Estimates of productivity growth over long periods of time are entangled with hedonics in a complex way that makes it difficult to make very strong statements. It's very possible that manufacturing productivity is being understated due to increases in the quality or variety of goods that is not reflected in inflation figures. Accelerating innovation makes hedonics adjustments more difficult, and the manufacturing sector is very susceptible to these changes.

This sounds pretty dang post-hoc. For example, the U.S. manufacturing output, as measured in billions of dollars, could have a downturn in 2009 for several reasons - I can count 5 previous downturns on the graph you use, and I'm pretty sure that not all of them are due to impending doom. So lack of any sort of account of ordinary variability is a big weakness of your argument. And maybe there was some kind of economic thingie around 2009 that impacted prices? Just throwing that out there.

I guess I'd like this more if it were more explicitly speculative, but this whole chain of claims about both the decline and the key importance of manufacturing just seems like a castle in the air.