Antidumping Methodologies: Market Economy Inputs, Expected Non-Market Economy Wages, Duty Drawback; and Request for Comments

AGENCY:

Import Administration, International Trade Administration, Department of Commerce.

ACTION:

Announcement of Change in Methodology, Request for Comment

SUMMARY:

This notice addresses three methodologies of the Department of Commerce (“the Department”) in antidumping proceedings. First, the Department is revising its approach concerning the use of market economy inputs in the calculation of normal value in antidumping proceedings involving non-market economy (“NME”) countries. Specifically, the Department is revising its approach concerning cases where an NME producer sources an input from both market economy suppliers and from within the NME. Second, the Department is revising its methodology for calculating expected NME wages in antidumping proceedings involving NME countries. Third, the Department is requesting comments on its approach concerning the calculation of duty drawback adjustments to export price in antidumping proceedings when a respondent producer obtains an input both from domestic and foreign sources. On this latter issue, the Department is seeking comments on the methodology that should be used when the producer receives duty drawback on certain exports containing the input but not on other exports containing the input.

FOR FURTHER INFORMATION CONTACT:

Lawrence Norton with regard to market economy inputs, Shauna Lee-Alaia with regard to expected NME wages, and John Kalitka with regard to duty drawback, Office of Policy, Import Administration, U.S. Department of Commerce, 14^th Street and Constitution Avenue, NW, Washington DC, 20230, 202-482-1579, 202-482-2793, or 202-482-2730, respectively.

SUPPLEMENTARY INFORMATION:

Issue One: Market Economy Inputs

Background

In antidumping proceedings involving NME countries, the Department calculates normal value by valuing the NME producer's factors of production, to the extent possible, using prices from a market economy that is at a comparable level of economic development and that is also a significant producer of comparable merchandise. The goal of this surrogate factor valuation is to use the “best available information” to determine normal value. See section 773(c)(1) of the Tariff Act of 1930, as amended (“the Act”); see also Shangdong Huraong General Corp. v. United States, 159 F. Supp. 2d 714, 719 (CIT 2001). When an NME producer purchases inputs from market economy suppliers and pays in a market economy currency, the Department normally uses the average actual price paid by the NME producer for these inputs to value the input in question, where possible. See 19 CFR 351.408(c)(1); see also Final Determination of Sales at Less Than Fair Value: Oscillating Fans and Ceiling Fans from the People's Republic of China, 56 FR 55271, 55274-75 (October 25, 1991). When a portion of the input is purchased from a market economy supplier and the remainder from a non-market economy supplier, the Department will normally use the price paid for the input sourced from market economy suppliers to value all of the input,¹ provided that the volume of the market economy input as a share of total purchases from all sources is “meaningful,” a term used in the Preamble to the Regulations but which is interpreted by the Department on a case-by-case basis. See Antidumping Duties; Countervailing Duties; Final Rule, 62 FR 27296, 27366 (May 19, 1997) (“Final Rule”); see also Shakeproof v. United States, 268 F.3d 1376, 1382 (Fed. Cir. 2001) (“Shakeproof”). Such market economy input purchases must also constitute arms-length, bona fide sales. See Shakeproof, 268 F.3d at 1382-83.

Additionally, the Department disregards market economy input purchases when there is evidence that the prices for such inputs may be distorted or when the facts of a particular case otherwise demonstrate that market economy input purchase prices are not the best available information. For example, the Department disregards all input values it has reason to believe or suspect might be dumped or subsidized. See, e.g., China National Machinery Import & Export Corporation v. United States, 293 F. Supp. 2d 1334 (CIT 2003), as aff'd per curiam 04 Fed. Appx. 183 (Federal Circuit, July 9, 2004). The Department has also disregarded the prices of inputs that could not possibly have been used in the production of subject merchandise during the period of investigation or review. See, e.g., Final Determination of Sales at Less Than Fair Value: Certain Frozen and Canned Warmwater Shrimp from the Socialist Republic of Vietnam, 69 FR 71005, and accompanying Issues and Decision Memorandum, at comment 8 (December 8, 2004) (“Shrimp”). The Department has further rejected purchase prices from market economies when the input in question was produced within an NME. See Final Determination of Sales at Less Than Fair Value: Polyethylene Retail Carrier Bags from the People's Republic of China, 69 FR 34125 and accompanying Issues and Decision Memorandum, at comment 4 (June 18, 2004).

The Department published on May 26, 2005, August 11, 2005, and March 21, 2006, three notices in the Federal Register requesting comment on its market economy inputs methodology in NME cases (70 FR 30418, 70 FR 46816, and 71 FR 14176, respectively). In these notices, the Department requested comment on various proposals concerning the Department's approach in cases in which NME firms purchase a portion of a given input from a market economy and source the remainder domestically. In such instances, the Department must make a case-specific determination as to what the best available information is for valuing the input: the market-economy purchase price or another surrogate value. The guidance given in the Department's regulations, as described above, is “normally” to use the prices paid for the market economy portion of the input to value the entire input. While the regulations do not elaborate as to what circumstances are “normal,” the Preamble states that the Department will disregard market economy purchases if the volume involved is not “meaningful.” In response to the Department's March 21, 2006 request for comment, the Department received comments in April 2006 from the following six interested parties: (1) the Committee to Support U.S. Trade Laws (“CSUSTL”); (2) the United States Steel Corporation (“U.S. Steel”); (3) the American Furniture Manufacturers Committee (“Furniture Committee”); (4) Stewart and Stewart; (5) the Ministry of Commerce of the People's Republic of China (“PRC MOFCOM”); and (6) Trade Pacific.

