To avoid potential confusion, I've removed the Euro example from my correction. The main point of the correction (that ISO and ANSI values don't always coincide with Unicode in the high-8-bit range, and that only the Unicode values are guaranteed to work in numeric references) stands.

Since the Euro is likely to be mapped somewhere in one's native codepage, pasting or typing a Euro symbol directly is likely to leave the character improperly encoded. Best to encode the Euro as €. That's all people need to know here— the rest just clouds the issue.