The Washington Post is excited by a new poll of economists got up by the National Association for Business Economics. It shows, says the Post, "overwhelming support" for Hillary Clinton. "Overwhelming" might be a slight exaggeration on the Post's part—Clinton had 55 percent support, meaning that 45 percent of those polled ("nearly half," as the Post might say under other circumstances) don't support her. That 45 percent is evenly divided between Gary Johnson, Donald Trump, and the most neglected category in all of economics: "Don't Know."

I guess the Post thinks of this as a man-bites-dog story. Aren't all those business types supposed to be supporting the Republican presidential candidate? But at least three points need to be made about the poll and the Post's evident surprise at its results. And—to quote a famous business type—I'm the only one who can do it.

The first and biggest mistake in the poll is the most fundamental and comprehensive. It assumes the elementary myth of economics: that "the economy" is a giant machine—a hydraulic pump, maybe—with levers and buttons that policymakers, especially the U.S. president and the governors of the Federal Reserve, push and pull to orchestrate "inputs" and "outputs." When the pushing and pulling is done right, the machine hums happily.

Really, though, the economy-as-machine is just a metaphor (and a bad one) that has got out of hand. The machine has no independent, real-world existence. What we call "the economy" isn't a machine or even a system; it's the constantly churning accumulation of individual decisions made by several billion people in the course of a day, along with the many billions of consequences, large and small, foreseen and unforeseen, that those decisions entail.

Needless to say—or maybe not—this accumulation is very hard to monitor, much less manipulate. This is why economists usually fail when they try to predict the behavior, or even the direction, of "the economy," their beloved fiction. You could even say that the persistent belief in "the economy" as a giant machine responding to the touch of policymakers is the best evidence that macroeconomics is a load of hooey. Somebody needs to get these guys to read Ludwig von Mises. Or at least Friedrich Hayek.

Second, and less philosophically, academic economists are all Democrats. Okay, not all. As surveys have shown, the political beliefs of university professors generally range from left-liberal to Maoist. The same surveys show the highest concentration of non-Democrats to be in business and law schools and in economics departments. Sometimes these percentages of libertarians and crypto-Republicans rise to the very low double digits. But the vast majority of even business professors and economists are still Democrats. Their preference for Clinton isn't news.

And third, while most of them are Democrats, they are still human. As such, even assuming that "running the economy" were truly a job that a human being could perform, most publicly-aware people would rather the job fall to Hillary Clinton—a brazen liar, arrant misogynist, sleazy opportunist, and gross incompetent—than to Donald Trump, who self-evidently suffers from a serious personality disorder of the most unpredictable kind.

Like Hillary, he believes in the economy machine. Unlike Hillary, he relates all natural and manmade phenomena to himself as the center of the world. He proposes high tariffs, trade wars, massive public works projects, and expanding social insurance as the best way to get the machine roaring. If "the economy" didn't respond the way President Trump wanted it to, he'd take it the way he takes everything: very, very personally. What he'd do then, you wouldn't believe.