LWVUS joins comments with the National Network of Arab American Communities (NNAAC) opposing proposed census rule change

LWVUS joins comments with the National Network of Arab American Communities in opposition to proposed changes to the census residence rule.

November 2, 2026

Submitted electronically via www.regulations.gov

U.S. Census Bureau

U.S. Department of Commerce 4600 Silver Hill Road Designation: ADDEMO/FRN Washington, DC 20746

Re: Comment on Proposed Rule, "Decennial Census of the Population of Americans; Proposed Residence Criteria and Proposed Regulations for Demographic Questions," 91 FR 57524 (Sept. 10, 2026), Docket No. 260903-0005 (FDMS Docket No. USBC-2026-0628), RIN 0607-AA75

To the U.S. Census Bureau:

The National Network for Arab American Communities (NNAAC), a consortium of 36 Arab American community-based organizations in 13 states, submits this comment with the undersigned organizations in opposition to the proposed rule cited above.

Every use of the census depends on it showing the country as it is. From dividing the seats in the House of Representatives to deciding where a clinic should open, government works effectively and spends efficiently only when the data reflect reality. Data shaped to serve a political end do neither. The public pays for them twice, once to collect them and again for every decision they misdirect. Communities that federal data have already failed know that cost, whether they were undercounted, like Black and Latino residents, or made invisible through misclassification, like Arab Americans being counted as White.

The proposed rule fails that test in both of its parts. Proposed 15 C.F.R. part 60 would rewrite the residence criteria to remove millions of residents from the apportionment count because of their immigration status, and would tie every person’s place of residence to tax records and a three-month winter window. Proposed 15 C.F.R. part 70 would end a practice as old as the census itself by barring any question about race or ethnicity, and would foreclose a question about sexual orientation that no census has asked. Each part discards reliable information that people report about themselves and replaces it with administrative records that were never built to count a population. The result would be an undercount, unrepresentative data, a costlier census, and a less capable government. Those harms reach everyone, but they fall hardest on communities of color, immigrants, and low-income households. The Census Bureau should withdraw the proposed rule in full.

1. The Census Exists to Show the Country as It Is

The notice of proposed rulemaking asserts that “[t]he sole Constitutional purpose of the decennial census is to apportion the seats in the U.S. House of Representatives.”¹ The premise is wrong. The Constitution requires an “actual Enumeration” conducted “in such Manner as [Congress] shall by Law direct.”² Congress has directed by law a “decennial census of population” that serves many ends, including the tabulations that states use to draw legislative districts.³ The notice itself concedes that “demographic questions have been asked in every census since 1790.”⁴ The First Congress understood what the notice now disputes: a government cannot serve a population it has not measured.

Today the count divides the 435 House seats and the votes in the Electoral College. States use it to draw congressional, legislative, and local districts. Data from the census and the programs built on it guided more than $2.8 trillion in federal spending through 353 programs in fiscal year 2021, and states and localities use the same data to divide their own funds.⁵ Planners use it to size schools, transit, housing, hospitals, and emergency response. Businesses use it to decide where to open, hire, and invest. It supplies the denominators for disease rates, the baseline for civil rights enforcement, and the benchmark against which other federal surveys are designed and weighted. Congress built a decentralized federal statistical system of more than 100 agencies and programs on that foundation, and policymakers, businesses, researchers, and governments at every level rely on it.⁶

A flawed count cannot be repaired until the next one. The Bureau’s annual population estimates start from the decennial count and carry it forward for ten years, adjusting only for births, deaths, and migration. Every formula that runs on those estimates inherits the error. The Older Americans Act shows how Title III funds are allotted according to each state’s share of the national population aged 60 and older.⁷ When a state’s count of older residents is depressed, whether because households did not respond or because the Bureau removed residents from the population base, its allotment falls in every year of the decade. However, the actual number of older people who need meals, transportation, and in-home care does not fall with it.

That is why politicizing the census is expensive. Every household that does not respond must be pursued in person, the largest and costliest operation the Bureau undertakes, or have its information imputed from neighboring households and administrative records.⁸ The 2020 Census cost about $14.2 billion.⁹ Every error then misdirects public money for ten years. A rule that predictably drives down response and strips residents from the count does not “promote efficiency,” as the notice claims.¹⁰ It buys worse data at a higher price.

2. The Proposed Residence Criteria Would Replace an Actual Enumeration with a Status Test and Undercount the Country

Since 1790, the census has counted people where they live. Proposed part 60 abandons that principle in two ways. First, it would exclude from the apportionment count every foreign citizen who is not also a U.S. citizen or lawful permanent resident.¹¹ That exclusion violates the Fourteenth Amendment. It cannot be administered without either misusing confidential records or questioning every household about immigration status. And it will suppress response far beyond the people it intends to target. Second, it would redefine “usual residence” for everyone as the place a person “lawfully” spent the most days between January 3 and April 1, as “evidenced by” tax records.¹² That definition will misplace or miss low-income, mobile, and older residents in every state. Both changes lead to the same result: an undercount that touches every community, falls disproportionately on communities and regions with large immigrant, minority, and low-income households, and leaves the country with data that do not represent it.

1. Excluding Residents Because of Immigration Status Is Unconstitutional and Will Produce Undercounts and Unreliable Data

1. The Fourteenth Amendment requires the census to count every person

The Fourteenth Amendment provides that “Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State.”¹³ That sentence was written after the Civil War to replace the clause that had counted an enslaved person as three-fifths of a person. Its purpose was to take away the government’s power to decide who qualifies as a person. The proposed rule would take that power back, this time drawn along lines of immigration status.

