Those figures are from the 1970s, when presumably there was a lot more low-hanging fruit. Still, I wouldn't be surprised if they invest a decent chunk into computation aerodynamical work as part of their body design.

Part of the issue is the extreme constraints imposed on designers by road regulation, another is simply that the majority (75%?) of drag reduction is in a fishtail that needs to go on the container, the part of the truck that gets left behind at the loading dock, that is often not "owned" by the person who owns the truck.

Because the individual logistics companies are not developing the trucks. The aerodynamic properties of a truck are mostly determined by the model/brand, and this is a huge consideration at companies like Scania, Volvo or Mann. Not so much at the individual logistics companies.

They absolutely do design trucks for aerodynamics. Fuel economy is a huge avenue of competition for semi trucks.

The reality is just that with the drive cycles and costs and tradeoffs of implementing all this it's not worth it to go this far.

Second, this sort of "round the whole thing" approach has mostly been replaced with "the simulation says we can do a 90/10 if we just do X, someone print up a block and toss it in the wind tunnel and see how it does" type approach so the result tends to be more surgical modifications and use of dead air and less "smooth out the whole thing with fairings".

They do, look for some semi-trucks on long haul routes, they have trailer skirts and/or trailer tails.