ROI is a wonderful thing. But take it from someone who has spent decades in creating new knowledge and then a couple more in attempting to managing that knowledge as a valuable asset, linking knowledge to outcomes ranges from difficult to impossible.
The best that we came up with is to link occassional home runs to the outcomes and benefits that they generated. We found that a few results more than paid for an entire program. Now, if we could only identify which lines of research were most likely to lead to a home run...!!!
The down side of all this was that the agencies who benefited from the federal research did not want to publicly admit their savings for fear of having their budgets reduced.
Ain't social science interesting!