This hinges on the assumption that economists are correctly accounting for value, which seems unlikely. For example, what's the value of being able to repair your jets in a conflict, because the supply chain is local, and doesn't start in an enemy's factory? Comparative advantage tells us we should trade for things that our economy doesn't do well, but if your rival makes bullets better than you do, you're probably still going to make your own. But how do you value that? Security is an intangible asset that can quickly become pathological, as we see in other areas (too much is never enough, just think of the children).

Industries don't exist in isolation, so these effects propagate. If you can't make cars, you probably won't make good tanks. When we assess the value of a local car industry, how do we account for the "use it or lose it" nature of retaining knowledge built up by industry? Part of learning and skilling up is actually doing the thing you're learning about. It's no surprise that the country the West has skilled up in the pursuit of greater profits is now it's chief rival.

And of course, all of this is before we ask questions like what metric is used to determine the benefit. Efficiency in particular is rife with competing definitions that fit various niche use cases, and for which the underlying assumptions may not be obvious. E.g. thermodynamic efficiency is often calculated using the lower heating value of a fuel, the reasons for which are good, but typically left unstated. A layman comparing thermodynamic efficiencies where differing methods are used might draw an erroneous conclusion if they don't understand that there can be differences.

I have never had anybody make this reference before!

Sure, decades of research by think-tanks hell-bent on pumping the assets of their sponsors at any cost to others.

"Trade Wars are Class Wars" by Klein and Pettis is a good counterpoint.