2013-26266. Antidumping Proceedings: Announcement of Change in Department Practice for Respondent Selection in Antidumping Duty Proceedings and Conditional Review of the Nonmarket Economy Entity in NME Antidumping Duty Proceedings

AGENCY:

Import Administration, International Trade Administration, Department of Commerce.

ACTION:

Change in Practice to the Department's Respondent Selection in Certain Antidumping Duty Proceedings and Elimination of Conditional Review of the NME Entity.

SUMMARY:

The Department of Commerce (“Department”) is hereby refining its practice with respect to the methodology for respondent selection in certain antidumping (“AD”) proceedings. Specifically, the Department is making changes to its current practice in antidumping administrative reviews for (1) respondent selection; and (2) conditional review of the NME entity. Normally, the Department makes these types of changes to its practice in the context of its case proceedings, on a case-by-case basis.^[1] For these particular changes in practice, the Department sought comments in advance of making changes in practice. However, the Department expects to continue to consider, and make changes in practice, as necessary, in the context of its proceedings based upon comments from interested parties submitted in the course of such proceedings.^[2]

DATES:

Applicability date: The Department expects to apply these changes in practice in AD administrative reviews for which the notice of opportunity to request an administrative review is published on or after December 4, 2013.

FOR FURTHER INFORMATION CONTACT:

Shauna Biby, International Trade Analyst, Import Administration, U.S. Department of Commerce, at 202-482-4267.

SUPPLEMENTARY INFORMATION:

The Department is hereby refining its practice with respect to the methodology for respondent selection in certain AD proceedings. Specifically, the Department intends to select respondents by sampling where certain criteria are met in AD administrative reviews. Further, while considering issues related to respondent selection and sampling, the Department has also reconsidered its practice of “conditionally” reviewing the nonmarket economy (“NME”) entity. In an administrative review of an AD order, the Department's current practice is to consider the NME entity to be “conditionally” under review. This means that even absent a request for review of the entity, the entity will become subject to review if an exporter subject to the review does not demonstrate that it is separate from the entity, and the entity's entries will be potentially subject to a new cash deposit and assessment rate. The Department has determined to discontinue such conditional reviews. If interested parties wish to request a review of the entity, such a request must be made in accordance with the Department's regulations.

The Department notes that in June 2005, it requested and received comments on the timing of assessment instructions for AD orders involving NME cases.^[3] Many commenters expressed support for a practice that would not delay assessment instructions of certain entries based on the Department's conditional review of the NME entity.^[4] Although the Department did not revise its practice with respect to conditional review of the NME entity at that time, the Department's experience to date indicates that there is no ongoing benefit to be achieved in maintaining conditional review of the entity. Furthermore, by eliminating the practice of conditional review, the Department eliminates an unnecessary delay in liquidation.

The notice-and-comment requirements of the Administrative Procedures Act do not apply to interpretive rules, general statements of policy or procedure, or practice. 5 U.S.C. 553(b)(3)(A). Although the notice-and-comment requirements of the Administrative Procedure Act do not apply, the Department provided an opportunity for the public to comment on the Department's proposed refinement to respondent selection in a notice published on December 16, 2010; and for the public to comment on the Department's practice with respect to the timing of assessment instructions in NME cases in a notice published on June 21, 2005.

Sampling Methodology

Background

On December 16, 2010, the Department proposed a refinement to its practice regarding its methodology for respondent selection in AD proceedings.^[5] As explained in the Proposed Methodology, when the number of producers/exporters (“companies”) involved in an AD investigation or review is so large that the Department finds it impracticable to examine each company individually, the Department has the statutory authority to limit its examination to: (1) A sample of exporters, producers, or types of products that is statistically valid based on the information available to the administering authority at the time of selection, or (2) exporters and producers accounting for the largest volume of subject merchandise from the exporting country that can reasonably be examined.^[6] The Department has, to date, generally used the second option in proceedings in which limited examination has been necessary. One consequence of this is that companies under investigation or review with relatively small import volumes have effectively been excluded from individual examination. Over time, this creates a potential enforcement concern in AD administrative reviews because, as exporters accounting for smaller volumes of subject merchandise become aware that they are effectively excluded from individual examination by the Department's respondent selection methodology, they may decide to lower their prices as they recognize that their pricing behavior will not affect the AD rates assigned to them. Sampling such companies under section 777A(c)(2)(A) of the Tariff Act of 1930, as amended (the “Act”), is one way to address this enforcement concern.

