In Columbia, South Carolina, protesters register disapproval of the state’s voter i.d. law.Tim Dominick / The State / Getty
Ever since the contested 2000 presidential election, the way that American elections are run has become increasingly partisan and contentious. The 2016 elections ratcheted up the number of complaints by all parties, yet there is heated disagreement about the nature of the problem — let alone potential solutions.
For many years, the main complaint by the GOP has centered on alleged incidents of illegal fraud, in which it is claimed that ineligible people registered and cast ballots, for example non-US citizens and felons, or simply imposters voting more than once. Throughout the campaign Donald Trump stoked up the heated rhetoric by alleging that victory would be stolen from him. After he won the Electoral College vote, he claimed (falsely) that he also won the popular vote “if you deduct millions of people who voted illegally.” In fact, across the country, officials found next to no credible evidence for cases of voter fraud.
For Democrats, by contrast, the main problem has been framed as one of the suppression of voting rights designed to depress legitimate citizen participation. Civil rights organizations routinely criticize attempts by GOP state legislatures to tighten voter ID requirements and restrict polling facilities, making it harder to vote, especially for minorities and the elderly. Here the evidence about the impact of implementing stricter registration requirements in depressing the vote is somewhat clearer, although debate continues about the size of the effect, among other questions.
On polling day, journalists highlighted accidental failures in the nuts and bolts of electoral maladministration, including human errors and machine breakdowns in registration and ballotingThe New York Times reported that scattered problems occurred on November 8 in several polling places, with long lines in North Carolina, Virginia, New York, and Texas, sporadic breakdowns for the electronic register in Durham NC, and malfunctioning voter verification in Colorado.
This year’s highly consequential twist was the hacking of the computer server of the Democratic National Committee and the private emails of Hillary Clinton’s campaign chairman, John Podesta — an activity the CIA and FBI have pinned on Russia.
On the cybersecurity question, President Obama has ordered a report by the CIA and FBI. But more generally, given divergent claims by each party, is there independent and reliable evidence to support criticisms about the performance of American elections?
One useful tool is the Electoral Integrity Project (EIP), an independent academic project based at Harvard and Sydney Universities. For the last five years, we have conducted the Perceptions of Electoral Integrity survey in the US and around the world. EIP interviews political experts about various aspects of elections in the areas where they live, a technique commonly used for evaluating performance in the absence of directly observable indicators. The approach is similar to that employed for the highly respected Corruption Perception Index by Transparency international.
Tapping experts for their opinions about local electoral integrity
The core concept of “electoral integrity” refers to international standards and global norms governing the appropriate conduct of elections during the pre-election period, the campaign, polling day and its aftermath. For the 2016 US election, the Perceptions of Electoral Integrity staff gathered evaluations from 726 political scientists based at universities in each state. We asked respondents, two weeks after polling day, to evaluate electoral integrity in their own state.
The survey included questions touching on 49 core indicators of electoral integrity, which were then grouped into 11 categories reflecting all stages of the electoral cycle: pre-election, campaign, polling day and its aftermath. For example, the experts were asked whether elections were well managed, whether votes were counted fairly, whether boundaries for candidates’ districts were set fairly, whether newspapers provided balanced election news, if ballot boxes were secure, and whether women had equal opportunities to run for office, among many other issues.
Each category rating is standardized to 100 points. The dataset also includes an overall 100-point index created by summing all 49 indicators. We also gathered information about each state, including the partisan composition of state legislatures and the share of the vote in the 2016 presidential race, which allowed us to determine, among other things, whether voting integrity was correlated with support for either Hillary Clinton or Donald Trump.
The project also collected details about the individual experts, such as age, sex, and ideological positions, to test whether these characteristics were systematically associated with evaluations of electoral integrity.
Is the resulting data reliable? It can be argued that political scientists are not neutral judges, given the well-known academic bias towards supporting liberal values (and, often, liberal political parties). But the external validity of the Perceptions of Electoral Integrity Index has been widely tested in previous research and found to be strongly correlated with other standard sources of evidence, like Freedom House and Polity.
It is important to be cautious when interpreting absolute rankings, since the differences between states were often relatively modest. What’s more, the number of responses was limited in some states, such as Utah and North Dakota (although none of the states with limited responses are among the worst performing). Finally, as its name suggests, the survey measures expert perceptions of the integrity of the voting process. Experts’ opinions serve as proxies for the underlying phenomena. However, the opinions are experts are valuable in their own right; if many people think that there is fraud, this is a problem, even in the absence fraud. The PEI index has been widely cited by scholars and practitioners around the world, and has become the most comprehensive measure for comparing electoral performance from Australia to Zimbabwe.
We present here some initial results, but we are still sifting carefully through the evidence. Our results should eventually be compared with other data as it becomes available — including state performance indices (such as voting wait times and turnout rates), forensic analysis of precinct-level voting statistics, scrutiny of credible complaints and legal cases, surveys of poll-workers and local electoral officials, analysis of social media, and surveys of public opinion.
