What's even more interesting to FLG is that Matt also writes this:
Federal regulators started to relax this rule in the 1980s and it was firmly repealed in the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994. It seems to me that if you want to look at a piece of 1990s financial deregulation that set off waves of bank consolidation, this is where you want to look.
And also writes:
The “one state only” rule is totally arbitrary. There’s no principled reason you should be allowed to have branches in Manhattan and Buffalo but not Manhattan and Newark, or that it’s okay to work with a local branch of a very large bank in Yuba City but not in Utah. But it does limit the overall size of banks pretty severely. California, Texas, and New York banks would be kinda big and banks in most of the country would be quite small. And the fact that the rule is so plainly arbitrary makes it relatively easy to enforce. Since it simply piggybacks on existing lines rather than claiming to be some fine tuned optimization, it’s relatively difficult to game the system without people noticing.
To be clear, because he wasn't in his previous post on this, FLG thinks that 1) the relaxation of interstate banking was and is more to blame for the existence of megabanks than repeal of Glass-Steagall and 2) breaking up banks by states is a preposterous idea, but 3) it's less preposterous than trying to regulate financial companies into separate functions in a world where the nature of modern, computerized finance is blurring those lines constantly. Moreover, people seem to forget that Bear Stearns and Lehman were pretty much pure investment banks, meaning that they would've existed pretty much as they were, even if Glass-Steagall were in place. Ergo, even under Glass-Steagall you still have the exact same problem.