The Department requested comment on its market economy inputs practice for two reasons. First, the undefined nature of what constitutes a “meaningful” quantity of market economy purchases implies that the Department must currently make case-specific decisions as to whether to accept market economy purchase prices to value inputs. This creates unpredictability as to what values would ultimately be used in the dumping calculation. Parties can advocate accepting or disregarding the use of market economy purchase prices in individual cases, but do not have a concrete framework for doing so. Indeed, parties representing NME exporters have argued that market economy purchase prices nearly always constitute the “best available information” to use in the Department's dumping calculations, whereas parties representing domestic industry have argued that market economy purchases should almost never be used to value the portion of an input that was sourced domestically within the NME. This conflicting understanding as to when market economy purchases should be used to value an entire input is also evident in the submissions the Department received in response to its requests for comment on its market economy inputs approach. Absent an announced threshold as to what quantities are generally considered to be “meaningful,” parties would continue to argue this issue without the benefit of any clear guidance from the Department.

The Department's second reason for requesting comment on its market economy inputs approach was its concern that it may, in some cases, have used market economy input purchase prices to value an entire input even when these prices may not have been the “best available information.” While the Department has not had a specific threshold for what constitutes a “meaningful” quantity, the Department is concerned that accepting a market economy input value when the portion sourced from a market economy is too low may not constitute the best available information, particularly when no additional scrutiny is applied to ensure that the market economy price is representative of what the total price would have been had the firm purchased solely from market economy suppliers. This is a potential problem because the Department has greater confidence that the market economy purchase price is reflective of total purchase values of the input (and, thus, that it represents the “best available information”) when the proportion of the total volume of the input that is sourced from market economies is higher. To take an extreme example, where an NME exporter purchases all of a given input from a market economy supplier, the Department can be confident that this price reflects total purchase value of the input. Conversely, if an NME firm purchases a tiny quantity of the input from market economy suppliers and sources the rest domestically, the Department may have little or no confidence that this purchase price reflects the NME firm's overall purchases of the input. There might be numerous factors that could easily distort a single, small volume market economy purchase price, for example: sample sales, “bundling” of the purchase at a low price with other purchases at higher prices, limited quantities available on the market at an unusually low price, or brief plunges in the market price for the input. Of course, even a single purchase of an input might also, depending on the facts, be representative of what an NME exporter's purchases would have been had it sourced all of the input in question from the market economy source throughout the period of investigation or review. As a general rule, however, the Department typically rejects purchases of small quantities because “insignificant” quantities are less likely to be representative of a company's cost of sourcing the entire input. See Final Determination of Sales at Less Than Fair Value and Negative Final Determination of Critical Circumstances: Certain Color Television Receivers from the People's Republic of China, 69 FR 20594 and accompanying Issues and Decision Memorandum, at comment 12 (April 16, 2004).

This was the intended reasoning in the Preamble to the Regulations, which states that the Department “would not rely on the price paid by an NME producer to a market economy supplier if the quantity of the input purchased was insignificant. Because the amounts purchased from the market economy supplier must be meaningful, this requirement goes some way in addressing the commenter's concern that the NME producer may not be able to fulfill all of its needs at that price.” See Final Rule, 62 at 27366. By announcing a basic threshold of what constitutes such a “meaningful” quantity, and by making it high enough to reduce the chance of using a distorted price, without setting it too high to routinely prevent the use of market economy input prices, the Department can give greater effect to the intent of the regulations and improve its market economy inputs practice, to the benefit of all parties. This was the reasoning behind some of the proposals the Department put forward in its Federal Register notices soliciting comment on its methodology in this area. This decision, along with a discussion of the relevant public comments, is set forth below.

Statement of Policy

Drawing on the many submissions the Department has received in response to its requests for comment, the Department is now revising its methodology. While the Department may still consider amending its regulations to remove the regulatory requirement that the Department “normally” use market economy input prices to value the entire amount of such inputs, the Department is now establishing clearer guidance as to the circumstances in which it will accept market economy purchase prices to value an entire input. The Department is now instituting a rebuttable presumption that market economy input prices are the best available information for valuing an entire input when the total volume of the input purchased from all market economy sources during the period of investigation or review exceeds 33 percent of the total volume of the input purchased from all sources during the period. In these cases, unless case-specific facts provide adequate grounds to rebut the Department's presumption, the Department will use the weighted-average market economy purchase price to value the entire input. Alternatively, when the volume of an NME firm's purchases from market economy suppliers as a percentage of its total volume of purchases during the period of review is below 33 percent, but where these purchases are otherwise valid and meet the Department's existing conditions (described in the Background section above), the Department will weight-average the weighted-average market economy purchase price with an appropriate surrogate value according to their respective shares of the total volume of purchases, unless case-specific facts provide adequate grounds to rebut the presumption. In determining whether market economy purchases meet this 33 percent threshold, the Department will compare the volume that the producer purchased from market economy sources during the period of investigation or review with the respondent's total purchases during the period.² When a firm has made market economy input purchases that may have been dumped or subsidized, are not bona fide, or are otherwise not acceptable for use in a dumping calculation, the Department will exclude them from the numerator of the ratio to ensure a fair determination of whether valid market economy purchases meet the 33 percent threshold. This addresses the comment by Trade Pacific that the Department explain how it intends to calculate whether a given quantity of purchases meets the threshold, and ensures a fair comparison between acceptable market economy purchases and total purchases of the input during the period of investigation or review. Morever, because this 33 percent threshold constitutes a rebutable presumption, parties will have an opportunity to demonstrate that case-specific facts outweigh the presumption.

