Introduction

Voter fraud – the illegal interference with the election process by voters or officials – has been a persistent concern in American politics. Over the past 40 years, multiple investigations, court cases, and studies have examined the prevalence of voter fraud in U.S. elections. The evidence overwhelmingly shows that fraud is extremely rare and has had virtually no impact on outcomes in federal elections​. This report provides a decade-by-decade analysis from the 1980s through the 2020s, detailing documented fraud cases (allegations, investigations, and convictions) and assessing whether any election results were materially affected. It also evaluates the effectiveness of voter integrity laws in preventing fraud, and considers how stricter voting requirements (like voter ID and proof of citizenship) could influence both fraud prevention and legitimate voter access, including potential disparate impacts on certain demographic groups.

1980s: Isolated Fraud in Traditional Machine Politics

During the 1980s, reported voter fraud at the national level was minimal, though a few high-profile cases in local politics drew attention. The vast majority of votes were cast and counted without incident, and no federal election outcomes were overturned due to fraud in this decade. Notable incidents included:

  • 1982 Illinois (Chicago) – A federal investigation uncovered widespread fraud in Chicago’s Democratic machine during the 1982 gubernatorial election. Sixty-two people were indicted and 58 were convicted for election fraud (including ballot box stuffing and votes cast in the names of dead voters)​. While thousands of fraudulent votes were reportedly cast in Chicago, they did not change the overall outcome of the Illinois governor’s race, which the Republican candidate won despite the fraud. This case did highlight vulnerabilities in big-city political operations and led to prosecutions of precinct officials.
  • 1987 Chicago Mayoral Primary – Investigations by the Chicago Board of Elections and a watchdog group found that “tens of thousands” of fraudulent votes were cast in the Democratic primary for mayor​. This fraud occurred in the context of intense intraparty competition. Even so, the fraudulent votes did not ultimately prevent the reformist mayor, Harold Washington, from winning the primary. The incident underscored ongoing issues in Chicago’s election administration but was an outlier rather than representative of national trends.

Outside of such cases, voter fraud in the 1980s was virtually negligible on the national scale. Hundreds of millions of ballots were cast in three presidential elections (1980, 1984, 1988) and numerous congressional elections during this decade. Documented fraud cases numbered only in the dozens nationally, aside from the Chicago investigations. Academic analyses have found overall fraud incident rates in this era were well under one-thousandth of a percent​. No evidence suggests any major state or federal election in the 1980s was “stolen” through fraud – even the aggressive fraud uncovered in Chicago did not reverse election outcomes beyond that locality. This low baseline would set the stage for the following decades, in which fraud remained rare and isolated.

1990s: High-Profile Cases and Election Challenges

The 1990s saw a few high-profile voter fraud cases and election disputes, primarily at the state and local level. These incidents garnered headlines but remained isolated exceptions, and no congressional or presidential results were overturned due to fraud. Key examples from this decade include:

  • 1993 Pennsylvania State Senate (Philadelphia) – A special election for a Pennsylvania state senate seat was invalidated by a federal judge in 1994 after evidence showed the Democratic candidate “stole” the election through absentee ballot fraud​. Hundreds of fraudulent absentee ballots had been cast in Philadelphia to benefit candidate William Stinson. The court threw out those ballots and ultimately awarded the seat to the Republican candidate (Bruce Marks), flipping control of the state senate​. This was a rare instance of proven fraud actually altering an election outcome in the 1990s.
  • 1996 U.S. House Election (California 46th District) – Incumbent Rep. Bob Dornan challenged his loss to Loretta Sanchez amid allegations that non-citizens had voted. A House committee investigation found 748 improperly cast ballots, including 624 votes by non-citizens​. However, Sanchez had won by 979 votes, so the illegal votes (even if all for Sanchez) were not enough to change the result​. No criminal indictments followed; many of the non-citizens had registered by mistake or had become citizens by Election Day. The House ultimately dismissed Dornan’s challenge, confirming that the outcome was legitimate. This case illustrated that claims of fraud were taken seriously but also that even a few hundred improper votes did not swing the result of a federal election in this era.
  • 1997 Miami Mayoral Election – This city election became one of the most notorious U.S. fraud cases of the late 20th century. An investigation revealed a “well orchestrated” absentee-ballot fraud scheme in which nearly 400 fraudulent absentee ballots (including votes cast for dead or fictitious people) were submitted​. The fraudulent ballots tipped a close race – the incumbent mayor had actually won the in-person vote, but lost after bogus absentees were counted. A Florida circuit judge threw out the election, citing “a pattern of fraudulent, intentional and criminal conduct,” and ultimately removed the apparent winner (Xavier Suarez) from office​. The runner-up (Joe Carollo) was instated as mayor. This Miami case demonstrated that fraud could affect a local contest, though it required a concerted criminal conspiracy to do so. It prompted Florida to pass tougher anti-fraud measures for absentee voting in 1998.

