Tariffs by Fiat: The Widening Chasm between U.S. Antidumping Policy and the Rule of Law

But long before the row over Section 232, Commerce was routinely abusing the vast discretion afforded to it while administering the antidumping law. Oddly, Commerce is given the role of judge, jury, and executioner, and it also acts as a consultant to the domestic industries filing petitions—the complainants seeking duties on imports. Dozens of remands per year from U.S. courts instructing Commerce to change its assumptions, methodologies, or calculations speak to a routine abuse of this discretion.

In 2015, in keeping with the tradition that legislation to liberalize trade be considered in conjunction with legislation to make the trade laws more accessible to domestic industry, Congress changed the antidumping law anticipating the passage of legislation to implement the Trans‐​Pacific Partnership (TPP). Although the partnership never came to fruition, the changes to the antidumping law went forward, giving Commerce carte blanche to exercise discretion in a number of areas while remaining beyond the reach of the courts.

Among the changes was a new provision authorizing Commerce to adjust or recalculate a foreign company’s submitted cost data if it determines that “a particular market situation exists such that the cost of materials and fabrication or other processing of any kind does not accurately reflect the cost of production in the ordinary course of trade.” Under such circumstances, “the administering authority may use another calculation methodology under this subtitle or any other calculation methodology.”

The changes authorized by the Trade Preferences Extension Act of 2015 gave Commerce much broader discretion to decide whether, when, and to what extent it could reject, revise, and recalculate a respondent’s submitted cost data. As a rule of thumb, higher production costs mean higher antidumping duty rates. Alas, this is a huge problem Congress must fix.

By inhibiting commerce between foreign producers and U.S. customers, the antidumping law is a source of friction between the United States and its trade partners. Since most U.S. antidumping measures restrict imports of intermediate inputs, such as industrial chemicals and raw steel, the law also pits U.S. producers of these upstream commodities against their own customers: U.S. companies in industries that consume these inputs to produce their own outputs. Recent changes to the law exacerbate both sets of tensions.

In 2015, through the Trade Preferences Extension Act, Congress amended the antidumping law, weakening the statute’s fraying tether to the rule of law and making it more useful to industries seeking political leverage over foreign competitors and their U.S. customers. The changes were intended to increase the already‐​high likelihood that antidumping proceedings would result in affirmative findings of dumping and that the duty rates generated by the Department of Commerce would be higher (i.e., more restrictive of imports).

Among the changes sought and obtained by domestic protection‐​seeking industries and their lobbyists in Washington was a provision authorizing Commerce to adjust or recalculate a foreign company’s submitted cost data if it determines that “a particular market situation exists such that the cost of materials and fabrication or other processing of any kind does not accurately reflect the cost of production in the ordinary course of trade.”¹ Under such circumstances, “the administering authority may use another calculation methodology under this subtitle or any other calculation methodology [emphasis added].”²

Before this change, Commerce could adjust a respondent’s submitted costs only if those costs failed to reconcile to the company’s audited financial statements, the respondent was uncooperative or incapable of supplying the requested information, or those costs reflected transactions not made at arm’s length. The changes authorized by the Trade Preferences Extension Act gave Commerce much broader discretion to decide whether, when, and to what extent it could reject, revise, and recalculate a respondent’s submitted cost data. As a rule of thumb, higher production costs mean higher antidumping duty rates.

In an April 2017 administrative review of the antidumping order on oil country tubular goods (OCTG) from South Korea, Commerce made its first‐​ever affirmative finding of a particular market situation (PMS) under the new law, concluding that the production‐​cost data submitted by the South Korean respondents understated their true costs, warranting an upward adjustment to their reported costs, which resulted in significantly higher antidumping duty rates for U.S. importers of OCTG from South Korea.³

The details of this case and those of 11 subsequent antidumping cases in which a PMS was found support the conclusion that the new provision gives virtually limitless discretion to Commerce to act as it wishes, even if its actions are partial, subjective, or capricious, while remaining well within the law. When a law empowers the executive branch to take actions that are unjust and the judiciary’s standards of administrative deference preclude the possibility of remand or restitution, then it falls to Congress to fix what it broke.

The new PMS provision is a direct blow to the rule of law and a possible violation of the Constitution’s nondelegation doctrine. The standard that was established in the jurisprudence on the constitutionality of congressional delegation of authority is that Congress must provide “an intelligible principle” that serves to limit the executive branch’s exercise of discretion.⁴

The purpose of this analysis is not to argue that the PMS provision is unconstitutional but that it hastens a decades‐​long transformation of the antidumping law from a plausibly remedial tool into a convenient cover for brute‐​force protectionism. This outcome damages U.S. import‐​consuming businesses, U.S. consumers, U.S. exporters, foreign exporters, relations with trade partners, and the rule of law in the global trade system.

Observers of the world economy are familiar with the story of globalization in the second half of the twentieth century. The story usually begins with the post–World War II institutions established under U.S. leadership, including the General Agreement on Tariffs and Trade (GATT) in 1947. The 23 original contracting parties to the agreement were committed to establishing a rules‐​based trading system under which quotas, tariffs, and other barriers to trade would be reduced progressively and, if possible, eliminated. Over the course of five decades, through eight successful multilateral negotiating rounds, the GATT succeeded at deepening and widening the scope of trade liberalization, culminating in the creation of the World Trade Organization (WTO) in 1995.

In the nearly 60 years between the founding of the GATT in 1947 and the full implementation of the trade‐​liberalizing commitments made in the Uruguay Round (the last completed round) going into effect in 2005, tariff rates on industrial products in developed countries fell from an average of 40 percent to an average of 4 percent, and exports increased from $58 billion to $9 trillion.⁵ During that period, world gross domestic product increased more than sevenfold, from $3.9 trillion in 1948 to $28.1 trillion in 2000.⁶ But planted among the seeds that helped produce a flourishing global economy were those of protectionism. Less frequently mentioned are the domestic political tradeoffs that occurred—the deals that governments had to cut at home to secure the necessary support to ensure ratification of the trade agreements reached. In the United States, a pattern emerged whereby congressional support for the trade deals brought home by the president’s negotiators would be conditioned upon provisions ensuring that the government would have the tools it needed to assist workers and industries adversely affected by that trade liberalization.

The Trade Adjustment Assistance program is one example of this trend.⁷ The program established benefits for workers displaced or adversely affected by trade and was created in conjunction with the Trade Expansion Act of 1962, which authorized U.S. participation in the Kennedy Round of GATT negotiations. In subsequent years, when considering legislation to facilitate trade liberalization, Congress amended the program’s provisions to extend more benefits to more workers in more industries for longer periods of time. These tradeoffs—at least in principle—were efforts to apply the observation that trade liberalization creates more winners than losers and that its economic benefits exceed its costs. Effectively, the program was intended to compensate the losers from trade liberalization with some of the bounty that accrued to its winners.⁸

Yet the Trade Adjustment Assistance program wasn’t the only one borne of this logic. Those who followed this logic reasoned that since Congress was going to allow the president to negotiate tariff reductions and expose U.S. industries to more foreign competition, the executive branch should also be assured easy access to tools such as contingent tariffs to redress injurious imports. The aim was to work to lower tariffs for the general good but to ensure protection for domestic industries if (or arguably when) imports threatened their economic health.

At the founding of the GATT, U.S. negotiators insisted that language permitting the use of antidumping be included. A view widely held—or at least the excuse given—by them at the time was that antidumping rules made deeper trade liberalization possible because they assured Congress that fallback contingencies were available if general tariff cutting exposed domestic industries to too much competition. Without antidumping, the argument went, Congress might not support general tariff liberalization.

Evidence that Congress was serious about preserving and protecting the U.S. antidumping law emerged during debate in the Kennedy Round over the implementation of this legislation. During the round, modest changes to domestic antidumping admin­istration had been agreed upon. But Congress refused to accept the agreement’s disciplines on domestic antidumping laws and ultimately carved out those changes from being implemented in the legislation.

With the fate of the Tokyo Round (1973–1979) hanging in the balance, U.S. lawmakers—at the behest of certain domestic producers—passed the Trade Act of 1974, which expanded the definition of dumping and included rules permitting more‐​aggressive techniques for dumping‐​margin calculation. The new provisions for a cost test (which is described in greater detail below), more than any other change to the antidumping law before it, served to inflate dumping margins, which broadened the law’s appeal to import‐​competing industries. The Trade Act of 1974 also provided businesses a formal channel of access to anti­dumping and other trade policymaking by mandating the creation of a private‐​sector advisory committee system to ensure that producer viewpoints would be incorporated in the negotiation, monitoring, and implementation of GATT agreements and related policies.

The Trade Agreements Act of 1979, which implemented U.S. commitments agreed upon in the Tokyo Round, also made several important changes to U.S. antidumping administration, including the transfer of jurisdiction from the Department of Treasury to Commerce, which proved enormously consequential. Whereas the Treasury Department’s mission is to attend to the broader macroeconomic well‐​being of the United States, and its officials tend to regard import duties as a deadweight economic loss, Commerce sets its sights on promoting the narrow interests of domestic producers.

A study of the GATT and U.S. trade policy by the Congressional Budget Office included the following observations about the shift of antidumping jurisdiction from the Treasury Department to Commerce:

Additional legislative changes in 1984 that gave even stronger incentives to domestic industries to file more petitions weighed the antidumping scales even further in favor of domestic protection‐​seeking industries. The amendments were offered and agreed as part of a broader package of legislation that ultimately delegated authority to the president to negotiate what became the Uruguay Round (1986–1994).

In August 2002, with the Doha Round negotiations underway, Congress passed legislation extending fast‐​track trade promotion authority to the president. The legislation included warnings to negotiators against “weakening” the antidumping law, instructing the president to

As that brief history reveals, U.S. antidumping policy evolved to become more accessible and rewarding to U.S. import‐​competing industries. That the volume of antidumping filings and the frequency of affirmative rulings dramatically increased beginning in the 1980s suggests that domestic industry was responsive to those changes. With the definition of dumping expanded, domestic industry had been invited to play a greater role in antidumping policy formulation and oversight (through the advisory committee system) and the government agency most closely aligned with domestic producer interests was given oversight of the antidumping regime.