The practice described above is consistent with our current regulations directing the Department to “normally” use market economy input prices to value an entire input. While, as discussed above, the term “normally” is not defined in the regulations, it has been established in both the Preamble and through the Department's long-standing case precedent that the Department may decline to accept market economy purchases to value an input when the volume involved is insignificant. See, e.g., Preliminary Results of Administrative Review: Automotive Glass Windshields from China, 70 FR 24373, 24380 (May 9, 2005) (“Windshields”) (“{w}here the quantity of the input purchased from market-economy suppliers was insignificant, the Department will not rely on the price paid by an NME producer to a market-economy supplier because it cannot have confidence that a company could fulfill all its needs at that price.”). Windshields is representative of the Department's consistent standard that it will rely on market economy purchases to value an entire input only when the share of the input sourced from market economy suppliers, relative to the total volume purchased, is high enough that the Department has confidence that the market economy purchase price is reflective of the firm's total purchases of the input.

Accordingly, the Department's decision to introduce a flexible 33 percent threshold represents an extension of its previous practice. This standard of 33 percent is consistent with a threshold that the Department has defended, and the Court has upheld, as constituting a “meaningful” quantity in a prior case. See Shakeproof, 268 F.3d at 1382-83. However, the Department is now announcing what will generally constitute a “meaningful” or “significant” quantity, as opposed to making this determination on a strictly case-specific basis and without general guidance. Establishing a proportional, rather than absolute, threshold is also consistent with the logic described in Windshields, because the decision of whether to accept market economy input purchases to value an entire input rests on whether market economy purchases are reflective of what the total price would have been had the firm purchased solely from market economy suppliers.

Some commenters (including PRC MOFCOM) have argued that the Department's proposed policy statements provide solutions to what are only theoretical problems. These parties argue that even if “bundling,” price fluctuations or other factors that could distort market economy purchases exist, they have not been shown to be a problem in past cases and so there is no need for a remedy. The Department disagrees with this assertion. The Department cannot be privy to the circumstances governing every purchase of market economy inputs, nor can it be expected to conduct an analysis of each input market to see if given sales were representative of what the total price would have been had the firm purchased solely from market economy suppliers. Instead, the Department has always relied on the quantity of the input sourced from market economies as a proxy to gauge its relative confidence that the market economy purchase price is indeed reflective of the total volume of the input. The only difference is that the Department is now announcing a threshold, rather than making exclusively case-specific decisions.

On this point, PRC MOFCOM and others have argued that the Department should not establish a “bright line” threshold, that any threshold is arbitrary, and that the Department already has sufficient discretion to disregard market economy purchases that are not legitimate or bona fide. As described above, however, the Department is not introducing a rigid, “bright line” threshold, but rather a threshold that is amenable to interpretation in the light of case-specific facts and circumstances. Moreover, this threshold is not arbitrary, but is carefully crafted to balance two competing concerns; i.e. to ensure that market economy purchases are reflective of total purchases without contravening the regulatory requirement to “normally” accept market economy purchase prices to value an entire input when they are available.

In response to the Department's proposal to weight-average the market economy purchase price with a surrogate value when the share of market economy purchases falls below the Department's flexible threshold, PRC MOFCOM argued that there can be only one single source of the “best available information,” and if the market economy purchase price constitutes the best information for valuing the portion of the input sourced from market economy countries, it must also constitute the best information for valuing the entire input. The Department disagrees with this assertion, and considers that the “best available information” in cases in which a respondent purchases a given input both domestically and from market economy sources may be, depending on the circumstances, a weighted-average of a surrogate value and a market economy purchase price. The fact that a given price is valid for a (relatively small) portion of the input in question does not necessarily mean that it is representative of the firm's total purchases of the input. While market economy input purchase prices present a valid price for the market economy purchases that an NME firm actually made, and the Department will use these data, when possible, to value the portion of the input purchased from market economy sources, these prices may not always be the best available information for valuing the portion of the input produced within the NME. When the Department cannot be confident that this price is representative, however, if the price is otherwise valid (as in being bona fide, not subsidized, etc.), weight-averaging an appropriate surrogate value with the market economy purchase price would be the most accurate valuation of the input.

Other parties (including U.S. Steel, Stewart and Stewart, and CSUSTL) argue that except in rare cases, the Department should never accept market economy input purchases to value the portion of the input sourced domestically within the NME. Such a policy would contradict the applicable regulation, which clearly directs the Department to “normally” use market economy input purchases to value the entire input, even if the market economy purchases formed only a portion of an NME firm's total purchases of the input. The Department may consider a regulatory change in the future to grant it greater discretion in this area. Nevertheless, the Department disagrees with the assertion that market economy inputs never constitute the “best available information” just as it disagrees that these purchases always do so. Whether the best available information to value the NME-produced portion of the input is the price of the firm's market economy input purchases or another surrogate value is a decision that should guided by the relative shares of the two types of purchases, as well as by case-specific facts. U.S. Steel argues that “establishing a bright line threshold for market economy input purchases (i.e., more than 33 percent) would encourage respondents to manipulate the results so as to favorably affect the calculation of their dumping margins.” The Department does not agree that a change in respondents' behavior as a result of this policy, by itself, amounts to “manipulation.” Moreover, it is the Department's view that requiring parties, in most cases, to meet a 33 percent threshold actually reduces the opportunity for manipulation.

