Does the US have a
duplicate voter problem?

We ran entity resolution across 50 million publicly available voter records from 7 states. Nearly 400,000 potential duplicate registrations — and 61 confirmed fraud cases. Here's what the data shows.


The Dataset

50 million records. One identity resolution engine.

50M+
Voter profiles

Across 7 US states

7
States analysed

GA, FL, MI, NC, PA, OH, AR

~400K
Potential duplicates

0.8% of the voting population

61
Confirmed fraud cases

In Ohio and Pennsylvania

Data sourced from public voter registrations, September–November 2023. Analysis uses text similarity, geographic proximity, and temporal range matching.


The Method

How entity resolution finds voter duplicates

Unresolved — Raw Records
JAMES R. WILSON
Georgia · DOB 1974-03-12 · Active
James Wilson
Florida · DOB 1974-03-12 · Active
Jim R Wilson
Georgia · DOB 03/12/1974 · Inactive
3 separate records across 2 states — no link between them
Resolved — Unified Entity
JW
James R. Wilson
entity_vt_4k9nQr · confidence 94%
StatesGeorgia, Florida
Registration statusActive + Inactive
Name variants3 recorded
Source records3 linked

State Breakdown

Where duplicates are most concentrated

Florida
148,516 (1.1%)
Highest absolute count
Pennsylvania
80,142 (1.01%)
2020 margin of victory: 80,555 votes
Georgia
51,876 (0.73%)
2020 margin of victory: 11,779 votes
Michigan
(90%+) 90%+
Registration rate — suggests deceased voters still on rolls
Arkansas
(0.92%) 0.92%
County variation up to 2.06% (Searcy County)
North Carolina
Included in cross-state matching
Ohio
61 confirmed fraud cases with Pennsylvania

Key Finding

Why this is a data quality problem, not a conspiracy

Voters move states and counties — old registrations are rarely purged in time

Name variations across records (Jim vs James, hyphenated surnames) prevent simple deduplication

Death registries are not updated in real time, leaving deceased voters on rolls

Cross-state matching requires fuzzy matching — exact-match rules (like Georgia's) miss most duplicates


What Good Looks Like

What real-time entity resolution would change

Cross-state deduplication

Catch the same voter registered in two states when they move, not years later.

Real-time death registry sync

Remove deceased voters from rolls automatically as records update.

Fuzzy matching at scale

Resolve name variants, typos, and format differences across 50M+ records in hours.


Entity resolution works on any public dataset

See what Tilores finds in your data. Free tier includes 50,000 records — no credit card required.