How does the SAVE system work in practice for state election officials and what are its documented error rates?
Executive summary
The SAVE system is an immigration-status verification database used by federal agencies that some states and the Justice Department have sought to leverage to screen voter rolls for noncitizens, producing small but politically fraught match lists that election officials say are prone to false positives; there is no comprehensive, government-published “error rate” for SAVE in election use in the available reporting [1] [2]. Public examples show states removing or flagging a few dozen to a few hundred names after SAVE queries, while voting-rights advocates and former secretaries of state warn the process can misidentify citizens and create the risk of wrongful purges if not handled with safeguards [2] [1] [3] [4].
1. What SAVE is and how election officials use it in practice
SAVE (Systematic Alien Verification for Entitlements) is an immigration-status verification tool operated by U.S. Citizenship and Immigration Services that can confirm whether a person appears in federal immigration records, and state election offices or the Department of Justice have used SAVE or its outputs to compare against voter rolls to identify potential noncitizen registrants; reporting notes states have run voter data through SAVE or similar queries to generate lists of potential noncitizen voters [1] [2]. In practice, the use seen in reporting is not an automated purge: jurisdictions or federal actors produce lists of “potential” matches that local election officials must investigate, but those lists can be redacted, contested, or the subject of litigation over disclosure and privacy [2] [1].
2. The mechanics on the ground: matches, follow-up, and legal friction
When SAVE or DOJ queries produce a potential match, election officials face a multistep administrative process: examine the match, notify the registrant if law requires, seek documentary proof of citizenship if statutes mandate, and follow state law for removal or confirmation; courts and state law limit what data can be shared and how quickly officials can act, and some states have been sued after providing or withholding information sought by the Justice Department [2] [4]. Officials at conference briefings and in local reporting emphasize that lists require manual verification and that SAVE’s records do not automatically translate into an enforceable finding of noncitizenship—USCIS officials have said SAVE does not “refer a noncitizen to ICE” as part of that process [1].
3. Documented outcomes and sample error signals
Publicized outcomes so far are numerically small but politically consequential: Alabama reported removing 186 people from the rolls after running data through SAVE—with state officials saying 25 of those removed had voted—while Tennessee identified 42 potential noncitizen voters after similar checks [2]. Local officials have also reported apparent false positives in small batches, such as a Michigan clerk’s claim of finding 15 noncitizens on the rolls that reporting flagged as “may not be reliable,” illustrating how lists can overcall problems if not thoroughly vetted [2] [1].
4. What the reporting says about SAVE’s error rates and limits of the record
None of the provided reporting or federal guidance supplies a validated, system-wide error rate for SAVE when used against voter rolls; journalists and election experts repeatedly warn that lists of noncitizens are “prone to error” and that comparing those lists to voter registration databases can generate false positives absent careful cross-checking [1] [2]. Senator Alex Padilla and advocacy groups argue immediate or broad deployment could lead to wrongful purges and administrative chaos, but those critiques are policy and risk assessments rather than quantified error studies [3] [4].
5. Mitigations, audits and the administrative context
Election-administration authorities and security guidance emphasize verification, paper trails, and post‑election audits as the practical safeguards against erroneous removals: the Election Assistance Commission and cybersecurity agencies encourage state officials to use established verification procedures, risk-limiting and post‑election audits, and IT security practices when introducing external data into voter rolls [5] [6] [7] [8]. The reporting indicates the practical recommendation from election officials is to treat SAVE outputs as leads requiring human review, not as a turnkey removal mechanism [1] [4].