Roundtable #35: Before the Ballot: Persons, Felons, and Federal Power in the Making of the Electorate
Section I: Who Counts? One Person, One Vote, or One Vote, One Person
When planning the 2020 Census, the Department of Commerce attempted including a citizenship question but was blocked by the Supreme Court of the United States in Department of Commerce v. New York (2019) because the citizenship question was deemed pretextual and therefore unjustified. To bypass the ruling, President Trump issued Executive Order 13880, instructing various federal departments to provide the Department of Commerce with citizenship and immigration data. Although that Executive Order was overturned by President Biden, efforts to limit or completely stop the count of noncitizens in federal data persist. In his second term, President Trump instructed the Department of Commerce to stop including those residing in the country illegally in U.S. census data. Similarly, Republicans have called to stop including noncitizens in census counts for reapportionment. These actions would heavily affect reapportionment—the process determining the allocation of seats to each state in the House of Representatives every decade—because it is based on population data. The results of reapportionment also impact redistricting, further affecting representation. Efforts to exclude noncitizens in reapportionment, and subsequently redistricting, have brought forth debates on whether excluding noncitizens violates the U.S. Constitution. The Fourteenth Amendment requires the counting of all “persons” residing in each state and Supreme Court precedent confirms that congressional representation must be based on total population, not citizenship status or voting eligibility, meaning excluding noncitizens from reapportionment and redistricting should be unconstitutional.
Excluding noncitizens from census apportionment defies stipulations of the Fourteenth Amendment that outline how apportionment is total population based. Section 2 of the Fourteenth Amendment directly says, “Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State….” Nowhere in Section 2 does the Fourteenth Amendment command or allow states to allocate seats based on its citizen or voting population, indicating that apportionment must be based on total population data. The word “persons” being used is a clear indicator that the Constitution intended including more than just citizens. The Penalty Clause of the Fourteenth Amendment also uses the word “citizens” to determine who gets counted and who gets to vote, showing that the framers were deliberately aware of the difference between “citizens” and “persons”. Furthermore, Section 1 of the Fourteenth Amendment uses the word “citizens” when talking about special privileges and immunities. The framers could have used “citizens” in Section 2 for apportionment but deliberately chose “persons” instead. This distinction demonstrates how the framers of the Constitution wanted apportionment to include citizens and noncitizens. This is acknowledged by the U.S. Census Bureau which outlines those included in the Census as individuals who have a usual residence in the United States, comprising people with permanent residency status, work visas, student visas, or are undocumented. Because the language of the Fourteenth Amendment directs apportionment to be total population based which is acknowledged by the U.S Census Bureau to include citizens and noncitizens, choosing to exclude noncitizens in apportionment would be unconstitutional.
Moreover, the Fourteenth Amendment aimed to protect the rights of newly freed African Americans. It rectified the problematic nature of the Three-Fifths Clause by ensuring African Americans counted as whole people. Although slavery was outlawed and the Three-Fifths Clause became inapplicable, the Fourteenth Amendment ensured that individuals are not denied their right to representation by specifically amending the Constitution to require complete counting of everyone. Therefore, any attempt to exclude individuals in reapportionment must be done through a constitutional amendment. Despite the Court not having directly decided whether “persons” prohibits all exclusions, the term “whole number of persons in each State” reinforces counting each person, not each voter, when determining reapportionment. Also, avoiding the word “voter” means other populations who could not or still cannot vote are counted. Historically, nonvoters like children have always been counted. And before the Nineteenth and Fifteenth Amendments, women and free Black men were fully counted for apportionment purposes despite being unable to vote in most states.
As civil rights expanded, the Court has upheld that redistricting must equally represent population sizes and population must include the area’s total population. The Supreme Court asserted this in Evenwell v. Abbott (2016) where the Court ruled that the Equal Protection Clause of the Fourteenth Amendment allows states to draw their districts in consideration of total population, not just the population of those eligible to vote. It should be noted that the Court did not require states to use total population for redistricting just that it could—meaning that excluding noncitizens is not deemed unconstitutional. Despite this, the Court reasoned that representatives serve all the residents of their district, not just its citizens. From this, a principle of representational equality arises. If districts ignored total population and focused on voter population, those not included would be deprived of their right to equal representation which violates the Equal Protection Clause. To uphold representational equality, states should use total population for redistricting. This is further supported by Wesberry v. Sanders (1964) and Reynolds v. Sims (1964) where the Court determined that districts must contain roughly equal populations for federal and state districts respectively. Determining equal population should be based on total population. Otherwise, there would be representation disparities between districts with large and small noncitizen populations, depriving constituents of their equal protection. Although these cases did not specifically address noncitizens, their reasoning supports how redistricting must fulfill representational equality and therefore account for the total population. Meanwhile, the case Plyler v. Doe (1982) addresses how undocumented persons are still protected by the Equal Protection Clause as it affirmed the Clause applies to any person in a state’s jurisdiction, including undocumented persons. Assuming that undocumented persons are accounted for, then the broader category of noncitizens is also protected. Because noncitizens are considered “persons” and the Fourteenth Amendment requires that all “persons” be counted in reapportionment, noncitizens must also be included in reapportionment.
