The SAVE Act's Hidden Math Problem

Millions of Married Women Could Lose Their Vote

The Safeguard American Voter Eligibility (SAVE) Act — legislation that would require documentary proof of citizenship to register to vote in federal elections — has become one of the most contentious voting bills in recent memory. The House has passed a new version of the bill^[1], but as of this writing the bill remains stalled in the Senate, unable to clear the 60-vote threshold needed to break a filibuster. Supporters, including President Trump and House Republicans, have pushed hard for its passage; Senate Majority Leader Thune has repeatedly brought it to the floor, and Republican leadership is reportedly now eyeing the budget reconciliation process as an alternative path forward^[2][3].

Supporters of the bill frame it as a straightforward election-integrity measure: only citizens should vote, and requiring proof of citizenship simply confirms what should already be true of every registered voter. Opponents counter that noncitizen voting in federal elections is already a felony and vanishingly rare, and that the practical effect of the bill would be to strip eligible citizens of their vote because they can't easily produce the required paperwork, not because they aren't citizens^[4].

What counts as valid ID and why a name change breaks it

What about this bill would make so many people re-register to vote? The SAVE Act allows a passport or a birth certificate as proof of citizenship. The catch is that whatever document you use has to match the name on your voter registration. A passport does this automatically, since passports are reissued under your current legal name. A birth certificate does not. It carries the name you were born with. If your birth certificate says "Jane Smith" and your driver's license says "Jane Doe," you have no single document proving those two names belong to the same citizen.

A REAL ID-compliant driver's license is often cited as a workaround, but its adequacy under this bill has already been called into question. And press secretary Karoline Leavitt has said that anyone with a name change would simply need to re-register with corrected documentation. This solution is not simple, as the numbers below show, may not be logistically realistic at scale before November’s elections.

Sizing the problem: Married women who changed their names

It's worth pulling apart to understand who, specifically, is caught up in this bill. Adoptees whose names changed at adoption, disaster survivors who lost paperwork, domestic violence survivors who can't safely return for records — the list of affected groups is long, and by nature hard to size precisely. But we can get a useful baseline by looking at the single largest group the bill would affect: married women who changed their name and don't already hold a passport. Unfortunately, this number will only include women married to men, as the number of women married to women who also changed their name was not available.

Here is the calculation, step by step, using national data on marriage rates, name-change rates, and passport ownership:

That leaves roughly 17.4 million women who are registered to vote, married to men, changed their name at marriage, and don't hold a passport. This means they have no single document that ties their birth-certificate name to their current legal name. Every one of them would need to track down supplementary documentation (typically a marriage certificate) just to re-prove they are who their registration says they are.

Some small groups may have another ID that is acceptable, but these groups are nominal to the overall problem. For example, this total doesn't account for military ID and service records, which the bill would likely accept as valid in some cases. That population of married women is small, perhaps around 100,000, and reliable data on how many of them also changed their name isn't available. Excluding them entirely, or any other niche group with another uncommon valid ID, doesn't meaningfully change the conclusion: the number needing to re-register still runs well into the tens of millions once every affected group is added together. 

Could they just re-register?

Statista tracks voter registration changes across two-year election cycles; the maximum net increase in registrations over a full two-year cycle has historically been around 15 million nationwide. Even under generous assumptions that every one of those registrations could be redirected toward re-verifying existing voters, and that this entire surge happened in a single year rather than being spread across the cycle, boards of elections would need to more than double their normal capacity just to process this one subgroup.

The timeline makes this worse, not better. If the SAVE Act passes and takes effect before the November election, boards of elections would have only a few months to process a volume exceeding what they normally handle in a full two-year cycle. The bill's documentation standards are also more demanding than current registration requirements, meaning each case takes longer to review and fewer applicants will be able to complete the process online. That backlog compounds across every other affected group e.g., disaster survivors, domestic violence survivors, adoptees) pushing the total far past 17.4 million. The Brennan Center for Justice has estimated the total to be around 21 million.

The bill does not include funding to help boards of elections scale up to meet this demand. They would be left holding the bag. 

The bottom line

None of these numbers are exact: they're estimates built from the best available national data, and reasonable people could adjust the inputs at the margins. But even under conservative assumptions, the scale is the point: tens of millions of eligible citizens, disproportionately married women, would be pushed into a re-registration process that the country's election infrastructure has never had to handle at this volume, on a timeline that leaves little room for error. Additionally, poll workers will have to be updated on a more complicated procedure, and millions of voters will have to educated about the change or end up getting turned away at the polls. 

Further, the proponents of the SAVE Act wish to present this as solving the problem of noncitizens voting. Recent testimony to Congress presented data that there were 250,000 noncitizens registered to vote, though many likely did not cast a ballot^[5]. There are likely problems with this count, but, taken at face value, this still doesn’t justify the change. In a country with 174 million registered voters, this bill would disenfranchise 10% of eligible voters in favor of knocking .14% non-citizens off the voter rolls.

The math doesn’t add up.

If you have a view on this, your senators are the ones who will decide it.