By Spencer (2009)
The issue of a jobless recovery is getting a lot of attention recently.
I’ve found the best way to look at the issue is to compare the change in real growth and productivity over the long run. There have been three periods of different productivity trends in modern US economic history.
Prior to about 1973 productivity growth averaged 2.8%. In the second or low productivity era, running from 1974 to 1995, productivity growth slowed to 1.5% before rebounding to 2.4% since 1995.
But real GDP growth also slowed over this period. As a consequence, the ratio of real GDP growth to productivity growth fell from 68% in the early strong productivity to 50% in the weak productivity era before rebounding to over 80% in the most recent era. Basically, real GDP growth equals productivity growth plus hours worked or employment growth. A consequence of stronger productivity in an era of weaker GDP growth this suggests that each percentage point increase in real GDP growth generates a much weaker increase in hours worked or employment. Currently, a percentage point increase in real GDP growth now generates under a 0.2 percentage point increase in hours worked versus 0.3 in the pre-1974 era and 0.5 percentage points in the low productivity era.
But to a certain extent comparing productivity and real GDP is comparing apples to oranges. To be accurate one should look at productivity versus output in the nonfarm sector. GDP includes the farm sector of course, but also the nonprofit and government sectors where productivity is assumed to be zero.
If you look at what happened in the 1990s and early 2000s recoveries in the nonfarm business sector, you see that productivity growth significantly outpaced output growth in the early recovery phase of the cycle. As a consequence hours worked or employment fell, generating the jobless recoveries. It looks like the problem in these two cycles was much weaker growth rather than strong productivity.
This shift to an environment of stronger productivity and weaker real growth generated an interesting development that has received little attention among economists or in the business press.
This development was a secular decline in labor’s share of the pie. Prior to the 1982 recession there was a strong cyclical pattern of labor’s but it was around a long term or secular flat trend. But since the early 1980s labor’s share of the pie has fallen sharply by about ten percentage points. Note that the chart is of labor compensation divided by nominal output indexed to 1992 = 100. That is because the data for each series is reported as an index number at 1992=100 rather than in dollar terms. So the scale is set to 1992 =100 rather than in percentage points. But it still shows that labor payments as a share of nonfarm business total ouput has declined sharply over the last 20 years and prior to the latest cycle we did not even see the normal late cycle uptick in labor’s share.
If this chart gets a lot of attention it will be interesting to see how the libertarian and/or conservative analysts who keep coming up with all types of excuses to explain away the weakness in real labor compensation in recent years explain this away. If you really want to raise a stink you could look at this as a great example of the Marxist immiseration of labor that Marx believed was one of the internal contradictions of capitalism that would eventually lead to its self destruction.
additional chart in response to comments.