The statute requires that the sample be “statistically valid.” ^[7] The Department has interpreted this as referring to the manner in which the Department selects respondents.^[8] Therefore, to ensure the statistical validity of samples, in the Proposed Methodology, the Department proposed employing a sampling technique that: (1) Is random; (2) is stratified; and (3) uses probability-proportional-to-size (“PPS”) samples. Random selection ensures that every company has a chance of being selected as a respondent and captures potential variability across the population. Stratification by import volume ensures the participation of companies with different ranges of import volumes in the review, which is key to addressing the enforcement concern identified above. Finally, PPS samples ensure that the probability of a company being chosen as a respondent is proportional to its share of imports in the respective stratum.

The Department's Sampling Methodology

In general, the Department will normally rely on sampling for respondent selection purposes in AD administrative reviews ^[9] when the following conditions are met: (1) There is a request by an interested party for the use of sampling to select respondents; (2) the Department has the resources to examine individually at least three companies for the segment; (3) the largest three companies (or more if the Department intends to select more than three respondents) by import volume of the subject merchandise under review account for normally no more than 50 percent of total volume; and (4) information obtained by or provided to the Department provides a reasonable basis to believe or suspect that the average export prices and/or dumping margins for the largest exporters differ from such information that would be associated with the remaining exporters.^[10]

Accuracy of the Sampling Method

Many of the commenters who oppose the proposed methodology focus on the issue of accuracy, and query how a small sample can be “statistically valid” within the meaning of the statute. However, in a previous proceeding, the Department explained that the phrase “statistically valid” in section 777A(c)(2)(A) of the Act refers to the manner or process by which the sample is taken, not the sample results.^[11] In that proceeding, the Department explained that “the phrase `statistically valid sample' was added to the statute in 1994 merely to conform the language of the statute with that of the World Trade Organization (“WTO”) AD Agreement (Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994), and is not different in substance from the phrase `generally recognized sampling techniques' used in the Act prior to the URAA.” ^[12] The Department determined that the “statistical validity” of the sample “refers only to the manner in which the respondents are selected, and not to the size of the sample under review.” ^[13]

Statistical Validity of the Department's Sampling Method

The statistical tools in the methodology described herein satisfy the requirements for statistical validity. The population average (mean) dumping margin of concern to the Department is the export trade-weighted average dumping margin across all firms (exporters under review). Because this trade-weighted average margin, in turn, is equivalent to the stratum-weighted average of the stratum means, the estimation of the population mean equates to estimation of the stratum means. Each stratum mean is estimated on the basis of a PPS-based sample mean,^[14] which accounts for the variance in trade shares across exporters in the stratum and is, therefore, an unbiased estimator of the stratum mean in the sense that there is no systematic error associated with repeated sampling. Without PPS sampling, the sample mean would be over-weighted toward smaller-exporter margins and a bias would result. PPS sampling removes this bias.

Finally, stratification of the sample population into appropriate size categories, e.g., small, medium and large-sized exporters by import volume, ensures a maximum degree of cross-sectional representation of the population in the sample.

Definition of Sampling Population

Currently, the Department generally chooses companies for individual examination based on import volumes reported in case-specific U.S. Customs and Border Protection (“CBP”) import data. It also assigns an AD rate to all other companies that are not selected for individual examination. The Department currently does not require any evidence of shipment from a non-selected company before making its respondent-selection decision. However, in the sampling context, the existence of shipments will be required in order to both define the population, and if the company is selected, establish a dumping margin for the company. Therefore, the Department will normally use CBP data as the basis for the volume of subject merchandise and expects to define the population from which to sample as all companies named in a review with shipments of subject merchandise.

In NME cases, only those exporters who receive a separate rate will be included in the sample population. Companies that do not receive a separate rate will not be subject to review pursuant to the elimination of the conditional review of the NME entity practice described below. Therefore, in order to establish the appropriate sample population at the time of the sampling selection, it is necessary for the Department to make its determinations regarding the separate rate status of the companies under review before the sample is determined. For the purpose of constructing the sample rate, the Department expects that companies' separate rate status will remain unchanged once the sample is determined.

