You aint missing anything, Chris. Good reply.

I take your point about business rules and agree with it all - maybe I wasn't exactly clear in my question.

What wouldn't they just declare a currency datatype of, say, NUMBER(25,2) and have done with it. It could then accept both of their own CURRENCY and LCURRENCY datatypes, and have room to grow should the business rules change in the future. Are they relying on the database to prevent their application from inserting too large a value into the column? Is it a mnemonic definition to help developers to scale their application interface accordingly?

I can't see any storage or usability features of restricting the definition of 1 to (9,2) and another to (11,2) just because CURRENCY will always be a smaller amount than LCURRENCY. Why not just make them both (11,2)?

Maybe I'm confusing myself here...