These episodes show that while localized fraud conspiracies occurred in the 1990s, they were rare and usually detected. Most importantly, none of the disputed or fraudulent episodes at the state/local level cast doubt on national election outcomes. For example, despite scattered fraud claims, no U.S. presidential or Senate election in the 1990s was overturned or even seriously challenged on fraud grounds. The overall rate of voter fraud remained extremely low. In one oft-cited analysis, a political scientist found that from 1992 to 2000, only 0.0003% to 0.0025% of votes may have been fraudulent – effectively a few hundredths of a percent at most. The voter fraud that did occur had no material impact on federal elections in the 1990s.

2000s: Modernization, Scrutiny, and Continued Low Incidence

The disputed 2000 presidential election (though centered on ballot design and counting, not fraud) brought national attention to election integrity. In response, the 2000s saw significant election reforms – notably the Help America Vote Act of 2002 (HAVA) – aimed at modernizing voting systems and improving voter registration processes. These reforms, along with continued vigilance by authorities, kept the rate of voter fraud extremely low through the 2000s. No federal election outcomes were overturned due to fraud in this decade, although a few localized schemes came to light. Key developments and cases included:

  • Election Reforms and Oversight: HAVA mandated statewide voter registration databases, provisional ballots, and ID requirements for first-time voters who register by mail, all of which improved integrity. The U.S. Department of Justice also launched a “Ballot Access and Voting Integrity Initiative” in the mid-2000s to prioritize prosecution of election crimes. Despite intensive scrutiny, these efforts found little fraud. For example, a federal review from 2002–2005 found only an average of 8 federal voter fraud convictions per year​ – vanishingly small numbers relative to the tens of millions of votes cast each year. Multiple statewide audits and academic studies in the 2000s continued to conclude that voter fraud was isolated and infrequent​.
  • Local Fraud Schemes: A handful of conspicuous local cases were prosecuted in the 2000s, showing that while rare, fraud tended to involve small networks conspiring in local elections rather than individual voters at the polls. In Clay County, Kentucky, for instance, a group of eight county officials (including a judge and a school superintendent) rigged several local elections between 2002 and 2006 by buying votes and miscounting ballots; they were eventually convicted on federal charges​. And in East Chicago, Indiana, the 2003 mayoral Democratic primary was thrown out after an investigation found a “widespread and pervasive pattern” of absentee ballot fraud – 46 people were convicted in a conspiracy to buy votes and manipulate absentee ballots​. These schemes did change the outcomes of the specific local contests (the East Chicago case, for example, led to a new election and a different nominee), but they were limited to those jurisdictions.
  • Voter ID Laws Emerge: In the latter 2000s, states began passing voter identification laws, citing fraud prevention. Indiana passed the first strict photo ID law in 2005, and the U.S. Supreme Court upheld it in Crawford v. Marion County (2008), even though the state had not experienced significant voter impersonation problems. The Court acknowledged historic incidents of fraud but also noted the paucity of recent cases; nonetheless, it ruled that states could require ID as a prophylactic measure​. By the end of the 2000s, a handful of states had implemented photo ID requirements at the polls. These laws did not reveal any new troves of fraud; instead, they confirmed that in-person voter impersonation was virtually non-existent (as detailed further below). For example, a comprehensive investigation in 2012 found only 10 cases of voter impersonation in the entire country from 2000 to 2012, about one case per 15 million registered voters​.

Overall, the 2000s reinforced a clear pattern: voter fraud cases were exceedingly rare and isolated, usually involving absentee ballots or local officials when they did occur. The documented fraud rate was near zero in federal elections. One analysis summed it up well: across all elections from 2000–2014, researchers found only 31 credible incidents of in-person voter impersonation out of more than 1 billion ballots cast. In other words, an American was statistically more likely to be struck by lightning than to commit voter impersonation at the polls​. No evidence from the 2000s suggests any state-wide or national races were swayed by fraud; even the razor-thin 2000 presidential contest in Florida was decided by legal vote-counting disputes, not fraudulent votes.

2010s: Voter ID, Fraud Allegations, and a Rare Federal Race Do-Over

The 2010s saw the expansion of voter ID laws and heightened partisan debate over election fraud, especially after contentious elections in 2016. Despite the rhetoric, the empirical incidence of voter fraud remained extremely low. Courts and independent inquiries repeatedly found no widespread fraud in any national contest. However, one U.S. House race was overturned due to fraud in this decade – a notable first in modern times – and several states implemented new voting restrictions in the name of “integrity.” Key highlights include:

