The Safeguard American Voter Eligibility (SAVE) Act is currently being debated in the Senate after passing the House in April. The current version of the SAVE Act would require “documentary proof of US citizenship” such as a passport, military identification card (and military record of service), or birth certificate. It also restricts mail-in voting, requiring all mail voters to submit an application before receiving a ballot.

The problem the act presents, as others have documented—including Senator Dick Durbin (D-IL) below— is that citizenship documentation requirements can pose real barriers for millions of Americans. A driver’s license, even a Real ID, is insufficient. And roughly half of Americans do not have a passport. And if you changed your name, through marriage, for example, your birth certificate wouldn’t be sufficient either.


These barriers aren’t distributed evenly across the US either. In fact, residents of states that were more likely to vote for Donald Trump in the 2024 presidential election are less likely to have a passport and more likely to have changed their last name.

Using data from the Department of State and Census Bureau, the Center for American Progress estimates that about 146 million citizens do not possess a valid passport. Passport ownership varies widely across states, with the lowest rates concentrated in states that voted most strongly for Donald Trump in 2024. Less than a third of people in states like West Virginia, Mississippi, and Alabama own a passport, while passport ownership rates in bluer states like New Jersey, California, Massachusetts and New York exceed 70 percent.


States where Trump won by the largest amounts also have the lowest passport ownership rates, as shown below.


If voters want to use their birth certificate as a form of proof of citizenship, their current name will need to match the name on the birth certificate. This requirement may be especially problematic for married women, who, by a four-to-one margin, are likely to take their spouse’s last name.

The share of married (heterosexual) couples with different last names is also negatively correlated with Trump’s vote share—in other words, states where Trump won by larger margins tend to have more married couples with the same last names.


Cost and time are also factors in getting proper identification documents. In most states, amending a birth certificate or getting a new birth certificate can be done by mail. But it takes time (typically 3-8 weeks) and fees range from roughly $9 to $32, not including mailing or delivery service charges.

Getting a new passport is also costly. New passports cost $165 for adults and can take about 4-6 weeks to receive. Applying for a new passport must generally be done in person, which is increasingly difficult after the State Department closed libraries that processed such applications. Renewals are easier and can be done via mail but incurs a $130 processing charge.

What does all this mean for potential election results?

Aside from the stated goal of preventing non-citizens from voting—which is already extremely rare and there is no credible evidence of widespread occurrence—the SAVE Act would likely discourage or prevent eligible citizens from voting, particularly in states with lower passport ownership.

To get a sense of the scale, I did a simple thought experiment using data from the Congressional midterm elections in 2024. In that election, 221 Republicans were elected to the House of Representatives to the Democrats’ 214 election wins.

In those 221 election wins, Republicans averaged a margin of 32.5 percentage points (that is, the share of Republican votes exceeds the share of Democrat votes by 32.5 percentage points) compared with a 31.4-point margin for Democrats. Thirty-one races were decided by less than 5 points (17 won by Democrats) and 11 were decided by less than 2 points (5 by Democrats).

I simulate 2024 US House election outcomes under a scenario in which people without passports are unable to vote. For each district, I estimate excluded voters using the state-level share of residents with passports, assuming that 25 percent of those without passports obtain one before the election. Excluded voters are then removed from district vote totals, with Democrats and independents removed at the same rate and Republicans removed at a rate 1.5 times as high to reflect the lower passport ownership rates observed in Republican-leaning states. Election winners and margins are then recalculated based on the adjusted vote totals.

Putting this in practice, California’s 27th congressional district north of Los Angeles recorded 300,090 total votes in the 2024 House election, including 154,040 Democratic votes and 146,050 Republican votes, giving the Democrat candidate a victory margin of 2.7 percentage points. The Center for American Progress reports that about 72 percent of California residents have a passport. Under the simulation, the remaining 28 percent are assumed to lack passports, but 25 percent of those individuals obtain one before the election. This leaves roughly 21 percent of voters effectively excluded, or about 63,500 voters in total. Those excluded voters are then allocated across parties assuming Republicans are more likely to lack passports than Democrats or independents. That corresponds to removing roughly 17,000 Democratic votes and 25,000 Republican votes. The adjusted vote totals are then used to recompute the election outcome and margin, which increases the Democrat’s margin of victory to about 6.4 percentage points.

Of the nearly 148 million votes cast in 2024, more than 52 million votes would not be cast in this year’s elections, representing a 35 percent decline in votes as a result of the SAVE Act.

Under this simulation—which is admittedly sensitive to these assumptions—Republicans would win 209 House seats, Democrats 214, and 12 seats would switch from Republican to Democrat (none go the other direction).




Obviously, this simulation assumes passport ownership and differential statewide passport ownership between Republicans and Democrats. The simulation takes no additional behavioral responses into effect, such as voters dissuaded from voting at all, or whether voters might opt to get new or revised birth certificates. I am also not assessing the quality of new candidates or national trends for or against either party. I’m also not accounting for any changes to mail-in voting. I’m simply estimating how a narrow change could affect previous election results.

It’s unclear whether the SAVE Act will be passed as currently written. President Trump has made it a legislative priority, even suggesting he wouldn’t sign any other legislation (aside from Department of Homeland Security funding) until it is passed. But Democrats are certain to filibuster any vote, putting the entire question in the air.

No matter how the legislative process plays out, the SAVE Act could have substantial and measurable effects on voting participation across the country, potentially altering the composition of the US Congress.

Like what you read? Subscribe to my Substack to get posts like this first — plus join the conversation: https://jschwabish.substack.com/

March 19, 2026
126