How so? Well, to begin with, voter identification laws do not belong to the same moral or legal universe as Jim Crow. Their public purpose, as a curb to fraud, is potentially legitimate rather than nakedly discriminatory, and their effects are relatively limited. As Roberts’s majority opinion noted, the voter registration gap between whites and blacks in George Wallace’s segregationist Alabama was 50 percentage points. When my colleague Nate Silver looked at studies assessing the impact of voter ID laws, he estimated that they tend to reduce turnout by around 2 percent — and that reduction crosses racial lines, rather than affecting African-Americans exclusively.

A 2 percent dip is still enough to influence a close election. But voter ID laws don’t take effect in a vacuum: as they’re debated, passed and contested in court, they shape voter preferences and influence voter enthusiasm in ways that might well outstrip their direct influence on turnout. They inspire registration drives and education efforts; they help activists fund-raise and organize; they raise the specter of past injustices; they reinforce a narrative that their architects are indifferent or hostile to minorities.

This, I suspect, is part of the story of why African-American turnout didn’t fall off as expected between 2008 and 2012. By trying to restrict the franchise on the margins, Republican state legislators handed Democrats a powerful tool for mobilization and persuasion, and motivated voters who might otherwise have lost some of their enthusiasm after the euphoria of “Yes We Can” gave way to the reality of a stagnant, high-unemployment economy.