Calculating and Assigning Sample Rates

After examination of selected respondents by the sampling method, the Department will need to assign a rate to all non-selected companies. To do so, the Department will calculate a “sample rate,” based upon an average of the rates for the selected respondents, weighted by the import share of their corresponding strata. The respondents selected for individual examination through the sampling process will receive their own rates; all companies in the sample population who were not selected for individual examination will receive the sample rate.

Implementation of Sampling Methodology

The Department expects to implement the sampling methodology in the context of its administrative reviews by providing interested parties with notice of the schedule for submissions related to sampling on a case-by-case basis. The Department is publishing concurrently with this notice a proposed rule to amend section 351.301 of its regulations, “Time limits for submission of factual information,” to implement procedural changes, as needed, with respect to submissions related to sampling in antidumping administrative reviews.

In sum, the rule proposes to require interested parties to submit requests for the Department to conduct sampling in antidumping duty administrative reviews together with their comments on CBP data within seven days following the release of the CBP data, unless otherwise specified. The rule proposes that the submission include: (1) A request that the Department conduct sampling; and (2) factual information ^[15] and comments on whether this factual information provides a reasonable basis to believe or suspect that the average export prices and/or dumping margins for the largest exporters differ from such information that would be associated with the remaining exporters. Under the proposed rule, if an interested party were to submit a request for the Department to conduct sampling, all other interested parties will then have a ten-day comment period and a five-day rebuttal period to comment on the sampling request.^[16]

Apart from the proposed rule, in cases in which the Department determines to sample for respondent selection, it expects to conduct the sampling following the conclusion of the 90-day period for withdrawal of requests for administrative reviews under 19 CFR 351.213(d)(1). In cases in which the Department decides to sample, the Department does not expect to exercise its discretion to extend the 90-day period for withdrawal of review requests.

Comments and Responses

The Department received 18 comments on the proposed use of sampling for selecting mandatory respondents. A summary of these comments are presented below and have been grouped by the issues raised in the submissions. The Department's response follows immediately after each comment.

Issue: Statutory and International Requirements, Including That of “Statistical Validity”

Some commenters generally support the increased use of sampling, with several commenters noting that the proposed methodology is consistent with statutory requirements. Citing the Statement of Administrative Action (“SAA”) and previous instances in which the Department has sampled, several commenters note that the Department is only required to use a methodology “designed to give representative results based on the facts known at the time of sampling.” Further, the Department must contend with limited time and resources and has the discretion under the law to devise an appropriate sampling methodology. Other commenters note that the Department should retain as much flexibility as possible, and should not confine itself to one sampling methodology for all cases and industries.

Other commenters raised a number of concerns with whether the proposed methodology meets the Department's statutory and international obligations. Further, these commenters generally questioned whether the proposed methodology is “statistically valid,” arguing that the Department must make some finding about the degree of precision it will require. Specifically, there is no reference to size or “precision” of the sample in the proposed methodology. Some commenters asserted that a “statistically valid sample” is a higher standard than a “generally recognized sampling technique.” Moreover, “statistically valid” must “include the key ideas of the size of the sample and the relationship of the sample to the whole.” The core problem, some commenters noted, is that, in most cases, the Department does not have the resources to investigate the large number of companies that would be required to make the sample statistically valid. These commenters generally note that sample size cannot be fixed at the start, but rather one determines sample size based on three factors: the number of companies whose behavior is being measured, the margin of error likely to result, and finally, the “confidence” level desired.^[17] These commenters assert that 90 or 95 percent is a typical confidence level. In sum, sample size must be large enough to permit a statistically valid inference. The statute therefore provides an alternative: Choose the largest exporters. This method, the commenters assert, will normally yield the most accurate and comprehensive results.

With respect to the Department's international obligations, one commenter submitted that any respondent selection practice must comply with the Antidumping Agreement (“ADA”) Article 9.3, under which a company's margin is linked to its behavior, stating further that the proposed sampling methodology lacks any such link. Further, the selection process must not produce results that deprive respondents of the right to revocation under Articles 11.1 and 11.3 of the ADA. Companies not selected as mandatory respondent have no opportunity to assert these rights.

