The Trade Games Are Back

U.S. trade policy is truly fascinating. Probably more than in any other area of public policy, trade agreements are structured by corporate interests behind closed doors. Then when a deal is produced, the establishment media and economists insist that we have to support the deal behind the important principle of “free trade.” The opponents are treated as knuckle-dragging Neanderthals who just can’t understand how the economy works.

We got another episode in this long-running show last week when the United States International Trade Commission (USITC) came out with its assessment of the United States, Mexico, Canada Agreement (USMCA), also known as the new NAFTA. It came as a surprise to virtually no one that the USITC study showed economic gains from the deal. The study projected that the deal would lead to an increase in GDP of 0.35 percent when its effects are fully felt. However, the real impressive part of the story is how it got this result.

Before getting to the particulars, it’s worth putting some perspective on the character of the report. The USITC was obviously determined to make the USMCA look good. One reason this is clear is the failure to put a date for the year for which their projections are made. The model that the USITC used to project gains, the model from the Global Trade Analysis Project (GTAP), assumes a long period through which the economy gradually adjusts to the changes put in place from a trade deal. The projected impact refers to the end year, which is around 16 years in the future.[1]

Incredibly, the UISTC study projecting the impact of USMCA neglects to mention the end year for its projections, which presumably would be something like 2034. The end year is presumably left out because a gain of 0.35 percentage points of GDP over sixteen years looks rather pathetic. It comes to just over 0.02 percentage points a year, an increment that would be essentially invisible.

By contrast, if we look at the report the USITC did projecting the impact of the Trans-Pacific Partnership, the year for which the projections are made, 2032, is mentioned literally hundreds of times in the report. It is clearly identified in every table. There is no legitimate reason why the USITC would not clearly identify the year for which its projections are supposed to apply. The people drafting the report obviously made a conscious decision to go against standard practice in the economics profession, and their own past practice, and conceal the time period over which the economy is projected to adjust to the changes in the trade deal.

But ignoring this obfuscation, it’s worth examining the impacts projected in the report. In one area, the report is entirely consistent with standard economic theory. It projects that the rules that are designed to incentivize more auto or auto part production in the United States, will in fact have this effect. The projected increases in output and employment are modest, employment in the auto sector is projected to rise by 28,000 (3.6 percent of the total) when the impact is fully felt.

Following standard trade models, these rules raise the average price of cars in the United States and lead to small drop in sales. Other things equal, the USMCA provisions for the auto sector would lead to a small drop in GDP and employment in the United States as the gains in the auto sector are more than offset by the losses to other sectors from paying more for cars. (Remember there is no involuntary unemployment in these trade models.)

With the changes in manufacturing rules alone (primarily auto) the USTIC would have to report the USMCA as a net loser in GDP and employment. However, it was able to pull a rabbit out of a hat with the claim that greater certainty, primarily in the handling of digital products (Internet commerce and data), would add 0.47 percentage points to endpoint GDP. This compares to a loss of 0.12 percentage points without the reduction in uncertainty (see Table 2.6).

There are three points to be made about the gains from reduced uncertainty assumption:

1) The impact is large compared to standard projections of the impact from trade liberalization;

2) Uncertainty arises in all areas of economic activity, if reducing the impact in trade has large effects, that must be true in other areas as well. (Also increased uncertainty should imply correspondingly large losses.)

3) The USITC is implicitly assuming that we are locking in arrangements, especially in the case of digital products, that are optimal.

Starting with the size of this projection of the benefits from reduced uncertainty, it is more than twice the 0.21 percentage point gain in GDP the USITC projected from the TPP. This was a trade agreement with eleven other countries, five of which we did not have previous trade agreements.

By contrast, the U.S. economy has been tied to Mexico’s economy for more than a quarter century with NAFTA and it has been tied to Canada for more than three decades with the U.S.-Canada trade agreement. It is difficult to believe that there is all that much uncertainty that will be reduced with this trade deal.

Taking the second point, it is perfectly reasonable to argue that uncertainty makes investment more risky, and therefore less uncertainty will lead to more investment, but this cannot only be the case with uncertainty associated with trade rules. Anything that creates uncertainty in the economy, such as changes in tax laws, changes in regulations, or changes in currency values, should reduce investment and therefore reduce growth.

Economists have generally not modelled these changes this way. For example, few, if any, economists argued that the Trump tax cut would lower GDP by creating more uncertainty about the future direction of the tax code. Nor do we hear many economists raise concerns about changes in currency values (which can be as much as 20-30 percent, in real terms) leading to sharply lower investment and growth. If uncertainty really has this large an effect on output (remember, the reduction is occurring in a small area of the U.S. economy – digital products traded between the U.S., Mexico, and Canada), then economists have been seriously negligent in ignoring a major determinant of growth.

