Income Inequality

Greg Mankiw and Brad DeLong are debating the income inequality data from the Piketty and Saez study. Paul Krugman talked about rising income inequality in his last column. In response, Greg Mankiw says:

New Data on Income Inequality: In today's NY Times, Paul Krugman calls attention to the update of the Piketty-Saez data on income inequality, although Paul describes the data differently than I would.
Here is what I see: After rising substantially from 1986 to 2000, income inequality is essentially the same in 2004 (the most recent year of data) as it was in 2000.  Click on the data and see for yourself.

In response to Greg, Brad DeLong says:

New Data on Income Inequality: Greg Mankiw sees a little bit of good news in the latest income distribution data: after rising astonishingly rapidly from 1986 to 2000, income inequality in 2004 was no worse than in 2000...

I hope he's right, and that the trend of rising inequality has stopped--it is a very disturbing phenomenon, and further rises would be very worrisome indeed. But I can't be as optimistic as he is. He sees an essentially flat trend from 2000-2004. I see numbers for 1999 and 2000 that may have been transitorily boosted by high salaries paid during the dot-com bubble, and then a decreased in inequality from 2000-2002--a decrease that is then reversed in 2003-2004, which carries us up to bubble levels.

So my hope that we might not see 1999 and 2000 levels of income inequality again appears to have been vain.

We can debate what has happened the last few years and whether this year or that year had special circumstances such as the problems Brad notes with using 2000 as a base year. But in doing so, we shouldn't lose sight of the overall upward trend in inequality. Here's a few graphs from the study:

The Top Decile Income Share, 1917-2004

The Top 0.01% Income Share, 1913-2004

Average Real Income of bottom 99% and top 1% in the United States, 1917-2004

The little dip at the end is what the fuss is all about. But even if Greg's claim holds up to the types of qualifications Brad talks about, and there are good reasons to worry about using 2000 as a base year, the upward trend in income inequality since the 1970s is undeniable. And in any case, as the last graph shows, real income for the bottom 99% of the distribution has been flat since the early 1970s despite large gains in productivity.