A lot of research ends up supporting or augmenting existing technologies and industries -- so-called "applied" research. This research shouldn't be expected to yield major game-changing innovations.

A lot of the GDP growth over that period was fueled by technological capabilities. Which means a lot of the new research funding ended up going toward applied research supporting industries build around those new capabilities.

Given these trends, measuring against GDP is a far more reasonable metric than absolute dollar values.

Basically, research in the past made new industries possible while providing a bit of subsidy to existing industries, whereas research today is mostly government subsidy for product improvements in existing industries with a little bit of basic research on the side...