The COVID-AM blog is a partnership between the UMI 3157 iGLOBES and the Institut des Amériques, coordinated by François-Michel Le Tourneau, Deputy Director and Marion Magnan, researcher at the Institute. About the blog.

the census in the united states: the poor tallies of the pandemic

October 16, 2020


by François-Michel Le Tourneau, Deputy Director of iGLOBES and Senior Researcher at the French National Center for Scientific Research (CNRS)






The United States does not have a national registry for births, marriages and deaths.They use the census data collected every ten years to determine the allocations given to each state, whether it be at the political or the budget level. But the 2020 edition had to confront both political pressures and the pandemic. The consequences are likely to be huge and they will be felt in the next ten years to come.

Official icone of the 2020 census
Official icone of the 2020 census



The Constitution of the United States stipulates that a census must be carried out every ten years, and that the allocation of seats at the House of Representatives (the total number was definitively set at 435 in 1911) is changed in accordance with the results. The ten-year census is therefore much more than an administrative exercise. It has significant medium to long term repercussions on national political parity. For example, California had two representatives in 1849, 20 in the 30s and 53 today...


The census is also used to restructure the districts that attend to the election of national and local representatives. Usually done by bipartisan commissions, these restructuring operations (gerrymandering) are always extremely contentious. Specialists in statistics from each party closely study the census data in order to create constituencies that are favorable to their side, for example by creating districts with a large Black majority (since the census also records the skin color and the ethnic background) for the Democrats and essentially rural for the Republicans.

Official presentation of the census by the U.S. Department of Commerce that the Census Bureau reports to.
Official presentation of the census by the U.S. Department of Commerce that the Census Bureau reports to.

But the absence of a national registry means that the census is also heavily used for putting public policies in place. Resources for emergency services, the police, firemen, medical infrastructures, schools, the infamous Medicaid program, food aid programs for families and even construction projects for federal roads directly depend more or less on the census. In 2017, a study indicated they had compiled 132 programs that distributed over 675 billion dollars each year based on the census. For 2020, some commentators give a total at stake of over one and a half trillion dollars annually. For some of the programs in question, and because of the large interval between censuses, the allotments are reviewed on the basis of annual estimations of the population. But these estimations are based on birth, death and general data rate projections...mostly taken from the ten-year census.


So the role of the census is vital for the federated states, but also for cities and counties which depend on these resources to meet the expectations and needs of their population. One important detail that reminds us that the US is a nation of immigrants, the census (and the number of representatives, and all the policies that stem from it) counts all people present in the US, not just the citizens.




Several factors contributed to making the 2020 edition particularly difficult. First of all, though tradition requires that the census count everybody present in the United States (including yours truly), irrespective of their status, the Trump administration has tried to use it as an argument for its hard line on immigration. Under the guise of wanting to have a better view on illegal immigration, they wanted to add a question that asked whether the filer was an American citizen or not. This add was largely condemned as discriminatory: illegal immigrants might not fill out the census for fear that it might, in the end, help locate them and get them kicked out (even if the census answers are supposed to be anonymous). However, in August 2019, federal judges prevented the administration from adding the question, which is therefore not on the 2020 questionnaire. This led to a debate which, in and of itself, may have been a deterrent (of note the video by the Census Bureau above made sure to point out that federal agencies cannot use the data from the questionnaires to identify people...).

The pandemic has certainly not simplified the census data collection (Source: image from the website
The pandemic has certainly not simplified the census data collection (Source: image from the website

And yet, whether they are illegal or not, immigrants contribute to the needs of the states with respect to hospitals, police, social services, schools and, as mentionned, they are theoretically taken into account in the allocation of seats at the House and distribution of resources...And this is really where the current government had an ulterior motivation. First, deprive of federal resources municipalities and states who have large illegal populations, which are often by coincidence Democratic....A sort of reaction to the sanctuary cities that were put in place, areas where the local police does not cooperate with federal agencies fighting against illegal immigration. Then, avoid reinforcing or stabilizing the strength of some states which have a lot of immigrants, for example the insufferable Democratic State of California.


The second hit received by the census was of course the Covid-19 pandemic. The census is done in several ways: on-line, via call centers and by mail, but also door-to-door. This last method is particularly important to roster the marginal populations (the homeless, popular neighborhoods, isolated rural zones, Native American reservations...) because they are the ones who are the least able to use the other methods, either because of a digital divide, or simply because they don't have access to the census information, or lastly because they don't speak any English or not well (the on-line version of the census does let you fill it out in 13 languages, including Chinese, Haitian Creole, Polish and Tagalog). The direct collection of information was limited because of the lockdown and the restrictions put in place in different states.


The agency responsable for the census, the Census Bureau, hoped to compensate these difficulties by extending the deadline. But the Trump administration yet again intervened to, on the contrary, close it down sooner. Despite appeals and judgments in favor of extending the deadline, the Supreme Court ruled on October 13, 2020 that the administration was in its right on the subject. Instead of October 31, the count would be stopped on the 15th (Hawaii time zone). 


Why this obsession to shorten the deadline? While everyone is expecting that the census will not be as complete as the preceding ones, the government wants to have enough time to approve the results before the potential arrival of a Democratic administration. This because, as we've seen, the biases of an incomplete census could be an advantage to Republicans.




The census difficulties are going to have ensuing consequences. Lets take for example college towns where students living on campus (or in private university and Greek housing close by) constitute a large portion of their population. A number of students went back home in the Spring of 2020 because of the lockdown and the switch to remote learning. The census recommendation is that they declare "their usual residence", but this is likely to be fuzzy in their eyes since they don't know whether they will have in-person classes, or when, or whether the economic crisis will allow them to continue their university studies. So, many college towns could see their official population decrease significantly, and because of that, resources which had been allocated to them decrease drastically, when in a year or two the numbers on campus will be back to normal.

Is the census useful? This pamphlet was created to entice the citizens to fill out the census.
Is the census useful? This pamphlet was created to entice the citizens to fill out the census.

Other consequences have to do, as we've seen, with the allocation of seats at the House of Representatives, but also the restructuring of the districts. On this topic, the profile of the underevaluated populations means that the current situation favors the Republican party which does not want to change this map in order to obviously maintain an overrepresentation by the rural world, or, more generally and to simplify, create an overrepresentation of the white conservative vote compared to the progressive Latino vote (see article on the demographic changes and their electoral implications).


Finally, a bad census will have consequences on research and on the image the US has of itself. Demographers, geographers and politicians depend on this information and their interpretation of it to analyze the dynamics of the country. Admittedly, the censuses are never perfect instruments, but even so a 1% miscalculation on the final result would amount to 3.3 million people unaccounted for, which is far from negligeable, and even less when they are concentrated in certain regions and cities. Furthermore, these same 3.3 million people can represent a much more meaningful proportion of a specific category (the homeless, illegal Latinos, Native Americans), whose increasing numbers could therefore go unnoticed.


The pandemic and the short-term calculations of the Republican administration is therefore likely to cost dearly, not only to cities and Democratic states, but all to the United States as a whole.

François-Michel Le Tourneau is a senior research fellow at the French National Center for Scientific Research (CNRS), Deputy Director of the UMI iGLOBES and a member of the Institut des Amériques Scientific Committee.  His work focuses on settlement and use of sparsely populated areas, especially the Brazilian Amazon. He is particularly interested in indigenous people and traditional populations and their relationships with their territory. He has authored a number of papers in national and international scientific journals (list here on HAL-SHS, here on Researchgate or here on