  • Expansion of Voter ID Laws: After 2010, many states (largely with Republican legislatures) enacted strict photo ID requirements for voting. By 2019, 36 states had some form of voter ID law (photo or non-photo) on the books. Proponents argued these laws would prevent fraud, particularly in-person impersonation. Critics pointed out that such impersonation is extremely rare – on the order of thousandths of a percent of votes​ – and warned the laws could disenfranchise legitimate voters. Studies during the 2010s continued to find no appreciable level of impersonation fraud to deter. For example, a 2014 nationwide study found only 31 credible cases of impersonation in 14 years (2000–2014). Courts scrutinizing ID laws noted the lack of fraud: a federal judge in 2016 striking down one such law remarked it was “a solution in search of a problem,” since only two in-person fraud cases had occurred out of 20 million votes in that state over a decade​. Nonetheless, voter ID became a prominent feature of the voting landscape in the 2010s. (The effects of these laws on voters are addressed in a later section.)
  • 2016 Election and Aftermath: The 2016 presidential election spurred unsubstantiated fraud claims at the highest levels. After losing the popular vote, Donald Trump falsely alleged that 3–5 million illegal votes had been cast. These claims were investigated: Trump’s own Presidential Advisory Commission on Election Integrity (2017) sought evidence of large-scale fraud. No significant findings ever emerged. In fact, the commission was disbanded in early 2018 without presenting any proof of widespread fraud. Courts and state audits similarly found no basis for Trump’s claim​. For example, a 2016 research paper estimated the upper bound on double-voting in 2012 was just 0.02% (and likely much lower, mostly due to clerical error). In sum, exhaustive reviews around 2016 found at most a handful of illegal votes nationwide, far too few to affect the outcome of the presidential race or other federal contests.
  • 2018 North Carolina’s 9th District – A Rare Do-Over: In the 2018 midterms, North Carolina’s 9th Congressional District saw a genuinely egregious case of fraud that did alter the outcome. Republican candidate Mark Harris appeared to win by 905 votes. However, evidence surfaced that a campaign operative, McCrae Dowless, orchestrated an absentee ballot harvesting and tampering scheme in two counties​. Operatives working for Dowless collected hundreds of absentee ballots from voters (often filling them out or discarding them) to boost Harris’s tally. The scheme was uncovered through unusual absentee voting patterns (Harris mysteriously got 61% of mail-in votes in one county, despite only 19% of those voters being registered Republicans). The bipartisan state elections board investigated and refused to certify the result, concluding the fraud was substantial. This led to the entire election being overturned and re-run in 2019, marking the first U.S. House race in decades to be invalidated for fraud. Dowless and several associates were criminally indicted. Importantly, this case was an outlier – a coordinated fraud operation that managed to corrupt a few hundred ballots and still was caught. It did not indicate any broader trend in 2018’s elections, which were otherwise carried out securely.
  • Other Scattered Cases: Throughout the 2010s, isolated instances of voter fraud continued to be prosecuted, though none approached the scale of NC-9. Examples include a Massachusetts state legislator convicted of casting absentee ballots illegally for voters in his district (2012), a local election in Paterson, NJ voided due to mail-in ballot fraud (2020), and a Philadelphia political operative (a former congressman) convicted of bribing poll workers to stuff ballot boxes in local elections between 2014–2016​. These cases involved dozens or hundreds of votes at most, usually in low-turnout municipal contests. They were uncovered and prosecuted, affirming that safeguards (and whistleblowers) can catch fraud when it happens. Meanwhile, nationwide statistics showed voter fraud remained minuscule. One nationwide review of the 2016 election found only four documented cases of voter fraud out of 135 million votes​. Similarly, a 2019 academic study found no detectable increase in fraud in states with more lenient voting rules, reinforcing that widespread fraud was not occurring.

In summary, the 2010s continued the pattern: voter fraud was exceedingly rare, sporadic, and isolated. Apart from the NC-09 case (which itself was limited to a narrow scheme in one congressional district), there is no evidence that fraud affected any major election outcome in the U.S. during the 2010s. Courts reviewing claims in this decade – including numerous lawsuits after 2016 – repeatedly concluded that assertions of large-scale fraud were unsupported by evidence​. Election officials, including many from the president’s own party, affirmed the integrity of the vote counts. Thus, despite greater public perception of fraud risk (often driven by political rhetoric), the reality was that U.S. elections in the 2010s remained secure from nearly all fraudulent activity.

