海角大神

Here are some numbers that are good: 4.2, and 3.9. The first is the growth in US Gross Domestic Product for the second quarter of 2018, at an annualized rate. The second is the unemployment rate for August. Both are evidence of a really, really strong economy 鈥� though they鈥檙e not as historically unique as President Trump sometimes boasts.

Here are some numbers that are bad: 39.8, and 82. The former is the percentage of Americans who approve of the job Mr. Trump is doing, according to an average of major polls at time of writing. That鈥檚 ... low. The latter is the percentage chance that Democrats will win control of the House of Representatives in November鈥檚 midterm elections. That鈥檚 high, uncomfortably so if you鈥檙e an elected Republican lawmaker. (Both of these are from the data journalism site .)

So here鈥檚 the question: Why isn鈥檛 the demonstrably excellent economy making Trump more popular? Does this mean the聽famous adage about American politics, 鈥淚t鈥檚 the economy, stupid,鈥� is no longer true?

Well, we can鈥檛 really tell how the economy is affecting Trump鈥檚 numbers, for one thing. It could be keeping them higher than they otherwise would be. And the data show 鈥渋t鈥檚 the economy鈥� remains a reliable guide to the political fortunes of presidents in the modern era.

Except when it isn鈥檛.

Let鈥檚 step back a bit and try to put Trump鈥檚 numbers in a historical context. We鈥檒l use the unemployment rate, since it鈥檚 a simple number, it鈥檚 good for Trump, and it correlates pretty well with presidential approval ratings over time. (For the most part.)

The current rate is 3.9 percent, as we noted above. When it hit that level for other presidents, what was their approval rating? Philip Klinkner, a political scientist at Hamilton College, some numbers on this recently: Dwight Eisenhower had a 68 percent approval rate when the economy was humming along with only 3.9 percent jobless in September of 1956. Richard Nixon had a 63 percent approval rating when it reached that level in January 1970. Bill Clinton鈥檚 approval was 58 percent positive when unemployment was 3.9 percent in October 2000.

Those ratings were all a lot better than Trump鈥檚 current number.

But then there鈥檚 LBJ. In November of 1967, US unemployment was 3.7 percent, and Lyndon Johnson鈥檚 approval was 42 percent. That鈥檚 higher than Trump鈥檚 current number, but barely.

What was LBJ鈥檚 problem? External shocks. The deadly slog of the war in Vietnam dominated front pages. Riots were burning out the core of major US cities. Some white voters disapproved of the administration鈥檚 push for civil rights. At times, Washington seemed to be losing control of the country, or at least to be out of touch. Johnson鈥檚 job approval peaked at around 80 percent when he took office following John F. Kennedy鈥檚 assassination, but then basically began a years-long slide, bottoming out at 35 percent in August 1968.

海角大神/File

President Lyndon Baines Johnson in 1965. Presidential approval ratings in the modern era tend to track closely with the economy 鈥� except when they don't.

In Trump鈥檚 case, the external shock is arguably Trump himself.

Republicans approve of Trump鈥檚聽personality and actions in office, seeing him as a truth-teller who makes liberals and mainstream news reporters uncomfortable, and who is taking steps to do what the country has long needed. Many also believe the media aren鈥檛 covering positive news enough. In the latest CNN poll, 82 percent of GOP respondents approved of the job the president is doing.聽

But many independents and most Democrats view Trump as someone with little regard for the truth or the norms of democracy. The same CNN survey has his approval rating among self-described independents at 31 percent. Among Democrats, it鈥檚 five percent. Talk about a partisan split.

So, what are the implications for the 鈥淚t鈥檚 the economy, stupid鈥� truism? The phrase, of course, was made famous in Bill Clinton鈥檚 1992 presidential run, when force-of-nature aide James Carville hung a sign to that effect in the campaign war room.

Well, a few years ago, in the pre-Trump era, Pew Research took a thorough look at how economic numbers affect presidential approval, and concluded 鈥渢he public response to all [modern] presidents has been shaped to some degree by rising or falling unemployment.鈥�

But the correlation is closer when unemployment gets worse 鈥� and particularly when it gets worse precipitously. When it gets better, and changes more slowly, the relationship between the measures is murkier.

Political scientists still consider the economy one of the 鈥渇undamentals,鈥� those underlying forces that can affect the outcomes of American politics more profoundly than most actual campaigns. But Trump has clearly upended some of their calculations on this subject, as he has so many other things.