I've been asking this exact question of blockchain enthusiasts and evangelists for over a year now, and have never received a clear response.

I understand the difference between a distributed ledger and a database, but in all use cases I've heard mentioned as ripe for blockchain derived innovation, not once has a blockchain really been necessary or even an improvement over a traditional database based solution.

To give you an idea of how buzzwordy this all is, my friend works in venture capital and was trying to explain to me how blockchain was going to change the world after he got back from a blockchain conference, saying that it would "revolutionize VR/AR, AI, etc...", technologies that have absolutely nothing to do with one another. I just laughed.

I suspect it boils down to the international trade part. Which party would have control over the centralized database? Each participating nationality might prefer to not have that database under the control of any of the other nations. In any case, the centralization represents a single point of failure with potentially a fairly large risk.

I don't necessarily believe this myself, but here goes:

A database under control of any one entity is prone to undue influence by hacking or bribing or other leverage. An append-only database that is controlled by consensus of all club members is safe from malfeasance on the part of a minority. There is value in that. Of course this has nothing to do with mining coins and stuff.

Now you can argue the society already solved these problems with rigorous administrative processes around existing databases. I would reply that blockchain-based process may end up being cheaper and/or faster to settle and reconcile. In the same way the computing previously done by humans was turned over to machines. It may have looked like the problem is under control and there is nothing to solve here, but it turned out there was a lot to be gained from dramatically increased computational power.