Also, I think the point of the paper is that growth gets harder, not easier. From that perspective R&D spending should have increased massively in constant currency to keep pace, not decline even slightly.
I think that in an increasingly specialized workforce, the management team has less bandwidth to apply to R&D. One stump hit and it's done for. Successfully leading R&D has little payback in the larger economy; leading M&A is better understood. This is a corollary of the general "tower of Babel" problem.
Remember that Amazon is slightly a tribe of madmen to most business types in large companies.