host: gary burtless, a response before we go to calls?uest: the question about whether there has been a rise in inequality in consumption is much more doubtful than diana suggests. spending patterns are based on a survey called the consumer expenditure survey. there is no survey conducted by the united states government that has seen such a terrible decline in the quality of responses in the last 30 years. in the mid-1980's, 80% of the consumption of the u.s. household was recorded by responses in this survey, and that number, by the middle of the last decade was down to just 60%. we of gone down to a missing almost 40% of the consumption, and no one knows how that extra consumption is divided among the very affluent and the low- income. if you cannot trust this to tell you the overall trend in average consumption, you can not trust it to tell you the distribution across well-to-do, and less of the when households. guest: i am not saying the service it is perfect, but what is interesting is the consistency -- the two and a half times rate