≡ Menu

Autor, Boudreaux, and Henderson Converging

David Autor responded in a comment at EconLog to this recent letter of mine in which I was critical of a point that he made in his September 4th, 2021, New York Times essay on labor supply and wages. I posted that letter both here at Cafe Hayek and in the comments section at EconLog.

Prof. Autor’s response is fair and welcome. Some welfare payments – most notably, the earned income-tax credit – can actually increase the labor supply and, thus, lower the wages that employers of low-skilled workers pay to these workers. In my letter I ignored this possibility. What especially caught my eye in Prof. Autor’s New York Times piece was his mention of “Medicare, Medicaid and food stamps” – each of which decreases the supply of labor (at least in a non-literally-subsistence economy such as ours). While it’s true that Prof. Autor also mentioned refundable tax credits as among the means-tested sources of income available to poor Americans, he included it in a longer list of these other sources – other sources that, again, decrease the supply of labor.

Earlier in his NYT essay Prof. Autor expressed skepticism that very generous unemployment benefits reduce people’s willingness to work. This skepticism suggests a belief that the supply of low-skilled labor is rather unresponsive to changes in government welfare. After all, if the amount of labor that people supply remains largely unchanged when government payments for not working rise, the amount of labor that people supply should remain largely unchanged when government payments for working rise. (For the record, I believe that availability of such government payments has a substantial impact on the labor supply.)

When Prof. Autor wrote, in his NYT essay, that “one reason companies can pay such low wages is that you’re [the taxpayer is] paying for the things their low-wage workers can’t afford,” I assumed that he believes that the source of these artificially low wages is workers’ willingness to settle for lower wages because taxpayers are paying for much of what these workers consume. My assumption here makes sense given that Prof. Autor seemed to rule out welfare payments affecting wages by shifting the supply curve of labor. Put aside Prof. Autor’s expressed skepticism that the recent surge in unemployment benefits reduce the labor supply. Even with that, I would have been less likely to misconstrue him as believing that the labor supply is unresponsive to government welfare had he pointed out that, while programs such as the EITC might increase the labor supply, programs such as Medicare and food stamps decrease it. Yet again, as his NYT piece reads, there’s no indication that Prof. Autor attributes changes in workers’ wages to changes in the supply of labor brought about by government welfare payments. He mentioned only an impact on wages.

Nevertheless, the good news is that David Autor and I – and David Henderson, for that matter –  are converging. Prof. Autor agrees that increased government welfare paid to people for not working reduces the amount of labor supplied at each wage. That’s what I take from his statement that “other forms of non-contingent benefits may shift it [the supply curve] the other way.” I’m genuinely happy to be corrected in my understanding of Prof. Autor’s meaning. I apologize for this misunderstanding. My defense is that I’d have been less likely to slip into this misunderstanding had he, early in his NYT piece, not suggested that labor supply is largely unresponsive to government welfare payments, and, later in his piece, indicated his awareness that programs such as the EITC can increase the supply of labor and that other government programs, such as food stamps, can decrease it. Again, as his NYT piece reads, the most we learn about Prof. Autor’s belief about the effect of government welfare payment on the supply of labor is that that effect is negligible. It’s good to see that Prof. Autor thinks this effect could be substantial and that his reason for not pointing this effect out was space limitations.

Next post:

Previous post: