Political Fact Checks
Is Voter Fraud a Real Problem?
1,619 cases on Heritage’s list since 1982, about 7 per 10 million federal ballots, and what 2020 found.
Crime & Justice
Which Side Commits More Political Violence?
Deaths and attacks by side in the counts of CSIS, Cato, New America and ADL, and what moves the totals.
History & Myth-Busting
Did the U.S. Political Parties Switch?
Roll-call votes, congressional seats and the Southern Strategy in their own words, tested against what both sides claim.
Do Voter ID Laws Reduce Turnout?
The claim is that requiring photo identification (ID) at the polls stops people voting, minority voters most of all. An often-quoted figure says Latino turnout was 10.3 points lower in states with strict photo ID laws. This page sets that figure beside a national study of voter records, studies of single states and a government audit.
Jump to the verdict ↓
What this covers
This page is about whether requiring photo ID at the polls reduces how many people vote, in the United States. It covers fraud only as far as is needed to say what ID can and cannot address; the evidence on how much fraud exists is on the voter fraud page. It is not about mail ballots, registration purges, or polling-place closures.
The short answer
Not by much. Measured falls in turnout after voter ID laws are small or absent. The largest study, of the voting records of nearly every voting-age American from 2008 to 2018, found none. Studies of Georgia, Rhode Island and South Carolina found falls of less than half a percentage point. A government audit found falls of about 2 to 3 points in Kansas and Tennessee in 2012. There, the fall was larger among Black registrants than among white registrants.
Which laws the studies are about
Thirty-six states have laws asking for or requiring some form of identification at the polls, by the count of the National Conference of State Legislatures in September 2026.
The conference sorts the laws by two questions: whether the ID must carry a photo, and what happens to a voter who has none. Under a non-strict law, at least some voters without ID can cast a ballot that counts with no further action. For instance, they can sign an affidavit of identity or be vouched for by a poll worker. Under a strict law, the voter casts a provisional ballot that counts only after further steps, for instance returning to an election office within a few days with ID.
Fourteen states had strict laws in force in 2026, ten of them requiring photo ID. The studies on this page are mostly about the smaller group of states whose laws are strict.
Two ways of measuring turnout
Studies built on voter records, which show whether a named person voted, have found effects from none to about 3 points. The largest estimates, among them the 10.3-point figure, came from a study built on surveys matched to voter records. A replication found that the study’s data could not support firm conclusions.
The national study found no fall
In 2021 the Quarterly Journal of Economics published a study by Enrico Cantoni and Vincent Pons. The study is built on what the authors call “the largest individual-level dataset ever assembled to study voter participation”. It follows the registration and voting records of nearly every voting-age American from 2008 to 2018, the same people election after election. It compares states that adopted strict ID laws with states that did not. The dataset holds 1.6 billion observations.
The turnout and registration figures below come from what the authors call their most demanding specification. This model compares each voter’s own record in years with and without a strict law. Its fixed effects allow for lasting differences between states and between voters, and for anything shared by every state in an election year.
Cantoni and Pons 2021, most demanding specification
State, year and voter fixed effects · 2008–2018
Effect on overall turnout−0.1 points (95% CI −3.0 to +2.8)
Effect on overall registration−0.1 points (95% CI −2.3 to +2.1)
Non-white turnout, relative to white+1.4 points (95% CI −0.5 to +3.2)
Non-white voters contacted by a campaign+4.7 points
How to read it: each row is one quantity Cantoni and Pons measured, with the point estimate first. In brackets is the 95% confidence interval (CI), the range the data is consistent with. A negative number means turnout fell; a positive one means it rose. Every range in the turnout, registration and non-white turnout rows crosses zero, which means the study cannot tell an effect apart from no effect at all. The campaign-contact row is statistically significant at the 5 percent level.
Their own summary is that the laws have no negative effect on registration or turnout, overall or for any group defined by race, gender, age or party. Across all voters, the lower end of the turnout row’s confidence interval rules out a fall in turnout of more than 3.0 points. Among voters in neighbouring counties on either side of a state border, the authors rule out a fall of more than half a percentage point.