In 2015, a new trade promotion authority bill provided Congress another opportunity to amend the antidumping law. And, once again, the changes made the law more user‐​friendly for domestic industries filing petitions and more likely to generate affirmative findings of dumping and higher antidumping duty rates. To put these changes into perspective, a brief review of the basics of Commerce’s calculation methodology follows.¹¹

Under the antidumping law, dumping is defined as the sale of a commodity by a foreign company in the United States at a price that is less than “normal value.” For antidumping duties to be imposed, two major legal requirements must be satisfied: Commerce must find that imports are being dumped, and the International Trade Commission must find that dumped imports are causing or threatening injury to a domestic industry.

Typically, normal value is based on the price of the same or a similar product in a comparison market (normally the foreign producer’s “home” market). The magnitude of dumping (or the “dumping margin”) is calculated by subtracting the export price from normal value and dividing the difference by the export price.

Accordingly, if a foreign producer sells ball bearings for $10 per pound at home and for $8 per pound in the United States, its dumping margin is (10−8)/8, or 25 percent. If there are insufficient sales of the comparison product in the ordinary course of trade in the producer’s home market, then, typically, the producer’s sales in a third‐​country market are used to calculate normal value. If no such sales are available, then Commerce typically bases normal value on “constructed value,” which is calculated as the cost to produce the product sold in the United States plus allowances for expenses and profit.

But that straightforward‐​sounding exercise of comparing prices and calculating dumping margins is rife with subjective interference and methodological sleights of hand. Commerce maintains considerable discretion over various decisions that directly affect how the existence and magnitude of dumping is determined. These include which sales should be included in calculating average prices, what product models should be collapsed together and treated as a single product for purposes of calculating average prices, what expenses should be subtracted from gross prices before net prices are compared between markets, and how company‐​wide costs should be allocated to the subject merchandise. Commerce’s decisions about these questions, and many others, impact the outcomes.

Whether the foreign producer’s export price is compared to his home market price, his price in a third‐​country market, or constructed value, higher normal values translate into larger margins of dumping. This fact drives nearly all the arguments in an antidumping proceeding. Petitioners submit arguments to support the case for higher normal values and lower export prices, respondents argue the opposite, and Commerce usually agrees with the petitioners. Moreover, the methodologies and procedures adopted by the department have led many observers to question the agency’s impartiality. Some of those techniques and their margin‐​inflating effects have been documented in previous studies.¹²

One of the most egregious methodological distortions among Commerce’s antidumping protocols is what became known as the cost test. Introduced in the Trade Act of 1974 at the behest of the steel industry, the cost test was designed to eliminate from the calculation of the average home market price those sales made at prices lower than the full cost of production.¹³ Of course, when below‐​cost sales are eliminated, the result is that U.S. sales are compared with only the higher‐​priced (i.e., above‐​cost) home market sales.

What possible purpose—other than to generate higher dumping margins—could be served by excluding below‐​cost home market sales from normal value? In fact, the existence of below‐​cost sales in the home market demonstrates an absence of a sanctuary home market, which is supposed to be an island of artificially high prices and profits from which a foreign producer’s dumped export sales are subsidized. If home market sales at a loss are found in significant quantities, there must be no sanctuary market and thus no cash cow from which to cross‐​subsidize cheap export sales. But because of the cost test, it is precisely under these conditions that dumping margins become significantly higher than they otherwise would be.

The effect of the cost test on the dumping calculation can be dramatic. Consider a foreign widget producer who makes five sales of widgets in his home market (at prices of $1, $2, $3, $4, and $5) and fives sales of widgets in the U.S. market (at prices of $1, $2, $3, $4, and $5). Assuming one widget is sold in each transaction, the weighted‐​average price for a widget is $3 in both markets. The dumping margin for this comparison is zero. There is no price discrimination whatsoever.

But the cost test requires that the foreign producer’s home market sales made at prices below the full cost of production be eliminated before calculating the average home market price. The cost test has no impact on the eligibility of the foreign producer’s U.S. sales in calculating the average U.S. price, however.

Assuming that the unit cost of producing the widgets is $2.50, the home market sales made at prices of $1 and $2 would be eliminated, and the average home market price would then be calculated as $4. This generates a dumping margin of 33 percent despite the absence of price differences between markets. Empirically, the cost test is among the most significant causes of inflated dumping margins.

Of course, the cost test often eliminates from eligibility all home market sales of a given product. Under those circumstances, Commerce uses a constructed value to serve as normal value for comparison purposes. Constructed value is supposed to approximate the price that the product being sold in the U.S. market would have sold for in the home market, and it is calculated by adding an estimated amount for the selling expenses and profit made in the home market to the cost of producing the U.S. product. This process opens the door to more administrative mischief, which is reflected in the fact that constructed value comparisons almost always produce higher dumping margins than do comparisons based on prices.

In addition to the cost test, the use of constructed value, and many other methodological distortions, the practice known as zeroing is particularly onerous. Zeroing is a results‐​oriented method that serves to inflate dumping margins and has been the source of considerable controversy and legal dispute. It has been found to violate the World Trade Organization’s antidumping agreement on numerous occasions, but the United States has remained determined to carve out exceptions where it believes it is still entitled to zero. For example, Commerce continues to zero in situations where it finds evidence of targeted dumping (where there are patterns of price differences that would—allegedly—conceal dumping but for the practice of zeroing). It is worth noting that Commerce finds evidence of targeted dumping with greater frequency now that the condition must be found for it to zero.