While there have always been challenges to limiting the representation of noncitizens in reapportionment or redistricting, textual evidence, history, and precedent have reiterated that noncitizens can and should continue to be counted and represented. Regardless, political debates over immigration and representation are likely to continue and may reach the Supreme Court in the future as seen in Trump v. New York (2020) where the Trump Administration announced that it planned on excluding undocumented individuals from the 2020 Census. The case was dismissed as premature as no one individual had actually been excluded from the count, meaning that nothing had been declared as unconstitutional. However, the case reveals how reapportionment and redistricting could be affected in the future because three justices dissented from the dismissal, indicating they would have reached the merits and found the exclusion unconstitutional. The constitutional question remains unresolved, and the future of noncitizen representation may ultimately depend on whether the Court agrees to hear a case on the merits.
By Emily Song
Section II: Who Participates? The Unconstitutionality of Standardless Felon Disenfranchisement
In April 2024, the Nebraska legislature passed Legislative Bill 20, titled “Provide for restoration of voting rights upon completion of a felony sentence or probation of a felony”, which would repeal the state’s two-year waiting period for felon reenfranchisement and provide the automatic restoration of voting rights upon a felon’s completion of their sentence. Two days before LB 20 was meant to take effect, on July 17, 2024, Attorney General Mike Hilgers issued a nonbinding advisory opinion concluding that LB 53, the law imposing the two-year waiting period, and LB 20, the bill meant to rescind it, were unconstitutional because Nebraska’s constitution authorizes the state’s board of pardons to restore voting rights. After the American Civil Liberties Union (ACLU) and the law firm Faeger filed a joint suit with the Nebraska Supreme Court on October 16, 2024, the Nebraska Supreme Court entered an order instructing election officials to enact these laws. The six-month chaos that ensued in Nebraska’s government represents the inconsistencies within an individual state’s system, shown by a singular person's power to block the implementation of legislative reform.
This issue does not exist only within the confines of Nebraska: forty-eight U.S. states currently have differing laws banning people with felony convictions from voting. Approximately four million Americans, or 1.7 percent of the voting-age population, are disenfranchised. The Constitution does not require Congress to remain passive while access to the franchise is distributed unevenly through arbitrary and inconsistent restoration practices. Where state disenfranchisement regimes generate unequal access to voting both within individual systems and across the nation as a whole, Section Five of the Fourteenth Amendment empowers Congress to establish a national minimum standard for voting rights restoration.
Felon disenfranchisement in America is not a single policy but rather a fractured collection of fifty different ones, ranging from states that never strip voting rights to states that impose lifetime bans requiring personal intervention by a governor. Ordinarily, this kind of variance raises no constitutional concern: states are free to set their own election laws. A Floridian has no cognizable equal protection claim simply because Vermont chose differently—the Equal Protection Clause binds each state to treat people equally under its own laws, not to match its neighbors. The state of Vermont, for example, is one of two states in the entire country that never restricts felons of their right to vote. Under Vt. Stat. Ann. tit. 28, § 807, any person in Vermont is guaranteed their right to vote by absentee ballot “provided the person otherwise fulfills all voting requirements.” On the opposite end of the spectrum, Florida—with a disenfranchisement rate of 6.13 percent—forces ex-felons to jump through hoops if they want any chance at reinstating their right, requiring under Fla. Stat. §98.0751 the completion of every term of a sentence, including full satisfaction of restitution, fines, fees, and costs before restoration is even possible. This disparity is not merely interstate, rather it exists within Florida’s own system. In Jones v. Governor of Florida (2020), the Eleventh Circuit, sitting en banc, upheld this financial-completion requirement against a wealth-discrimination challenge, holding that felons retain no fundamental right to vote once disenfranchised and that rational basis review applies. Under that standard, a Florida felon who can immediately pay his fines and restitution regains his vote, while another Florida felon—convicted of an identical offense, subject to an identical sentence—remains permanently disenfranchised for the sole reason that he lacks the money to pay. The Eleventh Circuit found this rational; however, it is not equal. The court’s holding does not resolve whether such a scheme is just, only that it survives the most permissive standard of judicial review available. That gap, between what the courts will tolerate and what equal protection should demand, is precisely the space Congress is empowered to fill under Section Five of the Fourteenth Amendment.
Before the constitutional injury of the current system can be fully appreciated, it is necessary to recognize the special constitutional status that the Supreme Court has afforded the franchise. In Reynolds v. Sims (1964), the Supreme Court emphasized that the right to vote is “preservative of all rights” and held that the Equal Protection Clause prohibits electoral systems that dilute or unequally value citizens’ votes. Although Reynolds arose in the context of legislative apportionment rather than felon disenfranchisement, it reflects a broader constitutional principle: because voting is central to democratic self-government, arbitrary inequalities in access to and participation in the electoral process demand careful scrutiny. Voting occupies this uniquely protected constitutional status; the way access to it is administered should not be left to arbitrary or standardless discretion. Yick Wo v. Hopkins (1886) supplies this principle. In Yick Wo, San Francisco required laundries operating in wooden buildings to obtain permits from local officials. Although the ordinance was facially neutral, officials granted permits to virtually every non-Chinese applicant while denying them to nearly every Chinese applicant. The Supreme Court held that a law fair on its face may nevertheless violate equal protection when administered “with an evil eye and unequal hand.” The constitutional defect was not the text of the ordinance itself, but the arbitrary and discriminatory manner in which it was enforced against similarly situated individuals. Felon disenfranchisement laws are facially neutral in a similar sense. They do not, on their face, distinguish between individuals based on race or other suspect classifications. Yet the restoration of voting rights is often administered in ways that produce dramatically different outcomes among similarly situated individuals. All eleven Confederate states adopted post-Reconstruction disenfranchisement provisions to suppress Black voting, satisfying Yick Wo’s discriminatory intent standard for those states. The recognition of voting’s fundamental status, however, has come chiefly in the apportionment context; where felon disenfranchisement is concerned specifically, courts have applied a far more permissive standard of review, as Jones illustrates. That gap between voting’s fundamental status in principle and the deferential review it receives in this context is not a reason to abandon the argument, but the problem Section Five empowers Congress to address.
The inequality Jones exposes within Florida’s own system is only half of the constitutional problem, however. Even if each state’s disenfranchisement scheme were, standing alone, perfectly constitutional, the aggregate effect of fifty different systems operating side by side is itself a structural harm, independent of any single state’s rules, that Congress has the power to address under Section Five. The absence of any uniform standard means that a person's access to a fundamental right depends entirely on the accident of geography—the same underlying conduct results in permanent disenfranchisement in one state and none at all in another, determined not by any principled distinction but by which side of a state line a person happens to live on. Bush v. Gore (2000) confirms that this kind of standardless variation is itself a constitutional defect, independent of whether the underlying process is permissible at all. In this case, the Supreme Court held that once a state has granted the right to vote, it “may not, by later arbitrary and disparate treatments, value one person’s vote over that of another,” and found that Florida’s recount failed this standard not because a recount was impermissible, but because “the absence of specific standards to ensure its equal application” left the process to vary arbitrarily from county to county, and even from one recount team to another within the same county. Notably, the Court grounded this holding in the same principle established in Reynolds: that the right to vote can be violated by dilution or unequal treatment just as surely as by outright denial. The felon disenfranchisement patchwork presents the same structural defect the Court condemned in Bush, magnified fifty-fold. Where Florida’s counties could not agree on how to interpret a dimpled chad, the small piece of paper that is supposed to detach when a voter punches their selection, American states cannot agree on whether a felony conviction extinguishes the right to vote permanently, temporarily, or not at all—and the burden of that inconsistency does not fall evenly. Although the Court limited Bush to its “present circumstances,” the equal protection principle it articulated is not unique to recounts. If arbitrary variation in counting votes raises constitutional concerns, arbitrary variation in restoring the right to vote raises similar concerns.
The persistence of this structural inequality is not merely a policy failure but rather the kind of constitutional problem that Section Five of the Fourteenth Amendment empowers Congress to remedy without waiting for the courts to act first. Katzenbach v. Morgan (1966) epitomizes Section Five’s power: in the 1960s, New York required people to pass an English literacy test in order to vote. This disproportionately affected Puerto Ricans who immigrated to the United States and as a result Congress passed the Voting Rights Act of 1965. Section 4e prohibited states from denying Puerto Ricans their right to vote simply because they could not pass an English literacy exam. New York argued that Congress exceeded its constitutional authority by overriding the state’s voting qualifications. However, the Supreme Court decided that Section Five “is a positive grant of legislative power authorizing Congress to exercise its discretion.” Notably, no court had struck down New York’s literacy test as unconstitutional; Congress banned it anyway, because Section Five’s grant of power to Congress is broader than the judiciary’s power under Section One. This same principle is applicable to the national standard that Congress has the power to establish. A national standard for disenfranchisement law would protect felons’ Fourteenth Amendment rights under the Equal Protection Clause and would create a system where a person’s right to vote would no longer be determined by their geographic positioning.
