Authoritarian Warning Survey polls democracy experts on threats to democracy from American political leaders in 2017-18. Respondents are academic scholars who study democratic decline, political institutions, American politics, or countries that have recently experienced democratic erosion. From May 2017 through June 2018, we have received 656 responses from 3195 contacts (a response rate of 20.5%).*
We currently survey a randomly selected sub-sample of our respondent pool (~600 scholars) each weekday. Every day, we calculate a rolling average of these responses as a Democracy Threat Index. This is a live measure of democracy scholars' views on threats to American democracy. Our method weights responses by time and also continually analyzes the data for discrete breaks in threat levels that may indicate major events. You can read more detail about how we construct the Index here.
For comparison, we asked the same questions in June 2017 for five other countries: the United Kingdom, Canada, India, Poland, and Hungary. Separate respondents were chosen for each. All are experts on their respective country, with a rough balance of experts based in the country and living abroad. Across the five surveys, 71 experts responded (from 307 contacted, a response rate of 23.1%).
We acknowledge there are important concerns about ideological balance in academic surveys of this kind. See below for three aspects of our survey that help to mitigate this potential bias.**
Drawing on common patterns of recent democratic erosion, we ask about six categories of threats to democracy:
Treatment of Media: Leaders' treatment of media, respect for free press, and transparency
Executive Constraints: Effective constraints of executive against abuses of power; leaders' respect for judiciary, legislature, and rule of law
Elections and Treatment of Opposition: Respect for free and fair elections and legitimacy of opposition
Civil Liberties: Respect for core freedoms (such as speech, assembly, religion, and privacy)
Civil Violence: Use of violence, intimidation, or paramilitary organizations for political ends
Rhetoric: Speech by political leaders indicating democratic erosion or weak normative attachment to democracy
For each category, respondents can choose among five responses:
Within range of a normally functioning consolidated democracy
Moderate violations atypical of a consolidated democracy, but that don't yet threaten breakdown
Violations that signal significant erosion of democracy quality and warn of high potential for breakdown in future
Critical violations that seriously threaten near-term survival
Violations severe enough to make system non-democratic
For the Democracy Threat Index, these responses are re-scaled to 0-100. For the monthly results, we leave them as 1-5.
We also ask respondents about (1) the likelihood of democratic breakdown, (2) whether democratic quality and stability has improved or declined over the last 10 years (so the comparison is to a Republican president), and (3) what recent events or actions (if any) they consider most threatening to democracy.
Click here for the exact text of the survey.
Click here for the full survey results from May 2017-June 2018 (anonymized).
May-June U.S. Results: (Average) (Median) (Range)
See here for a fuller analysis of the results.
Treatment of Media (2.8) (3) (1-5)
One respondent pointed to "attacks on law enforcement institutions and freedom of press."
Others pointed to "attacks on media as 'enemy of the people'" and "the president's scathing statements about the press and the FBI, and the implied (or explicit) threats in them."
Executive Constraints (2.7) (3) (1-4)
One respondent warned of "efforts by the executive to undermine the judiciary and rule of law."
Another warned of the "continued undermining of central political institutions and norms by the executive."
One respondent warned that "the breathtaking claims of pardon power (abuses of current understanding pardon power, ability to self pardon, and potentially pre-pardon) pose an imminent threat to the rule of law in the country which, if they are not checked, would put the president and those acting on his behalf largely above the law."
Several respondents warned of the lack of legislative oversight, e.g., "in the face of the corruption of the administration."
Elections and Treatment of Opposition (2.4) (2) (1-4)
One respondent warned of "the refusal of the Trump administration to recognize and be concerned about Russian interference with our electoral process"
Another respondent warned of rhetoric "denying accuracy and legitimacy of elections."
Civil Liberties (2.2) (2) (1-4)
One respondent warned of the president's "increasingly confrontational foreign policy stances (e.g. on trade and Iran), given that foreign confrontation is often used a pretext to erode civil liberties."
Several others warned of overreach in immigration enforcement.
Civil Violence (1.8) (2) (1-4)
Rhetoric (3.3) (3) (2-5)
One respondent pointed to "continued demonization of law enforcement, Mueller investigation, & media."
One respondent warned that "the President's use of Twitter and other public statements raise the risk that citizens lose confidence in the democratic process."
However, one respondent argued that "Trump is the symptom, not the cause" and that the greatest threat is the "breakdown of democratic norms in the electorate, and specifically in the Republican party."
We also asked about respondents' guesses on the likelihood that American democracy breaks down (by their definition) within the next four years.
The average estimate was 15.3%.
The median estimate was 10%.
Estimates ranged from 0 to 75%. Only 4 of 79 answered 0.
Lastly, we asked whether "the quality and stability of American democracy" had improved or declined over the last 10 years (chosen so the comparison point is also under a Republican president). Respondents answered on a five-point scale, from "much worse" to "much better."
96.2% indicated decline for American democracy.
39.2% indicated that American democracy was "much worse."
Results for Other Countries
For each country, we list the average rating across the six threat categories, the percentage of all responses indicating behavior outside the norm for consolidated democracies (2+), the average predicted likelihood of democratic breakdown, and the percentage saying democratic stability and quality has declined. The U.S. in July is shown for comparison.
Country: Avg. rating / % outside norm / % breakdown / % decline
United States: 2.52 / 87.3% / 15.3% / 96.2%
United Kingdom: 1.23 / 20.0% / 2.6% / 40.0%
Canada: 1.21 / 20.6% / 0.4% / 5.9%
India: 2.17 / 75.8% / 15.5% / 81.8%
Poland: 2.28 / 75.6% / 23.5% / 84.6%
Hungary: 2.90 / 91.5% / 44.2% / 95.0%
Click here for our April 2018 survey results.
Click here for our March 2018 survey results.
Click here for our Feb 2018 survey results.
Click here for our Dec-Jan survey results.
Click here for our November survey results.
Click here for our October survey results.
Click here for our September survey results.
Click here for our August survey results.
Click here for our July survey results.
Click here for our June survey results.
Click here for our May survey results.
* All survey results are anonymous. The survey was supported by IRB certification #051706 (29200) at George Washington University.
** First, we made an effort to include conservative academics in our sample. Several answers indicate conservative participation (e.g., one respondent claims that "the real threats to American democracy took place during the Obama administration"). Second, the five other countries serve as credible comparison points, as four of these countries have right-of-center governments and all have academics that tilt left. Thus, any ideological bias should be roughly held constant. Third, we employ reasonably concrete questions. Rather than ask about general approval of current politics, we ask respondents to compare to other consolidated democracies and evaluate whether events threaten democratic survival. As a positive sign, only six responses (1.2% of total) had the value 5 (the greatest threat) in May-June, indicating that respondents are not amplifying the threat for effect.