The Department's flexible percentage threshold of 33 percent for accepting market economy purchase prices to value an entire input will improve the predictability and accuracy of the Department's analysis, while continuing to meet the Department's regulatory requirement to “normally” use market economy purchases to value inputs when they are available. Predictability will be improved because parties will have a clearer idea of when the Department will accept market economy purchase prices to value an entire input. The Department will be able to calculate more accurate dumping margins, because the threshold sets a reasonable ratio of the market economy-sourced portion to that produced in the NME so that the Department can be more confident in the representativeness of the market economy purchase prices. However, this threshold is also not set so high that it would contradict the regulatory guidance on this issue. Finally, the fact that this threshold represents a rebuttable presumption means that it will be flexible, allowing the Department to take into account any case-specific facts that may arise.

The approach detailed above will take effect for all segments of NME proceedings that are initiated after publication of this notice in the Federal Register.

Issue Two: Expected NME Wages

Background

With regard to its calculation of expected NME wages, the Department stated in its November 17, 2004 final determination in the antidumping duty investigation of sales at less than fair value regarding Wooden Bedroom Furniture from the People's Republic of China, that it would “invite comments from the general public on this matter in a proceeding separate from the (Furniture) investigation.” Final Determination of Sales at Less Than Fair Value: Wooden Bedroom Furniture From the People's Republic of China, 69 FR 67313 and accompanying Issues and Decision Memorandum, at comment 23 (November 17, 2004). On June 30, 2005, the Department published a detailed description of its methodology for the calculation of expected NME wages and a request for comment. See Expected Non-Market Economy Wages: Request for Comment on Calculation Methodology, 70 FR 37761 (June 30, 2005) (“Wage Rate FR”). The Department received comments on August 1, 2005, from the following six interested parties: (1) CSUSTL; (2) Grunfeld, Desiderio, Lebowitz, Silverman & Klestadt (“Grunfeld”); (3) Lacquer Craft Manufacturing Company, Ltd.; (4) Dorbest Limited; (5) PRC MOFCOM; and (6) the Ministry of Trade of the Socialist Republic of Vietnam (“VN Ministry of Trade”).

The Department's expected NME wages are currently calculated each year in two steps. First, the relationship between hourly wage rates (obtained from the International Labor Organization's (“ILO”) Yearbook of Labour Statistics, relying on data that has been reported within the six-year period described below) and per-capita gross national income (“GNI”) (obtained from the World Bank) from market-economy countries (the “basket of countries”) is estimated using an ordinary least squares (“OLS”) regression analysis. Second, the GNI of each of the countries designated by the Department as an NME is applied to the regression, which yields an expected hourly wage rate for each NME. For further information, see Wage Rate FR.

PRC MOFCOM and the other interested parties (excluding CSUSTL) (“PRC MOFCOM et al.”) argued that when the Department is valuing any factor of production, including labor, the Department is obliged to use data from economically comparable countries and that the inclusion of countries not considered economically comparable is in contravention of our statute, citing 19 U.S.C. § 1677b(c)(4) and Antidumping Duties; Countervailing Duties, Part II, 62 FR 27296, 27367 (May 19, 1997) (“Final Rule”). Finally, PRC MOFCOM et al. asserted that the Department's original intention was to limit the regression analysis to economically comparable countries, citing Antidumping Duties; Countervailing Duties Part II, 61 FR 7308 (February 27, 1996) (“Proposed Rule”).

Accordingly, these parties proposed that the Department revert to its former practice of valuing direct labor using a surrogate wage rate from a surrogate country selected in each individual proceeding, or an average of the wage rates for the countries designated by the Department as economically comparable to the NME at the outset of each proceeding. Alternatively, some parties proposed that the Department should estimate the relationship between wage rates and per-capita GNI only for countries that are economically comparable to the NME country in question, defined by either the Import Administration's Office of Policy or by the World Bank's national income classifications. These parties asserted that the inclusion of non-comparable countries is both distortive and contrary to the Department's statutory directive to use “economically comparable” surrogate values.

Alternatively, acknowledging that the Department has a stated preference for more data when valuing labor, these parties proposed that the Department expand its basket of countries to include all countries for which the required data are available.

Finally, some parties argued that the Department should use a generalized least squares (“GLS”) methodology for its regression analysis in order to account for heteroscedasticity in the data set.

CSUSTL argued that the Department is required to value all factors of production for a given respondent, and must therefore capture all labor costs experienced by the respondent. Accordingly, CSUSTL proposed that the Department change its practice to rely on “labor cost” figures from Chapter 6 of the ILO's Yearbook of Labour Statistics or, failing that, that the Department should only use data from Chapter 5 that captures “employee earnings” rather than both earnings and wages. CSUSTL also noted that in order to capture all factors of production and other costs, the Department's calculation of surrogate financial ratios must be adjusted according to the labor cost elements that are included in the Department's expected NME wage rates.

Statement of Policy

Section 733(c) of the Act provides that the Department will value the factors of production in an NME using the best available information regarding the value of such factors in a market economy country or countries considered to be appropriate by the administering authority. The statute only requires that when valuing the factors of production, the Department utilize, to the extent possible, the prices or costs of factors of production in one or more market economy countries that are at a level of comparable economic development. See Section 733(c)(4) of the Act.

While surrogate values for other factors of production are selected from a single surrogate country, the Department determined in its Final Rule that it would be more accurate to base estimated labor values on data from many countries, stating that “more data is better than less data, and that averaging of multiple data points (or regression analysis) should lead to more accurate results in valuing any factor of production. However, it is only for labor that we have a relatively consistent and complete database covering many countries.” See Final Rule at 62 FR 27367.

