Census undercounts mean less Medicaid money for most Southern states

Census undercounts will mean a loss of Medicaid funding for most Southern states. (Photo by Thomas Cizauskas via Flickr.)

The 2020 census faced an inordinate number of challenges, including delays caused by the coronavirus pandemic and multiple failed attempts by the Trump administration to add a citizenship question to the decennial survey. Demographers and policy analysts worried that the United States population would be significantly undercounted as a result.

But nationwide, the undercount has been estimated at only 0.5%, according to an Urban Institute report released this month — far less than many feared.

However, seven states had undercounts estimated at more than 1% of their population: Alaska, Georgia, Louisiana, Mississippi, New Mexico, New York, and Texas. Among the Southern states, the estimated undercounts were greatest in Mississippi, at 1.3%, and Texas, at 1.28%, according to the report, titled "2020 Census: Miscounts and the Fairness of Outcomes."

As a result, Texas is expected to lose out on $247 million in Medicaid funding over the next decade, while Mississippi is expected to lose out on $20 million for its Medicaid program. Other Southern states that will lose out on federal health care funding for the poor due to population undercounts are Alabama ($5 million), Arkansas ($10 million), Florida ($88 million), Georgia ($47 million), Louisiana ($46 million), North Carolina ($24 million), and South Carolina ($16 million), the report found. In considering how undercounts affect federal funding, the Urban Institute focused on Medicaid because it's among the federal programs that use census data most directly to distribute funds.

The reasons for the undercounts in Southern states are complex, according to Diana Elliott, the principal research associate at the Urban Institute's Center on Labor, Human Services, and Population and one of the report's authors.

"It's more that those states have populations that are historically hard to count, and maybe have not invested in get-out-the-count efforts as much as other states," Elliott said. "There's a lot of complexity involved in terms of what makes a population hard to count."

She noted that homeless people are hard to count accurately in the census, along with renters, undocumented people, young children, and people of color in both rural and urban areas. Increasing distrust of the government also contributes to undercounts.

"What you see in the South is that a lot of those factors that make groups and geographies harder to count kind of come together at a confluence in a lot of communities," she said.

Elliott and co-authors Steven Martin, Jessica Shakesprere, and Jessica Kelly used a microsimulation model to estimate a hypothetical full census count of the U.S. population, meaning no one was left out or counted twice. The microsimulation measured fairness, net accuracy, and quality. The Urban Institute model found that the 2020 census most likely had 4.1% omissions (undercounts) and 3.6% erroneous inclusion (overcounts), leading to the overall net undercount of 0.5%.

The Black population nationwide was undercounted by an estimated 2.45% and the Latino or Hispanic population by 2.17%, the study found. Children under 5 were undercounted by an estimated 4.86%, undocumented people by 3.36%, and renters by 2.13%.

Undercounts in these historically hard-to-count groups were expected, as was an overcount of the white population, Elliott said.

But even though they were aware of the risk of an undercount, Texas officials waited until August 2020 — four months after the census count began — to spend $15 million on an ad campaign to encourage participation, the Texas Tribune reported. And Mississippi spent less than $500,000 in 2020 to increase census participation, Mississippi Today reported.

"If states are concerned — and they should be concerned — about having an accurate count, building up networks and thinking about how to message the importance of the census and various Census Bureau efforts, it's really important to do that earlier rather than later," Elliott said.

To improve the 2030 census, the report called for more promotion at state and local levels, along with operational changes and better funding.

"If you want to have a community that gets the right investments of health care or vaccines or testing, those estimates are ultimately derived from census data collection efforts," Elliott said. "It's really important for states and communities to understand that these products and the Census Bureau's work is incredibly important for their communities and their residents."