In his September 4 guest essay in the New York Times, “Good News: There’s a Labor Shortage,” MIT economist David Autor presents a mix of good reasoning, error, and confusion.

Autor argues that a labor shortage (he really means “increased labor scarcity”) is driving up real wages at the bottom end of the pay scale and will likely drive them up further. His reasoning is good assuming all of his premises are right. But I doubt one of his premises: that paying millions of unemployed people more to be unemployed than to work does not cause a lot of people to stay unemployed. Co-blogger Scott Sumner dealt recently with what’s wrong with that claim.

Other parts of his reasoning are good. He writes:

For years, social scientists have warned that because of declining birthrates, retiring baby boomers and severe immigration restrictions, we’re approaching an era of labor scarcity.

But there are also two other claims, the first of which is, at best, confused and, at worst, wrong and the second of which is simply wrong.

Here’s the first:

There’s no future in working the fry station at White Castle.

It’s true that those who plan to work the fry station at White Castle shouldn’t plan on doing well in the labor market if they want that job long term. But surely Autor knows that the vast majority of people who take those jobs while young are not in those jobs 10 or even 5 years later. Those jobs are a stepping stone to better jobs and can teach young people some basic labor market skills: being punctual, taking responsibility, taking direction, and organizing their time, to name four important ones. That means that there actually is a lot of future in working the fry station at White Castle.

Imagine that Autor had said:

There’s no future in going to middle school, grades 6 to 8.

Everyone would see the problem with that reasoning. You won’t do well if you plan to be, and succeed in being, in middle school for the next 10 years. But everyone understands that middle school is a step on the way to better things. We can disagree about how much you learn in middle school, with co-blogger Bryan Caplan being on one end of that debate and me being almost on that end. But the point is that it’s not the final step.

Autor also writes:

Let’s say that you like cheap burgers and don’t lose sleep over the low wages burger flippers earn. Would it bother you to know that their well-being is subsidized by taxpayers? Members of the bottom fifth of U.S. households receive an average of $9,500 per person per year in means-tested government benefits like Medicare, Medicaid and food stamps. More than 20 percent of income for the poorest fifth of households comes from refundable tax credits like the earned-income tax credit, which supports low-wage families with children. So, one reason companies can pay such low wages is that you’re paying for the things their low-wage workers can’t afford.

The bottom line, he claims, is that “That burger isn’t as cheap as you think; it’s just that you’re paying part of your meal tab in federal taxes.”

Ignore the irony that the push to put the cutoff for means-tested benefits higher and higher has come mainly from the left. If Autor is arguing for tightening eligibility, then I’m with him, although I’m not sure he is.

Here’s the problem with his reasoning: he gets the shift in the supply curve wrong. Means-tested benefits tend to make workers less willing to supply hours of work at a given wage. So the programs he cites (other than the earned income tax credit, which is more complicated) shift the supply curve to the left, not the right, making wages higher than otherwise and making that burger more expensive than otherwise.

Seven years ago, Bryan Caplan dealt well with what’s wrong with an Autor-style argument about means-tested benefits leading to increased labor supply.