The US International Trade Commission (ITC) recently released a comprehensive analysis of the economic impact of all trade agreements the United States has entered into since 1985. Trade agreement negotiation, passage and implementation has been controversial throughout this time period. The ITC report is significant as the first recognition by a federal government trade agency finding these agreements produce little benefit as well as negative consequences.
Key Points
- Negligible benefits for the U.S. economy resulted from the 16 trade agreements signed by U.S. between 1985 and 2020, refuting decades of extravagant claims for free trade agreements (FTAs) made by politicians.
- Confirmed the results of a growing number of prior studies using economic models that found trade agreements do not necessarily deliver benefits for large countries like the U.S.
- Income inequality often worsened, including detrimental effects for workers without college degrees, women, and people of color.
- Small and medium-sized businesses were harmed, while multinational corporations were helped.
- Small, poor nations benefit far more than rich, large nations from trade agreements.
- Economic models continue to assume no unemployment as a result of import penetration. When unemployment is estimated, model results are often negative.
ITC Findings
In 2015, Congress passed the Bipartisan Congressional Trade Priorities and Accountability Act in which the ITC was directed to report on the economic impact of all trade agreeements implemented since 1984. The ITC has now released that report, Economic Impact of Trade Agreements Implemented under Trade Authorities Procedures, evaluated 16 U.S. bilateral and regional trade agreements, from the US-Israel agreement of 1985 to the South Korea, Colombia, and Panama agreements which were implemented in 2012. Included in the study is the NAFTA agreement with Canada and Mexico (implemented in 1994), which has been by far the most consequential for the U.S. economy. The 390-page study also included separate “case studies” of specific effects of several smaller industry-specific agreements.
This wide-ranging comprehensive study, covering the 35-year period from 1985 to 2020, found that those 16 trade agreements yielded just $88 billion worth of additional value to annual U.S. gross domestic product, a miniscule amount, about half of one percent. These 16 agreements were found to have added just 1.6% to our exports and 3.4 percent to our imports. Note that if exports are responsible for production, income, and jobs, and imports are conversely responsible for lost production, lost income, and lost jobs, a rise in imports at double the rate of a rise in exports suggests a net loss of jobs.
If the ITC estimate is accurate, one company, Google, adds more to U.S. GDP than all 16 trade agreements combined. And Google was founded by two guys in t-shirts in a garage in California with a few lines of software code and a stack of computers in the garage and a fan pointed at the computers to prevent them overheating. Those 16 trade agreements, on the other hand, required hundreds of officials negotiating for years, often at the finest hotels in the world. The result was that in exchange for the estimated $88 billion of annual benefit, the U.S. suffered huge costs in terms of mass unemployment and social devastation from Youngstown, Ohio to Oakland, California, and hundreds of points in between, as factories shut and millions of workers were laid off due to increased imports.
But perhaps the most damning part of the ITC report was the various comments on the effects of trade agreements on income distribution for different groups within the U.S. working population. The ITC’s analysis reflected the findings of a number of other recent economic studies on the U.S. impact of trade agreements, especially NAFTA. The ITC’s conclusions, coming from a bipartisan agency of the federal government, are worth quoting:
“In U.S. localities that had industries which had been protected by tariffs on Mexican goods, findings show that (1) there was lower wage growth for blue-collar workers in those U.S. industries and localities, (2) lower wage growth for blue-collar women than men, (3) a larger increase in unemployment among nonwhite workers than white workers, and (4) outward migration of workers without a high-school degree from those localities.” [1]
In other words, NAFTA inflicted pain on the weakest members of society, made an unequal society even more unequal, and the benefits, such as they were, accrued primarily to the affluent and multinational corporations. And there is no reason to believe that these trends have stopped. On the contrary, they continue today.