It is also worth mentioning that the Trump administration has created an enormous amount of uncertainty with its seemingly arbitrary imposition of tariffs, threats of larger ones, and efforts to derail the WTO dispute adjudication mechanism. If the reduction of uncertainty associated with the USMCA would add 0.47 percent to GDP, the increased uncertainty created by these actions would probably subtract several times this amount.

The third point is that assuming gains from locking in current arrangements implicitly assumes that these arrangements are optimal. This means that we are not precluding better arrangements by having the rules in the USMCA.

That seems a rather heroic claim in the case of digital products. Is anyone prepared to claim that we now have optimal rules governing the behavior of Facebook, Google, and other Internet giants? Perhaps at some future point we will decide that these companies have to be regulated as monopolies, with the government imposing strict rules on privacy, data transfers, and crossovers to new lines of business. That may still be possible under the rules of the USMCA, but it is virtually certain that these companies would look for protection under the terms of the deal.

At the very least, it will be more difficult to alter the regulatory structure for Internet companies under the USMCA, which is precisely the point of reducing uncertainty. If we are not confident that we want to lock in current rules, then this is a big step in the wrong direction. And, it is important to keep front and center that the USMCA rules apply not only to our trading partners, they apply to the United States as well.

This point is important to remember in other areas as well, most importantly pharmaceuticals. The deal locks in stronger patent-related protections than currently exist in either Mexico or Canada. While the USMCA rules may not be stronger than what currently exists in the U.S., they will preclude efforts to weaken these rules to lower drug prices. This is especially important in the case of bio-similar drugs, which are excluded from the market for ten years under the terms of the deal. That is less than the current twelve years in the U.S., but more than the eight years that most other countries consider adequate.

It is bizarre that when it comes to patent and copyright protection (yes that is “protection” as in protectionism) most trade economists are just prepared to assume that more protection is always positive, or at least neutral. These protections create enormous distortions in the market, typically equivalent to tariffs of many thousand percent. Almost all drugs would be cheap without patent and related protections, selling for ten or twenty dollars per prescription, rather than hundreds or thousands of dollars.

Patent and copyright protection also create all sorts of perverse incentives and lead to the sort of corruption that economists predict will result from tariff protection. The difference is that the corruption associated with a patent that raises a drug’s price by several thousand percent is many times greater than the corruption resulting from a tariff on steel or aluminum of ten or twenty-five percent.

Arguably, the opioid epidemic is in large part due to patent monopolies. Purdue Pharma would not have had anywhere near as much incentive to push OxyContin if it had been selling at generic prices from the day it was approved by the FDA.

It is not just patent and copyright protection that most “free trade” economists are happy to ignore in discussions of free trade. While we have worked hard to eliminate any barriers that might protect our manufacturing workers from low-paid competition in the developing world, we have done little or nothing to reduce the barriers that protect our most highly paid professionals such as doctors and dentists.

As a result, our doctors earn roughly twice as much as their counterparts in other wealthy countries. And, there is much more money at stake with doctors (roughly $300 billion a year or $900 per person), than with steel (roughly $80 billion). But for some reason, the protections that allow doctors to earn so much more than their counterparts elsewhere never raise as much ire as a tariff on steel or aluminum.

The long and short is that most economists, and most people in policy circles, are prepared to bless any trade deal as a step forward in the great pursuit of “free trade,” even if the pact was crafted by special interests who have no interest whatsoever in free trade. New York Times columnist Thomas Friedman was good enough to state this explicitly:

“I was speaking out in Minnesota — my hometown, in fact — and a guy stood up in the audience, said, ‘Mr. Friedman, is there any free trade agreement you’d oppose?’ I said, ‘No, absolutely not.’ I said, ‘You know what, sir? I wrote a column supporting the CAFTA, the Caribbean Free Trade initiative. I didn’t even know what was in it. I just knew two words: free trade.'”

Unfortunately, Friedman is typical of the policy elite. They will endorse anything that ostensibly advances “free trade,” even if the deal is just about giving more money to the corporations who were the main actors in drafting it. There are few people in policy positions who actually are interested in free trade, they just like to use it as a cover for policies that make the rich richer.


[1] On page 43, the USITC reports that full adjustment is assumed to be six years after the USMCA is implemented. This is hard to understand, since many of the provisions of the agreement will not even be fully in place at that point.

This post originally appeared on Dean Baker’s Patreon page.

Dean Baker is the senior economist at the Center for Economic and Policy Research in Washington, DC.