Federal election oversight collides with voter data privacy as the Department of Justice layers legacy civil rights statutes to demand full, identity-linked state registration databases.
Voter records have always told a story. In 1960, they told a story of exclusion: who was kept from the ballot and how. By 1993, they told a story of access: whether the promise of registration was actually reaching eligible voters. By 2002, it told a story of accountability: whether the elections system could function accurately, uniformly and reliably.
Today, the Department of Justice demands those same records to tell a different story, one written in the broad language of federal oversight but carrying deeply personal consequences for the voters whose information may be exposed. Now, the story of voter records turns on data privacy: whether statutes enacted to protect voter rights, expand access, and modernize elections can be read to permit access to complete, identity-linked voter information.
The Requests: The Department of Justice’s Statutory Layering
The Department of Justice’s demands seek both statewide voter-registration data and information about state list-maintenance practices. In letters to several states, the Department of Justice requested electronic copies of statewide voter registration lists, including “all fields” in those databases. For example, its August 14, 2025 letter to New York stated that the list must include each registrant’s full name, date of birth, residential address, and driver’s license number or the last four digits of the registrant’s social security number. The Department of Justice has also requested records concerning duplicate registrations, deceased voters, confirmation notices, inactive voters, and voters identified as ineligible because of non-citizenship, felony conviction, or incompetency. In other words, the Department of Justice is not merely asking states to describe their compliance efforts; it is seeking access to the voter data those systems contain.
The Department of Justice relies on a mixture of three distinct but overlapping authorities to obtain this information: The Civil Rights Act of 1960, the National Voter Registration Act of 1993, and the Help America Vote Act of 2002.
The Civil Rights Act of 1960: The Access Mechanism
Congress enacted Title III of the Civil Rights Act of 1960 against the backdrop of Jim Crow-era voter suppression, when Black citizens were systematically denied access to the ballot through discriminatory registration practices and arbitrary administration by local officials. In the Senate debates, supporters described the 1960 bill as adding “essential features” to existing civil-rights law toward ending discrimination in the “voting privileges” of citizens. However, the House Report identified the practical enforcement problem more directly, explaining that evidence of discriminatory voting practices was “practically impossible to assemble” because evidence of discrimination was in the hands of the state and local officials administering registration. Title III resolved this issue by requiring election officials to preserve records concerning applications, registration, tests and other acts requisite to voting, and to make them available to the Attorney General upon written demand.
The Department of Justice relies on that same preservation-and-access framework to support its demand for modern voter registration data. The Department builds on this authority by interpreting 52 U.S.C. §§ 20701 and 20703 to mean that states must produce materials that qualify as “records or papers relating to voting.” Further, the Department of Justice draws its authority from § 20703, which, in turn, authorizes the Attorney General to demand access to those records, provided that the request includes a statement of its “basis and purpose.” However, Title III does not define “records,” but the Department of Justice’s recent demands take an expansive view of that language, treating the complete statewide voter-registration databases with “all fields” as records necessary to assess compliance with federal voting laws.
That expansive understanding of “records” is especially significant because § 20703 is enforced through a narrow production proceeding, not ordinary civil litigation. In Kennedy v. Lynd, the Fifth Circuit analogized § 20703 proceedings to an order to show cause or a production order in aid of an administrative investigation. The court further emphasized that judicial review is limited to whether the demand was properly made and whether the requested materials fall within the statute’s scope. In that sense, the Attorney General’s role under the statute is “purely investigative,” rather than adjudicative.
Consistent with the court’s reasoning, the Department of Justice does not treat § 20703 as requiring a prior finding of a violation. Instead, the statute is understood to authorize broad access to election records for investigative purposes. However, § 20703 only supplies access to the records, but the Department of Justice’s stated reason for seeking these records is grounded in later-enacted voter-registration statutes.
