Purpose for using Flows and Custom Objects vs. Named Credentials (w/ AuthenticationProvider)

[dturkelFerguson]

Was there any specific reason (other trying to make things appear more configuration vs. code-based) why the all of the authentication/token management within SFDC for the SFCC tokens is done using Flows/Subflows + Custom Objects, versus using Named Credentials and possibly a custom AuthenticationProvider (with custom metadata)?

The flow-based solution feels like it brings a lot of overhead for capability that the platform provides when it comes to managing authentication tokens; I can't say for certain if performance suffers, but when I see the all of the flows involved and the query-based manner of retrieving the most recent valid token, it makes me wonder.