Free traders were blindsided by the ITC report. Some, including at the Peterson Institute, attacked the report openly, sniping at details while apparently lacking an understanding of the economics behind the report. Bloomberg News noted ruefully that when President Bill Clinton held out trade agreements as a route to “economic freedom” for the world, “Clinton’s exuberance didn’t exactly pan out in the way that many had once hoped.”[2]
Other Views Support ITC Criticisms
Many expect economists like those at the ITC to always speak in favor of free trade agreements while churning out large forecasts of their future benefits, usually in reports dense with obscure technical jargon. But the economics profession is evolving. They are, somewhat belatedly, beginning to see they are in danger of being seen as out of touch, not just by voters and the public, but by the politicians who hold the power over the thousands of federal and state official and advisory jobs that many economists covet.
Until recently, nearly all economists’ evaluations of trade agreements were produced with a specific type of trade model, known as a computer-generated equilibrium (CGE) model. This type of model starts with data from one year and generates future scenarios following from implementing a trade agreement. This type of model usually focuses on only one effect of a trade agreement, the cut in prices that follows a reduction of tariffs or elimination of non-tariff barriers. In my view, it magnifies the growth effects of the price cut far beyond what is justified for the U.S. economy. It specifically excludes the negative effects of a trade agreement, such as unemployment, and it also excludes any other forces or trends in an economy, such as a growth strategy or currency manipulation practiced by a partner country. In short, it is a model custom-designed to produce exaggerated positive results from FTAs.
The new ITC study is based on a different kind of study, known as a gravity model. A gravity model looks at the ongoing trends not just in a single year but in many nations over many years. By comparing U.S. trade with trade agreement partners with U.S. trade with non-FTA trading partners, over many years, they are able to isolate and measure the effects of the FTAs.
Economists’ use of gravity models to evaluate FTAs date back about 15 years. One important study was published in 2011 by economists James Anderson and Yoto Yotov[3]. Anderson and Yotov found that the FTAs implemented worldwide in those years produced an increase in what they call global efficiency (roughly similar to GDP) of 0.62%, very close to the ITC estimate for the U.S. They found that FTAs tend to double the rate of growth of the volume of trade for the member countries, as compared to their trade with non-members.
But here is the crucial part: using a gravity model, Anderson and Yotov found that FTAs benefit smaller, poorer countries more than larger, richer ones. “NAFTA has benefited Mexican producers disproportionately, at the expense of producers in the other two partners,”[4] they wrote. And further: “The biggest winners from the integration of the 90s are producers from relatively small European and Latin American economies that signed FTAs with large trading partners.”[5]
Anderson and Yotov’s model shows tiny gains for the U.S. and Canada from NAFTA, “less than 0.1% each,” they say. But we have to remember that this study, like all of these models assumes no unemployment—in other words, they assume that all resources pushed out of one industry move immediately into other industries at the same pay levels. It’s very likely that factoring in real-world unemployment would change that tiny gain into a loss.
A recent study of the U.S.-South Korea KORUS trade deal reaches similar conclusions. Economists at USC and Ohio State analyzed KORUS in 2018 using actual data on both economies and found that “total gross output (sales revenue) is estimated to incur a net loss of $143 million. Moreover, 34 out of 57 sectors of the US economy are estimated to incur gross output losses. This includes three advanced manufacturing sectors estimated to incur gross output reduction in excess of $175 million each…These results indicate the continued shift in comparative advantage away from U.S. manufacturing with respect to rising economies such as that of South Korea.”[6]
This study found that the three U.S. sectors that suffered the greatest loss in revenue were machinery manufacturing, electronic equipment, and autos and parts. The study found that the U.S. would increase its GDP by $45 billion, which seems strange if total gross output falls. The explanation is that imports rise by $1.5 billion, more than offsetting the fall in domestic production. In this economic model, the balance of payments is not a constraint.
Additional Effects Not in the Models
These new studies, like the ITC study, find widely varying results for different sections of the population and different industries. They find that the net benefits are typically so small that given the margin of error, they could just as well be net losers for a national economy as net gainers. They also tend to find that smaller, poorer countries gain much more from FTAs than larger, richer countries. For example, the Anderson-Yotov study identifies some of the bigger gainers from trade agreements as: Bolivia, Tunisia, Hungary, Poland, and Bulgaria. This matches common sense: if you join a club where the other members are big and rich, you have a much better chance of selling your products to their consumers and becoming rich like them. The rich country, on the other hand, is likely to experience downward pressure on key variables like wages. Economists call this effect factor price equalization. Groucho Marx’s insight on the topic was simpler, and funnier: “I wouldn’t join any club that would have me as a member.”