Accordingly, section 351.408(c)(3) of the Department's regulations provides that:

For labor, the Secretary will use regression-based rates reflective of the observed relationship between wages and national income in market economy countries. The Secretary will calculate the wage rate to be applied in nonmarket economy proceedings each year. The calculation will be based on current data, and will be made available to the public.

19 CFR 351.408 (c)(3).

The Department's regulations concerning the valuation of labor were promulgated as part of a public notice and comment process. In the Proposed Rule the Department explained the benefits of a wage rate derived from a regression analysis, which include fairness and predictability. The Proposed Rule states:

Moreover, use of this average wage rate will contribute to both the fairness and the predictability of NME proceedings. By avoiding the variability in results depending on which economically comparable country happens to be selected as the surrogate, the results are much fairer to all parties. To enhance predictability, the average wage to be applied in any NME proceeding will be calculated by the Department each year, based on the most recently available data, and will be available to any interested party.

See Proposed Rule, at 7345.

PRC MOFCOM et. al.'s comment that the Department should abandon its regression-based calculation of expected NME wage rates in favor of the use of a single surrogate value for wage rates would contravene the Department's regulations, which direct the Department to use regression-based labor rates. In addition, as the Department noted in the Proposed Rule, while there is a strong positive correlation between wage rates and GNI, there is also variation in the wage rates of comparable market economies. For example, the Department's November 2005 regression illustrates that the observed hourly wage rates for market economy countries with national incomes below US$1,000 ranged from US$0.23 to US$0.94. See http://ia.ita.doc.gov/wages/03wages/110805-2003-Tables/03wages-110805.html . Therefore, if the Department adopted this suggestion in a proceeding involving an NME country with a GNI under US$1,000, values for labor might range from US$0.23 to US$0.94, depending on which economically comparable country is selected as the surrogate. See Proposed Rule at 7345. The Department is able to avoid this variability through the regression-based methodology for estimating wage rates due to the availability of reliable wage rate data and the consistent relationship over time between wage rates and GNI. The Department relies upon what is, in essence, an average wage rate, indexed to each NME's level of economic development via its GNI. Under the Department's regression methodology, the value for labor will be the same in every proceeding involving a given NME. This enhances the fairness and predictability of the Department's calculations.

Similarly, restricting the basket of countries to include only countries that are economically comparable to each NME is not feasible and would undermine the consistency and predictability of the Department's regression analysis. A basket of “economically comparable” countries could be extremely small. For example, there were five countries with GNI less than US$1,000 in the Department's 2005 calculation. A regression based on an extremely small basket of countries would be highly dependent on each and every data point. The inclusion or exclusion of any one country could have an extreme effect on the regression results. As described below, the Department screens the available data every year to ensure that they meet a number of important data suitability criteria. Therefore, the number and composition of the countries in the basket may vary unavoidably from year to year. A larger basket minimizes this potential for dramatic year-to-year variability.

Relative basket size would not be such a critical factor if there were a perfect correlation between GNI and wages. If this were the case, a precise regression line could be derived from suitable data from only two countries. However, while there is a strong world-wide relationship between wages and GNI (the r-square for the Department's 2005 calculation was .92, indicating an extremely strong relationship between GNI and wages), there is nevertheless variability in the data. For example, in the Department's 2005 calculation, observed wages did not increase in lockstep with increases in GNI in the five countries with GNI less than US$1,000: Pakistan, with a GNI of US$520, had reported a wage of US$0.38 per hour while Sri Lanka, with a GNI of US$930, had reported a wage of US$0.34 per hour. As stated above, a larger basket minimizes the effects of any single data point and, thereby, better captures the global relationship between wages and GNI. More data is, therefore, better than less data for the purposes of the Department's regression analysis, provided it is suitable and reliable data.

For this reason, consistent with the regulation and the statute, the Department's methodology relies on a significantly larger basket of countries. This maximizes the accuracy of the regression results, minimizes the effects of the potential year-to-year variability in the basket, and provides predictability and fairness. Importantly, the Department notes that economic comparability is established in the regression calculation through the GNI of the NME in question, which ensures that the result represents a wage rate for a country economically comparable to the NME.

With regard to the use of an alternative regression methodology, the Department notes that in its Proposed Rule, the Department explicitly stated that it would utilize an OLS regression analysis. See Proposed Rule, at 7345. OLS regression analysis is a commonly used analytical tool that is a basic component of any statistical analysis package. Like all statistical tools, the OLS analysis has certain limitations and cannot account for all characteristics of any given dataset, including heteroscedasticity. One of the assumptions of the OLS regression analysis is that the variance of the error terms is constant across observations. If the variance of the error terms is not constant, the error terms are considered heteroscedastic.

The data set upon which the Department bases its regression analysis changes on an annual basis. The Department does not consider it prudent, especially in light of its stated intention to use an OLS analysis, to decide on a year-by-year basis whether or not the level of heteroscedasticity in a given year's data would weigh in favor of using a GLS regression analysis. Instead, the OLS regression analysis allows the Department to rely on a simple, easily-duplicated methodology that enhances the fairness, predictability and transparency of the Department's antidumping duty calculations, while also ensuring their accuracy.

With regard to the CSUSTL comment that the Department should rely on “labor cost” figures from Chapter 6 of the ILO's Yearbook of Labour Statistics, the Department notes that the ILO defines data under “Chapter 5b: Wages in Manufacturing” as wages and bonuses, i.e., pre-tax monetary remuneration received by the employee. This is the data set that the Department relies upon in its calculations of expected NME wage rates.

