After we published a column last week criticizing David Brooks and, to a lesser extent, Paul Krugman for the way they used numbers to explain income inequality, a small blogospheric firestorm erupted. The turmoil has died down, but in the aftermath Paul Krugman dedicated a notable post on his nytimes.com homepage addressing the sticky issue of statistics and opinion journalism. He writes, “Now that rising income inequality has become a big political issue, people are throwing around a lot of numbers. Some of these numbers are reliable, other aren’t. But how are readers to tell the difference?”
Good question. And even better, Krugman posits a partial answer.
It is unfair to assume that the average reader is well versed in the complex language and methods of economics. In a letter published on September 13, one Times reader wrote that Krugman’s and Brooks’s opposing columns on income inequality showed that “statistics can be used to prove anything, a fact that has been known for a long time.”
In our September 13 column, we criticized Brooks and Krugman for inadequately defining economic terminology and failing to disclose their specific sources for income statistics. Krugman, for example, made the following statement in his September 8 column: “the income of a typical household headed by a college graduate was lower in 2005 than in 2000.” But he neglected to mention where the statistic comes from or what he means by “typical household.” While economists might immediately recognize that Krugman’s number came from the census and that “typical” is a synonym for “median” (and not mean, or family of four), the average reader, his audience, does not.
In order to approach columns critically, readers should have access to the statistics columnists use. In his Web post, Krugman acknowledges as much: “On inequality, and in fact on many matters economic, it’s all too common to have numbers - some from unknown sources - flying in all directions.” He also lists and describes how to navigate some of the standard sources for economic data—the census, The Congressional Budget Office, and several academic analyses—and promises to post similar explanations for other sources as he introduces them in his columns.
Most importantly, Krugman makes the following pledge: “From now on I’m going to post sources for the numbers in each column on TimesSelect, with links where possible (it usually is.)” With this new policy, Krugman has accomplished two public services. First, he has enhanced his own journalistic accountability. And second, he has widened the scope of informed debate on economic issues to include his readers.
We join Krugman in his hope that “other economic commentators will follow the same practice.” Perhaps in addition to “saving us a lot of confusion,” Krugman’s new policy will mark the beginning of a trend in which opinion journalists use numbers more transparently and responsibly. In fact, columnists at the Washington Post already provide links to the economic data they cite in their articles.
In our critique of Krugman and Brooks on September 13, we accused Krugman of using an obscure—“not widely published” — number. We were wrong about that — it came from the census. But from now on, with Krugman at least, readers will be able to go straight to the source.