The Department's response: The Department addresses the majority of these issues herein and otherwise will address any particular circumstances as they arise on a case-by-case basis. Specifically, the statute requires that the sample be “statistically valid.” The Department has interpreted this as referring to the manner in which the Department selects respondents and not to the size of the sample or precision of the sample results. Therefore, to ensure the statistical validity of samples, the Department will employ a sampling technique that: (1) Is random; (2) is stratified; and (3) uses PPS samples. Random selection ensures that every company has a chance of being selected as a respondent and captures potential variability across the population. Stratification by import volume ensures the participation of companies with different ranges of import volumes in the review, which is key to addressing the enforcement concerns identified herein. Finally, PPS samples ensure that the probability of a company being chosen as a respondent is proportional to its share of imports in the respective stratum. The Department intends to address any further comments on the statistical validity of its sampling methodology on a case-by-case basis as they arise. Finally, the Department will address any specific concerns with respect to revocation as they arise on a case-by-case basis.

Issue: Clarifying the Rationale for Increased Use of Sampling

Several commenters asserted that the Department failed to define the objective of its sampling proposal nor had it described or explained what benefits it perceives from sampling, for example, how sampling would advance any statutory or policy objective. Noting resource constraints, one commenter urged the Department to recall its authority under the Act to simplify and streamline procedures, including the use of averaging and statistically valid samples. Further, these commenters generally asserted that the Department should maintain its preference for selecting the largest exporters based on volume, which will result in “dumping margins that more accurately reflect the pricing of subject merchandise in the U.S.”

The Department's response: As noted herein, the Department has, to date, generally chosen the largest respondents in proceedings in which limited examination has been necessary. One consequence of this is that companies under review with relatively small import volumes have generally been effectively excluded from individual examination. This creates a potential enforcement concern in AD administrative reviews because, as exporters accounting for smaller volumes of subject merchandise become aware that they are effectively excluded from individual examination by the Department's respondent selection methodology, they may decide to lower their prices as they recognize that their pricing behavior will not impact the AD rates assigned to them. Sampling companies under section 777A(c)(2)(A) of the Act is one way to remedy this enforcement concern. Therefore, the Department is exercising its discretion to use sampling in its respondent selection procedures.

Issue: The Use of CBP Data and Other Issues Regarding Import Shares for Purposes of Defining the Sample Population

Several commenters also raised issues regarding the use of CBP data. These comments generally focused on those instances where CBP data may be problematic due to, for example, fraud, miscalculations, or multiple affiliations of sellers and resellers. Some commenters urged the Department to consider greater use of quantity and value (“Q&V”) questionnaires, while others also recognized that Q&V questionnaires are time-consuming and will probably lead the Department to an incomplete picture of the industry, especially in large industries.

Some commenters argued that the Department should exclude producers with statistically insignificant export volumes (for example, less than two percent). Such companies' sales may not be bona fide sales, and selecting such companies may result in a skewed sample. These companies should be excluded from the sample pool while still assigning them the sample rate from that review. One commenter further recommended establishing a rebuttable presumption that entries accounting for less than one percent of the import volume are not bona fide sales.

The Department's response: For the reasons explained herein, the Department intends to follow its current practice of relying upon CBP data. Consistent with that practice, the Department will consider any specific problems or issues identified concerning the reliability of CBP data on a case-by-case basis. The Department recognizes that the use of Q&V questionnaires is time-consuming and not always necessary and therefore intends to use them only where warranted, such as AD investigations in non-market economy countries.

With respect to the proposal to exclude producers based on low export volumes, at this time, the Department does not intend to implement a general rule to exclude any respondents based on sales volumes, especially in light of utilizing the PPS methodology, which ensures that any single respondent is not over-represented in the sample population, as implementing such a singular approach would be inappropriate in many cases. But, the Department will consider comments raised by interested parties on a case-by-case basis and make determinations based upon the facts and circumstances in each case. The Department will consider all information and allegations regarding specific CBP data and other sales volume issues on a case-by-case basis.

Issue: Stratification

Commenters questioned whether the Department should forgo stratification, define the strata based on different criteria than proposed, as well as consider defining the population (and probability of selection) by production, by import volume rather value, and by whether the respondents requested a review or whether respondents were named in a request for a review. One commenter argued that the Department has no factual basis for using size as a basis for stratification, which “must be based on some relationship between the criteria used or the strata and the variable being measured.” If the Department wishes to stratify, it must base strata on variables relevant to margins. One commenter proposed bifurcating the population into two groups: (1) Those respondents who requested a review of their own entries; and (2) respondents requested by the domestic parties. Under this novel methodology, the Department would stratify and sample the two populations separately, and assign rates to individual strata.

