If President Trump is impeached without being removed from office, it’s hard to imagine people taking up arms in protest. (Evan Vucci/AP)

Speaking with Reuters reporters on Tuesday, President Trump offered a prediction. Or, perhaps, a warning.

"It’s hard to impeach somebody who hasn’t done anything wrong and who’s created the greatest economy in the history of our country,” Trump said when asked about the prospect of being removed from office.

"I’m not concerned, no,” he added. “I think that the people would revolt if that happened."

It’s hard to tell whether this is meant as hyperbole, but it’s worth assessing as a claim. Were Trump removed from office, what would the response be?

There are probably three ways in which this could be interpreted: as a political reaction against Democrats; as a series of coordinated protests, as seen recently in France; or, most dramatically, as an actual armed insurrection.

Let's consider those in reverse order.

There doesn’t seem to be any significant surge in interest in a revolt in the United States right now — unlike shortly after the election in 2016 when “revolt” surged as a search term on Google.

(Philip Bump/The Washington Post)

But, then, how good a metric is that? No revolt ensued, merely a #resistance.

The subject has come up since. Conservative columnist Kurt Schlichter earlier this year gamed out a full-scale national civil conflict, down to the level of anticipating the guerrilla tactics that would need to be used against military forces deployed to quell an insurrection. Schlichter is prone to hyperbole, but the article was still surprising: He’d clearly given this idea some significant amount of thought.

He’s probably not alone. It is a centerpiece of gun-rights advocacy that there might come a time that Americans would need to take up arms against a tyrannical government. This is a theme that’s central to the American identity, however distant the prospect might seem. Trump has even seemingly echoed this theme, saying during a campaign event in 2016 that perhaps “Second Amendment people” could keep a President Hillary Clinton from appointing pro-gun-control judges.

About 30 percent of Americans own guns, according to Pew Research Center, most of them handguns. Republicans are more than twice as likely to own guns as Democrats; 57 percent of Republicans live in a household with a gun. This overlaps with the density of Republican support in rural areas, where gun ownership is much more common.

In March, Monmouth University asked gun owners why they owned a firearm, including whether the need to defend against tyranny was a major reason to do so. More than half of Republican gun-owners said it was at least a minor reason for them to own a gun. Only about a quarter of Democrats agreed.

That said, it was also the option offered in the poll that generated the lowest amount of support. Most Americans who own guns aren't primarily concerned about fending off a modern King George.

(Philip Bump/The Washington Post)

There is also, of course, a very big difference between saying you would be willing to use your firearm to defend against tyranny and using your firearm to attack the police and military. We’ve seen protests in recent years involving armed conservatives facing off against law enforcement; think of the Bundy Ranch situation. That, happily, didn’t escalate into gunfire.

The idea of protests in the streets seems much more likely. We've seen a number of mass protest movements in the past several years that have altered the course of a country's politics, including in the Middle East and in Ukraine. The protests in France have knocked that country's leadership back on its heels. Protest can be effective — and the bar for participation is also much lower.

Recent polling about protests has largely focused on particular ones, meaning that views of the validity of the demonstrations has been colored by views of the politics of what’s being protested.

Here, the distribution of Trump supporters nationally would seem to be disadvantageous. The 100 counties where Trump had the largest margins in 2016 were, on average, 86 percent rural. The 100 counties where Clinton had the widest margins were 30 percent rural. About a quarter of counties Trump won were 100 percent rural, compared with 10 percent of counties Clinton won.

The protests that followed Trump's election and inauguration were facilitated in part by the density of Democrats in urban areas. That much of Trump's support is in more rural areas would seem to make the prospect of ongoing mass protests a bit less likely.

But now we come to the central question: How many Trump supporters would be willing to take to the streets — or to the ballot box — to express their displeasure? More than a few, certainly, but we have to consider the full context of the situation.

If Trump is impeached without being removed from office — certainly a possibility, as it was for Bill Clinton in 1999 — it’s hard to imagine people taking up arms in protest. When Clinton was impeached, it was seen in part as a de-escalation of tension more than an escalation.

If Trump is removed from office, things are different. But among the things that are different is the underlying politics of the moment.

Trump could be removed from office with a vote of two-thirds of the Senate: 67 votes. As of this writing, Democrats control 49 votes, a figure that will drop to 47 in January. In other words, Democrats would need the votes of 20 Republican senators to oust Trump.

How could that happen? Only if those Republican senators believed that the politics of the moment would reward their action.

In other words, if more than a third of the Republican caucus were to vote to remove a Republican president, it would mean that there had been a significant enough shift in support for Trump that those senators weren’t worried about the political repercussions of their actions. We’ve seen over the past two years that Republicans have generally stood fast by Trump, a reflection of the president’s broad popularity with his base. In the 2018 primaries, we saw that Trump’s blessing helped candidates win Republican nominations. Unless those things change, 20 Republican senators are unlikely to suddenly turn on the president, regardless of what he does.

It’s worth remembering that Richard Nixon’s approval among Republicans had dropped to 50 percent by the time he resigned — still much higher than Americans overall, but nowhere near what it had been the prior year and nowhere near what Trump enjoys now.

For Trump to be removed from office, there will almost necessarily have to have been a collapse in his support in his own base. That changes the dynamic of any possible reaction significantly. But it suggests that a broad-based popular revolt would not be a likely response. How many people would be willing to grab their guns and battle the National Guard if their own senators won't fight over a removal vote?

We will also note that Trump’s assertion that he “has done nothing wrong” is itself something to be evaluated for accuracy, but that’s a subject for another article.