The authors rule out a fall in non-white turnout, relative to white turnout, of more than 0.5 points. So on this design, the larger harm to minority voters that the claim is about does not appear.
The campaign-contact row is one the authors flag themselves. It comes from a separate model, without voter fixed effects. In states that passed these laws, the share of non-white voters who told a survey that a campaign had contacted them rose by 4.7 points. The authors suggest party mobilisation might have offset some direct effect of the laws on minority turnout. Their own estimate of that offset, from studies of canvassing, is about a third of a percentage point.
What the authors say the study cannot do
The laws were 4 to 14 years old when the authors wrote, so they cannot rule out effects that appear later. They note that enforcement could tighten and party mobilisation could fade. They do not see their results as “the last word on this matter”. They also note that the states which adopted strict laws had different turnout levels beforehand, though not different trends, which is what their design depends on.
Studies of single states found small falls
Most studies of single states match voter records to licence records, so they can see who lacked ID. They measure the fall among those voters against voters who had it. Studies of single states have found falls in turnout after ID laws, mostly small.
| State and election | Study | What it found |
|---|---|---|
| Georgia, 2008 | Hood and Bullock, 2012 | Turnout lowered by about 0.4 points overall; no racial difference found |
| North Carolina, 2016 primary | Grimmer and Yoder, 2021 | A 0.7-point fall among the 3% of voters without ID, against voters with ID |
| Rhode Island, 2016 presidential election | Esposito, Focanti and Hastings, 2019 | A 2.7-point fall among people without a licence; a 0.42-point fall in all votes |
| South Carolina, 2014 | Hood and Buchanan, 2020 | Turnout lowered by 0.19 percent; no racial difference found |
| Virginia, 2014 | Hopkins and others, 2017 | Turnout higher where more voters lacked a licence, possibly because of a state mailing to them |
A fall measured against voters with ID is not the same as a fall in total turnout. Cantoni and Pons point out that such a result is consistent with an overall effect that is negative, null or even positive.
In North Carolina the effect outlasted the law. After the law was suspended, voters without ID were 2.6 points less likely than voters with ID to vote in the 2016 general election.
The Government Accountability Office (GAO) ran its own comparison in 2014. It set Kansas and Tennessee, which began requiring photo ID in 2012, against four states that made no change: Alabama, Arkansas, Delaware and Maine. Between the 2008 and 2012 general elections, turnout fell further in Kansas than in the comparison states, by an estimated 1.9 to 2.2 points. In Tennessee the estimate was 2.2 to 3.2 points. The GAO says its estimates are limited to that election in those two states and do not apply to other states or periods.
In Kansas and Tennessee, turnout fell further among Black registrants than among white registrants: by 3.7 points more in Kansas and 1.5 points more in Tennessee. The GAO found no consistent equivalent for Asian American or Latino registrants, which it said suggests the laws did not have larger effects on those groups.
Officials in Kansas, Tennessee and Arkansas disagreed with how the GAO chose the states it compared, and Kansas and Tennessee questioned the reliability of one of its three data sources. The GAO answered that no evaluation outside an experiment can account for every unobserved factor, and stood by its method.
The GAO also reviewed ten earlier studies, all covering general elections before 2008. Five found no significant effect on turnout. Four found significant decreases, and one found an increase.
Where the 10.3-point figure comes from
The 10.3-point figure comes from a working-paper version of a study by Zoltan Hajnal, Nazita Lajevardi and Lindsay Nielson. The working paper used validated turnout, checked against voter records, from national surveys of the 2008 to 2012 elections. For general elections, its model put Latino turnout 10.3 points lower in states with strict photo ID laws, and turnout among multiracial Americans 12.8 points lower.
The article published in the Journal of Politics in 2017 does not contain either figure. With no other factor taken into account, it says, Latino turnout was 7.1 points lower in strict ID states than in other states in general elections. The article’s model then allows for other state election laws, the contests on the ballot, how competitive the state was, and voters’ age, education and other traits. It predicts that Latinos are “10% less likely to turn out”.