Ranking electoral integrity and malpractices across US states
Figure 1 shows how experts evaluated the 2016 elections across all 50 US states and DC, on a 100-point scale. The patterns show that the south remains the region that local experts believe has the lowest levels of electoral integrity. The Supreme Court essentially ruled that voting restrictions in the South are a bygone problem when it blessed elimination of Section 5 of the Voting Rights Act, which had required states with a history of racial discrimination to get Department of Justice approval before changing voting laws. But evidence from these expert evaluations suggests that that decision may have been unduly optimistic. By contrast the Pacific West and New England were the regions that experts view most positively with respect to the quality of their elections.
But state performances varied even within major regions. Overall, the states scoring as worst in the perceptions of electoral integrity index in this election were Arizona (ranked last), followed by Wisconsin, Tennessee, Oklahoma, and Mississippi. Several of these states had also been poorly rated previously in the 2014 Pew Election Performance Index. On the other end of the spectrum, the states that experts rated most highly in electoral integrity were Vermont, Idaho, New Hampshire, and Iowa.
The exact reasons underlying these poor ratings still need further study. Problems can arise at any stage of the electoral process — not just at the ballot box. Figure 2 looks at how experts evaluated state performance across 11 distinct stages of the 2016 contest. (It also shows the states ranked, from 1 to 50.) The stages with the cleanest bills of health across all the states include the vote count, the voting process (such as whether fraudulent votes were cast), and the role played by electoral authorities. Greater areas of weaknesses, in experts’ eyes, include the how district boundaries are determined, state electoral laws, and the issues of campaign media and money.
Some issues related to electoral integrity are being debated while others are ignored
Clearly some of these issues are already on the mainstream reform agenda. Allegations of voter fraud received massive attention during campaign, as did changes to state electoral laws allegedly to prevent such fraud. Bernie Sanders made a campaign issue of the need to reform the role of money in politics, as did, in a different way, Donald Trump.
“Fake news,” which can lead voters astray on the issues and candidates, has gotten a great deal of attention. But there are also broader issues about the mainstream media that ought to raise serious concern. Reports by Harvard’s Shorenstein Center and other media watchdogs have highlighted a lack of substantive policy discussion during the campaign, dubious false equivalency standards in much journalism, and the overwhelmingly negative tone of news coverage.
The media has been examining its conduct, to some degree. By contrast, the issue of gerrymandered district boundaries, regarded by experts as the worst aspect of US voting procedures, was never seriously debated during the campaign. The practice ensures that representatives are returned time and again based on mobilizing the party faithful, without having to appeal more broadly to constituents across the aisle, which in turn exacerbates the bitter partisanship that plagues American politics. Gerrymandering through GOP control of state legislatures has also led to a systematic pro-Republican advantage in House districts which is likely to persist at least until 2022. In 2016, House Republicans won 241 seats out of 435, or 55 percent, although they won only 49.1 percent of the popular vote.
Where malpractices occurred, were they in states won by Trump or Clinton?
So how far does the performance of states relate to party control of state legislatures? Figure 3 shows expert assessments of each of the stages during the electoral cycle compared with which party controlled the statehouse. According to the expert evaluations, states controlled by Democrats usually had significantly greater electoral integrity than Republican-controlled states, across all stages except one. That stage involved the official declaration of the results. (That result may have been a byproduct of protests in several major cities following the outcome, reflecting challenges to the legitimacy of the Trump presidency.) The partisan gap was statistically significant on the issues of gerrymandered district boundaries, voter registration, electoral laws, and the performance of electoral officials.
Given this pattern, not surprisingly there was a clear tendency for Trump to win more states with electoral malpractices (see Figure 4), while Clinton won more states with electoral integrity. We do not claim, as we do not have sufficient evidence, that Trump won these states because of malpractices. But the correlation is clear.
Elsewhere our research has demonstrated that the 2012 and 2014 American elections ranked poorly in comparison with many other countries, with the United States scoring the worst in electoral integrity among Western democracies. The US also ranks 52nd out of all 153 countries worldwide in the cross-national electoral integrity survey. The comparison is even worse for the issue of district boundaries, where the US score is the second lowest in the world.
This new comparison of US states gives us a more detailed picture of the reasons behind this dismal performance.
The consequences of poor electoral practice are significant: They include the erosion of public faith in political institutions and lower confidence in democracy. It’s clear America needs a bipartisan commission to tackle the real shortcomings in US elections. If the federal government remains paralyzed, states can implement reforms on their own, where possible. Such reforms should include strengthening cybersecurity for official records, protecting voting rights, limiting partisan gerrymandering through independent commissions, improving fair and accurate campaign communications, reforming the Electoral College, and cleaning up campaign finance.
Whether any of these urgent reforms can be implemented in the current climate of bitter partisanship remains to be seen. But countries that fail to reach a consensus about the legitimacy of the basic electoral rules of the game, especially those with deeply polarized parties and leaders with authoritarian tendencies, are unlikely to persist as stable democratic states.
Pippa Norris is the McGuire Lecturer in Comparative Politics, Laureate Professor of Government and International Relations at the University of Sydney, and director of the Electoral Integrity Project. She is the author of Why Electoral Integrity Matters (2014) andWhy Elections Fail (2015). Holly Ann Garnett is a PhD candidate at McGill University. Max Grömping is a PhD candidate at the University of Sydney. More details and data are available at www.electoralintegrityproject.com.