Antidumping calculations typically involve dozens, hundreds, or thousands of price comparisons, which are used to generate a single margin of dumping for the subject merchandise. When individual sales are made at dumped prices (i.e., when the U.S. price is less than normal value), the magnitude of the difference between prices is given full weight. But when sales are made at nondumped prices (i.e., when the U.S. price is greater than normal value), the magnitude of the difference between prices is entirely disregarded. It is set to a value of zero. Accordingly, when calculating the overall weighted‐​average margin of dumping (the duty rate that will apply to imports) from the multitude of price comparisons, the ability of the value of the nondumped sales to offset the value of the dumped sales is foreclosed, resulting in an average dumping margin that is much greater than mathematics and logic permit.¹⁴

The standard methodological distortions described above don’t come close to exhausting the list of channels through which agency bias infects U.S. antidumping policy. This tendency toward subjectivity is in the agency’s DNA. According to its website, the mission of the International Trade Administration’s Enforcement and Compliance is to “safeguard and enhance the competitive strength of U.S. industries against unfair trade through the enforcement of U.S. AD and CVD trade laws.”¹⁵

When fulfilling that mission requires the agency to “conduct AD and CVD investigations and administrative reviews to determine if imports are being sold at less than fair value or benefitting from unfair subsidization” while also “counsel[ing] U.S. industries on how to petition the U.S. government to seek relief from injurious and unfairly traded imports,” there can be no question that the game is rigged.¹⁶

That the Commerce agency in question changed its name from the relatively innocuous Import Administration to the baton‐​wielding Enforcement and Compliance reveals its cavalier attitude toward even maintaining the appearance of objectivity. Indeed, the actual evidence of its bias is in abundance.

Despite being afforded enormous latitude to administer the law with bias, Commerce still manages to run afoul of those limits with great frequency. Hundreds of court rulings have found the agency’s administration of the law to have breached the bounds of legality. A Cato study on the subject identified a pro‐​petitioner bias in the agency’s decisionmaking, finding that the agency “routinely exploits gray areas in the law to favor the domestic interests that seek protection—and, according to the verdicts of U.S. courts, sometimes violates the law in the process.”¹⁷ In the 18‐​month period reviewed, the Import Administration was required to publish “19 antidumping redeterminations pursuant to court orders to revise its assumptions or calculations to become compliant with the law. In 14 of those redeterminations, the revised antidumping rates were lower than those originally calculated,” which indicates that Commerce’s original—and illegal—decisions came at the expense of foreign respondents.¹⁸

According to interviews with former congressional and Commerce staff conducted by the Office of Inspector General:

This [pro‐​steel industry] bias is illustrated by the actions of career Commerce Department officials through whom must pass all Department of Commerce [antidumping] determinations in steel cases. The members and staff of the congressional Steel Caucus meet with them regularly to discuss ongoing antidumping and countervailing duty proceedings pending before the Department of Commerce. At some of these meetings, these officials have shared advance draft investigation results with the congressional Steel Caucus well before they were announced in final form, allowing the Steel Caucus to “comment” on them. Time and again high‐​level officials within the agency have exerted pressure on lower level Department of Commerce staff conducting investigations of foreign steel producers to rerun calculations and alter methodologies, resulting in increased AD/CVD tariffs.¹⁹

This kind of collaboration between DOC officials and steel industry representatives in Congress and on K Street is a feature of the U.S. antidumping regime. That closeness explains many of the most important developments in antidumping policy, including the statutory changes that produced the new PMS rules.

In June 2015, the Obama administration was wrapping up negotiation of the Trans‐​Pacific Partnership and Congress was debating legislation to grant trade promotion authority to the president. In keeping with tradition, Congress was simultaneously considering legislation to protect industries and workers expected to be adversely impacted by the eventual trade liberalization. The PMS provisions at issue were borne of the Trade Preferences Extension Act, which became law in conjunction with the Bipartisan Congressional Trade Priorities and Accountability Act (i.e., the trade promotion authority bill). Congress also passed the Trade Adjustment Assistance Reauthorization Act of 2015 updating the Trade Adjustment Assistance program.

The steel industry, including its arsenal of lawyers and lobbyists, had long been pushing legislation to make it even easier for industries to bring antidumping cases and win relief. They sought rule changes to give Commerce more flexibility to be creative with its dumping‐​margin calculations and more wiggle room to exercise its discretion without running afoul of the courts. The promiscuously worded PMS provision ticked all those boxes.

To appreciate that “reform” was exclusively about further weaponizing the antidumping law, it is worth noting some of the other changes made by Congress. The legislation gave Commerce authority to reject a respondent’s submitted data in favor of using “adverse facts” available while excusing the agency from having to corroborate that the surrogates chosen as adverse facts were reasonable. The amendments provide Commerce discretion in applying adverse facts available when foreign respondents are determined to be uncooperative. Under those circumstances, it may assign a dumping margin from any prior proceeding under the same antidumping order, including the highest such rate or margin, and is relinquished of any obligation to corroborate the accuracy of the adverse facts available used.

