Hacker News Reader

Comment by eru | Hacker News Reader

And that infamous graph is silly and bogus.

The graph uses different counts of inflation adjustment for the two lines. It introduces more unnecessary complications. That's how you lie with statistics.

Instead: just look at the labour share of GDP; that is the sum of all nominal wages divided by nominal GDP. Available for the US at eg https://fred.stlouisfed.org/series/LABSHPUSA156NRUG

Because we are looking at a ratio of nominal values, we don't need to adjust for inflation: the influence of the price level naturally cancels.

For the US the labour share of GDP has stayed between 60 - 65% in the last 70 years.

For that ratio to stay so constant, productivity growth and wage increases must have approximately kept pace with each other.

View on HN

The core of the issue seems to be that the graph I posted represents the experience of the median rather than the average worker. The data you suggest as more representative is simply the result of placing CEOs together with factory floor workers. I would argue that for the general perception of the population, the median pay is far more significant than the average one. You can read more here if you are interested: https://www.epi.org/blog/nothing-misleading-about-this-typic...

Yes, other people have also noticed that the graph is rubbish, and EPI is trying to defend it. Badly.

I don't mind using median instead of averages, if it's clearly labelled. The version of the graph you linked first was almost not labelled at all. However, even in their apology / defense they are still not fixing the mixed-up deflators. They merely claim that we should ignore this problem.

However following links in the meek defense leads to https://www.epi.org/productivity-pay-gap/ which has a bit more information than what you first linked. (I might go through it later. But on a first skimming, it's seems to be still pretty light on actual details about the numbers and where they came from.)

> If the fruits of economic growth are not going to workers, where are they going?

> [...] And it went into higher profits (i.e., toward returns to shareholders and other wealth owners).

That's wrong. The capital share of GDP hasn't really budged. (And in this case, distribution doesn't matter for our discussion. Since we only care that it's not going to workers, not which shareholders are getting what.)

However, the share of GDP going to land rent is disturbing. But EPI doesn't want to talk about NIMBYs, I guess.

"The graph uses different counts of inflation adjustment for the two lines. It introduces more unnecessary complications. That's how you lie with statistics."

What are the two different adjustments being uses? You're making some bold hand-wavey claims without showing your work.

"Because we are looking at a ratio of nominal values, we don't need to adjust for inflation: the influence of the price level naturally cancels."

Again you're missing distribution! There's a reason they use a standard basket of goods. You're also completely ignoring imports and exports.

"Because we are looking at a ratio of nominal values, we don't need to adjust for inflation: the influence of the price level naturally cancels."

"For that ratio to stay so constant, productivity growth and wage increases must have approximately kept pace with each other."

Only on the aggregate. You can very easily find that executive and other high earner pay has consistently increased at a higher rate than the median workers. So if that ratio remained the same and the increases mostly went to the upper class, then there is less for the middle and lower class.

> What are the two different adjustments being uses? You're making some bold hand-wavey claims without showing your work.

See https://www.epi.org/blog/nothing-misleading-about-this-typic... for an admission by EPI themselves that they are mixing up deflators.

Oh, this is about the pay vs productivity graph? I wasn't even really looking at that one. I was most interested in figures 3 and 4 showing how the increase in wages were going mostly to high earners, thus the median and lower wage earners haven't seen much in real gains. If the gains were more evenly distributed, then this wouldn't be an issue at all.

That pay vs productivity graph is the only one on the link https://www.weforum.org/agenda/2020/11/productivity-workforc... in the comment I replied to originally.

Ok, well you pointed my comment to your response here.

Reply on news.ycombinator.com