The Department's response: The Department intends to stratify on the basis of volume, as this best meets the policy intentions described above; namely, creating the potential for individual examination for some of those respondents under review that otherwise would not normally be selected. Where circumstances warrant, especially in light of the enforcement concerns described herein, the Department may consider other characteristics by which to stratify on a case-by-case basis.

Issue: Whether the Department Should Limit Sampling to Reviews

The Department also received comments regarding the use of sampling in investigations as well as whether sampling should be the “default” method for respondent selection. At least one commenter argued that the Department should use sampling as the “default” procedure for respondent selection in administrative reviews. However, given the complexities and short time frames of investigations, the commenter recommends that the Department should establish deadlines under which petitioners must request sampling in investigations, with “selecting the largest” as the default procedure in investigations. Other commenters suggest only allowing sampling in investigations when doing so is requested in the petition. Another group of comments recommended that choosing the largest should remain the Department's “default” procedure for respondent selection, given the issues to which sampling gives rise. Many commenters urged the Department to retain its discretion in choosing its respondent selection methodology as the facts warrant.

The Department's position: Section 777A(c)(2)(A) of the Act provides the Department with authority to employ samples in both AD investigations and administrative reviews. The methodology described herein, however, was developed for purposes of administrative reviews. In large part, the enforcement concerns raised herein are not as salient in the case of investigations, where there has been no previous expectation of participating in (or being excluded from) a proceeding. Accordingly, the Department intends to consider sampling when the criteria described above are met in administrative reviews. Requests for sampling in investigations, for example, may give rise to other concerns that the Department has not yet considered. Therefore, the Department will address other requests for sampling as they arise in specific proceedings.

Issue: Whether the Department Should Reconsider Certain Aspects of the Proposed Methodology

The Department also received comments on the methodology itself, with some commenters arguing that the Department should retain the discretion to sample when selecting only two respondents, and other commenters arguing that three respondents is insufficient to meet the statutory requirements with respect to sampling. Further, the Department also received comments on the initially proposed 75 percent threshold, i.e., the percentage of imports represented by the largest respondents.

One commenter noted that the Department should use this limitation (i.e., the threshold) when sampling in investigations, but not in reviews, since this will not address the issues sampling is intended to remedy in industries dominated by a few large exporters. Another commenter noted that the Department has not articulated any rational basis to reject the greater coverage of 75 percent in favor of the lower percentage of imports likely to be covered by a sample. Rather, the Department should be required to individually examine a number of respondents proportional to the number of respondents in the population.

The Department's response: For the reasons described in greater detail earlier in the preamble and for purposes of this notice, the Department has determined to consider sampling when it can select a minimum of three respondents to examine individually and when the three largest respondents (or more if the Department intends to select more than three respondents) by import volume of the subject merchandise under review account for normally no more than 50 percent of total volume. The Department considers 50 percent to be a reasonable threshold because in these circumstances the agency would be able to calculate specific dumping margins for the majority of imports during a period of review. However, when selecting the largest respondents does not allow the Department to calculate dumping margins for the majority of imports, and the Department has the resources to review at least three respondents, the Department may choose to sample in view of the enforcement concerns discussed herein.

Issue: Respondent Characteristics

Several commenters noted that the Department should clarify what information it will consider with respect to variations in the population. Further, while the proposed methodology does acknowledge that significant differences in the population may affect the decision to sample, it does not address how the Department will assess these differences. In this vein, another commenter contended that the comments the Department receives in the proposed 10-day deadline should be used by the Department not only to determine whether to sample, but also how to sample. Several commenters warned against relying on the information presented in the comments as the basis to avoid sampling.

The Department's response: In general, the Department may consider sampling for respondent selection purposes in AD administrative reviews when (among other conditions) information obtained by or submitted to the Department provides a reasonable basis to believe or suspect that the average export prices and/or dumping margins for the largest exporters differ from such information that would be associated with the remaining exporters. Such a fact pattern supports the existence of potentially significant enforcement concerns, as variation in the dumping behavior of the population gives rise to concerns that a non-random means of respondent selection may systematically exclude certain dumping behavior. The Department has identified several types of information that a party may submit, including: Company margins from previous segments of the proceeding; market and company pricing information; the nature and structure of the foreign industry in question, including cost structure and/or actual pricing data; and the U.S. Customs and Border Protection import entry database. The Department may consider other information on a case-by-case basis.

