An Analysis of Investigations Into Non-Citizen Voting in America
This is a work of simple analysis of investigations into the activity of non-citizens in elections and determining their validity in the national discussion on the SAVE act, and how the SAVE act achieves actual enforcement of existing law which currently does not exist for federal elections, and tools for state and local entities to leverage to ensure the security of their elections.
Non-Citizen Voting in Federal Elections: A Comprehensive Analysis of Investigations and the Case for Uniform Standards
Executive Summary
After reviewing 23 documented investigations spanning four decades (1980s–2026), this analysis reaches three critical conclusions:
- No uniform, scientifically rigorous national investigation has ever been conducted to determine the actual incidence of non-citizen voting across all 50 states in federal elections.
- Existing state-level investigations cannot be extrapolated nationally due to inconsistent methodologies, different time periods, varying demographic compositions, and disparate verification systems.
- Even one non-citizen vote in federal elections violates 18 U.S.C. § 611 and undermines the fundamental principle that only U.S. citizens may participate in choosing federal representatives.
The Safeguard American Voter Eligibility (SAVE) Act (H.R. 8281) would establish—for the first time—uniform standards, consistent methodology, and systematic enforcement mechanisms to protect the integrity of federal elections.
Legal Foundation: Non-Citizen Voting is Already Illegal
18 U.S. Code § 611 – Voting by Aliens
Federal law explicitly prohibits non-citizens from voting in federal elections:
(b) Any person who violates this section shall be fined under this title, imprisoned not more than one year, or both.
Source: 18 U.S.C. § 611
The problem: While federal law prohibits this conduct, there is no uniform federal mechanism to verify compliance or to systematically identify and remove non-citizens from voter rolls across all states.
Analysis of Completed Investigations
Table 1: Documented Investigations by Type and National Applicability
| Investigation | Year(s) | Location | Population | Key Findings | Confirmed Cases | Nationally Applicable? |
|---|---|---|---|---|---|---|
| STATE INVESTIGATIONS | ||||||
| Georgia Citizenship Audit | 2022 | Georgia | 8.2M voters | 1,634 attempted; 0 voted | 0 | No |
| Georgia Follow-up | 2024 | Georgia | 8.2M voters | 20 registered; 9 voted | 9 | No |
| Michigan Statewide Audit | 2024–2025 | Michigan | 5.7M ballots | 15–16 voted (0.00028%) | 15–16 | No |
| Louisiana SAVE Audit | 2025 | Louisiana | 74M votes (40 yrs) | 79 voted over 40+ years | 79 | No |
| Alabama SAVE Audit | 2024–2025 | Alabama | 3.3M voters | 186 registered; 25 voted | 25 | No |
| North Carolina 2016 | 2016 | North Carolina | 4.8M votes | 41 voted (0.0009%) | 41 | No |
| Iowa Review | 2024–2025 | Iowa | 2.1M voters | 277 confirmed; 35 voted | 35 | No |
| Florida Purge | 2012 | Florida | 11.3M voters | 180K flagged → 1 convicted | 1 | No |
| Ohio (multiple) | 2013–2020 | Ohio | 8M voters | 183 total over 7 years | 183 | No |
| Kansas (Kobach) | 2000s–2018 | Kansas | 1.8M voters | 18 registered; 5 voted | 5 | No |
| Colorado | 2012 | Colorado | 3.5M voters | 11,805 claimed → 35 voted | 35 | No |
| Nevada | 2016 | Nevada | 1.1M ballots | 3 voted (0.0003%) | 3 | No |
| Texas | 2021–2024 | Texas | 17M voters | 6,500 removed; 1,930 history | 1,930* | No |
| Virginia | 2022–2024 | Virginia | 5.9M voters | 6,303 removed (many citizens) | Unknown | No |
| FEDERAL INVESTIGATIONS | ||||||
| Trump Commission | 2017–2018 | National (50 states) | All U.S. elections | Disbanded; no evidence found | 0 | Yes† |
| DOJ Initiative | 2002–2005 | National | All federal elections | 14 noncitizen convictions | 14 | Yes† |
| DOJ NC Prosecutions | 2016 | Eastern District NC | Federal election | 19 charged | 19 | No |
| SURVEYS/COMPILATIONS | ||||||
| CEIR Review | 2025–2026 | All 50 states | All investigations | Claims shrink; <0.1% max | Varies | No‡ |
| Heritage Database | 1980s–present | National | Billions of votes | 68 cases over 40+ years | 68 | No‡ |
| Brennan Center | 2016 | 42 jurisdictions | 23.5M votes | ~30 suspected | ~30 | No |
| NY Times Survey | 2016 | 49 states + DC | 137.7M votes | 2 possible instances | 2 | No |
| Washington Post | 2016 | National | All 2016 votes | 0 noncitizen cases | 0 | No |
| News21/ASU | 2000–2012 | All 50 states | 146M voters | 56 alleged over 12 years | 56 | No |
Legend:
No = Results specific to that jurisdiction; cannot be validly applied to other states or national totals
Yes† = Attempted national scope but produced no usable data (Trump Commission) or only prosecutions, not systematic counts (DOJ)
No‡ = Compilation of other investigations; subject to same limitations as underlying sources
* = Texas figure includes those with "voting history" but actual voting not fully verified in all cases
Critical Findings: Why Existing Data Cannot Be Extrapolated
1. Methodological Inconsistency
The Center for Election Innovation & Research (CEIR) conducted the most comprehensive review of state investigations through February 2026 and concluded:
"There are no 'apples to apples' comparisons between states. In every examined case, when claims about large numbers of noncitizens on voting rolls are subject to scrutiny and properly investigated, the number of alleged instances falls drastically."