A year later, in the same journal, Justin Grimmer, Eitan Hersh, Marc Meredith, Jonathan Mummolo and Clayton Nall replicated the published article. The rules for matching survey respondents to voter records had changed over time and differed between states, so the errors in measured turnout changed over time too.
In most state-years the method overstated turnout by about 25 points. In 15 states the method understated turnout by about 10. Where votes had not been checked against the records, as in Virginia before 2012, the coding implied turnout close to zero. Models fitted to years before the laws existed produced apparent effects of laws not yet passed.
The replication’s conclusion
Their conclusion: “When errors are corrected, one can recover positive, negative, or null estimates of the effect of voter ID laws on turnout, precluding firm conclusions.”
That is not a smaller effect. It is a statement that this data and this design cannot answer the question in either direction.
Two of the original authors, Hajnal and Lajevardi, replied in the same issue with John Kuk, under the title “We All Agree: Strict Voter ID Laws Disproportionately Burden Minorities”. They argued that the replication’s own reanalysis confirmed their core finding, “that strict voter ID laws discriminate”. That is a claim about who bears the burden, and it does not restore the 10.3-point figure.
A separate study shows who a strict ID law would stop from voting. In 2016 a court order let Texans without the required ID vote if they showed another document and signed a declaration saying why they lacked it. Bernard Fraga and Michael Miller linked about 16,000 of these declarations to the Texas voter file.
In their January 2021 draft, Black voters were 11.4% of Texans who voted with ID in 2016 and 16.1% of those who voted without it. Latino voters were 19.8% of those with ID and 20.7% of those without. Their race is estimated from each voter’s surname, address, sex and date of birth, not recorded. Under the strict law in force in 2014, the authors say, the Texans who voted without ID would have been turned away.
Download this section
The 10.3-point figure, where it comes from, and what a replication in the same journal found.
What photo ID can address
Photo ID at a polling place stops one thing. It stops a person turning up and voting as someone else. It does not touch mail ballots, registration, or anything done by an official.
Justin Levitt, a law professor, tracked every specific, credible allegation of voter impersonation he could find from 2000 to 2014 and published the list in 2014. He found about 31 incidents, some involving more than one ballot, in a period when general and primary elections alone drew more than a billion ballots. He bet that some of the 31 would turn out to be errors.
The Heritage Foundation, a conservative think tank, keeps a database of proven cases of election fraud. It says the database is a sampling, not a complete count.
The Heritage counts and their dates
In 2017 the database was reviewed by the Brennan Center for Justice at New York University School of Law. The center argues against the strictest ID laws. The database then listed 749 cases involving almost 1,100 people. Ten of the cases were in-person impersonation at the polls. At its update of 12 December 2025 the database listed 1,619 cases from 1982 to 2025, 34 of them impersonation. Impersonation was 1.3 percent of the cases in 2017 and 2.1 percent in 2025.
What photo ID reaches, and what it does not
Each figure is a count of cases, not a rate · Heritage as of 12 December 2025
In-person impersonation, credible allegations 2000–2014 (Levitt), against more than a billion ballotsabout 31
In-person impersonation cases in the Heritage database, 1982–202534
Absentee ballot cases in the same database, 1982–2025387
All cases in that database, 1982–20251,619
How to read it: these counts were collected over decades and are not on the same denominator, so they cannot be compared directly. Photo ID can address the two impersonation rows. The row it cannot address, absentee ballots, is the larger one, in a database Heritage built to show “the vulnerabilities in the election system”.
In 2008 the Supreme Court upheld Indiana’s ID law in Crawford v. Marion County Election Board. Justice Stevens’s lead opinion, joined by two other justices, says the only kind of voter fraud the law addresses is in-person impersonation at polling places. It adds that the record contains no evidence of any such fraud ever occurring in Indiana.
The opinion upheld the law on other grounds, including Indiana’s interests in modernising elections, in deterring fraud seen elsewhere and in a 2003 Indiana mayoral primary, and in public confidence. For confidence it quoted the 2005 report of the Carter-Baker commission, chaired by former President Jimmy Carter and former Secretary of State James Baker. The report said: “The electoral system cannot inspire public confidence if no safeguards exist to deter or detect fraud or to confirm the identity of voters.” The next section tests the confidence argument.
