Income Inequality and Statistical Lies

The NY Times (of all publications) is making the case this week that income inequality ain't so bad, and the Economist has latched on in its free exchange blog. (Some of my readers might be shocked that a pinko-liberal-hippie-commie rag would say such a thing. Personally, I like the NY Times, and I think it's a heckuva lot more objective than some of the alternatives out there, while still actually saying something.) The basic case is that household income statistics don't tell the whole story, and that's true enough. Their answer: consumption statistics, an equally silly figure. They note that among the bottom 1/5 of incomes, consumption (on average) is about twice their taxable income. So what? If retirees, students, recipients of government transfers, and the disabled are consuming more than what they report as taxable, isn't that kind of the point? Should we just tear down programs so the statistics will better conform with the "true" extent of income inequality? Hmm. In addition, most of the disparity in the consumption averages versus the income averages comes from change in net wealth. All it really takes is a handful of really wealthy retirees who earn nothing and spend a million dollars to throw that part of the statistical accounting totally out of whack.

Basically, what I'm saying here is that you can't just look at income, and while consumption may be a better measure of welfare for "most" families, you can't focus on that either. Statistically, the means don't justify the ends (of the distribution). In this case, you have to treat zero-income households differently (in statistics we call it "censored" data), which means you have to actually do some dirty work with the data.

Bang!