You can make a copy of unicode.4hf in your working dir and delete the line with ’. It will be converted as a character then.
Michal - the question is, should we do that in the sources? On the theory that with -cunihtf -utf8, characters should be output, not entities. It is not logical to output an entity for ' and not for `, after all. However, I have no particular clue how to affect the output only in the case of those options, and I don't think we should change the default output ... -k