In response to the nomination of Alan Krueger to head the Council of Economic Advisors, National Review Online's "The Corner" blog has published an embarrassingly ignorant attack on minimum-wage research that Krueger did with David Card back in the 1990s. The post reheats the very stale talking points pushed by fast-food lobbyist Richard Berman circa 1995 and leaves NRO readers with the completely false impression that the story ended in 1995 and did so badly for Krueger. NRO readers deserve to know the full story, if for no other reason that they can therefore avoid embarrassing themselves by repeating NRO's astonishingly incomplete account of events.
In case you weren't paying careful attention to all this back in the 1990s: In 1992, New Jersey implemented a state-level minimum wage that raised the state rate above the federal minimum and, importantly, above the minimum wage then in effect just across the state line in Pennsylvania.
Card and Krueger (C&K), who were both teaching at Princeton at the time, got the idea to compare fast-food employment in restaurants on either side of the New Jersey-Pennsylvania border, before and after the increase. If the minimum-wage increase was a job-killer, employment should have suffered in New Jersey relative to Pennsylvania, they reasoned.
Before the increase went through, C&K organized a telephone survey of wages and employment at a randomly selected sample of fast-food restaurants on either side of the border. They followed up with the same restaurants after the increase and assembled a final sample with 410 establishments. To the consternation of conservatives everywhere, their meticulous analysis of the results, published in the peer-reviewed American Economic Review in 1994, found "no indication that the rise in the minimum wage reduced employment" in New Jersey relative to Pennsylvania.
Fast-food lobbyist Richard Berman jumped into action. Working with a subset of restaurant owners in the same areas originally surveyed by C&K, Berman, together with economists David Neumark and William Wascher, collected payroll records to create their own, smaller, sample of 235 restaurants in the two states. Neumark and Wascher (N&W) made a great deal at the time that their payroll-based data was of much better quality than C&K's telephone-survey data, even though the N&W dataset was substantially smaller and not randomly collected. Most significantly, N&W argued (pdf) that the C&K data showed "substantially more variability over the period between their surveys than do the payroll data," an argument the NRO piece repeats today.
The NRO blog post leaves readers with the impression that after N&W published their analysis as an NBER Working Paper in 1995 that C&K's reputation lay in tatters (or should have). But, two crucial pieces of subsequent information are missing from the post.
The first is that after substantial delay, N&W eventually granted C&K access to the payroll data. C&K flagged serious problems in the N&W data (which C&K rightly refer to as the Berman, Neumark and Wascher data, since the fast-food lobbyist was instrumental in pulling the initial sample together). If you are a fan of great technical writing, or just enjoy watching a devastating take-down, read C&K's discussion of the problems with the Berman, Neumark and Wascher data in this 2000 paper in the American Economic Review.
The second and more important point, however, is that after 1995 C&K replicated their earlier work using what is called ES-202 data. These data are employment and earnings reports filed every quarter by essentially every business operating in the country (the data are used for the unemployment insurance system and are required by law). N&W's critique was that the C&K data weren't reliable because of reporting errors inherent in telephone surveys. The ES-202 data are as reliable as the payroll data — and have the added advantage that they constitute a virtual census of fast-food restaurants operating in the relevant parts of New Jersey and Pennsylvania. These new data answered the main objection raised by N&W and substantially increased the information available for the analysis.
What did the new C&K paper, which was published in the American Economic Review in 2000, find?
"Based on all the evidence now available, including the BLS ES-202 sample, our earlier sample, publicly available BLS data, and the BNW sample, we conclude that the increase in the New Jersey minimum wage in April 1992 had little or no systematic effect on total fastfood employment in the state, although there may have been individual restaurants where employment rose or fell in response to the higher minimum wage."
In the end, a careful examination of all of the data — the original C&K survey data, the non-random N&W payroll sample, and a dataset built from nearly every fast-food establishment in the relevant parts of New Jersey and Pennsylvania — entirely vindicated C&K's initial findings. But, NRO is content to repeat fast-food industry talking points — from 1995.