A muscular retort from Larry M to Matt Y

July 29th, 2015 at 3:34 pm

One of the more important pictures of inequality is the one you see here showing the growing gap between productivity growth and some version of the pay or income of the median worker or family. Larry Mishel and I noted the divergence beginning the early 1990s, recognizing it as a symptom of growing inequality.

Over at Vox yesterday, Matt Yglesias complained that the Hillary Clinton campaign’s use of the figure was “misleading.” As Larry points out today, in what I (and I’m admittedly in his camp on this) found to be an awfully muscular set of arguments, it’s not misleading at all if you know what’s actually represented in the figure.

Read Larry’s post, but I’ll just amp up two points he makes, and, not to pile on (Matt Y’s economic takes usually make a lot of sense to me), give Matt some credit too.

Most importantly, as Jordan Weissman gets here, the figure does not use overall average compensation for the whole workforce. That would, as Larry explains, basically assume away the part of the gap induced by the average diverging from the median, a known symptom of growing inequality. [If you want to see this for yourself, go to Google images and type in “mean median lognormal distribution” and then do the same for “mean median normal distribution;” you’ll see that the mean is “to the right” of the median in the lognormal (more unequal) case, yet coincident with the median in the normal (equal distribution) case.]

Larry’s using the wage (to which he adds benefit costs) of blue-collar manufacturing workers and non-managers in services, a BLS series that a) represents about 80% of the workforce, b) goes back far enough that you can use it for this, and c) moves roughly like the wage of the typical, or median worker.

Second, Matt convinces himself that the difference in the price deflators–the one for productivity versus the one for compensation–not inequality, is the big story in the productivity–wage gap. Larry’s data shows this not to be the case. The deflator difference–the fact that the product deflator grows more slowly than the consumer one–explains about a third of the gap since 1973. That’s not nothing; Matt’s not wrong to raise it, but more recently, since 2000, it explains only 15% of the gap. I take Larry’s point that the more recent result is more policy relevant, especially in a campaign context.

Finally, Matt has an important, if gnarly, insight that is ignored by some who play in this analytic sandbox. As noted, the prices of many intermediate goods–inputs into the process of creating the goods and services that comprise GDP–have been falling faster than many consumer goods. So, while computers and machine tools and steel bar have been getting relatively cheaper, the prices of child care and college have been growing faster than average.

So, if you want to make workers look better off, deflate their wages with a product deflator.

However, as Matt emphasizes, that would be nonsense. We must deflate wages with the stuff in the consumers’ market basket, not the producers’. As I’ve often said, to the guffaws of my nerdy friends, “people can’t eat machine tools.”

Anyway, good for HRC for highlighting this important figure and for the muscular Larry M, as shown below (h/t: AM), for his strong-man argument in defense of it.



Print Friendly, PDF & Email

3 comments in reply to "A muscular retort from Larry M to Matt Y"

  1. spencer says:

    I always think it is important to point out that about two-thirds of the difference between the CPI
    and the PCE deflator is the point that the PCE deflator does not contain shelter — that is accounted for elsewhere in the GDP data.

    If you use the PCE deflator you are in effect measuring the cost of living without shelter– one of the largest and fastest growing components of the consumer basket.

    People, especially right wing people tend to claim that the entire difference stems from the substitution effect and that is just plain incorrect.

  2. Smith says:

    Michel doesn’t just claim Yglesias is wrong about price-deflators because other factors account for larger portions of the wedge, especially recently. In a 2012 paper, he makes a case that the very notion of dismissing the effects of price-deflator differentials is wrong. He says even that portion of the wage productivity gap accounted for by the lower inflation rate for the price of airplanes and machine tools, vs. food and shelter, doesn’t alleviate, ameliorate, or explain away the problem wage productivity divergence. In his words:

    “Rather than dismiss or set aside this terms-of-trade factor that accounts for about a third of the growth of the productivity-median compensation gap, it deserves serious inquiry and a full explanation. Unfortunately, little research has been done in this area. Saying that changing terms of trade drives the productivity-pay gap is really more of a description than an explanation. Regardless of the cause, the implication is that the “typical” worker is not benefiting fully from productivity growth.”