Vernellia R. Randall, A Colorblind Census in a Racist Society: We Cannot Eliminate Racial Disparities by Refusing to Count Them, Racism.org (September 14, 2026).

 

The Census Bureau is accepting public comments through October 13, 2026, on a proposal that would exclude millions of immigrant residents from the population used for congressional apportionment and remove race and ethnicity questions from the decennial census questionnaire. This proposal would weaken political representation, resource allocation, civil-rights enforcement, public-health research, redistricting, and our ability to document racial inequality. Public opposition must become part of the administrative record before the government portrays the deliberate suppression of essential racial data as a neutral administrative change.   Submit a public comment through Regulations.gov

 

The Trump administration has proposed removing the portions of the 2030 United States Census questionnaire that collect information about race and ethnicity. The proposal has not become final policy, but its implications are enormous. Race-related questions—although the categories have changed considerably and often reflected the racism of their time—have appeared in every decennial census since 1790.

This is not a technical dispute over a government questionnaire. It is a struggle over whether the government will continue collecting the population data needed to expose racial inequality—or deliberately make that inequality harder to prove.

To eliminate a racial disparity, we must first establish that it exists. We need to know who is affected, where the disparity occurs, how severe it is, and whether it is improving or worsening. We also need reliable information to determine whether the actions taken to correct the disparity are working.

If the government removes race and ethnicity from the census, racism will not disappear. The government will merely make its consequences harder to prove and easier to deny.

 


One Human Race—But Racism Is Real

Race is not a biological division of humanity into separate and unequal species. There is one human race.

But that truth does not resolve the problem. Race is also a social, legal, and political construction. Throughout American history, governments have assigned rights, burdens, privileges, and disabilities according to racial classifications.

Race helped determine who could be enslaved, who could become a citizen, who could vote, who could own property, who could marry, who could attend a particular school, who could live in a particular neighborhood, and who could enter the country. Race shaped access to employment, medical care, public accommodations, government benefits, and legal protection.

Race does not have to be biologically real to have real consequences. American law made race matter. It used race to distribute freedom, citizenship, property, political power, education, housing, and legal protection. The law helped create the racial categories whose continuing consequences the government now proposes to stop measuring through the census.

“There is only one human race” answers a biological question. It does not answer the social, legal, or public-health question of whether people assigned to different racial groups receive different treatment and experience different outcomes.

We do not collect racial data to prove that biological races exist. We collect it to document what racism does.


What I Learned From African Descendants in Europe

Before the 2001 World Conference Against Racism in Durban, I worked extensively with African descendants living in Europe. I saw firsthand what happens when governments refuse to collect racial and ethnic population data.

Several European countries resisted collecting such data. Their argument was that there was only one race—the human race—and that governments therefore should not recognize or record racial distinctions.

“There is only one race—the human race” sounded enlightened. But in practice, it gave governments a respectable-sounding reason not to document racism.

African descendants knew they were being discriminated against. They saw it in employment, housing, education, health care, policing, and other areas of public life. But when they described what was happening, government officials and members of the general population demanded statistical proof.

Those same governments had refused to collect the population data necessary to produce that proof. They created the absence of evidence and then used that absence to deny the existence of racial inequality.

African-descendant organizations were left to construct their own evidentiary record. They had to rely on individual accounts, community surveys, local studies, and research conducted with limited resources. That evidence was valuable, but it could not fully substitute for comprehensive population data.

Small studies were easily attacked. The sample was too limited. The research covered only one city. The participants supposedly were not representative. The findings could not be generalized. Some factor other than racism might explain the results.

The communities experiencing discrimination were forced to bear the financial and practical burden of proving what their governments had refused to measure.

My work in Europe taught me an enduring lesson: refusing to collect racial data does not produce racial equality. It produces plausible deniability.


The Health Caucus at the World Conference Against Racism

I attended the 2001 World Conference Against Racism and headed the Health Caucus. One of our objectives was to persuade countries to move toward collecting racial and ethnic population data.

Health made the need especially clear. If a country does not have reliable information about the size, location, age, income, and other characteristics of its racial and ethnic communities, how can it determine whether those communities experience higher rates of disease, disability, maternal death, infant mortality, undertreatment, or premature death?

A raw number is not enough. Suppose 1,000 Black people die from a particular illness. That number has little meaning unless we know the size and relevant characteristics of the Black population. We need a denominator to calculate a rate. We also need comparable information about other populations before we can determine whether a disparity exists.

Without reliable population data, governments cannot determine the scope of a health problem. They cannot allocate resources intelligently. They cannot identify the communities most affected. And they cannot determine whether a policy reduced, preserved, or worsened a disparity.

