25 March 2012

Manufacturing and Industrial R&D

UPDATE 3/37: In the comment Matt H. (Thanks!) points us toward a just-released  data update from AAAS -- here in PDF -- which has industrial R&D by total and as percent of GDP.

It's a common assertion among those who argue that manufacturing is special -- the decline of manufacturing in the economy will cause a commensurate decline in industrial research and development. Today that argument appears in the LA Times in an op-ed by
The United States also has put at risk its greatest asset: the return on its intellectual capital. We have let China learn by doing, South Korea innovate by manufacturing, India build new capabilities in design and research and development — much of it on the back of initial American innovation.

With manufacturing gone to China, for example, R&D followed Apple to Foxconn. Applied Materials set up a major R&D shop in China, where solar cells are being manufactured. GE, Texas Instruments, Cisco and others established major R&D and design centers in Bangalore, India.

Why? Because you can't do R&D offshore from a distance. The "look-see-do" of innovation depends on close ties to the manufacturing process. Proximity to manufacturing is the key to other higher-value activities — design, engineering and R&D. And with that, jobs.
One big problem with such arguments is that they lack a basis in actual data.

The graph at the top of this post shows data collated by the AAAS on industrial R&D (the type that Narayanamurti is worried about) which indicates that the period with the most precipitous drop in manufacturing jobs (after 2001) did not see a commensurate decline in industrial R&D. (See this Batelle report in PDF which says that US 2010 industrial R&D exceeded $400 billion, indicating that the trend shown by AAAS continued). The fact that other countries are building their R&D capacities is a good thing and does not diminish the US in any way.

The idea that the decline of manufacturing as a proportion of the economy and net jobs has also resulted in a decline in industrial R&D makes for a great talking point. Unfortunately, it is just not supported by the evidence.