2020s: Post-2020 Audits, New Laws, and Ongoing Security

The 2020s have thus far been dominated by the fallout from the 2020 presidential election – an event that tested the electoral system like no other in modern history. The COVID-19 pandemic led to unprecedented levels of mail-in voting in 2020, and afterwards, unfounded allegations of fraud reached a fever pitch. Exhaustive recounts, audits, and court cases, however, all arrived at the same conclusion: the 2020 election was secure and free of any widespread fraud. Nonetheless, the perception of fraud (the so-called “Big Lie”) has driven a wave of new state laws tightening voting rules. Here is an overview of the early 2020s:

  • The 2020 Election – Most Secure in History: The 2020 general election saw record turnout (~158 million votes) and the expansion of mail voting. Despite fears, there is no evidence that these expanded voting methods led to significant fraud. On the contrary, a consortium of federal agencies (DHS’s Cybersecurity & Infrastructure Security Agency, etc.) declared the 2020 election “the most secure in American history.” Dozens of state audits and recounts confirmed the results. The Associated Press undertook an exhaustive review of voter fraud allegations in the six battleground states that decided the presidency – Arizona, Georgia, Michigan, Nevada, Pennsylvania, and Wisconsin – and found fewer than 475 potential cases of voter fraud out of over 25 million ballots cast in those states. That is approximately a 0.0019% fraud rate in those contested states, and even those 475 cases were only allegations (the number of actually fraudulent votes could be smaller after investigation). Crucially, 475 votes would not have changed the outcome in any of those states – the closest state (Georgia) was decided by about 12,000 votes, far beyond the scale of possible fraud. In many states, the number of confirmed fraudulent ballots in 2020 was in the single digits. For example, a Pennsylvania analysis found only a handful of double votes or votes cast in deceased persons’ names after 2020. Courts across the country (over 60 lawsuits) rejected claims of fraud or irregularities due to lack of evidence, and even President Trump’s own Justice Department stated it found no fraud on a scale that could change the election’s outcome. In short, while 2020 saw more allegations of fraud than ever, the actual incidence of fraud remained at essentially trace levels – similar to prior years, on the order of a few hundred at most questionable ballots nationwide, out of 160+ million cast. No credible evidence has emerged to suggest any outcome in the 2020 presidential or congressional races was affected by fraud.
  • Election Integrity Legislation Post-2020: In response to the widespread (but unfounded) distrust sowed after 2020, many states moved to change their voting laws in 2021 and 2022. Legislatures in states like Georgia, Texas, Florida, and Arizona passed bills under the banner of “voter integrity.” These laws tightened rules on mail-in ballots (e.g., stricter ID requirements to request or return absentee ballots, limits on drop boxes, banning of unsolicited mail-ballot applications), reduced early voting periods in some cases, and gave states more authority to purge voter rolls. At the same time, other states moved in the opposite direction, expanding voter access with mail voting or automatic registration. The net effect has been a patchwork of rules in the 2020s. It is too early to fully assess the impact of these new laws on fraud or turnout, but early indications (from the 2022 midterms) are that election outcomes remained unaffected by fraud. The 2022 elections saw robust security and very few fraud cases reported. For instance, Texas’s attorney general – in a state that enacted strict new voting rules – identified only 16 people statewide who voted illegally (all of whom were ineligible felons) out of nearly 9 million votes in 2022, and even those appeared to be mistakes rather than intentional fraud. Similarly, Georgia’s post-2020 fraud investigations have yielded scant results: an audit of 15,000 signatures on mail ballots in Cobb County, GA found no fraudulent votes – 0% – and only two technical errors. In summary, early data from the 2020s suggest that new voting restrictions have not revealed any hidden fraud epidemic; they have mainly served to make voting rules stricter despite the already high security of U.S. elections.
  • Current Fraud Cases: Voter fraud has not vanished entirely, of course – but it remains exceedingly rare. In the 2020s so far, confirmed cases continue to trickle in at very low levels. For example, in Florida’s 2020 election review, state officials referred 75 potential double-voting cases for investigation (out of 11 million ballots in Florida). Ohio found 75 possible cases of individuals voting in two states in 2020 (out of ~6 million votes)​. Missouri officials flagged 4 potentially fraudulent votes in 2020 (out of 3+ million). These numbers are microscopic – on the order of 0.001% of votes or less – and even many of these “cases” can turn out to be clerical errors or address changes rather than willful fraud. A few local elections have experienced issues: in Paterson, NJ (2020), a city council election was invalidated due to a small group engaging in mail-in ballot theft (about 900 fraudulent ballots were involved) – an isolated incident resulting in new local elections and criminal charges. And as recently as 2023, a mayoral primary in Bridgeport, CT was voided by a judge after video evidence showed an individual apparently stuffing a ballot drop box with multiple ballots (an investigation is ongoing). While concerning, these events are outliers and have been caught by oversight; they do not signify any systemic problem. The overarching pattern in the 2020s remains that voter fraud is exceptionally uncommon and promptly dealt with when discovered. Indeed, a prominent expert, Justin Levitt (now an DOJ official), summed it up: out of hundreds of millions of votes cast in recent years, only a few hundred allegations of fraud arise, and only a fraction of those are substantiated – meaning the fraud rate often rounds to effectively 0.000%.

In conclusion for this section, the early 2020s reinforce the lesson of prior decades: U.S. elections have robust integrity, and instances of fraud are isolated needles in the haystack of votes. Even amid record turnout and novel voting methods, there has been no credible indication of fraud at a scale capable of swinging statewide, let alone national, races.

