Indeed. Were you typing into a text editor and saving the raw file in UTF-8 the quotes would just be quotes. " " ". Unless I have nested quotes, I do not use the character entities for them— even though my colleague is charset-lazy and takes my UTF-8 pages and stuffs them into Latin-1 pages and database. Lots of characters get lost in there if I don't manually write them out as entities (and if they will ever become converted to XML don't use the named entities because there are only 4 that also work in XML from HTML, use the numeric entities), but " isn't one of them. They are the same character in utf-8 and latin-1. Windows 1252 is just weird though.
The problem is you are copying the em dash from an encoding that is not UTF-8.