The text of the Fourteenth Amendment leaves no room for it. Its drafters used three different words in the same section, “persons” for the count, and “inhabitants” and “citizens” for the clause that penalizes states for denying the vote. They knew the difference and chose the broadest term for apportionment. Congress in 1866 debated basing apportionment on voters or on citizens and rejected both, with full knowledge that, as Representative Roscoe Conkling put it, “the number of aliens in some States is very large.”¹⁴ As Senator Jacob Howard explained when he introduced the Amendment, “[n]umbers, not voters; numbers, not property; this is the theory of the Constitution.”¹⁵ The Supreme Court, without dissent, relied on that history in 2016 to confirm that “representatives serve all residents, not just those eligible or registered to vote.”¹⁶ Congress has written the same command into statute. Apportionment must rest on “the whole number of persons in each State” as ascertained by the census, drawn from a “tabulation of total population by States.”¹⁷ The Bureau’s discretion over how to conduct the count is not discretion over whom to count.

The notice concedes that the Fourteenth Amendment “rightly rejected the view that individuals of African ancestry are not full and equal ‘persons,’” but contends that the Amendment otherwise preserved a Founding-era meaning of “persons” limited to “inhabitants” who were “members of the body politic.”¹⁸ The argument fails on its own sources. The original Apportionment Clause counted women, children, indentured servants, and, at three-fifths, enslaved people. The Founding generation treated none of them as members of the body politic. The authorities the notice cites define an inhabitant by domicile and the intent to remain, “[t]he Domicil and the animus habitandi,” not by citizenship.¹⁹ The notice’s own quotation from Vattel defines inhabitants “as distinguished from citizens.”²⁰ The category the rule targets did not exist for the Founders to exclude, because the first federal laws restricting immigration came almost a century later. The 1820 and 1830 censuses separately tallied “foreigners not naturalized” and included them in the apportionment totals.

Every census since has counted residents without regard to citizenship or immigration status. The Supreme Court has held under the same Amendment that, “[w]hatever his status under the immigration laws, an alien is surely a ‘person’ in any ordinary sense of that term.”²¹ Apportionment encompasses all persons, including noncitizens and others ineligible to vote.²² And the rule does not even follow its own theory. It excludes refugees, asylees, and workers and students on visas, all of whom live here with the government’s express permission.

Franklin v. Massachusetts does not help the Bureau. There, the Supreme Court upheld counting overseas federal personnel in their home states because “usual residence” has “been used broadly enough to include some element of allegiance or enduring tie to a place.”²³ The Court used that idea to include people who kept ties to a state during a temporary absence. It did not authorize excluding people who live in a state. And the tie the Court described is a tie “to a place,” not loyalty to a sovereign. A person who has lived, worked, paid taxes, and raised children in the same city for fifteen years has an enduring tie to that place by any measure.

2. The Bureau cannot determine immigration status without misusing confidential records or questioning every household

The proposed regulatory text does not say how the Bureau would sort residents by immigration status. The notice offers two routes. Both fail.

The first is record linkage. The notice describes the Bureau’s “enormously broad access to administrative records to determine an individual’s legal status,” drawn from the Departments of the Treasury, Health and Human Services, Homeland Security, and State, the Postal Service, “certain state, local, and tribal records; and third-party commercial data.”²⁴ Those records reach the Bureau under laws that rest on one condition: statistical use. The Privacy Act defines a statistical record as one “not used in whole or in part in making any determination about an identifiable individual.”²⁵ Tax return information may be used only “to the extent necessary in” structuring censuses and “conducting related statistical activities authorized by law,”²⁶ a limit the notice acknowledges.²⁷ Congress has defined a nonstatistical purpose to include any use of identifiable data that “affects the rights, privileges, or benefits of a particular identifiable respondent.”²⁸ The Census Act bars the use of census information “for any purpose other than the statistical purposes for which it is supplied.”²⁹ Deciding, one named person at a time, whether an individual will be represented in Congress is a determination about an identifiable individual. It is not statistics.

Nor can the records do the job. They were captured by different agencies, for operational purposes, at different moments. Immigration status changes. People naturalize, adjust status, and win asylum between the date a record was created and Census Day, and no federal file records who is a U.S. citizen today. Linkage fails most often for the same populations the rule targets.

For example, the same Arabic and other non-English names are routinely transliterated several different ways across agencies’ files due to clerical errors. The rule offers no notice to a person classified out of the count and no way to contest the classification. A naturalized citizen matched to a stale record would simply vanish from their state’s apportionment population.

The exclusion is also overbroad on its own terms. It reaches everyone “whose legal status is less durable and indefinite in length than lawful permanent resident status.”³⁰ The Bureau asks “whether other forms of legal status are materially similar to lawful permanent resident status.”³¹ We decline to rank our neighbors. The question has no constitutional answer, because the Fourteenth Amendment does not sort persons by the durability of their papers. Every person who resides in a state must be counted.

The second route is a question. The notice states that the Bureau “could, in addition, include a question about legal status on the short-form decennial census questionnaire.”³² Because the records cannot do the job, a question is the likely outcome, and the proposed rule is built to accommodate one. Proposed § 70.2(c) makes most remaining demographic questions voluntary, but proposed § 70.2(d) exempts “[q]uestions that seek information related to the application of the residence criteria.”³³ An immigration status question would therefore be mandatory, and the notice reminds readers that failing to answer a census question “constitutes a federal crime.”³⁴ The same notice calls race too “(potentially) sensitive” to ask about.³⁵ The notice claims that a status question “would accord with substantial past practice,”³⁶ but no census has ever asked every person in the country for their immigration status. When the Department of Commerce last tried to add a citizenship question, the Supreme Court set the decision aside because the stated rationale “seems to have been contrived.”³⁷

3. The exclusion will suppress response, undercount communities, and shift political power

The census works only when people believe that answering is safe, confidential, and separate from immigration enforcement. The Bureau’s own research, using its randomized 2019 Census Test, found that a citizenship question lowers self-response in many households with noncitizens, U.S.-born Hispanics, and naturalized citizens, while also increasing the number of household members left off the forms that are returned. Households with at least one Individual Taxpayer Identification Number (ITIN) holder responded at 27.5 percent, compared with 70.4 percent for households of U.S.-born non-Hispanic White residents, and the question widened that gap. Those households omitted members at 6.0 times the comparison rate without the question and 10.4 times with it.³⁸ The Supreme Court accepted in 2019 that lower response from noncitizen households was “the predictable effect of Government action on the decisions of third parties” in holding states had standing to challenge a citizenship question.³⁹ The effect here would be larger. This rule announces that the government will assemble tax, health, immigration, postal, state, and commercial records to decide who counts. A household does not need to see a question on the form to be afraid of that.

The notice suggests in a footnote that the Bureau “expects to continue” publishing statistics “separate from those for apportionment.”⁴⁰ That is no answer. A household that does not respond is missing from every census product: the redistricting file, the base for ten years of population estimates, and the controls used to weight the American Community Survey. Most affected households are mixed. Citizens, lawful permanent residents, and people with other statuses live under one roof, and when the household goes uncounted, all of them do. The result is fewer resources and less data for entire states and localities, and especially for the racial and ethnic groups concentrated in them. The starting point is already unequal. The 2020 Census undercounted Hispanic residents by 4.99 percent, Black residents by 3.30 percent, and American Indian and Alaska Native residents of reservations by 5.64 percent, while overcounting non-Hispanic White residents by 1.64 percent.⁴¹ The proposed rule would widen those gaps.

The exclusion also denies the equal protection that the Fifth Amendment requires of the federal government.⁴² The House is meant to provide “equal representation for equal numbers of people.”⁴³ The notice itself relies on Wesberry.⁴⁴ The proposed rule would move House seats and Electoral College votes away from states where immigrant communities, which are predominantly Latino, Asian, Black, and MENA, make up a larger share of residents, and toward states with whiter populations. An analysis of the narrower exclusion attempted in 2020 found that California, Florida, and Texas would each end up with one fewer seat, and Alabama, Minnesota, and Ohio with one more.⁴⁵ This rule reaches far more people. Every resident of a losing state, including every U.S. citizen, would hold a diluted share of representation. The notice does not examine that effect. It instead pairs the exclusion with a rule that eliminates the race and ethnicity data needed to measure it. A reallocation of political power along racial lines, adopted alongside a ban on the data that would reveal it, cannot be defended as neutral.

2. The Proposed Definition of “Usual Residence” Will Miscount Low-Income, Mobile, and Older Residents

In 2020, a person’s usual residence was the place where the person “lives and sleeps most of the time,” as the person reported it.⁴⁶ Proposed § 60.2 replaces that standard for every resident: “A person’s ‘usual residence’ is the residence at which they have lawfully spent the greatest number of days during the census enumeration period,” and it “should be consistent with, and evidenced by, their tax records (e.g., tax returns, W-2 forms).”⁴⁷ Each element of that definition is a new source of error.

”Lawfully.” Only days on which a person was “physically and lawfully present” at a residence count.⁴⁸ The rule never defines the term. If it refers to immigration status, it duplicates the exclusion in proposed § 60.4 and spreads that provision’s chilling effect to the residence question. If it refers to lawful occupancy, it makes every enumeration turn on a legal question no census taker can resolve: whether a tenant in an unpermitted unit, a family doubled up beyond a lease’s occupancy limit, a relative not named on the lease, or a tenant contesting an eviction is “lawfully” present. These arrangements are common in low-income and immigrant communities, and households living in them will have reason not to report who lives there. Where the rule is silent, the Bureau would fill the gap later, without public comment.

Tax records. A tax record shows where a form was sent. It does not show where a person lives and sleeps. Many people do not appear in tax records at all, including people whose income falls below the filing threshold, older adults living on Social Security, and workers paid in cash.

Children appear only on someone else’s return. Others appear at a stale or different address: recent movers, students listed at a parent’s home, survivors who have fled a shared filing address, and service members whose W-2 forms reflect a state of legal residence rather than the community where they are stationed. The notice admits that its reliance on tax records “will necessarily be situation dependent.”⁴⁹ That is an admission that there is no standard. The rule does not say what happens when a person’s own response and a tax record disagree, or which one prevails.

The January 3 to April 1 window. The notice selects January 3 because it is the start or midpoint of a congressional term.⁵⁰ That date has nothing to do with where people live. A greatest-number-of-days test over three winter months would count older adults who winter in warmer states there, rather than in the communities where they live most of the year and use services. It would misplace seasonal farm, construction, and tourism workers, people who move during the window, and people cycling through temporary housing. Because one hour of presence counts as a day and a person “may spend a certain day at more than one residence,”⁵¹ a child who moves between two parents’ homes could tie, and the rule has no tiebreaker.

The people the rule leaves out. The 2020 residence criteria gave specific instructions for college students, residents of nursing facilities and group homes, people in shelters, and people experiencing homelessness.⁵² Proposed part 60 addresses none of them. A person without a usual residence is counted only if found at a “transitory location” such as a hotel or campground.⁵³ The rule says nothing about a person sleeping outdoors or in a shelter. The people with the thinnest paper trails are housing-insecure and doubled-up families, low-wage and seasonal workers, older adults, survivors of domestic violence, and students. And some of these people are the ones the census already misses most often.⁵⁴ This definition would miss more of them.

3. Eliminating Race and Ethnicity Questions Would Blind Government to the Country It Serves

Proposed § 70.2(a) provides that “[n]o question about race, ethnicity, or sexual orientation shall appear on the short form decennial census questionnaire or any questionnaire used for the enumeration of the population.”⁵⁵ No census has ever asked about sexual orientation. Every census since 1790 has asked about race. The notice would end that practice by regulation on the theory that the census “should be colorblind and should not be distorted in any way by questions about immaterial personal characteristics, such as race.”⁵⁶ Race and ethnicity are not immaterial to who dies in childbirth, whose neighborhood is split between districts, or who is turned away from a job or a home. A government that stops measuring disparities does not end them. It only stops seeing them.

3. Nothing Can Replace the Census as the Source of Race and Ethnicity Data

The decennial census is the only comprehensive, standardized, self-reported source of race and ethnicity data covering the entire population. The rest of the federal statistical system uses it as the reference standard. Agencies rely on census benchmarks to build sampling frames, weight survey responses, produce population estimates, and test the quality of their own records.

Changes in how the Bureau collects and processes race and ethnicity data carry through to every statistic that uses census denominators.⁵⁷ Other federal surveys depend on the same foundation. The National Health Interview Survey, for example, has used decennial census race and ethnicity data to design its sample and Census Bureau population estimates by race and ethnicity to weight its results.⁵⁸

The notice asserts that “[d]ata from administrative records can fill information gaps.”⁵⁹ For race and ethnicity, they cannot. Administrative records are collected to run programs, not to produce statistics. They use inconsistent definitions, vary in completeness, and often contain no race or ethnicity information at all. They carry forward the old categories that erased communities in the first place, such as Middle Eastern or North African communities. And assigning a person a race or ethnicity from a government file takes away the person’s ability to say who they are, which is the principle on which federal race and ethnicity standards rest.

The notice also points to the American Community Survey (ACS).⁶⁰ The ACS was designed to complement the census, not replace it.⁶¹ It samples about 3.5 million addresses a year, cannot produce block-level data, and yields wide margins of error for small populations. Its estimates are weighted to population controls by race and Hispanic origin that derive from the decennial census.⁶² Removing race and ethnicity from the census would degrade the survey the notice offers as the substitute. For MENA and other communities, the ACS is no substitute at all. The Bureau announced in 2024 that the 2027 ACS would use the Statistical Policy Directive No. 15 (SPD 15) question, including the MENA category, and then reversed course in May 2026.⁶³

Federal law depends on these data. The redistricting file that Congress requires the Bureau to deliver to the states has carried race and ethnicity to the block level because enforcing the Voting Rights Act of 1965 requires it.⁶⁴ Agencies, courts, and advocates use census race and ethnicity data as the baseline for enforcing Title VI of the Civil Rights Act of 1964, the Fair Housing Act, and equal employment opportunity requirements.⁶⁵ Even a “colorblind” Constitution forbids intentional racial discrimination, and discrimination cannot be detected or proven without data.

Removing the questions would not remove the need. It would shift the work to states, localities, federal agencies, health systems, researchers, grant recipients, and community organizations, each collecting what it can through separate, fragmented, and less reliable methods. The Bureau certifies that the information “would remain available or could be obtained in other ways,”⁶⁶ without saying how. The result is higher cost, duplicated effort, and data that cannot be compared. That is the opposite of efficiency.

4. The Proposed Rule Ignores Statistical Policy Directive No. 15 and Contradicts Itself

The rule discards the governing federal standard without mentioning it. In 2024, after years of research, testing, and public comment, the Office of Management and Budget (OMB) revised SPD 15, the standard that governs how every federal agency collects and reports race and ethnicity data.⁶⁷ The revision combined race and ethnicity into a single question, required detailed categories by default, and added MENA as a minimum reporting category. The 2030 Census was to be the first census to use it.⁶⁸ The notice asserts that racial and ethnic categories are arbitrary and that the need for the data has diminished.⁶⁹ OMB considered and answered that objection two years ago. It recognized that the categories are sociopolitical constructs rather than biological ones and reaffirmed that they remain essential to producing accurate, comparable, and useful federal statistics and to evaluating how well federal programs serve the public.⁷⁰ The notice never mentions SPD 15. For MENA communities, the proposed rule would remove their category from the census before it has been used once.

The rule calls the data useless and keeps collecting them. The notice describes race as an

“immaterial personal characteristic” and cites the “reduced need for or usefulness of this information.”⁷¹ Yet proposed § 70.4 preserves the Bureau’s ability to collect the same information through the ACS, through other surveys, and “from other sources.”⁷² If the data have no value, there is no reason to keep gathering them. If they have value, there is no reason to remove them from the only source that reaches everyone.

The burden rationale is unsupported. The notice claims that dropping the questions would reduce respondent burden and raise response rates.⁷³ The sources it cites compare long questionnaires to short ones. None isolates the effect of a race or ethnicity question.⁷⁴ SPD 15 already reduces two questions to one. And the same notice contemplates adding a mandatory question about immigration status, the kind of question that the Bureau’s own research shows drives response down. The Bureau has not shown through any rigorous analysis that the claimed savings outweigh the public cost of lost information, weaker statistics, and new collection burdens pushed onto everyone else. It declares that the rule “contains no information collection requirements” while asking the public to estimate the burden reduction for it.⁷⁵

5. Prohibiting a Question the Census Has Never Asked Shows That the Rule Is a Political Instrument, Not a Scientific One

Nothing in the proposed rule exposes its politics more plainly than the ban on a sexual orientation question. No decennial census has asked one, and none has been proposed for 2030. Banning it therefore delivers neither benefit the notice claims for its restrictions, a lighter response burden and a lower risk of disclosure.⁷⁶ It simply settles in advance a question that federal law leaves to research, testing, and reporting to Congress. For 2030, the first of those reports is not due until April 2027.⁷⁷

The Department has used that time to dismantle the process instead. In February 2025, it disbanded the Census Bureau’s scientific and 2030 Census advisory committees.⁷⁸ Federal agencies had asked for sexual orientation and gender identity data, and Congress paid for a large-scale test on the ACS. The Bureau then dropped the questions from its testing, and the government admits it has released no results.⁷⁹ Banning a question before the evidence is in, while withholding the evidence that exists, is the opposite of how a statistical agency should work. It would let census content turn on which communities are disfavored rather than on what the country needs to know.

