The only reason UTF-16 was so widely used is as a backwards-compatibility measure. All these systems originated in the UCS-2 era, or had to be compatible with ones which did. Then UTF-16 came along, and slotted in where UCS-2 was. This is why you have such sloppiness about treating surrogate pairs as two characters, for instance: the systems were made for UCS-2 and only poorly updated to the UTF-16 reality.

Any modern system would use UTF-8.