However, this argument must confront a limitation the court imposed three decades after Katzenbach. In City of Boerne v. Flores (1997), the Court held that Section Five legislation must be remedial or preventive rather than a vehicle for Congress to redefine the substance of a constitutional right, and that any such legislation must display “congruence and proportionality” between the injury Congress seeks to remedy and the remedy it imposes. A national minimum standard for felon disenfranchisement survives this test on both counts. It is remedial, not substantive: it does not declare that felons possess an unqualified right to vote, and it does not disturb the states’ underlying authority to disenfranchise felons in the first place—a question Richardson v. Ramirez (1974) confirms Section Two of the Fourteenth Amendment leaves to them, holding that felon disenfranchisement does not violate the Equal Protection Clause. A national minimum standard does not contest that holding or argue that disenfranchisement is unconstitutional; it instead standardizes the administration of restoration. Concretely, such a standard could require states to automatically restore voting rights upon a felon’s release from prison, without conditioning restoration on payments of fines, fees, or restitution—the same financial-completion requirement Jones upheld under rational basis but never deemed unjust. The rule would eliminate the arbitrary, standardless variation that Jones, Yick Wo, and Bush each condemned in their own context.
Taken together, these cases reveal two distinct constitutional injuries: unequal treatment within state restoration systems and structural inequality produced by the national patchwork of disenfranchisement laws. Although courts have largely tolerated these disparities, Section Five empowers Congress to address them. Under Katzenbach and Boerne, Congress may establish a national minimum standard for voting rights restoration that remedies arbitrary inequality while preserving state authority over disenfranchisement itself. If Congress does not act, the current system will continue producing the same inequalities and the courts, as per Jones, will continue to tolerate it under rational basis review.
By Andrew Nadim Elkareh
Section III: Who Holds the Roll? Federalism and the Demand for State Voter Data
On June 16, 2026, Axios released a report detailing Immigration and Customs Enforcement’s (ICE) access to voter files in Texas and North Carolina: on request, Homeland Security Investigations (HSI) received the voter rolls to continue “actively rooting out and investigating election fraud.” However, there exist concerns around the intentions of the Department of Justice (DOJ), as Democracy Forward Senior Oversight Counsel Dan McGrath told Axios, “using ICE to pursue a problem this rare should concern everyone.” In light of this increased federal access to state documents, federalist boundaries are left in jeopardy. To examine recent legal precedent addressing this issue, this article analyzes the outcomes of lawsuits filed by the DOJ against states that failed to provide sensitive voter registration information to the federal government. Based on the outcomes of these lawsuits, it stands that while the federal government can monitor state actions under Title III of the Civil Rights Act of 1960 (CRA), it cannot examine voter rolls that exist as documents created by the state rather than from a third party. In addition to a statutory collision between the CRA, the National Voter Registration Act (NVRA), and the Help America Vote Act (HAVA), and the purposivist interpretations by the courts, the DOJ lacks the jurisdiction to demand the voter rolls.
In late 2025, the DOJ demanded the voter rolls of “at least 43 states.” While states like Texas and Florida complied, thirty states either refused the demand or provided only publicly available information. The DOJ then sued them for failure to comply. Their arguments depended on three statutes: the National Voter Registration Act (1993), the Help America Vote Act (2002), and Title III of the Civil Rights Act (1960). As of June 30, eleven courts have already dismissed the case. In Michigan’s District Court, the case United States v. Benson (2026) was dismissed for failure to state a claim. The DOJ then appealed to the Sixth Circuit Court of Appeals and “challeng[ed] only the dismissal of its Title III claim.” However, their appeal was again dismissed; several courts have now deferred to this case as precedent. Because of its influence on recent court decisions, this article will examine the opinion and its analysis of the DOJ’s arguments.
