In 1960, the American South mostly voted for Democratic candidates and was solidly Democratic at the congressional level. By 2000, it was solidly Republican in presidential politics and mostly Republican at the congressional level. The transformation, clearly, had something to do with race and the Civil Rights Movement. But a provocative new paper suggests that Ku Klux Klan activity — as opposed to the broader phenomenon of racism — played a small but meaningful role in the process. Counties with more Klan activity saw a more dramatic shift toward the GOP, a shift that stands up to a range of reasonable statistical controls and suggests that the Klan was really a difference maker.
The evidence for the Klan's causal role
Notre Dame's Rory McVeigh, Brandeis's David Cunningham, and Yale's Justin Farrell looked at county-level presidential voting data from 1960 to 2000 in ten southern states, and coded each county based on whether a Klan chapter was established there between 1964 and 1966, when the organization was growing in response to the Johnson administration and Warren Court's increased vigilance on civil rights. Here are the counties identified:
(McVeigh et al, 2014)
The researchers wanted to capture the role that the Klan organizations themselves played, which is a tricky methodological problem to solve. There are a lot of other reasons why some counties might have seen more growth in support for Republicans, some of which could also explain the presence of a Klan chapter. For example, maybe a county that's just particularly racist both was fertile ground for the Klan and was less likely to vote for Democrats when they started backing civil rights — both had the same cause, but the Klan chapter didn't cause the increase in Republican support.
To take into account underlying racial animus, McVeigh et al controlled for the counties' levels of support for George Wallace, who made an explicitly segregationist third-party presidential bid in 1968. They also controlled for the level of support for Goldwater; his run galvanized Southern support for Republicans before the 1964-66 growth in Klan chapters and so is another factor that needs to be taken into account. Finally, the study controlled for changes in counties' racial makeup (counties whose black populations grew presumably saw less growth in Republican support), their level of economic prosperity and education (since Republican support is positively correlated with income and education), and whether the county had an NAACP chapter (which could have provoked the KKK to launch a branch in response).
They conclude that having a Klan chapter present was associated with a 2 percent bigger increase in Republican support from 1960 to 1972, a 3.7 percent bigger increase from 1960 to 1980, a 4.9 percent bigger increase from 1960 to 1992, and a 3.4 percent bigger increase from 1960 to 2000.How could this have worked?
The Klan used to ally with Democrats
Theodore Bilbo, perhaps the single most noxiously racist public official of the post-Civil War era. (Library of Congress)
In the midcentury South, the organized political expression of white supremacist politics was the Democratic Party. Indeed, number of prominent Democratic politicians — including Senate Majority Leader Robert Byrd, Supreme Court justice and Senator Hugo Black, and Mississippi governor and Senator Theodore Bilbo — were members of the Klan. But in the course of the 1960s, the northern wing of the Democrats joined with Republican elected officials (almost all of them northern) to pass the Civil Rights and Voting Rights Acts.
After that, southern presidential politics rapidly re-aligned with newly enfranchised black voters supporting Democrats and most whites voting for GOP candidates. If the Klan was successful in suppressing African-American turnout or in pulling white people into the electoral process, that would boost the fortunes of Republican candidates.
And at least in some cases, the Klan actively supported Republican candidates. "Certainly, generating support for specific Republican presidential candidates or the Republican Party in general was not a primary goal of the Klan," the authors write, but "while the Klan was perhaps best known for its violent tactics in the 1960s, the movement did invest significant energy in attempting to influence voting outcomes … Klan members advocated for Goldwater’s Republican candidacy in 1964 while incessantly criticizing Democratic incumbents’ intensifying support for civil rights."
Are we sure?
Even the best statistical methodology in the world can't control for all relevant factors, and it's totally possible that McVeigh and his coauthors missed something that explains why Republican performance would be higher in county states without the Klan itself having an effect. This is just one study, and shouldn't be treated as settled fact. But the fact that the effects hold up against a litany of controls ,and are actually stronger in 1992 and 2000 than they were in 1972, is nonetheless striking.
More from vox.com: