“The Cat Came Back” is a gritty folk song about Mr. Johnson’s unwanted old yellow cat, who just kept coming back no matter what. As economic researchers, we have been tracking down a cat — an old economic research finding too good to be true that, just like Mr. Johnson’s cat, “wouldn’t stay away.”
The research finding, by two Princeton economists, seemed to promise that the minimum wage could be raised while increasing the employment of low-wage workers.
Reporting on a natural experiment in New Jersey and neighboring Pennsylvania, David Card and the late Alan Krueger wrote in 1994, “We find that the increase in the minimum wage increased employment.”
This finding, contrary to long-standing tenets of economics, was widely heralded as the defeat of a false trade-off. We could raise wages and employment among low-income workers — an actual “free lunch.” And this finding would seem particularly relevant today, as 21 states began 2020 with higher minimum wages.
But after six years of data duels, the finding was discredited — or, at least, its strongest claim was. The authors reworded their shocking claim to this much milder form: “The increase in New Jersey’s minimum wage probably had no effect on total employment in New Jersey’s fast-food industry, and possibly had a small positive effect.” That’s a long way down from “the increase in the minimum wage increased employment.”
Their original paper fell out of news and popular discussion. But then, the cat came back.
That’s where our work started. We asked: In the popular media, how well did the revised 2000 conclusion catch up with the less defensible 1994 conclusion? Our answer is “not very well.”
We tracked media mentions of the original finding and found that after falling totally out of circulation from 2003 to 2008, the old conclusion was increasingly cited as true. It enjoyed a revival, peaking in 2015 and 2016 as the topic was front and center in candidate Hillary Clinton’s presidential campaign before fading again.
The old, erroneous finding actually surpassed the new one in media mentions as recently as 2016. Clearly, many people wanted it to be true.
These citations occurred in a wide variety of outlets, from blogs to mainstream journalism, in both news and opinion columns. New York Times columnist Paul Krugman cited the original paper’s conclusions instead of the revised one, 15 years after the revision had been published.
When we saw Krugman’s name pop up in a 2015 opinion piece, we naturally wondered whether he meant “increased employment” or “probably no effect.” His wording was not totally clear, but he linked to the older 1990s “increased employment” conclusion from his 2015 New York Times column.
Even a scholarly journal of the American Economic Association fell prey to the appeal of a costless increase in the minimum wage, as a team of scholars led by Russell Golman said in passing in a 2016 paper, “Raising the minimum wage in New Jersey increased rather than decreased youth employment in the fast food industry.” They cited the 1994 paper — ignoring the dramatic revision that had been public for 16 years.
Our point is not that only one side in the minimum wage debate cherry-picks its studies. Economists are human, too: Confirmation bias knows no ideology. But the case of the Card-Krueger study struck us, even without formal statistics or large numbers to work with, because of the prominence of the conclusion and the later modification.
With another presidential election coming up, we as economists hope that informed voters of all persuasions will be on the lookout for economic policy cats that are just too good to be true — and keep coming back.
M. Scott Niederjohn is the senior vice president for cooperative education and economic development at Lakeland University in Plymouth, Wisconsin. William C. Wood is the department head in economics at James Madison University in Harrisonburg, Virginia.