Whether ID laws have reduced fraud
One study has tested it. Cantoni and Pons counted cases before and after strict ID laws in the Heritage database and in a second one, compiled by News21, an investigative project by university students. They found no significant fall in either. Both databases list very few cases, 0.02 to 0.08 a year for every 100,000 residents. The authors note that a law can change how many cases are detected as well as how many happen. They also found no effect on how common people thought impersonation was.
Download this section
What polling-place identification reaches, set against what it does not, in counts rather than rates.
The strongest case for the laws
The case does not depend on impersonation being common. Heritage itself says it is “not making any definitive claims about the extent of election fraud”.
A safeguard against a rare crime is not worthless because the crime is rare; locks on doors are not refuted by low burglary rates in a neighbourhood. The national study finds no turnout cost, so the safeguard appears close to free. Advocates point to other democracies that require identification to vote. And the burden, where it exists, falls on a step that most citizens complete routinely for banking, travel and employment.
Two of these four arguments hold up: the rarity argument is coherent, and the national study’s result is real.
The international comparison is weaker than it looks, and it fails on a detail rather than on principle. The countries it names either supply ID widely or do not require photo ID. So the United States is being compared with countries that require ID and supply it.
Mexico’s voter card is free, and 99 percent of the people on its electoral roll hold a valid one. India’s Election Commission said in 2019 that more than 99 percent of electors held its photo identity card. Canada does not require photo ID in federal elections and lets a voter without ID be vouched for. Britain brought in photo ID at polling stations in May 2023. It created a free Voter Authority Certificate for voters without other photo ID.
The argument that ID is a routine step runs into evidence on cost. A free card is not a free process when the documents needed to obtain it are not free. In Texas, the free election certificate required documents such as a birth certificate, which cost $2 to $3 until 2015. The district court relied on an expert’s estimate that Black registered voters were 305% more likely than white voters to lack accepted ID, and Latino registered voters 195% more likely. The appeals court upheld the finding of a discriminatory effect built on that estimate.
One more point for the laws, and the sharpest, comes from the Indiana case. There the district judge found that the challengers had not produced evidence of a single individual resident who would be unable to vote because of the law. The judge also rejected as unreliable an expert estimate that up to 989,000 registered Indiana voters lacked acceptable identification. A claim that a law disenfranchises people is owed at least one person it disenfranchised, and in that record nobody produced one.
Texas later produced such a record: the court-ordered declarations of 2016 named about 16,000 voters who, by Fraga and Miller’s account, would have been turned away under the strict law.
The confidence claim, tested
Proponents also argue that requiring ID increases public confidence in elections. Cantoni and Pons found no significant effect of strict ID laws on whether people thought an election fair, or on how common they thought impersonation was.
A survey study of 2014 by Bowler and Donovan found confidence split by party. Republicans were more confident in strict photo ID states than in states without ID laws, and Democrats less. Across about 30 national surveys, Sances and Stewart found a consistent link between voting for the winner and confidence in how elections are run. In a post on their paper, they argue that an ID law may shift confidence only through its effects on who votes and who wins.
The verdict: what the evidence shows
Not by much. Measured on voter records, the effect of ID laws on turnout runs from no detectable fall, in the national study of 2008 to 2018, to about 2 to 3 points. That larger effect was in Kansas and Tennessee in 2012. Most studies of single states found falls of less than half a percentage point.
The burden is not spread evenly. In Kansas and Tennessee the fall was larger among Black registrants than among white registrants. Texans who voted without ID in 2016 were disproportionately Black and Latino. Studies of Georgia and South Carolina, though, found no racial difference. The often-quoted 10.3-point fall in Latino turnout comes from a working paper and cannot be relied on.
Five common claims, against the evidence
Claim“Voter ID laws suppress turnout.”
Small where foundThe largest study, of nearly every voting-age American from 2008 to 2018, found no fall. Most studies of single states found falls of less than half a percentage point. A government audit found falls of about 2 to 3 points in Kansas and Tennessee in 2012.
Claim“Minority voters are hit hardest.”
