Labor Share follow-up

Recently, I got a copy of Alberto Cairo's The Truthful Art, subtitled Data, charts, and maps for communication. According to the first blurb on the first page -- the first thing you come to --

The Truthful Art is both a manifesto and a manual for how to use data to accurately, clearly, engagingly, imaginatively, beautifully, and reliably inform the public.

This isn't a book report. I'm gonna skip the first 120 pages. Cairo tells a great story about Richard Thaler. Thaler is the guy responsible for "behavioral" economics. He invented the "nudge". Because of him, when you're filling in a form, the answers they want you to pick are preselected and set as defaults. Because of him, when you're looking for something to watch on Netflix, if you pause long enough to read the episode blurb, the damn thing starts to play. Thanks, Dick.

Alberto Cairo:

In his book Misbehaving: The Making of Behavioral Economics (2015), University of Chicago's Richard H. Thaler recounts an anecdote that may be useful for any teacher. At the beginning of his career as a professor, Thaler made many of his students mad by designing a midterm exam that was deemed too hard. The average score, on a scale from 0 to 100, was 72. He got a lot of complaints about it.

Thaler decided to run an experiment. In the next exam, he set the maximum score to 137 points. The average ended up being 96 points. His students were thrilled.

Thaler kept the 137 mark in subsequent exams and also added this line to his syllabus: "Exams will have a total of 137 points rather than the usual 100. This scoring system has no effect on the grade you get in the course, but it seems to make you happier." It certainly did. After he made this change, Thaler never got any pushback from students again -- even if he told them beforehand that they were going to be tricked!

"It's hard to picture 96 in comparison to 137 in your head", Cairo says. But "72 versus 100 ... is easy." I suppose his point is that to use data effectively, you need to make it easy for people to understand. Good tip.

Alberto Cairo concludes that section with these words:

It turns out that Thaler's second exam was harder than the first one. A score of 96 out of a maximum of 137 is a 70 percent score, in comparison to the 72 percent average of the first exam. But even if you're aware of that -- because you know how to transform a raw score into a percentage -- 96 over 137 still feels higher than 72 over 100...

(Emphasis in the original.)

In a book titled The Truthful Art, Alberto Cairo points out that 96/137 "feels" like it is more than 72/100, even though it isn't. He wants me to be truthful, but he's showing me a scam that worked. I don't know what to do with that.

My impulse would be to say no, 96/137 is not more than 72/100, and how you "feel" about your 96 is mathematically incorrect. You and I might get stuck in that disagreement until we quit talking for good, and that would be the end of that.

Okay. Suppose I'm trying to lay out my ideas on the economy. If you can't see my first point, how can I expect you to follow my logic to the end to see if it makes sense? The first point is the simplest one. If we can't agree on the first point, it's over.

You're not helping me, Cairo.


When you look at Labor Share, you're finding an answer to questions like "How much did our side get?" and "How much did the other side get?" The other side, apparently, is measured by something called Capital Share. (I read that somewhere.)

So I dunno, does it make sense to say we got 100 in 2012? Doesn't that mean the other side got zero? And how about back in 1947 at the start of those graphs, where Labor Share was around 110% or 115%? Wow! We were bleeding them dry in those days!...

No. Obviously not. The 100 on the Labor Share graphs is not 100 percent. It's not percent. I know we're talking about "shares" here. And I know that "shares" is percent. If I get half of something and you get the other half, then we each got 50%. I know how it works. But for Labor Share, no. It's not percent. It's an index.

Hey, I didn't make it up.

At BLS they offer the Monthly Labor Review. And in the February 2017 issue they look at Estimating the U.S. labor share. In that article there is a footnote that says

See “LPC databases,” Labor productivity and costs (U.S. Bureau of Labor Statistics), https://www.bls.gov/lpc/data.htm. Click the link that says, “Download the complete Major Sector Productivity and Costs dataset.” A zip file that contains the annual and quarterly datasets will open. This is the sole location on the BLS website where data on labor share levels can be found...

This is the sole location on the BLS website where data on labor share levels can be found. Percentages.

If you go there and get that, you get a large, well-organized file of quarterly data (and another, with annual data). Here's quarterly Labor Share for NonFarm Business:

Graph #1

And now (because these numbers show percent) I can subtract the values from 100, and get the other guys' share:

Graph #2

I'm not real comfortable calling it "capital share". I've seen it called that, but only a couple times, and nothing technical. But here it is.