6. The Privacy Rationale Is a Problem of the Department’s Own Making

The notice presents privacy protection and demographic data as competing goals. They are not. Federal statistical agencies have balanced the two for decades through statutory confidentiality protections and controlled access.⁸⁰ The article the notice cites for its disclosure-risk argument describes a framework for weighing confidentiality against the usefulness of data, not a reason to stop collecting them.⁸¹ Disclosure risk arises when data are published, not when they are collected, and the Bureau has tools to manage it. Rather than eliminate race and ethnicity questions, the Bureau could:

• Continue using disclosure avoidance methodologies.
• Expand restricted-access research environments.
• Use controlled-access data enclaves such as Federal Statistical Research Data Centers.
• Increase cell suppression and aggregation in public data products.
• Expand tiered-access systems that provide different levels of detail to researchers, federal agencies, and the public.

The existence of disclosure risk does not establish that demographic questions should be removed. It establishes the need for appropriate privacy protections. The Department of Commerce has been weakening those protections. On June 4, 2026, without public comment, it issued Department Administrative Order (DAO) 216-26, Disclosure Avoidance for Statistical Products, which directs that “[n]oise infusion shall not be used for any statistical product.”⁸² Noise infusion is the method the Bureau used to protect the 2020 Census, and the Bureau reads the order to bar data swapping and severely limit synthetic data as well.⁸³ The leading associations of data users and population scientists warned that the order would produce “less privacy for our personal information, or less usable data, or both.”⁸⁴ Three months later, the Department cites the disclosure risk that remains as a reason to stop asking about race and ethnicity, and proposed § 70.3(b) would bar publishing demographic data whenever doing so “necessitates introducing uncertainty into the population count data at either the national or the state level.”.”⁸⁵ The Department took away a privacy tool and now points to the gap it created as the reason to take away the data. An agency cannot manufacture the problem it offers as its justification.
 

The rule would also increase the privacy risk it claims to reduce. It removes one question from the form while building, behind the form, what the notice calls “a vast and expanding universe of records” that is “nearly exhaustive of the civilian Federal government’s holdings,” supplemented by “third-party commercial data” and a Bureau that is “continually searching for new sources.”⁸⁶ To decide each person’s status and residence, the Bureau would link tax, health, immigration, passport, postal, state, tribal, and data-broker records for every person in the country. That is a far greater concentration of personally identifiable information than any census questionnaire has produced, and a far greater target for breach and misuse. Replacing a voluntary answer with a compiled government file does not protect anyone’s privacy. For any community that has lived under government surveillance, a compiled file on every resident confirms the fear that keeps people from answering the census at all.

4. The Proposed Rule Undermines the Federal Statistical System

Both parts of the proposed rule conflict with the framework Congress built for federal statistics. Federal statistics are widely regarded as the nation’s most authoritative source of information. The population and demographic data that flow from the census are a strategic national asset that supports statistical work across the government, evidence-based policymaking, program administration, academic research, and private-sector decisions.⁸⁷ That asset is valuable only if it is accurate and trusted. Part 60 would corrupt the population counts that anchor the system by excluding residents, suppressing response, and tying residence to records that do not measure it. Part 70 would remove the demographic benchmark the system uses to check everything else.

Together they would leave every agency, state, and business that relies on federal statistics working from a less accurate picture of the country, and they would spend the public’s trust in the Bureau’s independence for no statistical gain.

5. Communities Already Missing From the Data Show What Unrepresentative Data Cost

The communities that federal data already miss show what the proposed rule would cost. Some have been erased by classification. Arab Americans and other MENA populations are one example. For decades, federal data have folded them into the White category despite their distinct health, economic, linguistic, cultural, immigration, and civil rights experiences. That left policymakers, researchers, public health officials, and community organizations without the information needed to serve them. Others have been undercounted census after census, among them Black, Hispanic, and Native communities, young children, and renters. The proposed rule would go further on both fronts. Part 70 would eliminate self-reported race and ethnicity data for every community, not only those already misclassified. Part 60 would remove residents from the apportionment count by design and miss others through a residence test built on records, deepening undercounts that already fall hardest on communities of color. Both result in data that do not represent the country, and decisions that miss the people left out.

7. Health

Race and ethnicity data show where disease, risk, and gaps in care are concentrated, which lets health systems act earlier and at lower cost. Detailed data revealed that Black women die from pregnancy-related causes at substantially higher rates than White women, allowing agencies to target interventions and track progress.⁸⁸ For Arab Americans, the data are missing.⁸⁹ Research documents significant mental health burdens tied to discrimination, acculturative stress, and conflict-related trauma.⁹⁰ Because federal data group Arab Americans with the White population, those disparities are hidden from the agencies that fund and plan care, and research on MENA health remains thin.⁹¹ The same aggregation distorts the statistics for White populations. The residence criteria compound the harm from the other direction. Census Bureau population estimates supply the denominators for official health statistics such as cancer incidence and death rates.⁹² Population measures also help determine where community health centers are established.⁹³ People left out of the count still get sick, give birth, and come to the emergency room. A depressed count produces clinics planned for a population smaller than the one that walks through the door.

8. Education and Economic Mobility

Race and ethnicity data identify disparities in educational attainment, income, employment, homeownership, and business ownership, and they show whether public investments reach the people they were meant to reach. Detailed data documented persistent racial gaps in intergenerational mobility and informed policies to close them.⁹⁴ For example, classifying Arab Americans as White has left schools, workforce agencies, and lenders without the data to see barriers facing MENA students, workers, and business owners, or to evaluate whether programs are working for them. The residence criteria strike the same programs through their funding.

Education, workforce, and economic development allocations run on population estimates anchored to the decennial count. The tax-record test falls hardest on low-wage, seasonal, and cash-paid workers and their children, the people those programs exist to serve. A school does not enroll fewer children because the census missed them. It serves them with less.
 