Issue: Timing

Several commenters contended that the Department should clarify that the clock for the 10-day comment period should start running when parties have all the information necessary to submit comments (i.e., after the deadline for seeking separate-rate status, no-shipment status, Q&V/CBP data is complete, etc.). The same commenters proposed establishing a 40-day deadline for submitting and clarifying no-shipment and separate-rate information, with a 10-day comment period following that.

One commenter proposed waiting to sample until the window for withdrawing review requests has expired (currently 90 days from initiation), while another commenter proposed amending 19 CFR 351.213(d)(1) to be 60 days from initiation or 15 days following the deadline for filing. However, these commenters also noted that the Department should retain discretion to adjust this deadline on a case-by-case basis, keeping the deadline at 90 days for cases where sampling is not employed.

The Department's response: The Department expects to clarify many of these timing issues by giving interested parties notice of the procedural requirements during the course of the particular proceeding, and will address any concerns as they arise on a case-by-case basis. In addition, the Department is promulgating an amendment to section 351.301 of its regulations to address procedures for submissions related to sampling in administrative reviews. With respect to withdrawal of review requests and its potential impact on the timing of sampling, in cases where the Department determines to employ sampling for respondent selection, it will conduct its sampling following the conclusion of the 90-day period for withdrawal of requests for administrative reviews under 19 CFR 351.213(d)(1). In cases where the Department decides to sample, the Department expects that it will not exercise its discretion to extend the 90-day period for withdrawal of review requests. In this way, the Department preserves the ability of firms to withdraw their review requests during the first 90 days of the review as required by section 351.213(d)(1) of its regulations, but also ensures that later withdrawals do not adversely impact the Department's ability to conduct its sampling in a timely manner given the time constraints for completion of administrative reviews.

Issue: Rate Assignment

One commenter maintained that the Department should assign each stratum's rate to the members of that stratum and should not average the rates together to calculate and assign a population-wide average rate; each stratum's rate is predictive of the behavior of members of that stratum, and averaging the rates together does not yield representative results for any member of the population.

The Department received a range of comments regarding the inclusion of adverse facts available (“AFA”), de minimis and zero rates in the sample rate, including that: (1) The Department should include all AFA, zero, and de minimis margins in the sample rate; (2) the Department should include AFA rates and exclude de minimis/zero rates; and (3) the Department should exclude all total AFA, zero, and de minimis margins, but should include margins based on partial AFA in the sample rate.

Several commenters submitted that the Department should use the weighted average of all calculated rates where there is at least one rate not based on AFA. Recognizing that there is no statutory directive when no calculated rates are available, this commenter noted that Court of International Trade and WTO precedent require the Department to “consider the significance” of zero and de minimis rates. However, these commenters and others further argued that international obligations are unambiguous with respect to this issue: AFA cannot be included in all-other or sample rates. Article 6.8 and Annex II list limited situations in which AFA may be applied, and that is only when a party does not cooperate.

The Department's response: As noted above, the aim of the sampling methodology is to obtain the population average (mean) dumping margin which is the trade-weighted average dumping margin across all firms under review. The Department considered the approaches suggested by the commenters, but found that the methodology described herein remains the most appropriate approach. The Department intends, however, to address any comments on how to assign rates on a case-by-case basis as they arise within a particular proceeding. Thus, in assigning all non-selected companies a rate, the Department will calculate a “sample rate,” based upon an average of the rates for all selected respondents, weighted by the import share of their corresponding strata. In line with the Department's practice heretofore, the Department will include all rates in the sample. Therefore, consistent with the statute, the Department will assign one rate to all respondents in the sample population that were not individually examined. The Department will address any further issues as they relate to the facts of specific proceedings on a case-by-case basis.

