Glenn Kessler has a difficult job. He is asked to assess claims that often arise in the middle of heated political debates. Inevitably his judgements will leave some unhappy. I sometimes fall into the unhappy camp, as is the case today with his assessment of Public Citizen's claims on the job impact of the Korea-U.S. Free Trade Agreement. He gave Public Citizen a whopping four pinocchios in saying that the deal led to the loss of 85,000 jobs.
Before getting to the substance, let me say that I have known Kessler for years, and I'm sure he reached this assessment in good faith. I have discussed many issues with him over the years (including a related trade issue) and I have always felt that he was trying to determine the facts of the situation. In this case, I just think he got it wrong.
He raises several objections to Public Citizen's number. Just briefly listing them, he is concerned about:
1) the relationship between the estimates of jobs per dollar of exports and jobs per dollar of imports;
2) the years chosen as endpoints for the Public Citizen analysis;
3) the nature of the counterfactual (i.e. what would have happened in the absence of the trade deal); and
4) Public Citizen's exclusion of re-exports from the calculation.
There is some validity to all of these concerns, but none of these issues individually or collectively undermine the basic story in the Public Citizen report.
Starting with the estimates of jobs per dollar, these are always very crude. Public Citizen took their estimate of jobs per dollar from a report from International Trade Commission which calculated how many jobs would be generated by a projected increase in exports associated with the U.S.-Korea Free Trade Agreement. Kessler points out that this estimate is from 2007 so it would be out of date by 2012, when the deal first took effect. Furthermore, it was an estimate of jobs per dollar of exports, not imports.
Both of these points are true, but not likely of much consequence. The original estimate of jobs per dollar of exports is at best a crude approximation rather than a carefully constructed number. Certainly no one could rule out that the true number in 2007 may have been 10 percent higher or 10 percent lower. Ideally we would have separate estimate for the jobs lost in the industries competing with imports, but it is unlikely to be hugely different. (Most exports and imports are manufactured goods.) Furthermore, U.S. imports tend to be somewhat less capital intensive than U.S. exports, which means that there are likely to be more jobs lost per million dollars of imports than are gained per million dollars of exports.
On the second point, Kessler takes issue with where Public Citizen decided the trade agreement's impact would first be felt. There will always be some ambiguity on this sort of question because businesses will start responding once they are sure the deal will go into effect, even before it actually is in place. Whatever date one picks, it is pretty hard not to see a clear pattern of rising deficits around the time of the deal. Here's the data from the Commerce Department going back to 2007.
Source: U.S. Department of Commerce.
The deal was ratified in late 2011 and went into effect in March of 2012. Whatever date we want to pick as the point at which the deal first started having an effect on trade, there seems no way of escaping the fact that there was a large increase in the deficit after that point.
The next point has to do with the counterfactual, what would have happened in the absence of a trade deal. Kessler points out that Public Citizen was not prepared to claim that we would have had 85,000 more manufacturing jobs in the absence of the Korea trade deal.
Public Citizen's reservations on this point are appropriate. These sorts of calculations are always made assuming that other things are equal. In the real world other things are never exactly equal. In the absence of the trade deal with Korea perhaps we would have seen larger trade deficits with other countries. Alternatively maybe the deficit with Korea would have risen by pretty much the same amount even without the deal. Both are plausible stories, but if anyone thinks that these or other counterfactuals are appropriate then it is incumbent on them to produce the evidence to make their case.
Finally, Kessler objected to Public Citizen's decision to exclude re-exports from its calculations. The issue here is that some of the goods that we export to Korea actually were produced in other countries but just get transshipped through the United States to take advantage of the trade agreement. For example, Mexico may start sending trucks through the United States to Korea so that they can enter without paying a tariff, as opposed to whatever tariff Korea would apply to imports from Mexico.
There are two problems with this calculation. First, to take an example from the United States Trade Representative, some items are not pure re-exports. For example, we may import the truck from Mexico, but then install mirrors in the United States before exporting it to Korea. In principle we should include the value of the installed mirrors in our calculations of exports to Korea. As a practical matter we generally don't have the detailed data that would allow us to calculate the value of the installed mirrors. In this case it would be much more accurate to exclude the value of the truck altogether than to count the entire price of the truck as an export from the United States.
The other problem is that some of the items exported from Korea to the United States will also be re-exports from other countries in the region. The logic would be the same, they are sending goods through Korea in order to take advantage of the fact that these goods can enter the United States without paying a tariff. By not excluding such goods from imports the Public Citizen calculation would be overstating the rise in imports from Korea. My guess is that the amount of such re-exports from Korea is much smaller than the amount of re-exports from the United States for the simple reason that U.S. tariffs with most countries were already very low.
Fortunately it doesn't make much difference in the case of Korea. The data in the figure above show the balance with all exports and imports.
In conclusion, it is hard to see the Public Citizen calculation as four pinocchio stuff. They largely followed the rules of Washington trade debates in setting up their calculations. There are plenty of grounds for criticism of Washington trade debate standards and certainly places to question the specifics of Public Citizen's calculations, but this is the stuff out of which trade policy has been made for decades.