So says JLF head John Hood in his column today. He also says its not that big a deal either way:

Keep in mind that while North Carolina has no requirement, many other states and countries do. Scholars have studied their effects on voting. Most have found nothing of consequence. In March 2011, for example, State Politics & Policy Quarterly published a study by Oakland University political scientists Roger Larocca and John Klemanski that examined several different election-law changes, including voter ID, same-day voter registration, no-excuse absentee voting and early voting.

Using data from the 2000, 2004, and 2008 presidential elections, they found that same-day registration and no-excuse absentee balloting resulted in higher voter turnout, all other things being held equal. Early voting actually resulted in lower turnout, for reasons best explained and analyzed another day (I find them at least plausible). As for voter ID, the authors wrote, “we find no evidence to suggest that voter identification regulations are associated with lower turnout among any age cohort in any of our three elections.” If anything, there is a small positive effect on turnout.

Those on both sides of the issue ought to ponder the available research. If you think that requiring voters to show identification will disenfranchise large numbers of citizens who would otherwise cast legitimate ballots, the research findings are inconsistent with your thesis. However, if you think that requiring voters to show ID will deter large numbers of non-citizens or felons who would otherwise cast fraudulent ballots, the same research findings are inconsistent with your thesis, as well.

Realistically, voter ID requirements are an insurance policy against the possibility that an extremely close election might be stolen by voter fraud.