I often think the same when I hear people pitching blockchain based identity solutions (especially for the developing world). They correctly identify a problem: lack of a provable identity presents hurdles for poorer people interacting with the economy. But they fail to understand the root causes or the specifics of the problems. The software industry has a habit of viewing problems in other fields/professions as "simple" if you throw technology at them; it's almost a collective Dunning-Kruger effect.
I would slightly quibble with your point that "Incidentally, this is the problem with all the startups claiming the blockchain will revolutionise the integrity of global logistics markets". Blockchain may allow automation of some of the paper processes for bills of lading etc and the nature of the logistics industry with lots of counterparties involved in one transaction might make a distributed system attractive. It wont make it much more fraud resistant than the current paper processes, but it could make it faster and more transparent which could be of value.