First, you are missing the point. It’s not ‘total costs of labor for corporations.’ It is income received by employees — as opposed to wages or salary. More and more of the income employees have been receiving has been in the form of benefits. Only because of an accounting anomaly is this income not counted.
Because wages and prices were controlled during WWII, FDR’s administration allowed benefits not to be taxed as income — to give unions something they could bargain for and thus justify their dues. This ‘emergency’ measure logically should have been rescinded after the war, but was not. Nevertheless, money paid by employers to subsidize benefits they receive really is income that otherwise would be paid as wages or salary.
You are right though that the growing contingent workforce of contractors and freelancers is often not getting extra benefits. (A smart contractor has to charge enough to pay for her own benefits.)
Second, regarding the mystery of lagging productivity growth, a useful article in Quartz summarized three main explanations:
- We may not be measuring productivity correctly.
- Businesses haven’t been investing in becoming more productive.
- The productivity slowdown is really an innovation slowdown.
You can take your pick.
Personally, I’m skeptical about #3. It seems to overlook, among other things, the explosive impact AI and advanced robotics are starting to have on the whole economy.
I think the mis-measurement problem is a bigger factor than some believe. Think about Linux — that OS produces immensely valuable services. Users pay little or nothing for them. But the work of developing and maintaining the software is contributed by a legion of unpaid volunteers. Traditional labor productivity measures only count work performed by employees. The Linux Foundation has estimated the value of the OS as $5 billion, but having no employees, that would imply infinite productivity. (Actually, the estimate is based roughly on what you would have had to pay programmers to write all that code, considering the total number of hours needed and average programmer compensation.)
Another manifestation is Baumol’s ‘cost disease.’ Baumol presumed that as the economy shifted to segments of services — from performing arts to education to healthcare — where productivity is immune to improvement, or maybe even measurement, costs would rise. But those sectors are not as immune to innovation and efficiency as Baumol and his colleagues believed.
For instance: “Baumol and Bowen pointed out that the same number of musicians is needed to play a Beethoven string quartet today as was needed in the 19th century; the productivity of classical music performance has not increased.” Not so. Recording and then radio enabled the same number of musicians in the 20th century and beyond to entertain a vastly greater number of people than they had when people had to go to the musicians’ venue to hear a performance. MOOCs are not as transformative as some think, but they do enable a professor to teach his course to tens of thousands of students instead of just dozens.
As for #2, I think that the migration of both money and talent to ‘financial engineering’ over the last couple of decades has put too much resources into intangible, often illusory ‘investments.’ The digital profits from ever more exotic financial derivatives are decoupled from any real economic ‘product.’ And too, very low interest rates enable companies to borrow money cheaply to boost stock value by buying back shares — again without producing anything more than numbers.