The NVRA and HAVA: The Compliance Playbook
The Department relies on the National Voter Registration Act of 1993 (NVRA) and the Help America Vote Act of 2002 (HAVA) as part of its authority to obtain voter registration data. The NVRA was enacted to “establish procedures that will increase the number of eligible citizens who register to vote in elections for Federal office,” “[enhance] the participation of eligible citizens as voters,” “protect the integrity of the electoral process,” and “ensure that accurate and current voter registration rolls are maintained.” House supporters described the bill as bringing democracy “into the 21st century” by removing historical deterrents to registration.
To balance expanded access with election integrity, the NVRA requires states to conduct a general program that makes a “reasonable effort” to remove voters who have become ineligible, due to circumstances such as death or a change in residence. In its recent requests, the Department of Justice specifically invokes 52 U.S.C. § 20507(i)(1) under the NVRA, requiring states to make available “all records concerning the implementation of programs and activities conducted for the purpose of ensuring the accuracy and currency of official lists of eligible voters” for public inspection.
Even though the Department of Justice may demand these records without a formal suit, the NVRA does not create a general right to obtain every field in a state voter-registration database. The NVRA’s disclosure provision is limited by its subject matter; it is tied to records concerning the implementation of list-maintenance programs and activities, not necessarily all personal identifiers contained in a statewide voter file. In that sense, the NVRA provides transparency, but not automatic access to the full substance of every voting record.
HAVA adds the modern database layer. Enacted after the 2000 Bush v. Gore election, HAVA was designed to modernize voting systems, improve voter access, create the Election Assistance Commission, and establish minimum standards for the administration of federal elections. The Act requires states to maintain centralized, computerized statewide voter registration databases containing identifying information, such as driver’s license numbers or the last four digits of Social Security numbers, to verify voter eligibility and maintain accurate voter rolls.
In its requests, the Department of Justice cites to 52 U.S.C. § 21083(a)(5)(A)(i), arguing that the Department must be able to review voter registration data in order to determine whether states are complying with the federal requirements under both the NVRA and HAVA. But HAVA explains why the data exists; it does not independently provide the Department of Justice with the same access under the Civil Rights Act or the NVRA. Rather, HAVA authorizes the Attorney General to bring civil actions to enforce its requirements, which requires the Department to allege a violation or the need for injunctive relief.
DOJ’s Integrated Statutory Approach
The Department of Justice does not rely on a single statutory source of authority to obtain voter registration data. Although each statute imposes distinct limits on federal authority, their combined use allows the Department of Justice to construct a broader theory of access than any one statute would support on its own.
The Civil Rights Act of 1960 provides the access mechanism for the voting-related records. It is unclear whether “all fields” is necessary to fulfill this, or whether states’ redacted lists may suffice. The NVRA supplies list-maintenance obligations and a public-inspection provision, but that provision is limited to list-maintenance programs that may not require “all fields” in order to comply. Lastly, HAVA requires states to maintain centralized voter registration databases with identifying information, but it does not independently create a comparable front-end demand power.
The Department of Justice’s demands rely on drawing these separate “pieces” together. By using the NVRA and HAVA’s compliance requirements as their “basis and purpose,” the Department uses the later-enacted statutes to define the practical scope of what it claims may be obtained under the Civil Rights Act. The result is a form of statutory layering: legal constraints that appear meaningful in isolation have become less protective when these statutes are invoked simultaneously. And that is where the privacy problem begins—not with any one statute alone, but with what their combined use allows the Department of Justice to demand.
Fragmented Privacy Protection
States are not merely refusing federal oversight; they are asserting that voter registration databases contain sensitive personal information that state law often requires them to protect. The data that states are required to collect (and protect) is different in kind from the paper registration records contemplated when the Civil Rights Act of 1960 was enacted. As a result, the states’ privacy interests are not incidental to the statutory dispute, they are central to it.
