Scott Lincicome reviews public-opinion data to show that there has been no upsurge in protectionist sentiment. “Trump’s protectionism,” he writes, “drives (and is not itself driven by) the opinions of a significant portion of the electorate — an electorate that, when confronted with the actual implications of Trump’s policies (i.e., higher prices, harmed businesses, or foreign retaliation), moves toward the freer trade position” (emphasis in original).
As useful a corrective as his study is, I’d add two caveats. The first is that certain arguments for trade barriers — e.g., we have to erect them to fight unfair trade practices abroad — have probably always made sense to most people, even if there has been no uptick in this sentiment.
The second is that a policy need not be popular to be politically useful. Let’s say that a significant number of voters who have little attachment to the Republican party listen to protectionist rhetoric from a Republican politician and vote for him on the ground that he, at least, is trying to do something that will benefit them and their communities and to be tough on countries that merit it; and let’s say, also, that the group of voters who take the opposing view — who consider protectionism a reason to defect from the Republicans they would otherwise support — is smaller. I’m not saying that these things are true, just that they could be. You would not detect this effect by looking at the broad trend of public opinion. (You might not even find it by asking people if they consciously based their votes on trade policy.)
To get a little less abstract: Protectionist policies are among the very few things Republicans have offered to provide specific help to distressed Rust Belt communities, and even if those policies are not nearly as helpful to them as advertised, and even harmful to the national interest, they may help get Republican votes compared to making no offer at all.