The Department also notes that the ILO defines “earnings” under Chapter 5 of its Yearbook of Labour Statistics as being inclusive of “wages,” and as including both bonuses and gratuities. The Department agrees with CSUSTL that, in order to ensure that its calculation of expected NME wage rates accurately reflects the remuneration received by workers, it should rely on “earnings,” not “wages.”

Chapter 6 data, on the other hand, includes all costs to the producer related to labor including wages, benefits, housing, training, etc. As described below, the Department is already capturing as much of such labor costs as possible in its financial ratio calculations. The Department notes further that significantly fewer countries report Chapter 6 labor data than report Chapter 5b labor data. As of August 2006, 15 market economy countries had reported 2004 Chapter 6 data, while 65 market economy countries had reported 2004 Chapter 5b data. Chapter 6 therefore results in a significantly smaller basket of countries for which reliable data is available and may not accurately capture the global average of costs associated with labor.

The Department agrees with CSUSTL, however, that in order to ensure that labor costs not included in the ILO defined “earnings” are accounted for in its calculation of normal value, it is best to adjust, where possible, the surrogate financial ratios employed by the Department to value overhead expenses, selling, general and administrative (“SG&A”) expenses, and profit. Accordingly, it is the Department's practice to categorize all individually identifiable labor costs not included in the ILO's definition of “earnings” under Chapter 5 of the Yearbook of Labour Statistics as overhead expenses. See Folding Metal Tables and Chairs from the People's Republic of China: Final Results of Antidumping Duty Administrative Review, 71 FR 2905 (January 18, 2006) and accompanying Issues and Decision Memorandum, at comment 1. Such adjustments are fact-specific in nature and subject to available information on the record. Specifically, where warranted, individually identifiable labor costs in the surrogate financial statements which are not included in “earnings” are categorized as overhead or SG&A expenses for purposes of the Department's calculation of surrogate financial ratios.

Finally, the Department agrees that the basket of countries upon which the regression is based should be expanded to include all countries for which data are available in order to ensure accuracy and fairness. All such data must meet the Department's suitability requirements described below, which include contemporaneity and that the data cover both men and women and all reporting industries in the country.

Under its practice heretofore, the Department includes data from Chapter 5 of the ILO Yearbook of Labour Statistics that has been reported within five years of the Base Year, thereby considering a total of six years of data. (As described below in Attachment 1, the “Base Year” is the year upon which the regression data are based and is two years prior to the year in which the Department conducts its regression analysis.) In the course of reviewing its methodology, the Department has concluded that the inflation of data up to five years potentially reduces the accuracy of the calculation. Wage data that are potentially six years old may not represent the wage dynamics in labor markets today. The Department believes that, given the significant availability of more contemporaneous data, inflating old data is no longer necessary in order to achieve an acceptably large basket of countries. For example, over 50 countries reported suitable data within one year of 2003. The Department expects that the number of countries that meet the Department's suitability requirements will increase over time, as a greater number of countries report wage data to ILO in a reliable manner.

Therefore, in its revised methodology, the Department will only rely on ILO wage data that have been reported within one year prior to the Base Year, thereby considering a total of two years of data.

Revision of Methodology

Pursuant to the comments received and the Department's analysis thereof, effective for the 2006 calculation of expected NME wage rates, the Department will make the following revisions to its methodology:

1. The Department will only use earnings data reported in Chapter 5b of the ILO statistics.

2. The basket of countries upon which the wage regression is based will include data from all market economy countries that meet the criteria described below and that have been reported within 1 year prior to the Base Year.

3. Each year, the Department's annual calculation of expected NME wage rates will be subject to public notice prior to the adoption of the resulting expected NME wage rates for use in antidumping proceedings. Comment will be requested only with regard to potential clerical errors in the Department's calculation in light of its stated revised methodology.

Accordingly, the Department intends to publish its 2006 expected NME wage rates on its website in the autumn of 2006, together with a notice in the Federal Register requesting comment with regard to potential clerical errors in light of the revised methodology described below. The Department intends to finalize its calculations within one month thereafter.

The Department's methodology is described in full in below.

The Expected NME Wage Rate Methodology

The Department's regulations generally describe the methodology by which the Department calculates expected NME wages:

For labor, the Secretary will use regression-based wage rates reflective of the observed relationship between wages and national income in market economy countries. The Secretary will calculate the wage rate to be applied in non-market economy proceedings each year. The calculation will be based on current data, and will be made available to the public.

19 CFR 351.408 (c)(3).

In accordance with Section 351.408(c)(3), the Department annually calculates expected NME wages in two steps. First, the Department uses an ordinary least squares regression analysis to estimate a linear relationship between per-capita GNI and hourly wages in market economy (“ME”) countries. Second, the Department uses the results of the regression and NME GNI data to estimate hourly wage rates for NME countries.

There is usually a two-year interval between the current year and the most recent reporting year of the data required for this methodology due to the practices of the respective data sources. The Department bases its regression analysis on this most recent reporting year, which the Department refers to as the “Base Year.” For example, the Department relied upon data from 2001 to calculate expected NME wages in 2003, i.e., the “Base Year” for the 2003 calculation was 2001. In practice, the “Base Year,” i.e., the year upon which the regression data are based, is two years prior to the year in which the Department conducts its regression analysis.

