in reply to Re^2: decoding a UTF-16B string found in an email subject
in thread decoding a UTF-16B string found in an email subject
According to all docs I've found, a BOM is not necessary, and when a BOM is not present then big-endian is supposed.
However the string you give seems to be little-endian (as is the case in the problem that got me to this page...).
If you s/UTF-16/UTF-16LE/ then your string gets decoded correctly.