Subject: Re: Why Has U.S. Manufacturing Productivity Stagnated_
One possible partial explanation, picking up on his note #4: there were only so many quick gains to be had in the modern era.

Manufacturing productivity (output per hour worked) has indeed risen much more slowly than overall output per hour worked in the US economy since around 2011. But if you look at a longer data series, manufacturing productivity growth blew the doors off total economy productivity growth 1987-2011.

A speculation...the productivity boom just hit manufacturing sooner than elsewhere in the economy. Perhaps computers, telecoms and other automation are great increasers of productivity, but they were put to work in manufacturing before other industries found a way to take advantage of them. Maybe manufacturing was the low hanging fruit. Somewhat consistent with this is that productivity growth in manufacturing has also slowed in other regions in the last 10-15 years (UK, EU, Japan, Canada), so it's not a US-specific trend. So the only thing that needs explanation in the US is not the slowdown, but why the slowdown in the US has been more pronounced than in other countries...a much smaller mystery, though perhaps more mysterious.

In addition to the things mentioned in the article you linked, here are some other explanations that have been floated, cribbed from the Economist, addressing the US-specific more pronounced effect:
* America is known to have laxer antitrust enforcement than its peers; perhaps scrutiny was especially needed in the manufacturing sector.
* Maybe American manufacturing was more advanced when robots arrived on the scene, so had less to gain.
* Because America’s software and internet sectors have been so lucrative, talent has been diverted away from older industries.

As an aside---
I'm leery of a lot of figures related to US capital investment trends, including machinery and robotics. If a US firm builds a factory inexpensively which requires a lot of staff to run, that's one baseline with modest capex and low labour productivity--let's assume the marginal cost of subwidgets is $100 each. If they build a highly automated factory in the US, that's higher capex and higher productivity, a trade-off that gets their marginal cost per subwidget down to only (say) $75 after the bigger up-front cost. But if their supplier builds a factory in an Asian country, and sell the subwidgets to our US firm, that's zero capex in the US, but their marginal cost per subwidget might be only $50. Note that the capex is counted at zero in the accounts of the US and of the US firm, but really just took place elsewhere. And note that the third option was by far the best for the US firm: their cost per widget decreased the most, but their capex requirement was the lowest. This is a huge increase in functional productivity--it just doesn't show up when you look too locally.

To overgeneralize, manufacturing capex has only tumbled in rich countries if you get misled by looking at one country at a time. Global manufacturing capex has risen strongly...the activity just moves around. The US stats look bad, but that's because they're the wrong stats. Within the US, the place that this type of "net" productivity benefit might show up is in things like, say, higher US corporate operating margins. Which, as noted in other threads recently, have been weirdly wonderful lately.

A corollary of big changes in outsourcing is that almost all national statistics of capex, employment, compensation and output are at risk of being distorted by this effect. Sometimes to the point of uselessness. For example, one factor in the fall in manufacturing employment is that manufacturers don't employ custodians any more, they hire cleaning firms. Apply this single notion to a thousand different activities.

Imagine trying to calculate the labour productivity of a neighbourhood by adding up the capex and payroll and sales of the firms located there. It might have been just fine so long as everything was local or the town employer was the dominant factor, but any trend you found recently would be meaningless if there were even a single factory more or fewer. In the modern era, looking at just one country, no matter how big, has the same problem.

Jim