At Durban, we had to explain the difference between biological race and the social operation of racism. There may be one human race, but racial classifications still affect where people live, the air they breathe, the work they perform, the medical care they receive, and the burdens of stress they carry.

We lobbied hard. Ultimately, the Durban Programme of Action urged governments to collect, compile, analyze, disseminate, and publish reliable statistical data concerning people subjected to racism and racial discrimination. It recognized that governments need disaggregated data to assess conditions, measure progress, identify social gaps, and develop policies to combat racism. It also recognized the need to protect privacy, human rights, and the principle of self-identification.

That was a major victory. After persistent lobbying, we secured international recognition that governments could not fight racism while refusing to collect evidence of its existence.

Twenty-five years later, the United States is considering moving in the opposite direction.


America’s Complicated History of Counting Race

The United States has collected information related to race since the first federal census in 1790. That history is deeply compromised by white supremacy.

Early census categories reflected slavery, Native dispossession, racial hierarchy, and unequal citizenship. Under the Three-Fifths Clause, enslaved Black people were counted to increase the political power of the people who enslaved them. They were counted for representation but denied any voice in the government their bodies helped enlarge.

The Constitution did not use the words “slave” or “Black” in the Three-Fifths Clause. It referred to “three fifths of all other Persons.” But the legal and political meaning was unmistakable: enslaved people increased a state’s representation in Congress without receiving any corresponding political rights.

Census classifications changed repeatedly as the nation’s racial ideology, immigration policies, political interests, and understanding of identity changed. Some categories were degrading. Some imposed identities that communities did not choose for themselves. Government data has also been misused against vulnerable populations.

We should not hide that history. Racial data has served two very different purposes. It has been used as an instrument of racial control, but it has also become an instrument of racial accountability.

Today, racial and ethnic population data helps researchers, advocates, agencies, and courts examine inequality in health, education, employment, housing, voting, environmental exposure, wealth, incarceration, and access to government services. It allows us to compare outcomes across communities, locations, and time.

The history of misuse supports strong confidentiality protections, voluntary self-identification, community consultation, transparent categories, and strict limits on the use of individual information. It does not justify destroying the population data needed to identify discrimination.

The answer to abusive data collection is responsible data governance—not government-imposed ignorance.


Civil Rights Require Evidence

Civil-rights laws do not enforce themselves. Courts and government agencies require evidence.

Racial and ethnic population data can help determine whether electoral districts dilute the voting strength of racial minorities. It can reveal whether schools discipline Black children more severely than white children. It can expose racial segregation in housing and education. It can show disparities in mortgage lending, unemployment, incarceration, environmental exposure, maternal mortality, and access to medical treatment.

Census data does not, standing alone, establish every element of a legal discrimination claim. But without reliable population data, many civil-rights claims become much more difficult—and far more expensive—to prove.

Census information provides an essential national foundation. It supplies population estimates and geographic detail against which other evidence can be evaluated.

Eliminating race and ethnicity questions from the census would not erase every racial statistic collected by federal agencies, states, universities, or private researchers. That is not my claim. My claim is that without authoritative census population data, those other datasets become more vulnerable to attack.

Researchers need reliable population denominators to calculate rates, compare communities, evaluate trends, and determine whether a disparity is widening or narrowing. Smaller studies can be dismissed as unrepresentative, geographically limited, or methodologically inadequate. The government can continue demanding rigorous proof of racial inequality after weakening the population framework necessary to produce that proof.

Without that foundation, civil-rights litigation and policy research become more difficult and more expensive. Large institutions are in a better position to defend themselves. Communities experiencing discrimination face a heavier burden of proof.

That is not a neutral redistribution of costs. It transfers power away from the people seeking accountability and toward the institutions being asked to account for racial disparities.


Health Data Shows Why This Proposal Is So Dangerous

Where Black people experience worse health outcomes, the cause is not some biological defect in Blackness. Racism shapes exposure, resources, treatment, stress, and access to care.

Racial health disparities often reflect conditions racism has produced: segregated neighborhoods, environmental hazards, inadequate insurance, unequal access to specialists, dangerous working conditions, food insecurity, provider bias, undertreatment, chronic stress, and generations of unequal access to wealth and medical care.

National averages can conceal these conditions. A program may appear successful overall while Black maternal mortality remains unacceptably high, an Indigenous community lacks access to treatment, or a particular Asian American population experiences a serious health burden hidden inside an overly broad category.

Disaggregated data allows us to ask questions that an overall average cannot answer: Who benefited? Who was left behind? Did the disparity narrow? Did the policy reach the people facing the greatest risk?

If the government cannot or will not answer those questions, it can declare success without demonstrating equity.

A policy that cannot be measured can be praised regardless of whether it works.


Statistical Invisibility Is Not Colorblindness

The Census Bureau’s proposal declares that the census should be “colorblind” and should not be “distorted” by questions about supposedly immaterial personal characteristics such as race.

Calling race “immaterial” does not make racism immaterial.

There is a fundamental difference between refusing to discriminate and refusing to investigate whether discrimination exists.

A formally race-neutral government can administer policies with profoundly unequal racial consequences. American history is full of examples. After slavery, facially neutral laws helped preserve white control. After Brown v. Board of Education, officials often replaced openly racial commands with school boundaries, housing policies, funding systems, and administrative decisions that maintained segregation without expressly mentioning race.

Modern racial inequality frequently operates through geography, accumulated wealth, institutional discretion, criminal enforcement, educational boundaries, environmental policy, and access to health care. As discrimination becomes less explicit, reliable data becomes more important, not less.

Colorblindness becomes a tool of white supremacy when it allows the government to create or tolerate racial inequality while withdrawing the population data needed to measure its consequences reliably.


The Production of Official Ignorance

This would not be an innocent failure to gather information. The government already knows why racial population data matters. It knows that the information is used to enforce civil-rights laws, identify health disparities, evaluate public programs, and protect against the dilution of racial minorities’ political representation.

Ending census collection with that knowledge would be a deliberate decision to withdraw a tool essential to racial accountability, whatever justification the administration offers for doing so.

The danger is not simply that the government will know less. The danger is that the government will deliberately reduce its capacity to know and then use that manufactured ignorance to deny responsibility.

The cycle is familiar:

  1. Stop collecting racial population data through the census.
  2. Demand rigorous proof of racial disparities.
  3. Reject smaller studies as incomplete or unrepresentative.
  4. Declare that there is insufficient evidence of discrimination.
  5. Refuse corrective action because the problem has not been proven.

This is the institutional production of ignorance.

It gives government and powerful private institutions greater protection from scrutiny. It forces civil-rights organizations and marginalized communities to spend more money producing less comprehensive evidence. It makes disparities affecting smaller populations especially difficult to detect. It allows racial inequality to be dismissed as anecdotal even when communities see and experience it every day.

The issue is not whether racial categories are perfect. They are not. The issue is whether imperfect but carefully collected information is better than blindness. Plainly, it is.


From Durban to the 2030 Census

In Durban, we fought to establish that recognizing one human race does not require blindness to racism. We fought for governments to collect the information necessary to determine whether racial disparities existed and whether government action was reducing them.

We understood that data alone would not end racism. Numbers do not enforce themselves. Statistics do not redistribute power, provide medical care, desegregate schools, or compensate communities for generations of stolen labor and wealth.

But without reliable data, governments can deny the problem, misdirect resources, conceal failure, and avoid accountability.

The United States has long possessed racial population data that African-descendant communities in parts of Europe struggled to obtain. That data has helped Americans document racial disparities for generations. The Trump administration now proposes to weaken a central part of that national data system by removing race and ethnicity questions from the 2030 Census.

We should improve racial and ethnic data, not eliminate it from the census. We should strengthen privacy protections, respect self-identification, refine overly broad categories, provide meaningful information about multiracial identities, and produce better data concerning communities too often hidden inside national averages.

What we must not do is confuse statistical erasure with racial justice.

Removing race from the census will not remove racism from American life. It will not close the racial wealth gap, prevent discriminatory policing, reduce Black maternal mortality, protect voting rights, desegregate neighborhoods, or ensure equal medical treatment.

It will make those injustices harder to prove. It will also make the remaining data easier to challenge.

I have already seen what happens when governments refuse to collect racial and ethnic population data. Racism does not disappear. Government accountability does. The people experiencing discrimination are told to prove what the government has chosen not to measure. Their testimony is dismissed as anecdotal, their studies are attacked as inadequate, and the government’s deliberate ignorance becomes its defense.

A colorblind census in a racist society is not progress. It is statistical erasure—and statistical erasure protects white racial power.


Sources

 Vernellia R. Randall, Professor Emerita of Law, University of Dayton School of Law.  This article is the sole intellectual and scholarly work of Vernellia Randall. ChatGPT was used only as a drafting aid comparable to a research or editing assistant. All concepts, analysis, legal reasoning, interpretations, and conclusions are entirely the author’s own, and the author assumes full responsibility for the content.