Issue: Replacement Respondents and the Use of Voluntary Respondents

Several commenters noted that the Department should address the potential need to replace a respondent. In such an event, one commenter suggested, the Department could rank all respondents in each stratum, and simply go down the list to replace a respondent. Alternatively, the Department can “re-run” the selection within that stratum. One commenter warned against “re-shuffling” the strata after a withdrawal, noting that the sample methodology need only be based on the facts known to the Department at the time of selection. Another commenter asserted that replacement of a respondent must be achieved through the PPS selection methodology in the affected stratum, “otherwise the sample will be skewed and any pretense of statistical validity will be further undermined.” It was also noted that, if the Department waits to sample until the population is set (after withdrawals and separate-rate applications), the issue of whether to replace respondents should not regularly occur. One commenter stated that inclusion of smaller companies increases the likelihood of non-cooperation and that the Department must increase the number of companies sampled in order to accommodate this eventuality. A number of commenters requested that the Department provide explicit guidelines for its selection of one or more additional mandatory respondents where a company initially selected does not cooperate.

With respect to voluntary respondents, several commenters contended that the Department should not alter its current voluntary respondent practice. Further, voluntary respondents should receive their own rates and those rates should not be used in the weighted average rate. At least one commenter contended that the Department should not allow for voluntary respondents when sampling, but stated that if any voluntary respondents are examined, those rates should not be included in the sample rate.

A number of commenters submitted that increasing opportunities for voluntary respondents provides a means to meet the Department's legal obligations, and that the Department's current policy of examining no voluntary responses whenever it has determined to limit the number of respondents ignores its own statute and international obligations. In general, these commenters urge the Department to encourage voluntary participation and be liberal in accepting voluntary respondents.

The Department's response: Prior to selecting its sample, the Department intends to establish the population from which to draw its sample by first accounting for withdrawals of requests for review and also the separate-rate status of respondents in NME cases. However, the exact replacement procedure, when replacement is considered, as well as whether the Department will accept any specific requests for individual-examination by voluntary respondents, will depend, as it must, on the facts of the specific case. In addition, the Department finds the comments, such as the impact of company size on the sample, to be speculative at this point, but will consider such comments raised by interested parties in the course of its proceedings on a case-by-case basis.

Review of the NME Entity

Background

While considering the many issues involved in sampling in administrative reviews, the Department determined that one of the issues that may impact the use of sampling in future segments is the Department's review of the NME entity in its administrative reviews. Specifically, in proceedings involving NME countries, the Department has a rebuttable presumption that the export activities of all companies within the country are subject to government control and, thus, imports from all companies should be assessed a single AD rate (i.e., the NME-entity rate).^[18] It is the Department's practice to assign this single rate to all exporters of merchandise in an NME country subject to an AD investigation or review unless an exporter can demonstrate that it is sufficiently independent in its export activities, on both a de jure and de facto basis, so as to be entitled to a “separate rate” (i.e., a dumping margin separate from the margin assigned to the NME entity). The Department analyzes each entity exporting the subject merchandise that applies for a separate rate under a test first articulated in Sparklers,^[19] and further developed in Silicon Carbide.^[20]

Exporters named in the initiation of an AD administrative review that do not establish that they are independent of government control are considered part of the NME entity. In such instances, it has been the Department's practice to consider the NME entity under review, even if no request for review was made specifically for the entity.^[21] Under this practice, the assessment rate for entries from exporters that are part of the NME entity is not determined until the final results of the review. Thus, the Department typically does not instruct CBP to liquidate entries for any exporters whose deposits were made at the rate of the NME entity pending the final results of the administrative review. As a result, importers with entries from exporters that are part of the NME entity, but that were not named in the initiation of the review, must nevertheless wait until the final results of review before final liquidation. However, in most cases, the assessment rate is not different from the cash deposit rate at the time of entry for such imports. Consequently, the Department's conditional review practice has resulted in the delayed liquidation (often over a year after the date of initiation) of NME entity entries, even though the NME entity rate is unlikely to change when the NME entity is under review.

Statement of Practice Regarding Review of the NME Entity

The Department will no longer consider the NME entity as an exporter conditionally subject to administrative reviews. Accordingly, the NME entity will not be under review unless the Department specifically receives a request for, or self-initiates, a review of the NME entity.^[22] In administrative reviews of AD orders from NME countries where a review of the NME entity has not been initiated, but where an individual exporter for which a review was initiated does not qualify for a separate rate, the Department will issue a final decision indicating that the company in question is part of the NME entity. However, in that situation, because no review of the NME entity was conducted, the NME entity's entries were not subject to the review and the rate for the-NME entity is not subject to change as a result of that review (although the rate for the individual exporter may change as a function of the finding that the exporter is part of the NME entity).