Ostensibly to reduce time and resource burdens, another change gave Commerce license to limit the number of foreign companies or “respondents” it investigates or reviews in each case. Those limits meant that most companies wishing to defend themselves against dumping allegations with their own company‐​specific data and arguments were now going to be denied that opportunity, relegating them to the arbitrary fate of the “all others” rate. That rate is usually calculated as the average of the rates of the companies that received their own rates (excluding those that are de minimis or based on adverse facts available).

The legislation also included language forbidding the International Trade Commission—the agency tasked with determining whether the domestic industry is materially injured or threatened with material injury by reason of dumped imports—from concluding that if an industry is profitable, it is not suffering injury. Of course, profitability is perhaps the strongest indication that an industry is “not materially injured,” but its evidentiary value was to be discounted under the new provisions.

Most disconcerting of the new provisions was the language that bestowed unbridled authority upon Commerce. Section 773(e) of the Trade Preferences Extension Act reads:

“[I]f a particular market situation exists such that the cost of materials and fabrication or other processing of any kind does not accurately reflect the cost of production in the ordinary course of trade, the administering authority may use another calculation methodology under this subtitle or any other calculation methodology [emphasis added].”²⁰

In other words, whenever Commerce determines that a PMS exists, it may reject a respondent’s actual cost data and use whatever data it wants to use instead. The statute plainly states that the administering authority can use “any” calculation methodology, which is the kind of latitude that invites mischief that cannot be disciplined by the courts.²¹

Before the Trade Preferences Extension Act, the concept of a PMS was described in Section 773(a) of the Tariff Act of 1930 as a market situation in the exporting country that “does not permit a proper comparison with the export price or constructed export price.”²² Under those conditions, which might have included the absence of a viable home market for the “like product,” the law directed Commerce to choose a viable third‐​country home market to serve as the comparison market. Absent a viable third‐​country home market, it was instructed to rely on the respondent’s cost data to calculate a constructed value estimate of what those home market prices should be.

The Trade Preference Extension Act changed the statute to no longer require Commerce to examine third‐​country home markets if it determines that a PMS exists in the home market. Under the revised law, the agency is instructed to go straight to its cost‐​based, constructed value approach and, having found a PMS, is empowered to use any calculation methodology to construct normal value.

According to trade lawyers with Sidley Austin, the 2015 amendments were enacted

with the explicit intent to increase the likelihood of affirmative injury determinations by the International Trade Commission and to grant the Department of Commerce greater discretion to augment the dumping margins and subsidy rates applied to foreign manufacturers and their U.S. importers. The latter, in turn, is likely to cause the reviewing courts to grant greater deference when evaluating the lawfulness of the Department’s determinations.²³

When it became law in 2015, the PMS provision was expected to be used as the basis for a new methodology to estimate dumping margins in cases involving China. According to the terms of China’s WTO accession protocol, the exceptional nonmarket economy methodology Commerce uses to calculate dumping margins in cases involving China was supposed to end on December 11, 2016, and a new methodology to deal with the Chinese economy’s unique features was presumed necessary. But the United States opted not to change its nonmarket economy approach to China, and Commerce chose to deploy its shiny new PMS provisions for the first time in a case involving oil country tubular goods from South Korea.

The details of this case make it abundantly clear that Commerce exercised its broad discretion to reject the respondents’ submitted production‐​cost data not because there was a PMS that distorted comparisons of export prices and normal value but because inflating the prices of South Korean OCTG in the United States with high antidumping duties was the outcome desired by the domestic steel industry, steel‐​state senators, and the White House.²⁴

But the evidence that politics intervened at the expense of South Korean OCTG producers and their U.S. customers begins with the original antidumping investigation in 2014.²⁵ In its preliminary determination, Commerce found no evidence of dumping by the South Korean producers. The case should have and would have been closed but for the fact that the steel industry lobby and certain members of Congress were unhappy about the outcome. In the months between the preliminary and final determinations, it was pressured to redo its analysis, reconsider facts on the record, and do whatever was necessary to ensnare the South Koreans in an antidumping order.²⁶

Commerce bowed to that pressure and reversed its decision to reject the petitioners’ proposed surrogate values for profit. In the preliminary determination, the department found those profit estimates to represent “neither production nor sales in the market under consideration,” that the estimates were “based on a research paper prepared by a student at the University of Iowa School of Management,” and that the paper included “a disclaimer statement regarding its accuracy.”²⁷ Nevertheless, Commerce found its way clear to ignoring those obviously disqualifying characteristics for the final determination and accepted and applied as factual the student’s estimate of 26.11 percent profits, which magically produced affirmative dumping margins ranging from 9.89–15.75 percent.²⁸

Had Commerce not submitted to that political pressure, South Korean OCTG producers would not have been subject to the antidumping order. Now ensnared under the order, they were set up for the injustices visited upon them in the first administrative review, when Commerce—once again under pressure from Congress and the White House—was compelled to invoke the PMS provisions.