9. Civil Rights Enforcement

Enforcement of the Civil Rights Act, the Voting Rights Act, the Fair Housing Act, and equal employment opportunity requirements runs on numbers. Agencies and advocates identify discrimination by comparing how a group is treated with its share of the relevant population. Without standardized race and ethnicity data, that comparison cannot be made. The absence of MENA data, for example, has made it harder to identify patterns of discrimination against MENA communities, document disparate impact, and demonstrate the need for a remedy under civil rights and voting rights laws. MENA Americans are often not perceived as White and do not see themselves as White, yet the data record them that way.⁹⁵ The residence criteria damage the other half of the comparison. Enforcement depends on an accurate denominator. A count that omits entire categories of residents, or misses the households most afraid to respond, removes the baseline against which unequal treatment is measured.

10. Political Representation and Accountability

Government answers to the people it can see. Black communities have been undercounted in every census, and the National Urban League demonstrated how the estimated 2 million Black residents missed in the 2020 Census count was approximately equal to the population of three congressional districts.⁹⁶ In Detroit, the nation’s largest Black-majority city, researchers found that the 2020 Census missed about 8 percent of occupied homes in the neighborhoods they audited, a rate that, if it held citywide, would mean roughly 50,000 residents were uncounted.⁹⁷ Arab Americans have faced a different kind of invisibility. Because federal data have classified them as White, elected officials and agencies have no standard measure of whether constituent services, translated program materials, or language assistance at the polls reach Arab American residents, and little pressure to answer for the gap. Likewise, the lack of data makes it nearly impossible in most states to assess how drawing Congressional districts affects Arab American voting power. The proposed rule would extend and exacerbate both losses for these and other communities. By removing residents from the count in Part 60 and by removing self-reported race and ethnicity data for everyone in Part 70, this rule would only exacerbate and spread such harms.

6. Conclusion

The census serves the country only when it shows the country as it is. The proposed rule would make the 2030 Census less accurate, less representative, less trusted, and more expensive. Part 60 would exclude residents the Constitution requires the Bureau to count, suppress response in the communities that are already hardest to count, and tie every person’s residence to records that do not measure it. Part 70 would end 240 years of data on race and ethnicity, discard SPD 15 before its first use, and push the cost of replacing that information onto every other level of government and onto organizations like ours. Everyone would pay for the result. Communities of color, immigrants, and low-income households would pay the most.

NNAAC and the undersigned organizations urge the Census Bureau to withdraw the proposed rule in full, to count every person residing in each state where that person lives and sleeps most of the time, to count each person for apportionment and not use the census in any way to determine immigration or citizenship status, and to ask about race and ethnicity as laid out in SPD 15 on the 2030 Census.

We appreciate the opportunity to comment. For questions regarding this comment, please contact NNAAC Director, Rima Meroueh, at [email protected].

Respectfully submitted,

National Network for Arab American Communities (NNAAC) Joined by:

AADA ACCESS

Access California Services

African Communities Together (ACT) Amel Association Internacional

Arab American Civic Council

Arab American Family Services (AAFS) Arab American Heritage Council (AAHC) Arab-American Family Support Center Beurre Roux Inc

Center for Civil Justice

Center for Law and Social Policy (CLASP) Civic Ark

Community Service Society of New York Communities United for Status & Protection (CUSP)

Culturingua

East Bay Sanctuary Covenant Elmahaba Center

Equality California LatinoJustice PRLDEF

League of Women Voters of the United States Majdal: Arab Community Center of San Diego National Arab American Bar Association National Immigration Law Center

Palestinian American Community Center Praxis Health Empowerment

Prevention Institute

Promote The Vote Advocates

Serving at-risk families everywhere, Inc. Somali Family Service

SOMOS Mayfair

Syrian Community Network UndocuBlack Network

¹ Decennial Census of the Population of Americans; Proposed Residence Criteria and Proposed Regulations for Demographic Questions, 91 Fed. Reg. 57524, 57524 (proposed Sept. 10, 2026) (to be codified at 15 C.F.R. pts. 60,

70).

² U.S. Const. art. I, § 2, cl. 3.

³ 13 U.S.C. § 141(a); Act of Dec. 23, 1975, Pub. L. No. 94-171, 89 Stat. 1023 (codified as amended at 13 U.S.C. § 141(c)).

⁴ 91 Fed. Reg. at 57527 (quoting Dep’t of Com. v. New York, 588 U.S. 752, 769 (2019)).

⁵ Ceci A. Villa Ross, U.S. Census Bureau, Uses of Decennial Census Programs Data in Federal Funds Distribution: Fiscal Year 2021 (June 2023), https://www.census.gov/library/working-papers/2023/dec/census-data-federal-funds.html.

⁶ Taylor R. Knoedl, Cong. Rsch. Serv., R48161, The Federal Statistical System: An Overview (Aug. 19, 2024), https://www.congress.gov/crs-product/R48161; see also infra notes 58, 92 and accompanying text.

⁷ 42 U.S.C. § 3024(a).

⁸ U.S. Gov’t Accountability Off., GAO-19-140, 2020 Census: Additional Steps Needed to Finalize Readiness for Peak Field Operations 1, 3 (Dec. 10, 2018), https://www.gao.gov/assets/d19140.pdf.

⁹ U.S. Gov’t Accountability Off., GAO-21-478, 2020 Census: Innovations Helped with Implementation, but Bureau Can Do More to Realize Future Benefits 6 (June 14, 2021), https://www.gao.gov/assets/gao-21-478.pdf.

¹⁰ 91 Fed. Reg. at 57524.

¹¹ Proposed 15 C.F.R. § 60.4(b), 91 Fed. Reg. at 57529.

¹² Proposed 15 C.F.R. § 60.2, 91 Fed. Reg. at 57529.

¹³ U.S. Const. amend. XIV, § 2.

¹⁴ Evenwel v. Abbott, 578 U.S. 54, 66–67 (2016); id. at 101 (Alito, J., concurring in the judgment) (quoting Cong. Globe, 39th Cong., 1st Sess. 359 (1866)).

¹⁵ Id. at 67 (quoting Cong. Globe, 39th Cong., 1st Sess. 2766–67 (1866)).

¹⁶ Id. at 74.

¹⁷ 2 U.S.C. § 2a(a); 13 U.S.C. § 141(b).

¹⁸ 91 Fed. Reg. at 57525–26.

¹⁹ Id. at 57526.

²⁰ Id.

²¹ Plyler v. Doe, 457 U.S. 202, 210 (1982).

²² Garza v. Cnty. of Los Angeles, 918 F.2d 763, 774 (9th Cir. 1990) (citing Fed’n for Am. Immigr. Reform v.

Klutznick, 486 F. Supp. 564, 576 (D.D.C.) (three-judge court), appeal dismissed, 447 U.S. 916 (1980)).

²³ Franklin v. Massachusetts, 505 U.S. 788, 804 (1992).

²⁴ 91 Fed. Reg. at 57526.

²⁵ 5 U.S.C. § 552a(a)(6).

²⁶ 26 U.S.C. § 6103(j)(1).

²⁷ 91 Fed. Reg. at 57529 n.17.

²⁸ 44 U.S.C. § 3561(8).

²⁹ 13 U.S.C. § 9(a)(1).

³⁰ 91 Fed. Reg. at 57526.

³¹ Id.

³² Id.

³³ Id. at 57530.

³⁴ Id. at 57527.

³⁵ Id.

³⁶ Id. at 57526.

³⁷ Dep’t of Com. v. New York, 588 U.S. 752, 784 (2019).

³⁸ J. David Brown & Misty L. Heggeness, Citizenship Question Effects on Household Survey Response, 45 J. Pol’y

Analysis & Mgmt. e70004 (2026), https://doi.org/10.1002/pam.70004.

³⁹ Dep’t of Com., 588 U.S. at 768.

⁴⁰ 91 Fed. Reg. at 57528 n.5.

⁴¹ Press Release, U.S. Census Bureau, Census Bureau Releases Estimates of Undercount and Overcount in the 2020 Census (Mar. 10, 2022), https://www.census.gov/newsroom/press-releases/2022/2020-census-estimates-of-undercount-and-overcount.html.

⁴² See Bolling v. Sharpe, 347 U.S. 497, 499–500 (1954).

⁴³ Wesberry v. Sanders, 376 U.S. 1, 18 (1964).

⁴⁴ 91 Fed. Reg. at 57525.

⁴⁵ Jeffrey S. Passel & D’Vera Cohn, How Removing Unauthorized Immigrants from Census Statistics Could Affect House Reapportionment, Pew Rsch. Ctr. (July 24, 2020), https://www.pewresearch.org/short-reads/2020/07/24/how-removing-unauthorized-immigrants-from-census-statistics-could-affect-house-reapportionment/.

⁴⁶ Final 2020 Census Residence Criteria and Residence Situations, 83 Fed. Reg. 5525, 5526 (Feb. 8, 2018).

⁴⁷ 91 Fed. Reg. at 57529.

⁴⁸ Proposed 15 C.F.R. § 60.2(b), 91 Fed. Reg. at 57529.

⁴⁹ 91 Fed. Reg. at 57528 n.7.

⁵⁰ Id. at 57526.

⁵¹ Proposed 15 C.F.R. § 60.2(b), 91 Fed. Reg. at 57529.

⁵² 83 Fed. Reg. at 5534–36.

⁵³ Proposed 15 C.F.R. § 60.10, 91 Fed. Reg. at 57530.

⁵⁴ Nat’l Acads. of Scis., Eng’g & Med., Assessing the 2020 Census: Final Report 70, 76 (2023), https://doi.org/10.17226/27150.

⁵⁵ 91 Fed. Reg. at 57530.

⁵⁶ Id. at 57524.

⁵⁷ Elizabeth Arias, Carolyn A. Liebler, Marc A. Garcia & Rogelio Sáenz, Data Impacts of Changes in U.S. Census Bureau Procedures for Race and Ethnicity Data, 29 SSM – Population Health 101742 (2025), https://doi.org/10.1016/j.ssmph.2024.101742.

⁵⁸ Nat’l Ctr. for Health Stat., Variance Estimation Guidance, NHIS 2006–2015, at 2–3 (June 13, 2016), https://www.cdc.gov/nchs/data/nhis/2006var.pdf.

⁵⁹ 91 Fed. Reg. at 57527.

⁶⁰ Id. at 57527–28.

⁶¹ Cong. Rsch. Serv., R41532, The American Community Survey: Development, Implementation, and Issues for Congress (Dec. 19, 2014), https://www.congress.gov/crs_external_products/R/PDF/R41532/R41532.20.pdf.

⁶² U.S. Census Bureau, Understanding and Using American Community Survey Data: What All Data Users Need to Know 2, 6, 12–15 (Sept. 2020),

https://www.census.gov/content/dam/Census/library/publications/2020/acs/acs_general_handbook_2020.pdf.

⁶³ Donna M. Daily & Karen Battle, Implementation of SPD 15 in the American Community Survey, U.S. Census Bureau: Random Samplings (Oct. 31, 2024), census.gov; see also Agency Information Collection Activities; Submission to the Office of Management and Budget (OMB) for Review and Approval, 91 Fed. Reg. 27241 (May 14, 2026) (FR Doc. 2026-09582), and accompanying 2027 ACS Supporting Statement A (stating that "the Census Bureau will use the 1997 standards in the American Community Survey").

⁶⁴ See 13 U.S.C. § 141(c); 52 U.S.C. § 10301.

⁶⁵ See 42 U.S.C. §§ 2000d, 2000e to 2000e-17, 3601–3619.

⁶⁶ 91 Fed. Reg. at 57528.

⁶⁷ Revisions to OMB’s Statistical Policy Directive No. 15: Standards for Maintaining, Collecting, and Presenting

Federal Data on Race and Ethnicity, 89 Fed. Reg. 22182 (Mar. 29, 2024).