Following initiation of an administrative review when there is no review requested of the NME entity, the Department will instruct CBP to liquidate entries for all exporters not named in the initiation notice, including those that were suspended at the NME entity rate. This change in practice will eliminate the unnecessary delay in liquidation of entries from the NME entity.

Dated: September 30, 2013.

Paul Piquado,

Assistant Secretary for Import Administration.

Footnotes

1.  In the context of its proceedings, Commerce is entitled to change its practice and adopt a new administrative practice provided it explains the basis for the change, and the change is a reasonable interpretation of the statute. Saha Thai Steel Pipe Company v. United States, 635 F.3d 1335, 1341 (2011).

2.  In particular, under 19 U.S.C. 1677f-1(b), the authority to select “statistically valid samples rests exclusively with the administering authority.” Commerce must retain the ability to alter its sampling methodology in each case, as is clear from the above provision that Commerce “shall, to the greatest extent possible, consult with the exporters and producers regarding the method to be used to select exporters, producers, or types of products under this section.”

3.  See Timing of Assessment Instructions for Antidumping Duty Orders Involving Non-Market Economy Countries, 70 FR 35634 (June 21, 2005).

4.  See public comments received July 15, 2005, available at http://ia.ita.doc.gov/​download/​nme-assessment/​nme-assessment-timing.html.

5.  See Proposed Methodology for Respondent Selection in Antidumping Proceedings; Request for Comment, 75 FR 78678 (December 16, 2010) (“Proposed Methodology”).

6.  See sections 777A(c)(2)(A) and (B) of the Act.

7.  See section 777A(c)(2)(A) of the Act.

8.  See Brake Rotors From the People's Republic of China: Final Results and Partial Rescission of the 2004/2005 Administrative Review and Notice of Rescission of 2004/2005 New Shipper Review, 71 FR 66304 (November 14, 2006) and accompanying Issues and Decision Memorandum at Comment 1A (“Brake Rotors”).

9.  This sampling methodology has been developed for AD administrative reviews, not AD investigations, or countervailing duty investigations or reviews.

10.  This information may include for example: (1) Company margins from previous segments of the proceeding; (2) market and company pricing information; (3) the nature and structure of the foreign industry in question, including cost structure and/or actual pricing data; and (4) the U.S. Customs and Border Protection import entry database.

11.  See Brake Rotors, 77 FR 66304 and accompanying Issues and Decision Memorandum at Comment 1A.

12.  Id., (citing Statement of Administrative Action, H.R. Rep. No. 103-316, at 872 (1994)).

13.  Id.

14.  The sample mean is the arithmetic average of the data values in the sample. For a sample of ten numbers, the sample mean is (x₁ +x₂. . . .)/10. In the AD respondent sampling context, the sample mean for a stratum is the simple average of the dumping margins of the sampled respondents from the stratum.

15.  A detailed description of what this information may include is listed in footnote 10 under “The Department's Sampling Methodology” section of this Federal Register notice.

16.  In NME cases, parties must submit their separate rate applications or certifications no later than 60 days after the notices of initiation of the reviews are published, unless otherwise specified in the notices of initiation.

17.  “Confidence level” relates to the probability that a sample-based estimate falls within specified error limits of the estimated parameter value, and the range of values defined by an estimate plus or minus the specified error limit is a “confidence interval.”

18.  See 19 CFR 351.107(d) (providing that “in an antidumping proceeding involving imports from a nonmarket economy country, `rates' may consist of a single dumping margin applicable to all exporters and producers”).

19.  See Final Determination of Sales at Less Than Fair Value: Sparklers from the People's Republic of China, 56 FR 20588 (May 6, 1991) (“Sparklers”).

20.  See Final Determination of Sales at Less Than Fair Value: Silicon Carbide from the People's Republic of China, 59 FR 22585 (May 2, 1994) (“Silicon Carbide”).

21.  This practice was affirmed in Transcom, Inc., v. United States, 294 F.3d 1371 (Fed. Cir. 2002).

22.  In accordance with 19 CFR 351.213(b)(1), parties should specify that they are requesting a review of entries from exporters comprising the entity, and to the extent possible, include the names of such exporters in their request.