Election Integrity Measures: Effectiveness in Preventing Fraud

By the 2020s, a variety of election integrity measures are in place across the United States – from voter ID laws to advanced ballot security features – aiming to prevent or catch fraud. Evaluating their effectiveness requires understanding what types of fraud are most prevalent. Research consistently shows that the most common forms of fraud are absentee/mail ballot misuse and voter registration irregularities, whereas in-person voter impersonation is exceedingly rare​. Consequently, measures like improved mail-ballot verification and accurate voter roll maintenance tend to address more realistic vulnerabilities, whereas strict photo ID laws target a form of fraud that almost never happens. Here is an analysis of key integrity measures and how well they work:

  • Voter Identification (ID) Laws: Voter ID requirements are one of the most visible anti-fraud measures. As of 2024, 36 states require voters to show some form of ID at the polls (and most of these require a photo ID)​. These laws are intended to stop someone from impersonating another registered voter at in-person polling places. Effectiveness: Given how rare voter impersonation already is, ID laws have a very limited pool of fraud to prevent. In theory, they can deter or block the few would-be impersonators out there. For example, in Kansas, the Secretary of State credited an interstate database with catching a couple of double voters in 2016 who cast ballots in two states (those individuals were prosecuted). Such instances are exceedingly uncommon – on the order of a few cases out of millions of votes. Indeed, a comprehensive News21 database found only 10 cases of voter-impersonation fraud in the entire country from 2000–2012​. None of those cases resulted in an election outcome changing. Thus, while voter ID laws do stop that virtually non-existent fraud from occurring, their overall impact on fraud prevention is marginal simply because in-person impersonation was nearly zero to begin with. Moreover, voter ID laws do not address other forms of fraud. An ID at the polls does nothing to stop absentee ballot fraud, ballot box stuffing, or officials altering tallies. Notably, many of the proven fraud cases (e.g. Miami 1997, North Carolina 2018) involved mail ballots or election workers – scenarios unaffected by voter ID rules. As one analysis concluded, in states with strict ID, “the fraud that matters is organized fraud, and voter impersonation is practically non-existent”​. Bottom line: Voter ID laws likely prevent an almost negligible amount of fraud, but they can improve public confidence in elections. Their effectiveness is inherently limited to a fraud type that constitutes roughly 0.5% of fraud cases (see chart below), and evidence shows election outcomes in the ID era have not materially changed compared to prior periods of no ID – consistent with fraud having been minimal all along.
  • Mail Ballot Security Measures: Because a significant share of known fraud has involved absentee or mail-in ballots, many safeguards focus on this area. These include signature verification (comparing the voter’s signed envelope to their signature on file), witness requirements (some states require a witness or notary to sign absentee ballots), ballot tracking/barcodes, and limits on ballot collection (to prevent “harvesting” large numbers of ballots). Effectiveness: When properly implemented, these measures have proven fairly effective. Signature verification, for example, has caught instances of people attempting to vote another person’s ballot (those ballots are rejected as mismatches). States routinely reject a small percentage of mail ballots for flaws – e.g., 0.3% of mail ballots in 2020 were rejected nationally, often due to non-matching signatures or missing secrecy envelopes. Those rejected could include fraudulent submissions, though more often they’re innocent mistakes by voters. Importantly, robust processes for mail voting can deter fraud. The 2020 election was a major test: despite a huge expansion of mail voting, there were no systemic fraud issues; states like Colorado, Utah, and Oregon that vote almost entirely by mail have minuscule fraud rates (e.g., Oregon documented only about a dozen cases of proven fraud out of 100+ million mail votes over 20 years – roughly 0.00001%)​. Where mail ballot fraud has occurred (e.g. the North Carolina 9th case, or a Mississippi case of vote-buying via absentee in 2013), it often exploited weaknesses like lack of witness enforcement or lax ballot distribution. New laws in the 2020s have tightened these points – for instance, limiting who can handle someone else’s ballot and expanding criminal penalties for ballot tampering. These changes likely make a repeat of schemes like NC-09 even harder. Overall, measures targeting mail voting fraud (plus vigilant monitoring by campaigns and law enforcement) have kept detected fraud to a tiny fraction of a percent​. The trade-off is that overly strict rules (e.g., requiring notarization) may burden legitimate mail voters (discussed later). But from a pure fraud-prevention standpoint, states’ controls on absentee ballots have generally been effective – evidenced by the rarity of mail-ballot fraud given the tens of millions of postal votes cast each cycle.
