An op-ed in the Washington Post on Sunday included data purporting to show that the rich are grabbing an increasing amount of the income earned in this country.

Jacob Hacker of Yale University cites data from economists Thomas Piketty and Emmanuel Saez supposedly showing that “the share of national income held by the richest 1 percent of Americans–stable at about 32 percent throughout the middle decades of the 20th century–began to rise sharply in the 1970s and by 2002 had surpassed 40 percent.”

Actually, the Piketty Saez data show that it is the top 10 percent whose share supposedly rose from 32 to 40 percent.

More importantly, a forthcoming Cato paper by Alan Reynolds shows that the Piketty-Saez data, based on federal tax return information, is a deeply flawed source for such income “distribution” analyses.

For example, because of tax law changes since the 1970s, a huge share of business income that used to be reported on corporate returns is now reported on individual returns. Reynolds finds that much of the supposed rise of the share of income at the top is simply a result of this paper shuffling regarding where business income is reported.

This Piketty-Saez data has been frequently misused by Paul Krugman, the Economist magazine, and many other news outlets to draw grand conclusions about how the gains of U.S. economic growth have supposedly only gone to the those at the top. Reynolds study shows that that is probably not true. At least, his findings show that reporters and pundits should be very careful in using any data that claims to show changes in income shares over time.

Certainly, they should not use sloppy language like Hacker’s phrase “income held by the richest.” Income data shows an annual flow–it is not “held” like wealth. Note that “richest” also refers to wealth holdings, not annual income flows.