MixedA government audit found larger falls among Black registrants than among white registrants in Kansas and Tennessee, and Texans who voted without ID in 2016 were disproportionately Black and Latino. Studies of Georgia and South Carolina found no racial difference.
Claim“Latino turnout was 10.3 points lower in strict ID states.”
Do not quote itThe figure comes from a working-paper version of a survey study. A replication in the Journal of Politics, which published the study, found its data could give positive, negative or null estimates.
Claim“Voter ID laws prevent fraud.”
No measured effectImpersonation at the polls, the only fraud polling-place ID can stop, is 34 of the 1,619 cases in the Heritage Foundation’s database. The one study that tested ID laws against fraud counts found no significant fall.
Claim“ID laws increase confidence in elections.”
Not overallThe national study of 2008 to 2018 found no rise in the share of people who thought an election fair. A survey study of 2014 found strict photo ID laws went with more confidence among Republicans and less among Democrats, compared with states without ID laws.
Download the conclusion
The claim and the verdict above as one image, with the three findings that decide it. The image is the size social networks use for link previews, and anyone may download and post it.
What would change this conclusion
This article concludes that: (1) where turnout has been measured on voter records, ID laws have reduced it by little or nothing, from no detectable fall nationally to about 2 to 3 points in Kansas and Tennessee in 2012; (2) the fall was larger among Black registrants than among white registrants in Kansas and Tennessee, and Texans who voted without ID in 2016 were disproportionately Black and Latino; (3) the 10.3-point figure for Latino turnout cannot be relied on.
These conclusions would be falsified by:
• A study on voter records from elections after 2018 finding a fall in total turnout larger than the 3.0 points the national study rules out for all voters
• Records from other states showing that voters without ID, or those whose provisional ballots go uncounted for lack of ID, are no more often Black or Latino than other voters
• A survey-based estimate of the effect on Latino turnout that passes the replication’s tests, including its placebo test on years before the laws
Sources
Cantoni, E. and Pons, V. “Strict ID Laws Don’t Stop Voters: Evidence from a U.S. Nationwide Panel, 2008–2018.” Quarterly Journal of Economics 136(4), November 2021, pp. 2615–2660. Source for the 1.6 billion observations covering nearly every voting-age American from 2008 to 2018; the authors’ description of the dataset as the largest assembled to study voter participation; the most demanding specification and what it compares, and the separate model behind the campaign-contact figure; the −0.1 point estimates on turnout and registration with 95% confidence intervals of −3.0 to +2.8 and −2.3 to +2.1; the +1.4 point non-white relative effect with an interval of −0.5 to +3.2; the bounds of 3.0 points for all voters, 0.5 points in neighbouring border counties and 0.5 points for non-white turnout relative to white; the 4.7 point rise in non-white voters reporting campaign contact and the authors’ estimate of the offset; the limits the authors state; News21 as an investigative project by university students; the tests of fraud counts in the Heritage and News21 databases and of perceived impersonation and belief that the election was fair; and the account of earlier studies. The link is to the authors’ working paper, revised May 2021. nber.org
Huntington-Klein, N. “Lecture 8 Fixed Effects.” Lecture slides, 5 January 2021. Source for what fixed effects allow for: everything unique to a person, a country or another group that does not change, whether it can be measured or not. nickch-k.github.io
U.S. Government Accountability Office, GAO-14-634, “Elections: Issues Related to State Voter Identification Laws”, September 2014, reissued February 2015. Source for the comparison of Kansas and Tennessee with Alabama, Arkansas, Delaware and Maine; the estimates of 1.9 to 2.2 points in Kansas and 2.2 to 3.2 points in Tennessee between the 2008 and 2012 general elections, and the GAO’s statement that they apply only to that election in those states; the larger falls among African American registrants, 3.7 and 1.5 points more than among white registrants, and none found for Asian American or Hispanic registrants; the review of ten earlier studies; the objections of Kansas, Tennessee and Arkansas officials and the GAO’s reply. gao.gov
