If qualitatively they seem to cease to work well as welfare measures, then there is no reason to think they are good welfare measures! When you debunk widespread qualitative expressions of welfare with this "data", you are really debunking the quality of your measures! 10/