A regression analysis shows that the two variables — Jeffress's percentage of the county level vote in the first primary and Morrison's vote by county — explains quite extraordinarily nearly 90 percent in the county variance in yesterday's results. Moreover, while the pattern was not nearly as stark as in the First District, Jeffress had a turnout advantage. Specifically, voters showing up to vote in the competitive state Senate primary between Eddie Cheatham and Gregg Reep covering much of the southeastern part of the district, also voted overwhelmingly for Jeffress; Jeffress gained over 75% in most of those counties.
Jeffress did it the old fashioned way, gaining votes person-by-person. It was enough to pull off a strong showing in a low turnout election. The odds are not good that it will be repeated in the fall. The Fourth District contest will be a great experiment in the comparative power of old-style Arkansas politics (a personal campaign where money matters very little, practiced by Jeffress) and a new-style media driven politics practiced by newcomer Tom Cotton. The money is on the future, but it will be fun to watch.