1. Regression Analysis

The Department's regression analysis, which describes generally the relationship between wages and GNI, relies upon four distinct data series: (A) country-specific wage rate (earnings) data from Chapter 5B of the International Labor Organization's (“ILO”) Yearbook of Labour Statistics; (B) country-specific consumer price index (“CPI”) data from the International Financial Statistics of the International Monetary Fund (“IMF”); (C) exchange rate data from the IMF's International Financial Statistics; and (D) country-specific GNI data from the World Development Indicators of the World Bank (“WB”).

The wage rate data described above are converted to hourly wage rates and adjusted using CPI data to be representative of the current Base Year. The data are then converted to U.S. dollars using the appropriate exchange rate data. A regression analysis is ultimately run on these adjusted wage rate data and GNI. The following sections describe each data series and how it is used.

(A) Wage Data

For every country for which data is available and suitable (as described below), the Department chooses a single wage rate that represents a broad measure of wages for that country. The Department will choose data that is either contemporaneous with the Base Year or one year prior. Thus, the Department limits its selection of data to a two year period.

The ILO Chapter 5B database categorizes data under a number of parameters.³ The Department prioritizes these parameters in order to arrive at a single wage rate for each country representing the broadest possible measure of wages. As such, there are three criteria that all data must meet in order to be considered suitable for the Department's regression analysis.

First, under the category “Type of Data,” the Department will only use data that is reported in “earnings.”

Second, under the category “Sex,” the Department will only use data that cover both men and women.⁴

Third, under the category “Sub-Classification,” the Department will only use data that represent all reported industries. This is indicated in the database by a value of “Total” for the “Sub-Classification” parameter.

If there is more than one record in the ILO database that meet these three requirements, the Department will choose the data point from the Base Year over data from the prior year. At times, there is more than one data record in the ILO database that is both (1) reported in the same, most contemporaneous year and (2) meet the three required criteria above. In such cases, the Department chooses a single data point by prioritizing the following three parameters, described in greater detail below: (1) “Worker Coverage,” i.e., coverage of different types of workers; (2) “Type of Data,” i.e., the unit of time for which the wage is reported; and, (3) “Source ID,” i.e., a code for the source of the data.

For example, for the parameter “Worker Coverage,” the Department considers “wage earners” to be the best measurement for calculating expected NME wages and prioritizes such data over “employees,” “salaried employees” and “total employment,” in that order.

When the values for all parameters listed above are equal, the Department prioritizes data reported on an hourly basis over that reported on a daily, weekly and monthly basis, in that order, for the parameter “Type of Data.” Through this choice, the Department minimizes potential error due to converting daily, weekly or monthly wages to hourly wages.

When the values for all parameters listed above are equal, the Department prioritizes data classified under the International Standard Industrial Classification (ISIC) Revision 3 (ISIC Rev.3-D) over ISIC Revision 2 (ISIC Rev. 2-3). ISIC Rev. 3-D was revised in 1989 and is a more recent classification standard than the 1968 ISIC Rev. 2-3. See http://unstats.un.org/unsd/cr/family2.asp?Cl=2 and http://laborsta.ilo.org/applv8/data/isic2e.html.

Finally, when the values for all parameters listed above are equal, the Department prioritizes data with a “Source ID” value of “no value” over “1,” “2” and “3,” in that order.

The ILO data that are not reported on an hourly basis are converted to an hourly basis based on the premise that there are 8 working hours per day, 5.5 working days a week and 24 working days per month.

(B) CPI Data

Once hourly figures have been calculated based on the wage rate data discussed above, the wages are adjusted to the Base Year on the basis of the Consumer Price Index for each country, as reported by the IMF's International Financial Statistics. This adjustment is made for any wage rate data not reported for the Base Year.

(C) Exchange Rate Data

These inflation-adjusted wage data, which are denominated in each country's national currency, are then converted to U.S. dollars using Base Year period-average exchange rates reported by the IMF's International Financial Statistics.

Thus, using (A) wage data, (B) CPI data and (C) exchange rate data, discussed above, the Department arrives at hourly wages, denominated in U.S. dollars and adjusted for inflation for each country for which all the above data are available.

Finally, once the data have been converted to U.S. dollars per hour and adjusted for inflation, it is the Department's practice to eliminate values that could not possibly be reflective of actual wage levels or values that vary in either direction in the extreme from year to year (and which probably reflect errors in the original source data). For example, if a country is found to have average wage levels of US$0.01 per hour, the Department would eliminate that value as erroneous.

(D) GNI Data

The Department uses Base Year GNI data for each of the countries in the Department's analysis, as reported by the WB. GNI data are denominated in U.S. dollars current for the Base Year. The WB defines GNI per capita as equivalent to gross national product (“GNP”) per capita, which is “the dollar value of a country's final output of goods and services in a year divided by its population.”

The Department conducts its linear, ordinary least squares regression analysis using the Base Year wages per hour in U.S. dollars discussed above and Base Year GNI per capita in U.S. dollars to arrive at the following equation: Wage[i] = Y-intercept + X-coefficient * GNI. The X-coefficient describes the slope of the line estimated by the regression analysis, while the Y-intercept is the point on the Y-axis where the regression line intercepts the Y-axis. The results of this regression analysis describe generally the relationship between hourly wages and GNI.

2. Application of Regression Results to NME GNI Data

The Department applies the NME Base Year GNI to the equation presented above to arrive at an estimated wage rate for the NME. This is done for each NME.

Issue Three: Duty Drawback

Background

With respect to the duty drawback adjustment, the Department is directed by section 772(c)(1)(B) of the Act, which states that “{t}he price used to establish export price and constructed export price shall be -- (1) increased by … (B) the amount of any import duties imposed by the country of exportation which have been rebated, or which have not been collected, by reason of the exportation of the subject merchandise to the United States.”

