Over at the Weekly Standard, Mark Hemingway runs through some of the more egregious examples of media “fact checking,” which has become a staple of their political coverage but, as Hemingway shows, too often doesn’t live up to billing. In his piece, Hemingway approvingly cites my work with Jason Richwine on federal employee pay as a rebuttal to PolitiFact’s “fact check” on statements made by Senator Rand Paul.
As it happens, Jason and I have had two additional run-ins with fact checkers with regard to public employee pay, and both highlight a key problem with these columns: fact checkers have taken on more than they can really chew, and as a result outsource their fact checking to people who either may not be fully knowledgeable or who have their own dogs in the fight.
For instance, this time last year FactCheck.org—the granddaddy of the fact-checking game—published an article titled “Are Federal Workers Overpaid?” which looked to referee disputes between federal employees, who claim to be underpaid, and some conservative politicians, who argue they are paid double private sector levels. (Both are wrong, though conservatives are closer to the truth.) But FactCheck’s writers were apparently wholly ignorant of three decades of peer-reviewed economic research on public-private pay comparisons, nearly all of which conclude that federal employees are overpaid. Worse, FactCheck basically outsourced their analysis to a private pay consultant who thinks that wage regressions are something so exotic that most economists don’t understand them. What we ended up with was a letter to the editor after the fact, which doesn’t amount to much.
More recently, PolitiFact (Cleveland Plain Dealer edition) declared a study of Ohio public sector pay by Jason and me to be “mostly false,” by effectively turning over their analysis to Alicia Munnell of Boston College. Munnell concluded that Ohio public employees are overpaid by roughly 10 percent, which isn’t chicken-feed but is far less than the 43 percent that Jason and I found. Here’s the problem: Munnell’s own work on public sector employees uses a number of questionable assumptions while ours is more consistent with the academic literature. Had PolitiFact, say, contacted us, we’d have told them. But instead they simply published the article, just prior to Ohio’s big vote on repealing Governor Kasich’s reforms of public sector pay, leaving us to respond in a letter to the editor. This was one of very few occasions when I’ve been really, really ticked off with how a news outlet handled something.
Fact checking is good for newspapers’ business and in many cases serves the public, but it’s very easy for them to expand the franchise well beyond their ability to accurately adjudicate the facts. In particular, outsourcing their fact checking to presumed experts who themselves may need to be fact-checked is no way to go.