I appreciate the work that Glenn Kessler does as the writer of Washington Post's Fact Checker column. It's a difficult job. I don't always agree with his assessments, but I think he tries to be fair in his analysis. For this reason I was disappointed to see him max out with four Pinocchios over Donald Trump's trade representative Robert Lighthizer saying that NAFTA led to a government certified loss of 700,000 jobs.

According to Kessler, the basis for this figure is the 757,000 petitions for NAFTA-related trade adjustment assistance that were certified by the Labor Department between Jan. 1, 1994 and Jan. 1, 2001. Kessler raises three major objections to this figure.

First, he argues that the number is old. This is true, but it is difficult to see why that is relevant. There may have been some additional NAFTA related job loss in subsequent years, but that would make the number higher not lower. Complaining that the number is dated would be a bit like criticizing a figure for the number of traffic accidents in 1995. Presumably, there has been no major recalculation of the number of accidents that took place in 1995, so using the originally calculated number would be reasonable for most purposes, even though it is now more than twenty years old.

The second point is that the number could be overstated because the Labor Department was very generous in accepting petitions and likely gave assistance even in instances where the job loss had nothing to do with NAFTA. This is undoubtedly true, but there also had to be many cases where workers lost jobs due to NAFTA, who never filed a petition.

For example, suppose a factory closed in large part due to competition from a Mexican factory. While the factory workers might file successful petitions, would the workers at a restaurant that shut due to the loss of business from factory workers? The certification process is clearly open to errors in both directions. I don't know which type of error would be larger, but Kessler doesn't even consider the possibility that the certification process might have missed NAFTA related job loss.

The third issue is that Lighthizer doesn't take account of the jobs gained due to NAFTA. This is right, but it is not obvious why this makes his statement about job loss false. Kessler says on this point:

"As we have noted before, most mainstream economists do not believe that the number of jobs is significantly affected by trade policy; instead, trade changes the mix of jobs."

This is true, with an important qualification, but it is not clear it responds to Lighthizer. His claim was that workers lost jobs due to NAFTA. If these job losses were largely offset by other gains, the net impact on employment may be inconsequential, but the workers who lost their jobs are likely worse off as a result of the trade deal. (The qualification is that if the United States is suffering from a shortage of aggregate demand, as it did for many years following the collapse of the housing bubble, a trade deficit reduces demand and for this reason leads to fewer jobs in the economy as a whole.)

Furthermore, if NAFTA reduced the demand for non-college-educated workers and increased the demand for college-educated workers, as conventional trade theory would imply, this would mean real losses for those workers without college degrees, which would go along with gains for workers with college degrees. The impact may not have been that large, since our trade with Mexico is not that large (Canada is part of NAFTA, but we already had a trade deal with Canada before NAFTA), but Lighthizer's claim of 700,000 lost jobs is not all that large relative to the size of the labor force (150 million), as Kessler himself points out.

Lighthizer certainly can be criticized for making a bigger deal out the number and the impact of NAFTA than the facts warrant, but that hardly seems a four Pinocchio offense. In short, Lighthizer is certainly hyping the harm from NAFTA and arguably making it a bigger deal than he should, but his basic claims seem right. (I would add that NAFTA could have had a substantial impact on workers' bargaining power by making the threat to move a plant to Mexico very credible.)

Moving away from Kessler's piece, the Post has had an absolute love affair with NAFTA ever since the deal was debated by Congress in the early 1990s. This love affair led to the absurd claim in a 2007 editorial that:

"Mexico's gross domestic product, now more than $875 billion, has more than quadrupled since 1987."

This would have implied a 7 percent annual growth rate. China managed to do this over a two decade period, very few other countries have. According to the IMF, Mexico's GDP grew by 83 percent over this period, which is pretty far from quadrupling.

Incredibly, the Post has never bothered to issue a correction to this editorial. Newspapers typically append a correction in the Internet Age in case some reader happens to stumble on the piece. Apparently, the Post's editorial writers don't like to acknowledge their mistakes, a trait they share with Donald Trump.