The States’ Concerns
State election laws often distinguish between public voter information and protected voter information. For example, Delaware stated in response to the Department’s demand that its publicly available voter registration list includes names, addresses, party affiliation, voting history, legislative district information, and year of birth, but does not include Social Security numbers or driver’s license numbers. Likewise, North Carolina law excludes certain voter information, including dates of birth and Social Security numbers, from public disclosure. The limitations of state law reflect a basic premise: voter registration lists may serve public election-administration purposes, but not every field in those systems is public election information. States’ privacy concerns are not limited to any single data point. A name or address may appear on a public voter list. A date of birth may be available in some government records. The last four digits of a Social Security number or a driver’s license number may be collected for verification purposes. But when these data points are compiled into one centralized database, the privacy risk changes. Aggregation makes the information searchable, cross-referenceable, and more useful for matching individuals across government systems.
States have framed their privacy-based objections with a federalism undertone. The federal government is demanding state-held data that state officials are legally obligated to protect. In dismissing the Department’s suit against California, the district court warned that the Department’s theory risks the “unbridled consolidation of all elections power in the Executive” and emphasized that Congress, not the Executive, must clearly authorize the centralization of private voter information. Delaware raised a related but more statutory objection, explaining that the Department’s requests must be evaluated within the “statutory framework” Congress created, because each grant of federal investigatory authority carries its own limits and “checks and balances.” That framing matters because election statutes do not operate in a vacuum. They sit beside federal privacy laws that regulate how agencies collect, maintain, use, and disclose personal information.
The Privacy Act of 1974 and The Driver’s Privacy Protection Act of 1994
Federal privacy laws reinforce the states’ concern that the Department of Justice’s demands are not merely election records but rather requests for identity-linked personal information. States have invoked the Privacy Act of 1974 and the Driver’s Privacy Protection Act of 1994 to challenge demands for protected personal information and enforce limits on agency collection of personal data.
The Privacy Act of 1974 was enacted against the backdrop of growing concern over federal data collection, automated record systems, and the misuse of personal information by executive agencies. The Act regulates agency “systems of records” disclosure without consent, and requires agencies to maintain only information that is relevant and necessary to accomplish a statutory purpose. This matters because the Department of Justice is not asking for isolated voting records, but rather it is requesting “all fields” that include identifying information. Even with the Department’s authority to inspect election records, the Privacy Act shows that Congress was concerned with the exact danger raised by these types of requests, which is the federal accumulation and use of personal information in government databases.
The Driver’s Privacy Protection Act of 1994 (DPPA) reflects a similar recognition of privacy, but in a more specific context. The DPPA restricts state motor vehicle departments from disclosing personal information obtained in connection with motor vehicle records, subject to enumerated exceptions. The Supreme Court has described the DPPA as a regulatory scheme that restricts states’ ability to disclose driver’s personal information without consent, after Congress found that states had been selling that information to individuals and businesses. The DPPA may also implicate voter-registration data obtained from motor vehicle records. Thus, when the Department of Justice demands voter files, including driver’s license numbers, it is seeking information that Congress has separately treated as privacy sensitive.
To be clear, neither the Privacy Act nor the DPPA creates an absolute bar to disclosure. The Privacy Act allows disclosure without consent in limited circumstances, including for a published “routine use,” a written law-enforcement request, disclosure to Congress, or disclosure under a court order. The DPPA also permits disclosure of motor-vehicle information for enumerated purposes, including use by any government agency carrying out its functions. However, the Department of Justice’s reliance on these exceptions does not answer the harder question: whether “all fields” in a statewide voter-registration database are necessary to evaluate compliance with the Civil Rights Act, the NVRA, or HAVA.
Further, states have emphasized that these exceptions are not blank checks. New York, for example, objected to the Department of Justice’s demands by asserting the Department’s “system of records notices” did not cover the requested unredacted voter file. Delaware repeatedly asked the Department whether the voter registration information would be subject to the Privacy Act of 1974 and asked the Department to identify the system of records under which the information would be collected and maintained.
Read together, the Privacy Act and the DPPA do not necessarily bar the Department of Justice from obtaining voter information. But they do show that Congress has repeatedly treated personal identifiers as deserving of protection. That recognition matters when the Department of Justice relies on an older election-records statute to demand modern voter databases containing “all fields.”
