When IRC was developed it was expected that most users would be connecting from systems controlled by someone else and which prevented users from running their own identd servers or changing the reply. Ident responses could therefore be trusted as much as the remote system administrator could be trusted. Most people now have full control of their own systems and can run anything they like on port 113, so the identd response could no longer be trusted even if mIRC itself followed the specification, so mIRC doesn't bother.
The presence of an identd response can still be a useful indicator though, since most IRC drones and floodbots don't have them. If a channel gets attacked then it may be able to place a ban on *!~*@* and still retain most of its genuine users.
True identd is still used when IRC eggdrop/bouncer providers need to set up trusted connection agreements with IRC networks.