  • Voter Registration and Roll Maintenance: Another aspect of election integrity is ensuring that voter registration lists are accurate, so that ineligible names (deceased, moved, non-citizen, etc.) are not used to cast fraudulent votes. Federal law (the National Voter Registration Act) already mandates regular roll maintenance. In recent years, many states joined the Electronic Registration Information Center (ERIC), a multi-state consortium that cross-matches voter rolls to catch duplicates and outdated records. States have also adopted online registration and other modernizations to reduce errors. Effectiveness: These measures help prevent certain fraud scenarios, such as someone registering in two states and voting twice, or votes cast under dead voters’ names. Cross-state data sharing has, for instance, identified a small number of duplicate registrants (who are then removed or flagged). One analysis in 2017 using ERIC data found that out of 14.6 million voters in certain states, only 372 possible double votes or deceased votes were identified (0.0025%) – and those were referred for investigation​. Proper maintenance can thus nip even that tiny number in the bud. However, aggressive purging of rolls can accidentally remove legitimate voters, so it must be balanced (more on this in the next section). Regarding proof of citizenship at registration – a controversial measure some states have attempted – the effectiveness is questionable. Case study: Kansas’s proof-of-citizenship law (2013–2016) required new voters to show a passport or birth certificate to register. It was pitched as a way to block non-citizen registrations. In practice, it prevented over 18,000 eligible Kansans from registering (they couldn’t produce documents in time) while the evidence showed only 14 non-citizens even attempted to register, and just 3 actually cast a vote, in nearly 20 years​. A federal court struck down the law, noting the “magnitude of harm” to legitimate voters far outweighed the virtually non-existent benefit in fraud prevention. This illustrates that stringent registration rules can stop theoretical fraud, but in reality the occurrence of such fraud (non-citizens voting) was extremely rare – about 0.00003% of Kansas’s votes – whereas the burden on valid applicants was much larger. In general, then, current practices like database cross-checks and routine removal of known deceased voters have been sufficient to keep registration fraud low. There is little evidence that additional harsh requirements (like documentary proof of citizenship) add meaningful security; instead, they risk disenfranchising eligible voters while non-citizen voting remains “extremely rare” nationwide (one study put it at between 0% and 0.0004% of votes)​.
  • Election Audits and Equipment Security: Modern election integrity also involves post-election audits (like risk-limiting audits to verify vote counts) and secure voting technology (paper ballots or voter-verifiable paper trails to ensure results can be checked). These measures don’t prevent fraud at the moment of voting but serve to detect and correct any irregularities, whether due to malicious acts or error. Effectiveness: Post-election audits have become increasingly common (as recommended by the National Academies in 2018). They are highly effective at catching discrepancies. For example, Georgia in 2020 conducted a full hand recount/audit of every presidential ballot, confirming the machine count almost exactly and uncovering only a few hundred initially uncounted votes (due to clerical errors, not fraud) which did not change the outcome. Such audits would flag anomalies if, say, ballots had been stuffed or equipment hacked. Likewise, requiring paper ballot records in voting machines (a response to fears of machine tampering) ensures there is a physical record to check. Virtually all states now use systems that leave a paper trail. These measures have made it extremely difficult for any would-be fraudster to alter a count undetected. Notably, in the disputed 2020 election, statewide audits in places like Arizona (including a high-profile audit by a partisan group) found no evidence of fraudulent ballots changing the result. The robustness of auditing processes is a strong deterrent against attempting large-scale fraud, because the chance of getting caught is high and the window to act is short. In essence, today’s election security regime – from equipment certification, to chain-of-custody for ballots, to bipartisan observers and audits – means that large-scale manipulation would be very likely discovered, and so far, none has been.