In the review, petitioners based their claims of the existence of a PMS on four factors:

1. South Korean imports of Chinese hot‐​rolled steel were distorting OCTG costs.
2. The South Korean government was subsidizing domestic production of hot‐​rolled steel.
3. Strategic alliances between hot‐​rolled coil suppliers in South Korea and OCTG producers were distorting the cost of hot‐​rolled steel.
4. The South Korean government’s involvement was impacting South Korean electricity pricing.

In its preliminary determination, Commerce rejected the argument of the petitioners and found relatively low dumping margins for the South Korean producers. But once word of this outcome reached Capitol Hill and the White House, the agency once again came under political pressure to drop the hammer, culminating in an email from Peter Navarro, director of the Office of Trade and Manufacturing Policy for the president, to Commerce Secretary Wilbur Ross suggesting that antidumping duty rates in the neighborhood of 36 percent were appropriate and imploring Ross to get creative with the new PMS provision.²⁹

Once again Commerce submitted to that pressure and changed its analysis for the final determination, finding the existence of a PMS in the South Korean market. The justification for the reversed conclusion, despite no new information being placed on the record, was that for the preliminary determination Commerce had considered the four factors and had concluded that none individually constituted a PMS. However, in the final determination, Commerce noted that the statute was silent on whether the factors causing a PMS should be considered individually or collectively, so it decided that the four factors in aggregate constituted a PMS in the South Korean market. And, just like that, much higher dumping margins were calculated for more than 50 South Korean producers. The individual rate for the company NEXTEEL increased from 8.04–24.92 percent, and the “all others” rate (that applying to all but one of more than 50 other South Korean producers) increased from 5.92–13.84 percent.

In deciding how to adjust respondents’ costs to account for the PMS, Commerce noted that of the four contributing factors, the only element that it could estimate was the South Korean government’s subsidies to hot‐​rolled steel producers. A U.S. countervailing duty order against South Korean hot‐​rolled steel was already in effect, so Commerce used the highest company‐​specific subsidy rate found in the original investigation—a rate of 57 percent (which was based on a finding of adverse facts available)—as the basis for the upward adjustment to OCTG producers’ costs for the hot‐​rolled steel input.

When the problems with the PMS finding and the choice of calculation adjustment were raised by respondents, Commerce’s response was twofold. First, it asserted that the law allows it to do whatever it wants; therefore, the approach it took was reasonable. Second, noting that this was the first case to include a PMS finding since the enactment of the Trade Preferences Extension Act, Commerce indicated that it would continue to develop the concepts and types of analysis necessary to address future PMS allegations.³⁰ In other words, the agency retained the finding because it could, not because it was the right thing to do.

Since April 2017, Commerce has found a PMS in the final determinations of 12 cases, spanning 19 original investigations and admin­istrative reviews.³¹ Ten of these cases concern products made from hot‐​rolled steel. The common thread in those 10 cases is that they identify excess Chinese steel production as a factor warranting a finding of a PMS. In the 6 cases involving South Korea, all four conditions considered in aggregate to warrant a PMS finding in OCTG were cited. In the antidumping order on corrosion‐​resistant steel, a fifth condition—the purchases by South Korean corrosion‐​resistant steel producers of allegedly underpriced Japanese hot‐​rolled steel—was determined to be contributing to a PMS.

This raises some threshold questions. How can there be a PMS, which enables Commerce to depart from using respondents’ submitted cost data, when the conditions cited as indicative of that PMS are present in multiple markets? That would seem to suggest a “general” market, not a “particular” market situation. Chinese excess steel production capacity, which affects the global market, cannot be considered a PMS. That Commerce cites the effects of this overcapacity on the markets in South Korea, Thailand, India, and Turkey suggests that the market situation is distinctly not particular. How can a worldwide glut of steel be considered particular to any market?

Chinese overcapacity and excess production have impacts on global prices and global costs of producing finished steel products that consume hot‐​rolled steel in the production process and, by definition, cannot constitute a PMS. Moreover, Section 773(e) permits Commerce to use any other calculation methodology “if a particular market situation exists such that the cost of materials and fabrication or other processing of any kind does not accurately reflect the cost of production in the ordinary course of trade.”³² The findings of a PMS in these cases cover a period ranging from 2014 to 2018, which raises the question: Shouldn’t a market situation that persists over a five‐​year period be considered the “ordinary course of trade?”

The U.S. market was also affected by excess Chinese production, and the government offered a policy response. Instead of permitting U.S. OCTG producers (and the other downstream steel‐​using industries that sought antidumping protection) access to low‐​priced Chinese hot‐​rolled steel, which would have enabled them to better compete on price with OCTG producers, the United States chose to restrict imports of hot‐​rolled steel from China (as well as from virtually every other major source of that commodity). As of April 9, 2020, the United States has 19 active antidumping and countervailing duty orders on imports of hot‐​rolled steel from 14 countries and maintains 25 percent tariffs under Section 232 on imports from nearly every other country.

American trade restrictions caused U.S. prices for hot‐​rolled steel to rise well above the world market averages, and by artificially boosting the supply of hot‐​rolled coil outside the United States, those restrictions exacerbated the decline of world average prices, thereby conferring even greater cost advantages on foreign OCTG producers (and other pipe‐ and tube‐​product producers). When those disadvantages caught up to U.S. OCTG producers (and other pipe‐ and tube‐​product producers), tariffs were then imposed on imports of those downstream products.

