I don't really agree with your characterisation of ROI.
Every potential decision outcome has a ROI (which we mostly don't know in advance, though we can often guess). The investments for which we need governments are those that won't ever work in the private sector due to misaligned incentives (no company will pay out of its own pocket to educate young children, not because there's no ROI overall but because it's much too diffuse -- the ROI for that company is extremely low).
People can disagree about whether government should also make some investments that would also be made by self-interested people/companies.
ROI is very much a business term of art. And that's how it's mostly used here in this little venture-capital supported niche.
I agree one can broaden it to mean something much vaguer, using it metaphorically. But A) I think we at least have to explicitly distinguish that, and B) I think casting societal questions in business terms is a perilously wrong frame, one easily leading to all sorts of errors of thought, ones that have negative social implications.
Just as an example, we could take Elon Musk's recent swipe at funding to fight homelessness. There he uses a common business framing and ends up with some conclusions that are deranged from the point of view of people actually working on the problem. Which wouldn't matter much if he were some random internet commenter, but here his error could have a significant body count.
Every potential decision outcome has a ROI (which we mostly don't know in advance, though we can often guess). The investments for which we need governments are those that won't ever work in the private sector due to misaligned incentives (no company will pay out of its own pocket to educate young children, not because there's no ROI overall but because it's much too diffuse -- the ROI for that company is extremely low).
People can disagree about whether government should also make some investments that would also be made by self-interested people/companies.