I recently received an old document typed on Word for Mac. Accented characters got scrambled because they hadn't been converted from MacRoman to Unicode. I have no idea whether there is anything like an "encoding" or "charset" parameter in these old documents but I guess one can safely assume if it's from Word for Mac, it would be encoded in MacRoman.
I'm afraid 'old document by Word for Mac' a bit too broad. Could you open your file in any hex viewer and paste the first 16 bytes of it here?
fe 37 00 23 00 00 00 00 00 00 04 00 00 19 00 00
Interestingly, I created a file with version 0x23 via export converter, and it is converted properly. Could you attach the document here? If it is not suitable for the public view, you can send it to me to obfuscate it first.
Well, no document, no bug.