In summary, current voter integrity measures collectively have kept fraud to a bare minimum in recent U.S. elections. Voter ID laws address the rarest type of fraud and mainly boost confidence rather than stop significant numbers of illegal votes. Measures targeting mail ballots and registration integrity address slightly more prevalent issues and have been largely successful, as evidenced by the tiny fraud rates. And robust auditing and oversight add further insurance that if fraud does occur, it will be caught and remedied. The result is that American elections today have multiple layers of defense, such that the odds of undetected, outcome-altering fraud are effectively close to zero​. Indeed, one recent analysis by the conservative Heritage Foundation’s own database (which tracks fraud cases) concluded that over decades of elections, fraud cases were “a handful…amongst millions and millions of ballots cast, and none of them shifted the election outcome.”​

Impact of Stricter Voter Requirements on Fraud and Voter Access

Debates about election integrity often center on proposals for stricter voter requirements – such as mandating photo ID, requiring documented proof of citizenship to register, limiting early or mail voting, etc. The key questions are: Would these measures significantly improve fraud prevention? And at what cost to eligible voters in terms of convenience or disenfranchisement? Here we analyze projected impacts, drawing on studies and real-world data, and whether such laws have been shown to disproportionately affect certain demographic groups.

Fraud Prevention Benefits

Stricter requirements tend to target theoretical vulnerabilities, but evidence suggests the fraud reduction would be very small because current fraud levels are so low. For instance, requiring a government-issued photo ID for all voters might stop some instances of voter impersonation or double voting. However, as detailed earlier, impersonation is extraordinarily rare (roughly one case in 15 million voters)​. Even doubling or tripling security against in-person fraud would yield only negligible changes in the already tiny fraud rates. Similarly, proof-of-citizenship rules for registration could deter non-citizen voting, but reputable studies have found that non-citizens voting illegally is extremely rare (often effectively 0% in surveys)​. The Kansas example showed that a strict proof-of-citizenship law stopped tens of thousands of legitimate voters while only potentially preventing a handful of improper votes​. Other stringent measures, like culling voter rolls more aggressively to remove duplicates, might prevent a few double votes but run the risk of purging lawful voters in error (false positives). In short, while more stringent laws might catch a few additional illicit votes, the absolute number of fraud cases averted would likely be extremely small – on the order of a few dozen nationwide – given how uncommon fraud is to begin with​. The 2020 election provides a practical baseline: with relatively broad access (many states mailed ballots to all voters, many expanded early voting), there were vanishingly few fraud issues. If more requirements were layered on top, 2020’s fraud rate might have hypothetically dropped from, say, 0.001% of votes to 0.0005% – a difference that is practically undetectable and would not change any outcomes. Meanwhile, such requirements could have other consequences for voters.

Impact on Voter Participation and Access

Stricter voting rules often make it more cumbersome for legitimate voters to exercise their rights. Research has shown that laws like voter ID requirements and proof-of-citizenship can reduce turnout, especially among groups less likely to have the required documents. For example, a comprehensive Government Accountability Office (GAO) study in 2014 examined turnout in states that implemented strict photo ID. It found that turnout dropped by an estimated 2–3 percentage points more in those states relative to similar states without new ID laws. This decline was particularly pronounced among African-American voters and younger voters (18–23), who are less likely to possess driver’s licenses​. Other academic studies have had mixed findings – some detected smaller effects or no clear effect – but there is a consensus that certain demographics are disproportionately impacted by ID laws. Specifically, Americans who are lower-income, racial minorities, elderly, or students are less likely to have current, valid photo ID. One study estimated that nationwide about 11% of voting-age citizens lack government-issued photo ID, with higher rates among African Americans (possibly 25%) and seniors over 65 (18%), for example. Even more recent data from state surveys found gaps: in one state, 85% of White registered voters had a valid ID, versus 81% of Black registered voters. And certain groups such as transgender citizens face challenges if their ID does not reflect their current appearance or name, making it effectively invalid at the polls (around 42% of voting-eligible transgender Americans did not have ID that accurately reflected them, according to a 2020 Williams Institute report). Thus, a strict ID mandate can pose a hurdle for these populations.

Proof-of-citizenship requirements similarly have an outsized effect on certain groups. Naturalized citizens might not always have immediate access to papers like a naturalization certificate or passport, especially when registering through a DMV or voter drive. Women who changed their name might have mismatches in documents. Older individuals born in an era or place where birth certificates were not issued or recorded properly can struggle to produce proof. In Kansas, the thousands of would-be voters stalled by the proof-of-citizenship law included many older individuals; notably, some who had citizenship documents still weren’t registered due to bureaucratic snafus​. The law disproportionately impacted young new voters as well (since they were registering for the first time and had to submit paperwork).

Beyond ID and citizenship rules, cutbacks on early voting or mail voting (often justified by fraud concerns) can also hit certain communities harder. For instance, souls-to-polls drives used by Black churches rely on Sunday early voting – reducing early voting days may curtail minority turnout. Limits on mail ballot collection (ballot harvesting) might burden rural Native American communities, where designated community members often collect and mail ballots for those with no home postal service. While these policies are race-neutral on their face, courts have found some to be intentionally or effectively discriminatory. A striking example: North Carolina’s 2013 omnibus voting law (which included voter ID, reduced early voting, and ended same-day registration) was struck down in 2016 by a federal appeals court, which concluded the law targeted African Americans “with almost surgical precision” – in other words, lawmakers had data on racial voting patterns and crafted the law to hit practices used disproportionately by Black voters​.

Evidence from Jurisdictions with Stricter Laws

We can also look at places that implemented strict requirements to gauge the effects. Georgia and Indiana adopted strict photo ID in the mid-2000s; studies of Indiana’s 2008 election found some drop-off in turnout among groups without ID, though overall turnout also depended on campaign interest. Texas’s 2011 voter ID law was initially blocked after evidence showed hundreds of thousands of registered voters (disproportionately minorities) lacked the specific IDs; a softened version later went into effect. Turnout in Texas did not collapse, but analysis indicated that thousands of lawful votes were not counted because voters lacked ID or faced obstacles with provisional ballots. Kansas’s proof-of-citizenship (while in effect) prevented over 30,000 registrations (until courts intervened) – essentially disenfranchising those would-be voters for several election cycles. Arizona had a proof-of-citizenship rule for state elections that led to thousands of rejected registrations until it was adjusted. On the flip side, states with stricter absentee ballot rules (like Alabama requiring notarization or witnesses) see very low mail voting usage – voters either comply with the extra steps or choose not to vote by mail. It likely prevents some fraud (and also some voting). By contrast, states with no-excuse mail voting and ballot drop boxes (like Colorado, which mails every voter a ballot) have high turnout and extremely low fraud. This suggests that accessibility and integrity are not mutually exclusive – Colorado for example has automatic mail ballots, yet also strong verification (signature checks, etc.) and has virtually no fraud issues; it consistently ranks among the highest in voter turnout.