Hood, M. V. and Bullock, C. S. “Much Ado About Nothing? An Empirical Assessment of the Georgia Voter Identification Statute.” State Politics & Policy Quarterly 12(4), 2012, pp. 394–414. Source for the fall of about four-tenths of a point in Georgia’s turnout in 2008 and the absence of a racial difference. doi.org
Grimmer, J. and Yoder, J. “The durable differential deterrent effects of strict photo identification laws.” Political Science Research and Methods 10(3), 2022, pp. 453–469, first published online 2021. Source for the 0.7 point fall in North Carolina’s 2016 primary among the 3% of voters without ID, against voters with ID, and the 2.6 point gap between the two groups in the 2016 general election after the law was suspended. doi.org
Esposito, F. M., Focanti, D. and Hastings, J. S. “Effects of Photo ID Laws on Registration and Turnout: Evidence from Rhode Island.” NBER Working Paper 25503, January 2019. Source for the 2.7 point fall in turnout among people without a licence in Rhode Island’s presidential election after the law, and the 0.42 point fall in all votes. A working paper, not peer-reviewed. nber.org
Hood, M. V. and Buchanan, S. E. “Palmetto Postmortem: Examining the Effects of the South Carolina Voter Identification Statute.” Political Research Quarterly 73(2), 2020, pp. 492–505. Source for the 0.19 percent fall in South Carolina’s turnout in the 2014 general election and the absence of a racial difference. doi.org
Hopkins, D. J., Meredith, M., Morse, M., Smith, S. and Yoder, J. “Voting But for the Law: Evidence from Virginia on Photo Identification Requirements.” Journal of Empirical Legal Studies 14(1), 2017, pp. 79–128. Source for the higher turnout in 2014 in Virginia precincts where more active registered voters lacked a driver’s licence, and the authors’ suggestion that a targeted state mailing may explain it. doi.org
Hajnal, Z., Lajevardi, N. and Nielson, L. “Voter Identification Laws and the Suppression of Minority Votes.” Working paper, University of California, San Diego. Source for the 10.3 point Latino and 12.8 point multiracial general-election estimates, which do not appear in the published article. pages.ucsd.edu
Hajnal, Z., Lajevardi, N. and Nielson, L. “Voter Identification Laws and the Suppression of Minority Votes.” Journal of Politics 79(2), 2017, pp. 363–379. Source for the published article’s gap of 7.1 points in Hispanic general-election turnout with no other factor taken into account, the factors its model allows for, its model’s prediction that Latinos are “10% less likely to turn out”, and its survey-based design. doi.org
Grimmer, J., Hersh, E., Meredith, M., Mummolo, J. and Nall, C. “Obstacles to Estimating Voter ID Laws’ Effect on Turnout.” Journal of Politics 80(3), 2018, pp. 1045–1051. Source for the replication: the matching rules that changed over time and between states, turnout overstated by about 25 points in most state-years and understated by about 10 in 15 states, turnout close to zero where votes were not validated, the placebo test on years before the laws, and the conclusion quoted on this page. doi.org
Hajnal, Z., Kuk, J. and Lajevardi, N. “We All Agree: Strict Voter ID Laws Disproportionately Burden Minorities.” Journal of Politics 80(3), 2018, pp. 1052–1059. Source for the reply’s argument that the replication’s reanalysis confirmed the finding “that strict voter ID laws discriminate”. doi.org
Fraga, B. L. and Miller, M. G. “Who Do Voter ID Laws Keep from Voting?” Journal of Politics 84(2), 2022, pp. 1091–1105. Source for the court-ordered declarations of 2016, the linking of about 16,000 of them to the Texas voter file, and the finding that voters without ID were disproportionately Black and Latino. The shares of 11.4% and 16.1% for Black voters, and 19.8% and 20.7% for Latino voters, are from Table 1 of the authors’ draft of January 2021, which also describes how race was estimated. doi.org; draft: dropbox.com
Levitt, J. “A comprehensive investigation of voter impersonation finds 31 credible incidents out of one billion ballots cast.” Washington Post, 6 August 2014. Source for the about 31 incidents of alleged impersonation from 2000 to 2014, some involving more than one ballot, against more than a billion ballots in general and primary elections, and for Levitt’s bet that some would be debunked. washingtonpost.com
