top of page

Expert Survey on American Democracy: Dec 2017 - Jan 2018

Survey on American Democracy (November 2017)

Our survey results between December 2017 and January 2018 demonstrate mixed results, with some threat ratings improving and others worsening. From November to January, ratings improved on all six dimensions of democratic performance, reversing the deterioration from October to November. However, democracy experts still see American political behavior in 2017-18 as firmly outside the norm for consolidated democracies. Further, experts rate a 11.3% chance of democratic breakdown in the U.S. within the next four years and 93% of respondents believe that democratic quality has declined over the last 10 years. Both of these figures increased from November.

In general, the results indicate a high point of threat in November-December, but January saw a return to earlier (still high) levels. This may reflect a relatively quiet month for Trump or experts perceived warning signs late last year that abated.

From May through January, we polled 463 democracy experts on threats to American democracy. We previously asked respondents once a month. January is the sixth month after fully switching to a daily-updated rolling survey, although we will continue with our monthly update posts. (We summarize two months here since the December survey was limited due to holidays.) See here for a longer description of the survey methodology and sample, including our attempts to limit ideological bias. As a useful set of comparisons, we also asked the same questions about five other countries: the United Kingdom, Canada, India, Poland, and Hungary. Separate experts were chosen for each country, with a total of 71 respondents. As of January, the U.S. rates comparably to India and Poland on threats to democracy, although better than Hungary.

Survey Questions

We first asked about six categories that have comprised key warning signs of democratic decline elsewhere: (1) Treatment of the media, (2) Constraints on the executive against abuses of power, (3) Elections and treatment of the opposition, (4) Civil liberties, (5) Civil violence, and (6) Rhetoric indicating democratic erosion or weak normative attachment to democracy. We asked about political leaders’ behavior in these categories, but did not single out specific leaders.

Respondents graded each category from 1 to 5, with higher values indicating greater threat to democracy. To ease interpretation, the values were chosen to have fairly concrete and tangible meanings: 1 = Normal consolidated democracy, 2 = Moderate violations atypical of consolidated democracy, 3 = Significant erosion of democratic quality with potential for future breakdown, 4 = Near-term survival threatened, and 5 = Non-democracy. Intentionally, coding a 3 or above on any dimension is a high bar.

We also asked respondents about the likelihood of democratic breakdown, whether democratic quality and stability has improved or declined over the last 10 years, and what recent events or actions (if any) they consider most threatening to democracy.

Shifts Over Time

Shifts over time in expert survey of democracy (May-November 2017)

Results indicate declining concerns in January compared to November, with average threat ratings decreasing across all six categories. However, only four dimensions declined from December to January, with one increasing (Civil Violence) and one holding steady (Civil Liberties). However, we could also see this decline as a return to the already-high threat levels of August-September. An exception is for Civil Violence, where the threat has steadily declined (except for a small increase in January) since Charlottesville.

The above figure shows the average rating for each category from May to January. For comparison, the overall average for India is 2.17, Poland is 2.28, and Hungary is 2.90. The U.S. still averages above India and Poland in democratic threats, but is no longer approaching the estimated threat for Hungary.

Average predictions for democratic breakdown have generally held close to steady from July to January (10.3% to 11.3%), as well as the number indicating democratic decline over the past 10 years (about 85-93%).

December-January Results

Expert survey on democracy (December 2017 - January 2018)

The above figure shows the average response by category in December and January. Experts still see the greatest threat manifested in anti-democratic rhetoric, with several pointing specifically to the president. One respondent noted that "the President's use of Twitter and other public statements raise the risk that citizens lose confidence in the democratic process." Another pointed to "constant rhetoric by Trump that undermines trust in basic institutions of the American democracy.

Treatment of the media regained its place as the second greatest threat. One respondent claimed that "Trump's attacks on media ('enemy of the people, etc.) are troubling." Others pointed to the "the president's attitude that opposition & unfavorable press coverage is illegitimate" and "unrelenting attacks on media."

Constraints on executive power represents the next greatest source of threat, with one respondent warning of "Ongoing threats to media and judiciary. Lack of political will to hold officials accountable. Lack of civilian oversight over military." Another warned of "rhetoric surrounding Mueller investigation, talk of coups against president, unwillingness to follow rule of law."

Further behind are treatment of elections, with continuing concern about the questioning of election legitimacy. In contrast, concerns about civil violence have declined since August (when Charlottesville occurred) and concerns about civil liberties have declined since October.

When asked to identify the most threatening recent event (if any), the most common response was the president's attacks on the media. About 20% of respondents referenced this in some way. Other frequent responses were attempts to interfere with the Mueller investigation and anti-democratic rhetoric by the president.

Expert survey on democracy in six countries (November 2017)

How does the U.S. compare to other countries on the same questions? The above figure shows the average threat rating across the six categories for each country. Countries fall in three groups: the United Kingdom and Canada face the lowest threat level; the U.S. fits in a middle group with Poland and India; and Hungary faces the greatest threat. As of November, the U.S. rates worse than India and Poland, although narrowly. This comparison helps to validate the survey questions—not all democracy experts rate their country of expertise as being threatened or outside the norm, even in a Conservative-led UK. Rather, the U.S. jumps out as distinct from other consolidated democracies.

Expert survey on American democracy (December 2017 - January 2018)

The third figure shows the percentage who rated each category 2 or above (indicating behavior atypical of a consolidated democracy) and 3 or above (indicating erosion and future breakdown threat). Worryingly, majorities rate the U.S. as outside of the norm for stable democracies on every dimension but Civil Violence. Nearly all rate rhetoric, executive constraints, elections, and treatment of media as outside the norm.

This is the second time that a category besides rhetoric has a majority rating a 3 or above, which is true of media treatment and executive constraints. In general, respondents were cautious about assigning high values—only 6 response (1.0% of total) registered the highest threat category of 5. This reassures us that respondents are not amplifying their answers for effect.

We also asked respondents whether democratic stability and quality has improved or declined over the last 10 years (chosen so the reference point is another Republican president). 93.1% responded that American democracy has declined, with 36.6% saying it's "much worse." One respondent said it has improved.

Finally, we directly asked respondents about the likelihood of democratic breakdown (by their definition) within the next four years. Note that "breakdown" does not imply full dictatorship, only a sufficient erosion of democratic functioning. The responses averaged an alarming 11.3%, with a median of 10%. Responses ranged from a low of 0 (given by 10 respondents) to a high of 85%, with 14 answering 20% or higher. Note that 11.3% is orders of magnitude higher than what traditional models of democratic breakdown would predict for the U.S. given its stability and socioeconomic advantages.


The results give us clear reason to remain concerned about the future of American democracy. The past two months saw declines in threat levels in some respects, but only slightly and they remain at a high level. There is a near-consensus that American democracy has weakened over the last 10 years and is now outside the norm for consolidated democracies, especially in rhetoric, media freedom, and executive constraints. As discussed earlier, one should not view rhetoric being the highest-threat category positively since rhetoric can erode the norms holding democratic compacts together and often predicts later anti-democratic behavior.

Comparing the U.S. to other countries provides useful context. American democracy rates as considerably worse than democracy in Canada and the United Kingdom, and now worse than India and Poland (as of June). This is highly consistent with seeing American democracy as shifting downward in quality, with at least some chance of breakdown.

Indeed, the estimated 11.3% chance of democratic breakdown (within four years) is a major warning sign. However, the most likely downward path for American democracy remains the steady erosion of democratic norms and practices, as followed before by Venezuela, Hungary, Turkey, and many others. This presents threats that extend beyond the current administration and may imperil American democracy for years to come.


Michael K. Miller is an Associate Professor of Political Science at George Washington University.


bottom of page