Disproportionate Impact on Demographics

Virtually every study on voter ID laws has found that the burden falls unevenly across demographics. The Brennan Center and other researchers have documented that racial minorities, the elderly, students, and people with disabilities are less likely to possess current photo IDs or the underlying documents needed to obtain one​. For example, an analysis of Wisconsin’s strict ID law in 2016 found that Black and Latino voters in Milwaukee were far more likely to lack the required ID than white voters, contributing to a depressed turnout in minority neighborhoods. In Texas, the ID law initially did not recognize student IDs or federal employee IDs, but did allow gun carry permits – choices that tend to favor one demographic (older, rural, white) over another (young, urban, diverse). Although many states now offer a free ID to those who lack one, the process of obtaining it can be onerous (traveling to a DMV, providing birth certificates, etc.), effectively creating a new hurdle akin to a modern poll tax (in time and effort if not money). Courts have been attentive to these issues: when striking down or modifying laws, they often cite disparate impacts on African Americans, Latinos, or other protected groups. For instance, the Fifth Circuit Court of Appeals found Texas’s voter ID law in violation of the Voting Rights Act due to its disproportionate impact on minority voters (it noted that in one election at least 600+ votes were not counted because voters lacked ID, and minorities were far more likely to be affected).

Strict proof-of-citizenship laws also have clear disparate impacts: they primarily block naturalized citizens (who are disproportionately Latino and Asian-American immigrants) from voting if they can’t immediately prove citizenship. This is despite the fact that such voters are U.S. citizens; they simply might not have their papers on hand. In Arizona and Kansas, naturalized citizens were far more likely to have their registrations suspended for lack of citizenship proof, even though virtually all were eligible voters. In effect, these laws treated one class of citizens differently (naturalized vs. native-born). That is one reason courts have generally disallowed strict citizenship documentation rules at registration for federal elections.

Projected Impact

If more stringent voter requirements were implemented broadly (national voter ID mandate, universal proof-of-citizenship, etc.), the likely outcome is that a very small number of fraudulent votes would be prevented, but a much larger number of legitimate voters could be impeded. As an illustration: Suppose a national voter ID law prevented, say, 50 impersonation or double-voting attempts nationwide in a general election (a generous estimate). But if even 2% of the 150 million voting-age Americans lack acceptable ID (that’s 3 million people) and even half of them face difficulty voting as a result, that’s 1.5 million legitimate votes at risk. Even with efforts to provide free IDs, it’s estimated hundreds of thousands might not get one in time or could be discouraged by the process. This trade-off raises serious concerns. Election experts often argue that policies should be evidence-based – and the evidence is overwhelming that U.S. voter fraud is too rare to justify broad policies that could disenfranchise significant numbers of voters​. Many judges, in striking down laws, have echoed that sentiment: when virtually no impersonation fraud is occurring, a law that could deter thousands of voters is not a cure but a burden.

Conclusion on Stringent Measures

More stringent voter requirements provide only marginal gains in security in an already secure system, while potentially erecting barriers for segments of the electorate. They have been shown to disproportionately affect racial minorities, low-income voters, the elderly, students, and people with disabilities – the very groups who historically have faced obstacles to voting​. This doesn’t mean all fraud safeguards are bad – rather, the most restrictive measures tend to yield the least benefit. There are often less restrictive alternatives that enhance integrity without hurting access (for instance, using non-photo IDs or verification questions for those without ID, automatic registration to keep rolls updated, robust audit trails to catch fraud instead of blocking voters upfront). Policymakers thus must balance the extremely low prevalence of fraud against the fundamental right to vote for all eligible citizens. As the data show, policies that heavily target voter-side fraud (like ID laws) address a problem that is practically microscopic, while policies that target administrative integrity (like audits and good ballot tracking) have been very effective without disenfranchising voters. The evidence from the last 40 years suggests that the United States can maintain honest elections and high voter participation at the same time.

References

A variety of sources were used to compile this report. They are as follows.

Academic Studies & Nonpartisan Research

Government Reports & Data

Court Cases & Legal Opinions

News and Investigative Reporting

Maps and Visuals

Wikipedia Articles

Updates

Any updates to the article will be noted below.

Updated April 11, 2025 to fix a Heading tag.