⁶⁸ Updates to Race/Ethnicity Standards for Our Nation, U.S. Census Bureau (Dec. 20, 2024), https://www.census.gov/about/our-research/race-ethnicity/standards-updates.html.

⁶⁹ 91 Fed. Reg. at 57527, 57529 n.12.

⁷⁰ 89 Fed. Reg. at 22182–83, 22187.

⁷¹ 91 Fed. Reg. at 57524, 57527.

⁷² Id. at 57530.

⁷³ Id. at 57527.

⁷⁴ Id. at 57528–29 n.8.

⁷⁵ Id. at 57528.

⁷⁶ Id. at 57527.

⁷⁷ 13 U.S.C. § 141(f)(1).

⁷⁸ Census Scientific Advisory Committee (CSAC), U.S. Census Bureau (Mar. 6, 2025), https://www.census.gov/about/cac/sac.html; 2030 Census Advisory Committee (2030 CAC), U.S. Census Bureau (Mar. 6, 2025), https://www.census.gov/about/cac/2030cac.html.

⁷⁹ Answer ¶¶ 7, 10, 12–14, Freeman v. U.S. Census Bureau, No. 1:25-cv-07834 (S.D.N.Y. Jan. 12, 2026), ECF No. 26; Lauren Bouton & Elana Redfield, Williams Inst., Removal of Sexual Orientation and Gender Identity from Federal Data Collections: January 2025 to January 2026, at 6 & n.8 (Feb. 2026), https://williamsinstitute.law.ucla.edu/wp-content/uploads/Federal-SOGI-Data-Collection-Feb-2026.pdf.

⁸⁰ Knoedl, supra note 6.

⁸¹ 91 Fed. Reg. at 57529 n.9; Michael B. Hawes et al., Toward a Principled Framework for Disclosure Avoidance, Harv. Data Sci. Rev. (Aug. 20, 2025), https://hdsr.mitpress.mit.edu/pub/bux70n5u/release/3.

⁸² U.S. Dep’t of Com., Dep’t Admin. Order 216-26, Disclosure Avoidance for Statistical Products § 5.02(c) (effective June 4, 2026), https://www.commerce.gov/opog/disclosure-avoidance-statistical-products; 91 Fed. Reg. at

57529 n.13.

⁸³ 91 Fed. Reg. at 57527; Beth Jarosz, Jae June Lee & Chris Dick, Standard Deviations: How Will Methodology Restrictions Affect Demographic and Economic Data?, The Census Project (Aug. 13, 2026), https://thecensusproject.org/2026/08/13/standard-deviations-how-will-methodology-restrictions-affect-demographic-and-economic-data/.

⁸⁴ Ass’n of Population Ctrs. et al., Joint Statement on Commerce’s Disclosure Avoidance for Statistical Products

(June 17, 2026), https://apdu.org/joint-statement-on-commerces-disclosure-avoidance-for-statistical-products/.

⁸⁵ 91 Fed. Reg. at 57530.

⁸⁶ Id. at 57526.

⁸⁷ Knoedl, supra note 6, at 1-2.

⁸⁸ Working Together to Reduce Black Maternal Mortality, Ctrs. for Disease Control & Prevention (Apr. 8, 2024), https://www.cdc.gov/womens-health/features/maternal-mortality.html.

⁸⁹ Germine H. Awad, Nadia N. Abuelezam, Kristine J. Ajrouch & Matthew Jaber Stiffler, Lack of Arab or Middle Eastern and North African Health Data Undermines Assessment of Health Disparities, 112 Am. J. Pub. Health 209, 209 (2022).

⁹⁰ Nadia N. Abuelezam, Abdulrahman M. El-Sayed & Sandro Galea, The Health of Arab Americans in the United States: An Updated Comprehensive Literature Review, 6 Frontiers Pub. Health 262 (2018).

⁹¹ Nicole J. Fleischer & Katherine Sadek, Arab, Middle Eastern, and North African Health Disparities Research: A Scoping Review, 12 J. Racial & Ethnic Health Disparities 1397 (2025).

⁹² Population Denominator Data Sources, Ctrs. for Disease Control & Prevention (June 1, 2026), https://www.cdc.gov/united-states-cancer-statistics/technical-notes/population-denominator-data-sources.html.  ⁹³What Is Shortage Designation?, Health Res. & Servs. Admin. (Sept. 2026), https://bhw.hrsa.gov/workforce-shortage-areas/shortage-designation; Scoring Shortage Designations, Health Res. & Servs. Admin. (Sept. 2026), https://bhw.hrsa.gov/workforce-shortage-areas/shortage-designation/scoring.

⁹⁴ Raj Chetty, Nathaniel Hendren, Maggie R. Jones & Sonya R. Porter, Race and Economic Opportunity in the United States: An Intergenerational Perspective, 135 Q.J. Econ. 711, 711 (2020).

⁹⁵ Neda Maghbouleh, Ariela Schachter & René D. Flores, Middle Eastern and North African Americans May Not Be Perceived, nor Perceive Themselves, to Be White, 119 Proc. Nat’l Acad. Scis. e2117940119 (2022); Khaled A. Beydoun, Boxed In: Reclassification of Arab Americans on the U.S. Census as Progress or Peril?, 47 Loy. U. Chi.

L.J. 693, 718–23 (2016).

⁹⁶ Press Release, Nat’l Urb. League, Historic Census Undercount of Black Americans Robs Communities of Billions in Funding and Fair Political Representation (2021), https://nul.org/news/historic-census-undercount-of-black-americans-robs-communities.

⁹⁷ Patrick Cooney et al., Univ. of Mich. Poverty Solutions, Analysis of the Census 2020 Count in Detroit 1, 4 (Dec. 2021), https://sites.fordschool.umich.edu/poverty2021/files/2021/12/PovertySolutions-Census-Undercount-in-Detroit-PolicyBrief-December2021.pdf.