In contrast, the South Korean government’s response to Chinese overproduction of this commodity was not to restrict imports of Chinese hot‐​rolled steel through antidumping or other trade barriers but to provide relief in the form of subsidies to domestic hot‐​rolled steel producers. The aim was to mitigate the unfair advantage accruing to Chinese hot‐​rolled steel producers without penalizing its higher value‐​added, downstream, steel‐​using industries. In the PMS cases, petitioners have argued and Commerce has agreed that the benefits of those subsidies have passed through to South Korean OCTG producers and other pipe and tube producers by way of lower hot‐​rolled coil costs.³³

First, it certainly seems reasonable to conclude that some of the value of those subsidies passed through. But the PMS adjustment was predicated on the assumption that the full value of those subsidies passed through, which defies logic. It supposes that all the subsidies went exclusively toward reducing the costs of producing hot‐​rolled steel. This is speculative, and, as excerpted from Commerce’s final review in the South Korean corrosion‐​resistant steel case, “any subsidization benefit logically would be retained by the recipient of the subsidy and not translated into direct one‐​for‐​one reductions in price.”³⁴

Second, if it is going to be determined that the tertiary effects of upstream subsidies render a producer’s costs or sales outside the ordinary course of trade in a downstream antidumping proceeding, then where should the line be drawn? Under such expansive definitions, virtually any government action, reaction, or inaction—even in response to an action of a foreign government—can be cited as deeming a company’s sales or costs to be outside the ordinary course of trade, giving Commerce license to declare a PMS and go off the reservation with whatever calculation methodology it chooses.

In the U.S. antidumping case, South Korean OCTG producers were made to pay a price for a decision they didn’t make. The South Korean government’s subsidies of domestic hot‐​rolled steel production do not warrant a finding of a PMS. To the extent that those subsidies conferred a benefit on OCTG producers, they are properly considered under the U.S. countervailing duty law. But it isn’t too hard to understand why the U.S. OCTG industry would prefer to seek redress for these subsidies under a law that now gives Commerce carte blanche to produce steep, punitive tariffs.

The third factor Commerce found relevant to its finding—when aggregated with the other factors considered—was the existence of so‐​called strategic alliances among producers in the production supply chain. While admitting to an absence of evidence on the record in the OCTG case in the final review, Commerce nevertheless assumed such alliances existed and were part of the problem:

Although the record does not contain specific evidence showing that strategic alliances directly created a distortion in HRC [hot‐​rolled coil] pricing in the current period of review, the Department nonetheless finds that these strategic alliances between certain Korean HRC producers and OCTG producers are relevant as an element of the Department’s analysis in that they may have created distortions in the prices of HRC in the past, and may continue to impact HRC pricing in a distortive manner during the instant POR [period of review] and in the future.³⁵

Once again, Commerce seems to have gone to great lengths to find a way to reach its conclusion about “strategic alliances.” There are other implements in the agency’s toolkit to redress situations where input prices are suspected of not being made at arm’s length. Commerce tests for formal affiliations between parties, and it also uses a major input rule—requiring the use of the higher of production cost, acquisition cost, or market price—when the arm’s-length nature of intermediate‐​goods acquisitions is in question. In other words, there is a more measured way to adjust for a possible distortion than to lump uncorroborated speculation about it into a case for a PMS, where Commerce’s actions are unrestrained and unpredictable.

The fourth consideration cited as a contributing PMS factor is the allegation of subsidized electricity rates for South Korean industry. Commerce mentions that the largest South Korean supplier of electricity is government controlled, but then—in a tone of admission that the impact is nevertheless negligible—notes: “To be clear, our determination of a particular market situation in this review is not based solely upon any support from the government of Korea for electricity.”³⁶

Again, the proper place to seek redress of subsidies is under the countervailing duty law. It turns out, however, that South Korean electricity prices have not been found to be countervailable in previous countervailing duty cases. It seems, then, that Commerce, using its new PMS power, may have found some way to scratch South Korean producers for an itch that has been beyond its remedial reach in the past.

In all the PMS cases involving steel‐
consuming products from South Korea, Thailand, India, and Turkey, Commerce accounted for the supposed market “distortion” by upwardly adjusting the respondents’ reported hot‐​rolled steel costs by an amount derived from estimates of the subsidy rates calculated in U.S. countervailing duty cases. The United States has active countervailing duty orders on hot‐​rolled steel from South Korea, India, and Thailand but not from Turkey. Although subsidy rates were calculated for Turkish exporters to the United States, the case ended in a finding that Turkish exports were not a cause of material injury. Commerce was comfortable using a derivative of the rates calculated for Turkey to adjust their hot‐​rolled steel input costs for PMS purposes nevertheless. In the cases involving biodiesel from Argentina and Indonesia, the primary reason for Commerce’s findings was the existence of export taxes on the product’s main input—soya in the case of Argentina and palm oil in the case of Indonesia. Those taxes were perceived by Commerce as a subsidy to domestic biodiesel producers, and the adjustments for the so‐​called distortions were derived primarily from the export tax rates.

