Think e.g. about the economic effect of converting the stupid part of some country's population to smart people. It is enormous and would completely dwarf the 20Bn price which is what, a rounding error in the US Federal Budget?

Rounding error or not, it's not easy to get research funding in that kind of quantity. $20bn is about twice the annual budgets of Stanford, MIT, and Caltech (including the JPL) put together, or about NASA's annual budget (in late 2000s dollars) at the peak of the Apollo program; we're basically talking in terms of creating a major research university out of thin air and devoting it to one problem for two decades. I don't think it's unreasonable to suggest that you could develop some quite interesting things with that level of resources, although I'm not neuroscientist enough to speak of intelligence research as such.

In any case, contingent on such a program existing, I think I'd expect its bottom line to be dominated by implementation costs, not R&D. Education on a mass scale isn't cheap. Neither are most medical procedures.