Based upon this statutory language, the Department applies a two-prong test to determine entitlement to a duty drawback adjustment. That is, the party claiming such adjustment must establish that: (1) the import duty paid and the rebate payment are directly linked to, and dependent upon, one another (or the exemption from import duties is linked to exportation); and (2) there were sufficient imports of the imported raw material to account for the drawback received upon the exports of the manufactured product. See, e.g., Notice of Final Results of the Eleventh Administrative Review of the Antidumping Duty Order on Certain Corrosion-Resistant Carbon Steel Flat Products from the Republic of Korea, 71 FR 7513 (February 13, 2006) and accompanying Issues and Decision Memorandum, at comment 2 (“CORE from Korea”). Moreover, the courts have sustained the Department's traditional two-prong test. See, e.g., Wheatland Tube Company v. United States, 414 F. Supp. 2d 1271, 1287 (CIT 2006); Allied Tube & Conduit Corp. v. United States, 374 F. Supp. 2d 1257, 1261 (CIT 2005); Allied Tube & Conduit Corp. v. United States, 132 F. Supp. 2d 1087, 1093 (CIT 2001); Far East Machinery Co., Ltd. v. United States, 699 F. Supp. 309, 311 (CIT 1988); Carlisle Tire & Rubber Co. v. United States, 657 F. Supp. 1287, 1289-90 (CIT 1987).

The Department previously requested and received comments regarding its practice with respect to duty drawback adjustments to export price in antidumping proceedings. See Duty Drawback Practice in Antidumping Proceedings, 70 FR 37764 (June 30, 2005) and Duty Drawback Practice in Antidumping Proceedings, 70 FR 44563 (August 3, 2005). Among other things, the Department requested comments on the appropriate methodology to apply when duty drawback is claimed for some, but not all, exports incorporating the input in question. In past cases, certain parties have argued that the Department should allocate the total amount of relevant drawback received to total exports, regardless of destination, to ensure that the adjustment claimed on U.S. sales is not overstated. See, e.g., CORE from Korea, Issues and Decision Memorandum at comment 2.

Some parties argued, for example, for application of a “reasonableness” standard in this regard. They claim that, while an adjustment in the full amount of the duty drawback received should be made when the foreign producer can directly trace particular imported duty-paid inputs through the subsequent production process and into particular finished goods that are exported to the United States, this is an unlikely situation. Because it is more likely that exported goods may or may not actually have incorporated the imported input, a reasonable approach would involve allocating the drawback received to all exports that may have incorporated the duty-paid input in question. By doing so, these commenters claim, the Department would reasonably avoid excessive claims for drawback adjustments in antidumping calculations. These commenters further suggest that parties claiming favorable adjustments such as claims based upon duty drawback carry the burden of proof in this regard. See Statement of Administrative Action, H. Doc. 103-316, 103d Cong. 2d Sess., 829 (1994) (“{A}s with all adjustments which benefit a responding firm, the respondent must demonstrate the appropriateness of such adjustment.”).

The Department agrees with these commenters and proposes to modify its approach by limiting the duty drawback adjustment in certain circumstances. The Department generally agrees that it should allocate the total amount of duty drawback received across all exports that may have incorporated the duty-paid input in question, regardless of destination, to ensure that the adjustment claimed on U.S. sales is not overstated. Absent such a limitation, the Department is concerned that its current practice of permitting an adjustment to export price and constructed export price for all duty drawback received, whether or not it is related to U.S. sales, is an inappropriate application of its statutory authority to account for the effects of foreign drawback programs on price differentials between normal value and U.S. price. Furthermore, the Department is concerned that the adjustment could be manipulated by certain parties for purposes of obtaining a more favorable dumping margin. However, the Department will continue to permit a full adjustment for duty drawback received should the foreign producer claiming such adjustment demonstrate that it can directly trace the particular imported duty-paid inputs through the subsequent production process and into particular finished goods that are exported to the United States. The Department welcomes comment on this proposed methodology.

DEADLINE FOR SUBMISSION OF COMMENTS (on duty drawback): November 17, 2006.

Comments (Duty Drawback Issue Only)

Persons wishing to comment should file a signed original and six copies of each set of comments by the date specified above. The Department will consider all comments received before the close of the comment period. Comments received after the end of the comment period will be considered, if possible, but their consideration cannot be assured. The Department will not accept comments accompanied by a request that a part or all of the material be treated confidentially because of its business proprietary nature or for any other reason. The Department will return such comments and materials to the persons submitting the comments and will not consider them in development of any changes to its methodology. All comments responding to this notice will be a matter of public record and will be available for public inspection and copying at Import Administration's Central Records Unit, Room B-099, between the hours of 8:30 a.m. and 5 p.m. on business days. The Department requires that comments be submitted in written form. The Department recommends submission of comments in electronic form to accompany the required paper copies. Comments filed in electronic form should be submitted either by e-mail to the webmaster below, or on CD-ROM, as comments submitted on diskettes are likely to be damaged by postal radiation treatment.

Comments received in electronic form will be made available to the public in Portable Document Format (PDF) on the Internet at the Import Administration Web site at the following address: http://ia.ita.doc.gov/.

Any questions concerning file formatting, document conversion, access on the Internet, or other electronic filing issues should be addressed to Andrew Lee Beller, Import Administration Webmaster, at (202) 482-0866, e-mail address: webmaster-support@ita.doc.gov.