Posted by David O'brien on December 18, 2017
Atlanta’s recent mayoral election has completed its recount. The December 5 runoff election between two city council members saw Keisha Lance Bottoms beat Mary Norwood by just 832 votes (or 0.89 percent of the total votes). Norwood requested a recount (which Georgia law allows second-place finishers to do when the margin is less than 1 percent).
After the recount, Norwood had cut Bottoms’ lead by only 11 votes (0.0119 percent of the vote total), leaving a margin of 821 votes between the candidates. In order to change the outcome in her favor, however, Norwood would have needed a change in 833 votes, or about 0.9 percent.
These results are consistent with FairVote’s data, which show that recounts rarely change the outcome of elections. Of the 27 recounts that occurred between 2000 and 2015, the average margin swing shift only 0.0191 percent, or 282 votes. In Atlanta, a shift of that size would translate to about 18 votes. Of those elections where fewer than one million votes were cast (as was the case in Atlanta), the average margin shift was 0.0389 percent, which would translate to about 36 votes in the Atlanta race.
While the possibility of a legal challenge remains, for the moment the numbers are comfortably on Bottoms’ side.
Finally, while Atlanta’s mayoral recount is largely consistent with FairVote’s data, we should be careful when using that data to draw conclusions about city-wide races in general. The data comes from a survey of statewide recounts and there are often important differences between the those and city-wide elections like Atlanta’s, both in the size of electorates and in how elections are administered, that could sometimes lead to different results.
Recounts are an important tool to confirm the validity of elections, but as Mary Norwood has learned (and as Roy Moore soon might), they rarely lead to the result the challenger is seeking.