We would add four additional points effects of FTAs that are not in economic models but which add to their negative impact:
First, the forecasting models do not include trade cheating or other politically-managed behaviors. Models look at how trade liberalization and market forces shift resources into more efficient uses, assuming each participant behaves objectively and allows the free market to operate. But many participants in trade agreements don’t do that. They sign an FTA but continue to use currency manipulation or non-tariff barriers to prevent the rise in their imports the model (and the partner) would predict. That harms the economy of any nation that naively plays by the rules.
Second, what economists call “efficiency” is not what the public normally thinks of as efficiency. For example, beginning in the 1980s, the U.S. began removing its restrictions on imports of textiles and apparel. Today, over 90% of our consumption of these goods is imported, from places like China, Bangladesh, and Central America. To an economist, those nations are more “efficient” because their costs are lower. However, their production processes are actually less efficient in the sense in which most people understand the term. They generally use older, less modern equipment, and less of it, because their production is based on cheaper labor. So what many Americans think of as a race to the bottom in terms of labor wages and standards shows up in an economic model as greater efficiency.
Thirdly, none of the economic modeling allows unemployment to follow from increased import penetration. At CPA, we have integrated unemployment in affected sectors in our modeling of KORUS[7] and a potential U.S.-U.K. trade agreement[8]. In both cases it leads to a reduction in GDP for the U.S. Economists defend their models as academic exercises focused on the long-term, i.e. after the unemployed find new jobs, but this is disingenuous because mounting evidence shows the so-called long term can take a decade of more. And politicians are quick to seize on the results of the positively-biased models as evidence of benefits of trade agreements or increased trade, without warning voters of the intervening years of unemployment.
Finally, and perhaps most importantly, trade modeling, like most trade economics, is a static discipline. It focuses on comparing two different states of the economy at a single point in time. It ignores how resource allocation might impact an economy’s long-term growth over years or decades. For any nation, long-term growth is far more important than the allocation of resources at any single moment. The above-cited Korean study provides a good example. If the KORUS agreement reduced the U.S presence in electronics, machinery and auto parts, and led to an increase in industries like retail, hospitality, and construction, that might benefit GDP in the short term, but in the long term, it reduces U.S. strength in industries almost certain to grow relatively faster over the next decade and more. Over the long term, the potential loss to Americans’ income and the loss of high-quality jobs far outweighs a short-lived gain in prosperity from reallocation.
Economists’ studies and their models are finally moving towards a more realistic assessment of trade agreements and that is to be welcomed.
[1] USITC, Economic Impact of Trade Agreements Implemented under Trade Authorities Procedures, 2021 Report, June 2021, Publication Number 5199, pg. 18.
[2] Baschuk, Bryce, Debate Flares over Whether U.S. Trade Deals Pay Off, Bloomberg News, July 20, 2021.
[3] Anderson, James E., Yotov, Yoto V., Terms of Trade and Global Efficiency Effects of Free Trade Agreements 1990-2002, April 2011, NBER Working Paper 17003.
[4] Ibid, pg. 25.
[5] Ibid, pg. 25.
[6] Wei, Dan, Chen Zhenhua, and Rose, Adam, Estimating Economic Impacts of the U.S.-South Korea Free Trade Agreement, March 2018, pg. 3.
[7] Ferry, Jeff, CPA Testimony on Economic Modeling of Trade Agreements, Oct. 6, 2020. Available here.
[8] Gopalakrishnan, Badri Narayanan and Ferry, Jeff and Mayoral, Amanda, The Economic Impact of Tariff Eliminations in a U.S.-U.K. Free Trade Agreement: A CGE Model with Worker Displacement (July 27, 2021). Available at SSRN: https://ssrn.com/abstract=3894566.
