When the U.S. Census Bureau counted the state's largest county, it might have missed by a figure comparable to the population of Savannah, Fulton County demographers have suggested.
If that's what happened, Fulton would needlessly lose hundreds of millions of dollars in federal funding and have diminished representation at the state capital.
The 2010 census figure came in 126,635 below the Fulton Information Technology Department's population calculation -- a 14-percent discrepancy. For Atlanta, the number was 118,460 short of demographers' forecast, which was 28 percent off.
Fulton Commission Chairman John Eaves and Atlanta Mayor Kasim Reed plan to formally protest the census numbers, and staffers for both governments were poring over the data tract by tract to determine if census takers were amiss.
“The city of Atlanta showed some growth, but less than expected based on the census bureau's own projections over the past five years," Reed said. "Because of this, we plan to challenge these results."
During the Complete Count campaign, census officials estimated that each uncounted person would cost the county $1,700 per year in federal appropriations, including Community Development Block Grant funds and funds for schools, transit, public safety and aging services. With suburban and outlying counties outpacing them in growth, Fulton and DeKalb stand to lose seats in the state Legislature.
DeKalb County had a census population figure of 691,893, which was 58,000 fewer people than projected, and also is considering a challenge.
Fulton demographer Sandy Speer puts the county population at 1,047,216, a figure based on livable housing units, the average household size, vacancy rates, census estimates and estimates at military barracks, jails, dormitories and homeless shelters. The census bureau's most recent projection was slightly higher at 1,048,872.
The official 2010 census count, however, put the population at 920,581.
"We have, what we thought, were rock solid numbers," Speer said. "What came out were numbers that were shockingly low."
What makes the data suspect, he said, is that it lists Fulton as having 437,105 housing units. Tax assessors office records show 471,348 housing units.
Atlanta demographers found similar discrepancies. They forecast 538,460 people in the 2010 count, but the official number came back 420,003.
Based on that, Atlanta grew by only 3,529 people between 2000 and 2010, despite adding 37,648 new household units. City officials say the bureau's figures are off by 10,000 to 14,000.
Several things could have gone wrong, Speer said. Newly constructed houses and housing complexes could have been overlooked. Speer's department found large numbers of housing units missing from Atlanta and College Park in the census data.
With the high rate of foreclosures in Atlanta and south Fulton, Speer said, houses might have been empty and former occupants went uncounted, even if still living in the area with friends or family. The bureau also could have been shut out of gated communities or secured condo complexes.
Census officials doubt the oversights occurred. Wes Flack, Geographic Coordinator for the bureau's Atlanta Regional Office, said the discrepancies were more likely the result of the census bureau raising its Fulton estimates in the middle of the last decade, after the county, Atlanta, Roswell, East Point and Mountain Park mounted challenges.
"There are some people right now, I think, that are questioning whether those challenges to the population estimates might have significantly inflated the estimate," Flack said.
The bureau will start taking challenges to its 2010 count on June 1 through its Count Question Resolution program. Only alleged errors involving inaccurate boundaries, housing units with incorrect jurisdiction codes, units counted more than once or units left out because of processing mistakes will be considered.
Jerome McKibben, a senior demographer researcher who compiled the numbers for Atlanta, said challenging is “very specific and cumbersome.”
The city would have to challenge each individual census block unit it feels is in error, and there are between 30,000 and 40,000 blocks.
“You can’t cover the whole city,” McKibben said. “You are looking for problem areas.”
About the Author
