A Recount Probably Won’t Change The Winner Of The North Carolina Governor’s Race

The number of recounts (or requested recounts) of results from this month’s elections keeps climbing — but none of the affected races are close enough that a recount is likely to change the result.

In the highest-profile race that hasn’t been conceded, North Carolina Gov. Pat McCrory, a Republican, is requesting a recount of the state’s gubernatorial votes. McCrory’s Democratic opponent, Roy Cooper, has claimed victory. But McCrory has not given in, and his allies have challenged the results in dozens of counties. The Republican candidate in North Carolina’s auditor race also plans to seek a recount, according to the executive director of the state’s Republican Party. And Roque “Rocky” De La Fuente, who ran for president in some states, including Nevada, is filing for a recount there. Add it to the three states where Green Party presidential candidate Jill Stein is seeking recounts in the presidential election: Michigan, Pennsylvania and Wisconsin.

In all six of these races, the leading candidate’s margin over the apparent runner-up is far bigger than the typical amount by which statewide recounts swing races. In the closest race, the North Carolina state auditor’s race, Beth Wood leads Chuck Stuber by more than 5,000 votes. In the race with the widest margin, Donald Trump leads Hillary Clinton by more than 70,000 votes in Pennsylvania. But none of the 27 statewide recounts since 2000 has changed the margin in a race by even 2,000 votes, according to data compiled by FairVote, a nonpartisan group that researches elections and advocates for changes to the electoral system, for a report that it published on Wednesday. And recounts don’t always tighten races; more than half the time, the leader’s margin grew as a result of the recount. (See data on all 27 recounts at the end of the post.)

On a percentage basis, too, the recounts look unlikely to change the outcomes. The average recount since 2000 changed the margin by 0.02 percentage points; in the closest of the six 2016 races that might be recounted, the two top candidates are separated by 0.13 percentage points, a far wider margin.¹

Some of those calling for recounts are claiming that this year’s elections weren’t typical. If that’s the case, typical recount results might not apply. For instance, if voting machines were hacked or if large numbers of votes were cast by people who were ineligible (either because of felony convictions or because they cast early ballots and then died before Election Day), vote totals could have been affected by more than the amount that a typical recount uncovers. But there’s no evidence that any of that happened. And it’s not clear what recounts would uncover. Recount processes vary by state, and recounts often fall short of what advocates of full-scale, routine audits would like to see happen to ensure that only eligible voters participate and that their votes are counted accurately. Meanwhile, fewer than half of respondents to presidential exit polls said they were “very confident” in the vote count, and that was before President-elect Donald Trump claimed, with no evidence, that millions of people “voted illegally” for Hillary Clinton.