Scott Sumner Goes Marxist, Proposes Targeting Labor’s Share of Income

I’m joking of course. He’s still grinding the supply-side axe (though judiciously here, IMO). But you gotta admire a fellow when he follows the logic of the data where his own logic requires him to go. He’s just done three posts about Germany’s growth and unemployment rates through the great recession:

Annualized change, Q1 2006 – Q4 2012:

RGDP: 1.3%
NGDP: 2.4%

But the unemployment rate fell from about 12% to 5.4%.

Lackluster GDP growth coupled with a damned impressive drop in unemployment. How do you account for that in Scott’s long-argued model, where NGDP growth drives employment growth?

As he says, he buries the lede in his first post. Here it is, from the end of the post (I’m reversing these two paras to make it flow even better; emphasis mine, correction his):

Recessions are not caused by less spending; they are caused by less income going to workers.  Usually the two go hand-in-hand, but the German miracle tells us that when they diverge, it is employer employee income that matters most.

When I started blogging I assumed wage targeting would be politically impossible, and knew that NGDP targeting was already a well-regarded concept.  So I latched on to NGDP targeting.  But in retrospect I wish I’d latched on to aggregate employee income.  Call it “income targeting.”

Did I mention admirable? Speaking of the very argument he’s been making doggedly and insistently for years, he says:

I took some shortcuts, instead of staying true to the “musical chairs model.”  In the past I’ve often argued that a fall in NGDP causes unemployment because there is less income to pay workers, and yet hourly wages are sticky.  Some workers end up sitting on the floor.  The logic of that model suggests that the real problem is not unstable NGDP, but rather instability in a component of NGDP, namely total wages and salaries.

And speaking of the school of economics based on that thinking, a school that he’s been primarily responsible for creating and popularizing, he says:

Now that market monetarism riding high, I figured it was time for a vicious internecine struggle for the soul of market monetarism.  Consider this the first shot. 

The fruit doesn’t fall far from the tree, of course; in his second post he attributes the “miracle” largely to “structural”/supply-side factors:

Germany did lots of labor market reforms, which I’ve discussed in previous posts, and this opened up lots of low wage jobs.

The basic idea: the Hartz reforms beginning in 2003 resulted in more jobs, though often with shorter hours and/or lower wages, all netting out to a larger share of GDP going to employee compensation.

This is not crazy. But it’s incomplete. And he even acknowledges that in a nod to his commenters:

Many commenters pointed to various job sharing programs, or subsidies to keep workers employed during the recession. Libetaer used the metaphor of putting 2 workers in one chair.  The one chair reflects relatively meager growth in NGDP, and the two workers represent job sharing.

The key problem I find in his thinking is his glossing over a key word in the preceding: subsidies. He only discusses job sharing, which fits neatly in his musical-chairs analogy. (How about musical benches instead? When the music “stops” — national income is weak — workers squeeze in more tightly.)

But the Hartz reforms did more than make it easier for firms to hire cheap (create more chairs/bench space); they increased subsidies for low-wage and part-time jobs. This 1) makes it easier for employers to hire lower-productivity workers for low wages, 2) gives those lower-productivity workers sufficient incentive to take those jobs, and 3) increases the share of national income going to lower-income workers.

And since those workers have a high propensity to spend their income, all things being equal that distributional shift should mean there’s a higher average velocity of money, aggregate demand, NGDP, etc., all in a virtuous cycle. (See JW Mason here and me here.)

Scott says much the same thing from a different direction:

The welfare loss to a society from a 5% RGDP shock is much greater if 5% of workers lose their jobs, as compared to all workers staying employed, but working 5% less hard.

This is a statement about absolute utility, but (hence, because spending is driven by the desire for utility) it’s also a statement about different policies’ distributional effects on spending and aggregate demand. Because subsistence has (very!) high utility, cutting some subsistence incomes (which would surely be re-spent) results in a bigger hit to aggregate demand than cutting many marginal incomes (which are less likely to get re-spent). It’s straightforward Marginal Propensity to Spend out of income thinking.

I’m here to suggest that the same logic, rather inevitably, applies to subsidies for low-wage jobs (paid for by better-off taxpayers). We redirect disposable income at the margin from higher-income folks, and flow it into the hands of lower-income folks who will spend it on. Got velocity?

Which brings me back to the axe that I’m forever grinding: the best and most feasible structural labor-policy change we could make in the U.S. to improve long-term macroeconomic performance would be to greatly expand the Earned Income Tax Credit (while streamlining its tortured administration and — to increase its “salience” — delivering the credit on weekly paychecks as we do with payroll tax deductions).

If these subsidies were sufficient to make low-wage work/workers attractive for both employers and workers (and perhaps if they subsidized hourly compensation rather than annual family income), we might even be able to do what the Germans, with their generous low-wage subsidies, have been able to do: do without minimum-wage laws.

I bet Scott would like that.

And we still haven’t talked about national compensation/income targeting by the Fed, which promises to be a very lively discussion indeed. (I can just see Ben Bernanke slapping his forehead and looking heavenward.)

Cross-posted at Asymptosis.

Comments (3) | |