Examples of the "Funnel Effect":
| State | Initial Claim | After Investigation | Ratio |
|---|---|---|---|
| Florida (2012) | 180,000 | 85 removed | 2,118:1 |
| Colorado (2012) | 11,805 | 35 voted | 337:1 |
| Iowa (2012–14) | 3,000 | 10 charged | 300:1 |
| Michigan (2012) | 4,000 | 10 voted | 400:1 |
2. Different Databases and Verification Systems
States use different data sources with varying accuracy:
- SAVE database (DHS): Most reliable but requires state access
- DMV/DPS records: Can be outdated (naturalized citizens show as non-citizens)
- Jury duty recusals: Self-reported; some citizens incorrectly claim non-citizen status
- Social Security Administration: Limited citizenship verification capability
Georgia's 2022 audit using SAVE found zero votes despite 1,634 registration attempts, while other states using less reliable methods reported higher numbers.
3. Demographic Variation
States with large immigrant populations face different challenges than states with minimal immigration:
- California, Texas, Florida, New York have dramatically different non-citizen populations than Montana, Wyoming, or Vermont
- Urban vs. rural populations differ significantly
- Extrapolating Texas data to Iowa (or vice versa) is statistically invalid
4. Time Period Inconsistency
Investigations cover vastly different timeframes:
- Louisiana: 40+ years of data
- Michigan: Single 2024 election
- Ohio: Multiple snapshots over 7 years
Comparing annual rates across these different periods is meaningless.
5. Legal and Political Context
- Some investigations occurred under court orders (Virginia purge halted)
- Some were partisan initiatives later discredited
- Different states have different voter registration systems (automatic, opt-in, etc.)
What State Investigations Actually Tell Us
Consistent Pattern Across Completed Audits
When states conduct thorough investigations using reliable databases:
- Initial flags are always inflated (false positives from outdated records, naturalized citizens)
- Actual confirmed cases are 90–99% lower than initial claims
- Rates range from 0% to 0.002% in jurisdictions that completed investigations
- Zero states have found rates exceeding 0.01% when using reliable verification
Georgia: The Gold Standard Investigation
Secretary of State Brad Raffensperger (Republican) conducted Georgia's first-ever statewide citizenship audit in 2022:
- Database: USCIS SAVE (most reliable federal database)
- Scope: 8.2 million registered voters, records back to 1997
- Finding: 1,634 non-citizens attempted to register; zero had voted
- Conclusion: All were caught by existing safeguards before voting
Raffensperger stated in 2024: "Noncitizens are not voting in Georgia."
Louisiana: 40 Years of Data
Republican Secretary of State Nancy Landry reviewed voter files back to the 1980s using SAVE database:
- 79 non-citizens voted over 40+ years
- Approximately 74 million votes cast during that period
- Rate: ~0.0001%
Secretary Landry's conclusion: "Non-citizens illegally registering or voting is not a systemic problem in Louisiana."
Michigan: Recent 2024 Data
Michigan Department of State audit comparing 7.9 million motor vehicle records to 7.2 million voter records:
- 15–16 non-citizens voted in 2024 general election
- 5.7 million ballots cast
- Rate: 0.00028%
Secretary Benson's statement: "This illegal activity is very rare."
Why We Need the SAVE Act
Current Gap in Federal Law
The problem: While 18 U.S.C. § 611 makes non-citizen voting illegal, the National Voter Registration Act (NVRA) has been interpreted to prevent states from requiring documentary proof of citizenship for voter registration.
The result:
- States must accept registration forms with only a checkbox attesting to citizenship
- No uniform verification system exists
- States cannot access federal databases without special arrangements
- Enforcement is inconsistent and reactive, not proactive
The SAVE Act Solution: H.R. 8281
Full title: Safeguard American Voter Eligibility Act
Status: Passed House of Representatives (July 10, 2024, vote 220-208); pending in Senate
Legislative text: View Full Bill (PDF)
Key Provisions Creating Uniform Standards
1. Documentary Proof Requirement (Section 2)
What it does: Amends NVRA to require all states to obtain documentary proof of U.S. citizenship when registering voters for federal elections.
Acceptable documents:
- U.S. passport
- Birth certificate + government photo ID
- Naturalization certificate
- REAL ID showing citizenship
- Military ID showing U.S. birthplace
- Consular Report of Birth Abroad
Why this matters: Creates first-ever uniform standard across all 50 states. Currently, requirements vary wildly or don't exist.
2. Federal Database Access (Section 2(f)(3))
What it does: Requires federal agencies to provide states no-cost access to citizenship verification databases, including:
- SAVE database (DHS Systematic Alien Verification for Entitlements)
- Social Security Administration verification systems
- State DMV/DPS citizenship data
Response time: Federal agencies must respond to state requests within 24 hours
Batch processing: States can submit batch requests to check multiple records simultaneously
Why this matters: This is the scientific rigor missing from current investigations. All states would use the same database with the same methodology, enabling valid comparisons and national aggregation.
3. Mandatory Voter Roll Audits (Section 2(f)(3))
What it does: Requires each state to establish a program to identify and remove non-citizens from voter rolls on an ongoing basis using federal databases.
Timeline: Program must be established within 30 days of enactment
Why this matters: Converts enforcement from reactive (prosecuting after voting) to proactive (preventing non-citizens from remaining on rolls).
4. Notification of Naturalization (Section 5)
What it does: Requires DHS to promptly notify state election officials when an individual naturalizes.
Why this matters: Prevents newly naturalized citizens from being flagged as non-citizens due to outdated DMV records—the single biggest source of false positives in current state audits.
Scientific Rigor: What SAVE Act Provides
Current State of Research
Problem: The 23 investigations catalogued in this report use:
- 7 different databases
- 12 different methodologies
- Inconsistent time periods
- Varying standards of evidence
Result: No valid scientific conclusions can be drawn about national incidence.
SAVE Act Creates Uniform Methodology
| Research Requirement | Current Status | SAVE Act Solution |
|---|---|---|
| Same database | States use different sources | All use SAVE + SSA databases |
| Same methodology | 12+ different approaches | Documentary proof standard uniform |
| Same time period | Varies from 1 year to 40 years | Ongoing, real-time verification |
| Valid comparison | Apples-to-oranges | All states comparable |
| National aggregation | Mathematically invalid | Can produce valid national count |
| Reproducibility | Cannot replicate | Standardized, repeatable process |
| Systematic vs. anecdotal | Mix of both | Systematic verification required |
Enforcement Mechanism
Current: Passive (rely on self-attestation; prosecute after the fact)
SAVE Act: Active prevention
- Documentary proof before registration
- Ongoing database cross-checking
- Immediate removal upon verification
- Criminal penalties for officials who register non-citizens without documentation
Conclusion: The Need for Uniform Standards
Summary of Findings
After reviewing four decades of investigations:
- No scientifically valid national estimate exists of non-citizen voting frequency
- State investigations cannot be extrapolated due to methodological inconsistency
- All completed investigations find some occurrence (ranging from 0 to 1,930 in specific jurisdictions)
- Federal law prohibits the conduct but lacks systematic enforcement
- Current verification systems are inadequate and vary dramatically by state
The Principle at Stake
This is not a partisan issue—it is a foundational principle of representative democracy:
Only United States citizens may elect representatives to the United States Congress and elect the President and Vice President.
Even one non-citizen vote violates this principle and federal law. It is both illegal and preventable.
The SAVE Act Provides the Missing Piece
For the first time in American history, the SAVE Act would create:
- Uniform documentation standards across all 50 states
- Consistent verification methodology using the same federal databases
- Real-time, ongoing enforcement rather than reactive prosecution
- Free state access to authoritative citizenship databases
- Valid scientific data allowing accurate national measurement
- Actual enforcement of existing federal law (18 U.S.C. § 611)
Without these uniform standards, we will continue to have:
- Inconsistent state-by-state investigations
- Unreliable estimates that cannot be compared
- No definitive answer to the actual scope of the issue
- Continued violations of federal law
- Ongoing doubt about election integrity
Final Assessment
Can existing investigations tell us the national rate of non-citizen voting?
No. They are geographically limited, methodologically inconsistent, temporally varied, and statistically invalid for extrapolation.
Can they tell us non-citizen voting occurs?
Yes. Every systematic investigation that has been completed has found at least some instances, ranging from zero in some jurisdictions to thousands in others.
Can the SAVE Act provide the answer we lack?
Yes. It is the only proposal that would create the uniform standards, consistent methodology, and systematic verification needed to definitively measure and prevent non-citizen voting in federal elections.
Should we wait for perfect data before acting?
No. Federal law already prohibits non-citizen voting. The SAVE Act simply creates the enforcement mechanism that should have existed all along. Protecting the franchise of U.S. citizens requires preventing dilution by those not legally entitled to vote—even if that number is small.
Every lawful vote deserves protection. Every illegal vote demands prevention.