I've long wondered how productivity is measured, so I went and found the study these graphs are from: http://www.epi.org/publication/understanding-the-historic-

I've long wondered how productivity is measured, so I went and found the study these graphs are from: http://www.epi.org/publication/understanding-the-historic-divergence-between-productivity-and-a-typical-workers-pay-why-it-matters-and-why-its-real/ http://www.epi.org/publication/understanding-the-historic-di... A few things were interesting: a) Productivity is measured as roughly GDP/hours worked (with some tweaking) b) The biggest chunk of the gap is income inequality. The graph that gets used is for average non-supervisory worker in the private sector. The graph looking at all workers is much closer. c) The second biggest is that consumer goods are more expensive relative to what workers are producing. The term of art seems to be "terms of trade" d) The third biggest is more returns to capital, ergo profits, was the smallest, only contributing 8.9% of the difference; they note that this part is hard to measure, but the sources of error they describe are unlikely to bump this up beyond 10% === At the end of the paper they rebut some common arguments and a section I found particularly interesting was titled "Individuals’ productivity cannot be inferred from industry trends" about why using industry-level productivity statistics is a bad idea. Which pretty much argues that actual productivity is not important because worker's salaries will not rise in more productive industries. They bring up Baumol’s law/cost disease to explain why there is a decorrelation between sector productivity and pay, and I guess this also ties into the "terms of trade" issue. Your barber is no more productive than he would have been 100 years ago, but now he needs to be compensated at a level that makes it worthwhile for him to do it. === Looking at this, I don't really come away more convinced that what economists call productivity is what lay people would call productivity. Or that measuring "average productivity" in this way makes particular sense, rather than just being a convenient and available metric. I also find the graph they use misleading since what that graph is largely (implicitly) showing is that wage gains have not been evenly distributed, but it implies something different. === They have a whole other paper on what they see as policy issues that lead to this (globalization, minimum wage stagnation, fall of unions, economic rents to CEOs) and their recommendations: http://www.epi.org/publication/raising-americas-pay/ http://www.epi.org/publication/raising-americas-pay/