One of the first reasons the DOJ’s arguments were dismissed is due to a legal distinction in the language of Title III. For context, Title III mandates election officers “‘retain and preserve…[1] all records and papers which [2] come into h[er] possession,’” On a surface level, the CRA appears to expand federal power. But the legal distinction between “coming into” and “being in possession of” was a line Congress drew to ensure they did not encroach upon state sovereignty. As written in Title III, the Attorney General has the authority to demand documents states “come into” from third parties, such as poll taxes and voter applications. This is to ensure states complied with the CRA’s fair conditions, and did not modify or destroy documents that contained evidence of discriminatory practices. However, voter rolls exist as documents created by the state rather than received externally. As Judge Mathis found in United States v. Benson (2026), because the state created these voter rolls, they cannot be demanded under the CRA. He articulates the scenario with a metaphor about a baker. When a baker purchases the ingredients to bake a cake and “proceed[s] to bake cakes”, it would be unreasonable to say she has “come into” possession of the cakes she has created. Had the DOJ initially demanded individual voter registration applications (the ingredients), the outcome may have been different. For example, ICE’s recent access to voter information in Texas and North Carolina appears to follow a narrower path: rather than requesting state-created voter rolls, ICE accessed information on select individuals. This suggests that they relied on third-party voter documents that would fall within the CRA’s “come into possession” language; the purpose being, however, for immigration enforcement.
Another reason for the dismissal of the DOJ’s claims was the statutory collision that would occur had interpretations of the NVRA, HAVA, and CRA all been agreed with. The DOJ argued that because Michigan “refus[ed] to provide records regarding their list maintenance programs… to assess its compliance,” it violated the NVRA and HAVA. For context, these acts require voter information to be updated with “‘reasonable effort[s]’ to ensure accuracy”; Section 20510(a) of the NVRA also states that the Attorney General can “bring a civil action” if there appears to be a violation. However, the court found that to argue Michigan violates both these acts and the CRA would make the outcome contradictory. Compared to the NVRA and HAVA, the CRA requires the preservation of records to check for discriminatory practices. Interpreting all three statutes together to establish federal jurisdiction “would place Title III on a collision course with the NVRA and HAVA,” the court wrote in United States v. Benson (2026). Under the principles of statutory construction, courts avoid readings that place congressional enactments in direct conflict. However, the DOJ’s interpretation would do exactly that. How could an elected official both preserve voter rolls and actively modify them to “ensure accuracy”? The official would be placed “in violation of one federal law for trying to comply with others.” Essentially, by agreeing with the DOJ’s interpretations, elected officials are left with contradictory expectations.
Lastly, a reason the DOJ’s lawsuit was dismissed in California is due to the court’s purposivist interpretation of the law. First, the DOJ’s intentions for using the CRA directly conflict with its original purpose. The CRA was designed to protect racial minorities’ voting rights: Title IV explicitly states that if a person has “been deprived on account of race or color,” they would, empowered by the Attorney General, “be entitled…to an order declaring him qualified to vote.” However, in the lawsuit against California, United States v. Shirley (2026), instead of using the CRA to combat racial discrimination or voter fraud, Judge Carter writes that the data would be used “for broader immigration enforcement.” This rejection of the DOJ’s interpretation reflects the legal theory of purposivism, which holds that “a court’s statutory interpretation should reflect the statute’s original purpose.” The statute’s reach, therefore, does not extend beyond what Congress intended. Otherwise, statutes could be abused “without approval from Congress or Americans themselves.”
In conclusion, the Sixth Circuit decision in United States v. Benson reinforces the boundaries of federalism: the language of the CRA was written to grant the Attorney General access to third-party documents, but not documents created by the state. This was written as such to avoid stripping state sovereignty over election administration by requiring only what was necessary. Beyond the boundaries of Title III, the DOJ’s arguments also failed under the NVRA and HAVA due to a statutory collision with the CRA, and due to the court’s purposivist interpretation of the statutes.
Despite many of the lawsuits being dismissed, the chilling effect of the DOJ’s actions remains: compliance is less costly than litigation, putting pressure on officials to hand over documents voluntarily rather than considering the legal context. In a statement directed at Tim Walz, former Attorney General Pam Bondi asked him to help “bring an end to the chaos” by “turning over the records.” The chaos, however, started with the DOJ when states that refused the request were sued. A tracker the Brennan Center developed shows that several cases are still ongoing since last year. These circumstances create an incentive for elected officials to comply voluntarily, effectively removing the judicial check on the executive branch’s demands. ICE’s recent access to voter rolls may be a demonstration of this lack of monitoring. It is therefore crucial that the courts continue to reject the DOJ’s interpretation on a consistent and timely basis, both to strengthen the precedent established in Benson and to counteract the chilling effect on state officials.
By Emily Nicole Kam
Edited by Gabi Fabozzi
This piece was reviewed and finalized by Qizhen (Kiara) Ba.
The views in these articles are those of the individual authors and not of the Columbia Undergraduate Law Review.