The problem is not that privacy laws provide no protection; it is that the protection is fragmented. These Acts regulate certain uses, sources, and disclosures, but they do not clearly answer whether an old election-records statute should permit front-end access to an entire modern voter database. As a result, the Department of Justice’s statutory theory pushes through the gaps between these Acts, seeking complete voter data before any single privacy law clearly resolves the front-end access question.
The Problem: Same Access Authority, Different Data Consequences
The central problem is not whether the Department of Justice may ever obtain voter-registration records. Without a doubt, voter-registration databases are “records relating to voting,” but a record’s status as producible does not mean every sensitive field within that record must be disclosed unredacted. The Civil Rights Act, the NVRA, HAVA, the Privacy Act and the DPPA each speak to parts of the issue, but none squarely answer the question presented by modern voter databases: when a state record contains both election-administration information and sensitive identity-linked fields, who decides which fields must be disclosed, and under what standard?
That silence highlights the underlying tension of treating a modern database as one indivisible record, or as a container of information made up of separable fields. Courts interpreting the NVRA have recognized that disclosure of the voter-registration records does not necessarily require disclosure of every sensitive field within those records. In Public Interest Legal Foundation, Inc. v. Bellows, the First Circuit observed that “nothing in the text of the NVRA prohibits the appropriate redaction of uniquely or highly sensitive personal information in the Voter File,” and that proper redaction can help “assuage the potential privacy risks” created by public release. But the Department of Justice’s demands imply that the full database is treated as the relevant “record,” so redaction is non-production.
The danger of the statutory gap the Department of Justice attempts to use is that it allows the Executive to demand complete voter databases while leaving states with little meaningful recourse to protect citizens’ privacy interests. By treating anything short of an unredacted “all fields” database as non-production, the Department collapses the distinction between producing voting-related records and exposing every identity-linked data point within them. And once that data is transferred, the privacy risk is already set. Downstream safeguards may regulate use or disclosure, but they cannot fully undo the consequences of improper access, breach, or repurposing. Congress may need to ultimately close this gap, but until it does, statutory silence should not be misconstrued as congressional permission to centralize complete, unredacted sensitive identity-linked information in the Executive Branch.
These concerns are not merely theoretical. Recent federal data incidents show that sensitive information held by government agencies can be exposed, misused, or accessed through third-party systems. In 2023, the Centers for Medicare & Medicaid Services notified individuals that their personally identifiable information may have been exposed through the MOVEit file-transfer breach involving a federal contractor. In 2024, the Treasury Department disclosed a “major cybersecurity incident” in which hackers accessed Treasury workstations and unclassified documents through a compromised third-party software provider.
The Office of Personnel Management (OPM) breach makes the risk even clearer. A House Oversight Committee timeline traces the first known adversarial access to OPM’s network to July 2012 and identifies breach activity beginning in December 2013 that was not identified until April 2015. Even more troubling, during that period, adversaries did not merely access ordinary personnel records, they obtained deeply sensitive identity-linked information, including Social Security numbers, background investigation files and financial information. If highly sensitive federal personnel data can remain exposed for that long, then the compelled transfer of complete voter databases should not be treated as a low-risk administrative step.
These examples do not mean the Department of Justice will misuse voter data. But recent allegations involving Social Security data show why the concern is not limited to outside hackers. Sensitive government databases may also be vulnerable to improper internal access or unauthorized sharing, even where the agency disputes the allegations. The examples show why privacy protections cannot depend solely on promises of proper handling after the data is already in federal hands, and require limits before transfer.
Conclusion
The legal fight is not simply about voter record access; it is the scope of that access in the digital age. And more importantly, it requires protecting citizens’ privacy interest by ensuring that the Executive does not expand its power by layering statutes enacted at different moments, for different purposes, and with different assumptions about what voter records contain.
The current framework effectively addresses privacy after transfer, through confidentiality provisions, agency safeguards, or downstream limits on disclosure. But once a complete voter file is transferred, the privacy risk has already been created and remediation is minimal. In the digital age, disclosure is not merely procedural—it is the privacy event itself.
Today’s story of voter records matters because it risks every story that came before it: exclusion, access, accountability and the promise that the right to vote should empower citizens, not expose them.