Arguably, subsidies on upstream inputs reduce the costs of downstream production below the market average. But by how much? The answer depends on supply and demand elasticities in the supply chain. It depends on the supply of and demand for both the inputs and the final products. But in each PMS case, Commerce assumes that the downstream industry obtains all the benefit of the upstream subsidy, which is an assumption that does not hold up to scrutiny.

Moreover, if Commerce is going to account for the dispersed benefits of a foreign government’s public policies that happen to accrue to a group of producers in an eventual, tangential, and residual manner, then the agency should also weigh the effects of that foreign government’s policies—such as taxes or subsidies to producers of substitute products—that hurt producers in an eventual, tangential, and residual manner. Those tertiary costs should then be deducted from the tertiary benefits of the policy deemed to be contributing to a PMS. Instead, the assumptions employed in adjusting for a PMS all come at the expense of the foreign industry under investigation or review.

More recently, at the urging of petitioners in a few different cases—and because Commerce can use any calculation methodology to surmount the PMS problem—regression analysis has been used, ostensibly, to measure the impact of the underlying distortion on the costs of production.

In the case concerning corrosion‐​resistant steel from South Korea, Commerce relied upon a regression model designed by one of the domestic petitioners that sought to measure the impact of excess global steel production capacity on the average price of imported hot‐​rolled steel—the main input to the production of corrosion‐​resistant steel. While regression analysis is a useful tool for estimating the relationship between independent and dependent variables and for determining how much of a change in the dependent variable is explained by changes in the values of the independent variables, it is prone to misuse and abuse when in the wrong hands. In the current context, these aren’t statisticians testing hypotheses or scientists objectively interested in learning how changes in a factor x affect condition y. As a statistical model in the hands of Commerce, its underlying data are tailored by petitioners and replete with surrogate values alleged to be representing one variable or another, and the outcome of model sensitivity testing and the parameters used to determine significance levels is completely ignored. All this considered, regression analysis merely serves to put a veneer of legitimacy on a method that is only as valid as the underlying assumptions and representativeness of the data.

The slope of the relationship between global steel overcapacity and average import values for hot‐​rolled steel, as derived from a model with other independent variables (including aluminum, scrap, and iron ore prices and exchange rates) that may have little explanatory power and therefore inflate the beta for overcapacity, is a very crude metric for estimating the impact on South Korean hot‐​rolled steel input prices from the overcapacity that renders South Korea a PMS. But under the law, Commerce can use any calculation method to account for the PMS, so winning the argument that a specific approach is problematic or wrong, unfortunately, guarantees nothing.

Indeed, in 2020 petitioners in a case concerning cold‐​rolled steel from South Korea proposed that Commerce use benchmarks, as it does in nonmarket economy cases, to estimate the cost of inputs, such as hot‐​rolled steel.³⁷ In a world where hot‐​rolled steel prices are presumed to be affected by the overhang of excess global production capacity—after all, that is the predicate for the PMS finding in the first place—what benchmark, average, or surrogate would more accurately reflect the cost of hot‐​rolled steel than the actual acquisition or production cost of the South Korean cold‐​rolled producers?

Petitioners gave an answer. They recommended that Commerce use the U.S. steel industry’s average hot‐​rolled steel cost as a surrogate for the South Korean hot‐​rolled steel input costs.³⁸ Why? Because the U.S. cost for hot‐​rolled steel is about the highest in the world because of the wall of antidumping, countervailing, and Section 232 duties on imports from nearly every major producer in the world.

Given Commerce’s expansive statutory latitude and the growing distance between antidumping administration and the rule of law, anything is now possible.

Like the cost test, zeroing, and numerous other methodological sleights of hand employed in the process of “calculating” antidumping duties, a PMS finding opens the door to all sorts of discretionary mischief. Those who appreciate logic, mathematics, economics, objectivity, procedural fairness, and the rule of law should be appalled by the details of the U.S. antidumping regime.

The PMS provision is just the latest example of congressional surrender of its constitutional authority over tariff and trade policy to the executive branch. Notwithstanding this abdication and its consequences, the antidumping law continues to enjoy broad bipartisan support in Congress, where the sordid details revealed above are rarely given a passing thought. Instead, the law’s ongoing support is explained by its reputation as a tool that levels proverbial playing fields and protects upstanding American companies and their hardworking employees from predatory foreign firms.

But that is a fallacy. Typically, it is the law itself, not unfair trade, that creates an unlevel playing field. The law has become a commercial weapon used by U.S. companies against other U.S. companies. Antidumping has become a convenient channel through which domestic firms can saddle their competition (both foreign and domestic) with higher costs and their customers with fewer alternative sources while giving themselves room to raise their own prices, reap higher profits, and reinforce their market power.

Hopefully, this report will help stir an awakening to what has been happening and inspire Congress to fix what it broke.

Ikenson, Daniel J. “Tariffs by Fiat: The Widening Chasm between U.S. Antidumping Policy and the Rule of Law.” Policy Analysis No. 896, Cato Institute, Washington, DC, July 16, 2020. https://​doi​.org/​1​0​.​3​6​0​0​9​/​P​A.896.