Trendspotting: Breaking Down the 2020 Census

NewsWire: 3/18/22

• According to a new report, the 2020 census undercounted black, Hispanic, and Native American residents. It also overcounted white and Asian residents. (The New York Times)
  
• NH: Last week, the Census Bureau released its quality check of the 2020 census. If you saw this news, it was likely under a shocking headline: The census undercounted nearly 19M people.
• If this were true, it would be huge news. But it wasn't true.
• Shortly after, the NYT and other news outlets issued corrections. The census had not undercounted nearly 19M people, but incorrectly counted them in some way.
• What does this mean? According to the Census Bureau report, the 2020 census miscounted 18.8M residents, double-counted around 5.2M people, wrongly included another 2.0M, and missed others entirely. Many of the people whom the census originally failed to count were later added to the totals through imputation--basically, educated guesses for who was living in places that did not respond.
• The imputations, however, did not guess the race or ethnicity of the missing residents. As a result, while the overall population count (331.5M) was largely accurate, many minorities were missed.
• Undercounting minorities is an issue that the census has struggled with for decades. Blacks, Hispanics, and American Indians tend to be undercounted, while non-Hispanic whites and Asians tend to be overcounted.
• In 2020, the discrepancies were worse than usual. The Hispanic population had a net undercount of nearly 5.0% (vs. 1.5% in 2010); blacks, 3.3% (vs. 2.0%); American Indians and Native Alaskans, 5.6% (vs. 4.9%); and those identifying as “some other race,” 4.3% (vs. 1.6%). Non-Hispanic whites had a net overcount of 1.6% (vs. 0.8%), and Asians had a net overcount of 2.6% (vs. 0.08%).
• Undercounts can occur for many reasons and are more likely to take place with more transient or lower-income populations. Renters and young people, for instance, have also been historically undercounted. Overcounts tend to occur with the affluent: When someone has a vacation home and a permanent home address, for example, they can get counted twice.
• The larger-than-usual shortfalls spurred an outcry from civil rights groups. Some threatened litigation. Census data is used to apportion House seats and to distribute federal funding. But IMO, this response feels like political theater. While misidentifying race and ethnicity among large numbers of people is certainly not ideal, the distribution of seats and funds depends mostly on the total population size of a given district--which apparently, the 2020 census did get right. The misidentifying would only impact funding in rare instances where it's allocated specifically based on race or ethnicity.
• The bottom line: Behind the alarmist headlines, this is not much of a story. Yes, there were many undercounts and other errors in the 2020 census--but for most groups, they were not statistically different from those in 2010 and will not result in major political or financial consequences. The most important takeaway is that the overall population count was largely accurate. This, if anything, should be the headline. No one expected this after the unprecedented challenges and disruptions that the 2020 census faced.