Exclusive poll: Most Americans don't like Trump's China tariffs

Data: Survey Monkey online poll conducted Sept. 28–30, 2018 among 2,291 U.S. adults. Total margin of error is ±3.0 percentage points. Modeled error estimates: African-American women ±9.0, age 18 to 34 ±6, suburban white women ±6.5, “Never Hillary” Independent voters ±11.0, Rural voters ±5.5. Poll methodology; Chart: Harry Stevens/Axios

A narrow majority of Americans think President Trump's tariffs against China have been bad for jobs, and his overall trade policies are unpopular by a nearly identical margin, according to a new Axios/SurveyMonkey poll.

Why it matters: Trump has a chance to improve his approval numbers now that he's reached a new trade agreement with Canada and Mexico to replace NAFTA. But this poll, taken before the agreement was announced, shows how much ground he has to make up due to the unpopularity of the tariffs.

By the numbers:

  • By a margin of 53% to 42%, Americans think the trade dispute with China has been bad for jobs rather than good for jobs. That's almost an exact match with his overall trade numbers: 53% disapprove, 44% approve.
  • The tariffs are unpopular with four of the five voter subgroups Axios is tracking in the run-up to the midterm elections. Rural voters are the only group that believes the tariffs are good for jobs.
  • Those who think tariffs are bad for jobs think the trade fight with China is getting worse (69%), while those who think the tariffs are good for jobs think the dispute is staying about the same (59%) or getting better (26%).

What to watch: The 18% of Americans who approve of Trump but think the trade dispute is getting worse. (There's almost no movement in the other direction — only 1% of people who disapprove of Trump think the dispute is getting better.)

The bottom line: Trump's overall approval ratings may get better if Americans decide the new United States-Mexico-Canada Agreement is a good deal. But the China tariffs are unpopular enough that those who disapprove aren't likely to forget about them.

Methodology: This analysis is based on SurveyMonkey online surveys conducted Sept. 28-30 among 2,291 adults in the United States. The modeled error estimate  for the full sample is plus or minus 3 percentage points. Sample sizes and modeled error estimates for the subgroups are as follows:

African-American Women (n=147 +/- 9), Millennials Age 18 - 34  (n=395, +/- 6), White Suburban Women  (n=459 , +/- 6.5), NeverHillary/Independent voters  (n= 133, +/- 11), Rural  (n= 585, +/- 5.5). Respondents for this survey were selected from the more than 2 million people who take surveys on the SurveyMonkey platform each day. More information about our methodologyhere. Crosstabs available here.