Heritage Foundation, Election Fraud Map, last updated 12 December 2025. A conservative think tank’s compilation, cited for its own figures and statements. Heritage calls it “a sampling” that is “not comprehensive”, limited to cases that ended in a finding of wrongdoing. Source for its 1,619 cases from 1982 to 2025, the 34 of impersonation at the polls and the 387 of fraudulent use of absentee ballots (counted in the case file its search page uses), and the two statements quoted on this page. electionfraud.heritage.org
Mehrbani, R. “Heritage Fraud Database: An Assessment.” Brennan Center for Justice, 8 September 2017. Cited for its own count. Source for the database’s 749 cases involving almost 1,100 people in 2017, 10 of them in-person impersonation. Its voter ID page, the source for its position, says it supports “commonsense voter ID rules, not requiring ID that many eligible Americans don’t have”. brennancenter.org; voter ID page: brennancenter.org
National Conference of State Legislatures, “Voter ID Laws”, updated 3 September 2026. Source for the 36 states with ID laws, the strict and non-strict definitions, the 14 strict states in force in 2026, and the arguments proponents make. ncsl.org
Crawford v. Marion County Election Board, 553 U.S. 181 (2008). Source for Justice Stevens’s lead opinion, joined by two justices: that the only fraud the Indiana law addresses is in-person impersonation, that the record held no evidence of it in Indiana, the state interests it accepted, and its quotation of the Carter-Baker commission’s report of September 2005, with the commission’s two chairs; and for the district court’s findings, quoted in the opinion, that no individual unable to vote under the law had been identified and that the 989,000 estimate was unreliable. loc.gov
Veasey v. Abbott, No. 14-41127 (5th Cir., 5 August 2015). Source for the free election certificate, the $2 to $3 charged for the birth certificates needed to get one until 2015, and the estimate of the plaintiffs’ expert Stephen Ansolabehere, which the district court relied on, that Black and Hispanic registered voters were 305% and 195% more likely than white voters to lack an accepted ID; the panel upheld the district court’s finding of a discriminatory effect. law.justia.com
Bowler, S. and Donovan, T. “A Partisan Model of Electoral Reform: Voter Identification Laws and Confidence in State Elections.” State Politics & Policy Quarterly 16(3), 2016, pp. 340–361. Source for the 2014 finding that Republicans in states with strict photo ID laws were more confident in their state’s elections, and Democrats less. The comparison with states without ID laws is from the authors’ post on the paper on the LSE United States Politics and Policy blog, 17 February 2016. doi.org; post: blogs.lse.ac.uk
Sances, M. W. and Stewart, C. “Partisanship and confidence in the vote count: Evidence from U.S. national elections since 2000.” Electoral Studies 40, 2015, pp. 176–188. Source for the consistent link, across about 30 national surveys, between voting for the winner and confidence in how elections are run, as the abstract of the authors’ 2014 working paper states it. Their argument about ID laws is from their post on the paper on the LSE United States Politics and Policy blog, 11 February 2016. doi.org; post: blogs.lse.ac.uk
International comparison. Mexico’s Instituto Nacional Electoral says getting the Credencial para Votar is free (centralelectoral.ine.mx), and its figures of 1 October 2026 give 101,411,425 people with a valid credential among 102,453,671 on the electoral roll in Mexico (ine.mx). India’s Election Commission said in February 2019 that more than 99 per cent of electors held its photo identity card (pib.gov.in). Elections Canada accepts two pieces of ID without a photo, or vouching, in federal elections (elections.ca). A London council’s guide gives photo ID at polling stations from May 2023 and the free Voter Authority Certificate (redbridge.gov.uk). Cited for the design of those schemes, not for turnout figures.
Note on what the turnout figures are, and are not. None of the turnout estimates on this page counts people prevented from voting. They compare turnout before and after a law against states or voters the law did not affect, so they capture the net of any deterrence and any counter-mobilisation. The Texas figures are different in kind. They describe people who did vote, and what they had to do to vote without ID.
Cite this article
TruthBased.org. “Do Voter ID Laws Reduce Turnout?” Updated November 2026. https://www.truthbased.org/do-voter-id-laws-reduce-turnout