FiveThirtyEight

Our social media editor, Meena Ganesan, dutifully sends out a tweet each morning from the @FiveThirtyEight account detailing President Trump’s approval rating according to our constantly updating average of polls. The tweet isn’t automated, but it probably could be, since the number is always pretty much the same. For the past 66 days, Trump’s approval rating has been somewhere between 40.0 percent and 42.1 percent, according to our tracker. It’s been toward the higher end of that range recently — but that isn’t much of range.

Indeed, over the whole course of his presidency, the range Trump’s approval ratings travel in has been remarkably narrow. Upon his inauguration in 2017, Trump did not enjoy much of the “honeymoon effect” that new presidents typically do; instead, on Feb. 1, 2017, his approval rating was only 44.8 percent, according to our tracker. It’s never been higher than that since. But it’s also never been lower than 36.4 percent, a nadir that Trump hit in mid-December. While a few news events early in Trump’s term did seem to move his numbers — in particular, his decision to fire FBI Director James Comey and the various stops and starts in the Republican health care debate — those now seem more like the exceptions than the rule. Good or bad news for the president, almost nothing seems to change Americans’ views of Trump.

I was pretty skeptical last month, therefore, when, after the raid on former Trump attorney Michael Cohen’s office, The New Yorker’s Adam Davidson wrote an article claiming that we were “now in the end stages of the Trump Presidency.” In a critique of Davidson’s article on Twitter, I pointed out that a potential end to Trump’s presidency was heavily contingent on public opinion. A Republican-led Congress isn’t likely to impeach Trump and remove him from office so long as his approval rating among Republican voters is 85 to 90 percent. A Democratic-led Congress might try, especially if Democrats had comfortable majorities — but that would require a big Democratic wave in this year’s November midterms, which is hardly guaranteed. In short, Trump wouldn’t seem to be at that much political risk unless his popularity deteriorates even further. But since the hit he took after firing Comey, Trump’s approval ratings have mostly been unaffected by news stemming from the Russia investigation or other potential scandals, and in fact have increased slightly over the past few months.

But Davidson’s column was partly intended as a warning against what he sees as a complacent assumption that the center will hold for Trump, just because it (sort of) has so far. He compared Trump’s presidency to the Iraq War and the financial crisis, two matters on which the situation on the ground was much worse than the conventional wisdom generally acknowledged — until it was too late. Trump’s approval ratings could be a lagging indicator; “polls seem capable of entering a stable place for some time but not [staying] there forever,” he wrote in a follow-up tweet.

This is a complicated subject — but let’s ask a relatively simple, empirical question: If a president’s approval ratings are stable early in his term, does that tend to imply that they’ll also be stable later on? In other words, is it safe to assume that Trump’s approval ratings will continue to vary only within a narrow range from the high 30s through the low 40s, given that that’s what’s happened so far?

The short answer is no, that is not a safe assumption. There are several presidents whose approval ratings were steady early in their terms — not quite as steady as Trump’s, but steady — but then became volatile later on.

For each president, I’ve compiled Gallup approval ratings data (I’m using only Gallup data because it allows for apples-to-apples comparisons ) at three stages of his presidency: in the first 500 days of his first term (Trump will hit the 500-day mark on June 3), in the remainder of his first term, and (if applicable) in his second term. Then I looked at the 95th and 5th percentile of Gallup approval ratings that each president achieved during each period. The area between those two percentiles is the range that the president’s approval rating traveled within except for short-lived outliers. For instance, Trump’s 95th percentile in Gallup so far is a 43 percent approval rating, while his 5th percentile is a 35 percent approval rating — a difference of only 8 percentage points. Trump has occasionally been higher or lower than that but not on a sustained basis. (Note: When I use the term “range” throughout the remainder of this article, I’ll be referring to the difference between the president’s 5th and 95th percentile approval ratings, and not the absolute minimum and maximum.)

Early stability in approval ratings doesn’t always last

Low and high represent the 5th and 95th percentile, respectively. Johnson’s first term lasted fewer than 500 days, so his entire first term is covered within the “first 500 days” column and his entire second term is covered in the second term column. Trump’s numbers cover his first 473 days in office.

Source: GALLUP

Trump’s 8-point approval-rating range is the narrowest of any president to this point in his term. The previous record-holder was John F. Kennedy, whose range spanned between 72 percent and 82 percent, a 10-point spread. Lyndon B. Johnson, Richard Nixon and Dwight Eisenhower also had narrow approval-ratings ranges in their first 500 days; to a lesser extent, so did George H.W. Bush, Barack Obama and Bill Clinton.

But three of these presidents experienced significant volatility in their ratings later on in their terms — and in all cases, they wound up worse off for it:

By contrast, there are four presidents — Eisenhower, Kennedy, Obama and Clinton — whose numbers started out fairly steady and then mostly remained that way (or became even steadier) throughout the balance of their terms. And there are other presidents — Truman, Carter and arguably George W. Bush — whose approval ratings were volatile throughout their terms. Still, the correlations between one period and the next are pretty low ; a steady approval rating can sometimes become volatile or vice versa.

These charts show FiveThirtyEight’s approval rating algorithm as applied to past presidents (they include all polls rather than just Gallup polls):

Nonetheless, be careful if you think Trump’s demise is right around the corner. Even if there were a big shift in Trump’s approval ratings eventually, it could take a long time for it to come, perhaps only after he was successfully re-elected — as was the case for Nixon. Or, the surprise could be that Trump’s ratings shift way upward instead of way downward. (Since Trump’s approval ratings are low to begin with, he might have more room to go up than down.)

Also, it probably isn’t a coincidence that both Trump and Obama had narrow approval-rating ranges given how intense partisanship is these days. For Trump’s approval rating to get much higher than 50 percent, he would need significant crossover support from Democrats; for it to get much lower than 35 percent, he would need a significant number of Republican voters to turn on him. The theory that Trump’s popularity is bounded within a narrow range has a lot going for it, including that it has described the actual behavior of his approval rating well so far. Still, history tells us that presidential approval ratings behave in a relatively unpredictable fashion.

Filed under