Monday, July 20, 2015

Suggested by the post that was suggested by Ian Tarr

At the FRED Blog, Wage stickiness. I was interested because I found the link just at the same time I was writing my recent Sticky Notes post. (Ordinarily, "sticky" doesn't do much for me.)

In the FRED Blog post (which was "suggested by Ian Tarr" -- hence the title above) we find this:

The graph above shows two time series from the Bureau of Labor Statistics: unemployment (red line) and private industry wages and salaries (green line) from the employment cost index. Note that even when unemployment rapidly doubled, the green wages line continued to rise (albeit at a reduced rate). In other words, as the economy contracted and employers sought to cut costs, they almost exclusively opted to lay off workers rather than negotiate for lower wages.

As the economy contracted, they say, employers opted to lay off workers rather than negotiate for lower wages.

Well, maybe. But tweak that green line to show the employment cost index per worker and it tells a more interesting story:

Graph #2: Unemployment (red) and Employment Cost per Worker (green)
Cost-per-worker increases rapidly before and during the increase of unemployment. As soon as the "cost per worker" increase slows down, unemployment begins to fall. In other words, the dollars that employers plan to spend on employment buy more employment when the price of employment is rising slowly, than when that price is rising quickly.

Note that increasing unemployment (red) precedes the onset of recession. Also, accelerating per-worker employment costs precede increasing unemployment. I wonder if this relation holds for a longer time period...

Graph #3:Unemployment and Cost-per-Worker, Longer Term
On second thought, acceleration of per-worker employment costs and increasing unemployment seem to begin concurrently.

Data for the green line only go back as far as 2001, so it is hard to get a good feel for what that data says. But the green line increases more rapidly while unemployment is rising, and more slowly when unemployment is falling. When employment costs increase most slowly, unemployment falls most rapidly.

Not really unexpected, but it is nice to see the expected on a graph.


The FRED post anticipates a possible objection to their analysis:
Of course, it’s possible that inflation is cutting real wages even if nominal wages aren’t changing. However, when we adjust the wages data for inflation in the graph below (blue line), the pattern remains similar. Although real wages posted a slight decline several years after the recession hit, it pales in comparison to six years of elevated unemployment.

Okay. But what happens when we adjust employment cost per worker for inflation? The pattern becomes similar to that of unemployment:

Graph #5: Inflation-Adjusted Employment-Cost-Index per Capita (green) and Unemployment


So what is the meaning of all this?

No comments: