The following two things are somehow both true: First, the Trump administration’s response to the coronavirus crisis has been a dangerous mix of denial, disinformation, and delay, oscillating between simple incompetence and possible malevolence. And, second, his approval ratings are at or near their all-time high. Americans are getting sick, losing their jobs, and even dying, thanks in part to Trump’s early blunders–and yet his popularity has been on the rise. How can this be?

One group of people not puzzled by the seeming paradox: political scientists. It turns out that an approval bump due to major crises or wars is one of the most consistent patterns in American politics. “It would have been amazing if Trump didn’t get any rally,” said Matthew Baum, the author of several academic books about public opinion. The surprising thing, rather, is how small the bump has been. As of March 13, according to FiveThirtyEight’s tracker, Trump was at around 42 percent approval and 53 percent disapproval, precisely where he has hovered throughout most of his first term. Since then, those numbers have shifted to about 46 percent and 50 percent–a 6-point overall swing. “You would expect, given how unpopular Donald Trump was the day before this all happened, he would have a massive rally,” Baum said.

Read all of our coronavirus coverage here.

Franklin Roosevelt, for example, got a 12-point boost after Pearl Harbor in Gallup’s tracking poll; George W. Bush received the all-time polling bounce immediately after 9/11, nearly 40 percent. Since the 1970s, political scientists have referred to the phenomenon as the “rally-around-the-flag effect.” But the name, which evokes some kind of groundswell of patriotic fellow-feeling, is a little misleading. The mechanism by which disaster leads to higher approval is subtler. It’s less about how voters react to the crisis itself, experts say, than about the signals they get from political leaders and the media. When disaster hits, or war is launched, the opposition party becomes a bit less oppositional, dialing back its criticism of the president and giving mainstream journalists fewer negative quotes to pass along to readers. That sends a powerful message.

“Most partisan signals in Washington are not very credible to anyone other than the partisans of the two parties,” said Tim Groeling, a UCLA professor who coauthored a book on the topic with Baum. Republicans aren’t going to change their mind about Trump because Nancy Pelosi criticizes him, and vice versa. “But, when you have a ‘rally’ event, something that causes those partisan patterns to break down, you can get significant effects. When a Democrat says something nice about Donald Trump–three months after an impeachment based on partisan lines–that is very credible information to voters.”

The Trump approval bump shows these dynamics in action. It hasn’t been driven by Republicans; they already supported the president so uniformly that they were nearly maxed out. Rather, the change is coming from Democrats and independents, who started from a lower baseline and who are more likely to take their cues from the media and Democratic Party leaders.

“It’s not that people aren’t criticizing Trump,” said Adam Berinsky, a political scientist at MIT who studies public opinion. “It’s that criticism, to the extent that it exists, is being minimized relative to a typical week of the Trump presidency.” In Congress, Democratic leaders actually worked with the Trump administration and the GOP to pass a historic, $2 trillion economic recovery package. Joe Biden, the presumptive Democratic presidential nominee, has shied away from a full-throated attack on Trump’s handling of the crisis. And at the state level, Democratic governors like Andrew Cuomo, Jay Inslee, and Gavin Newsom have offered at least tepid praise for the federal response. Meanwhile, TV networks have been broadcasting Trump’s daily briefings live, giving the president an opportunity to appear serious and pushing the media to cover what he and his team are saying, rather than what the government has actually accomplished. “These blips in Trump’s approval have been independents and Democrats giving him another chance because of those credible messages,” said Groeling.

Trump’s rally effect is smaller than what past presidents might have seen given the severity of the crisis. It’s also far smaller than the approval increases of other world leaders, along with American governors. One reason is that American politics at the national level are more polarized than at any time since Reconstruction. As a result, there isn’t much wiggle room left–most Americans already have strong feelings about Trump that are hard to sway by external factors. Not many voters are still open to persuasion. “It’s a surprisingly small rally, under the circumstances, and I think that is entirely a testament to just how intensely polarized people are,” said Baum.

Still, a record-high approval rating during a reelection campaign is far from trivial. The big question is whether it will matter come November. On that point, political scientists are skeptical.

“History tells us that rallies don’t last,” said Baum. A long line of presidents enjoyed rising approval during times of crisis, only to sink into unpopularity after a few weeks or months. Even Jimmy Carter enjoyed a 13-point approval bump at the beginning of the Iran hostage crisis; three months later his rating was back in the gutter as he drifted to a landslide defeat in 1980.