Nick Szabo  🔑

20% of coins were already counterfeit by his day. He may have delayed the inevitable, but during the remainder of the 18th century, the problem only got worse, driven by craftsmen of Birmingham who also gave us much of the industrial revolution.

By the end of the 18th century merchants had mostly given up on high-value coins in favor of harder to counterfeit bank notes, moving away from trust minimization by substituting IOUs for the actual metal. A trust that in the 20th century would be heavily abused.

I do not see an essential difference between mints and banks in this sense. In both you need to trust third-party and potentially also transacting party depending on how easy was to counterfeit

The metal in coins can be validated. It's costly, not done at retail, but was often done by bankers and merchants. So people who dealt with coins on large and international scales could use coins in a trust-minimized way. But nobody can use bank notes in a trust-minimized way.

I see. You mean that in principle you can validate a metal's weigth, but it is harder to asses what is backing a banknote? If one could know bank assets this problem would be at least partially solved?

If the assets are trust-minimized collateral (i.e. bank doesn't have the ability to just transfer them out). Otherwise, the audit just tells you that the funds have not been absconded with or withheld from creditors *yet* -- useful but much less trust-minimized than Lightning.