IMPORT ADMINISTRATION POLICY BULLETIN Number: 92/3 Date of Issue December 15, 1992 Topic: The 10/90/10 Test for Sales Below the Cost of Production. Author: Mark Lunn Approved (Signed) ALAN DUNN Assistant Secretary for Import Administration Issue Whether to apply the "10/90/10" test for sales below cost on a such or similar basis or a model specific basis. Background In 1974, Congress added section 773(b) to the Tariff Act of 1930 ("the Act") to provide in certain situations for the disregard of home market or third country sales if such sales are made at prices which represent less than the cost of production of the merchandise in question. Section 773(b) requires that: Whenever the administering authority has reasonable grounds to believe or suspect that sales in the home market of the country of exportation, or, as appropriate, to countries other than the United States, have been made at prices which represent less than the cost of producing the merchandise in question, it shall determine whether, in fact, such sales were made at less than the cost of producing the merchandise. If the administering authority determines that sales made at less than the cost of production -- (1) have been made over an extended period of time and in substantial quantities, and (2) are not at prices which permit recovery of all costs within a reasonable period of time in the normal course of trade, such sales shall be disregarded in the determination of foreign market value. Whenever sales are disregarded by virtue of having been made at less than the cost of production and the remaining sales, made at not less than the cost of production, are determined to be inadequate as a basis for the determination of foreign market value under subsection (a) of this section, the administering authority shall employ the constructed value of the merchandise to determine its foreign market value. Since 1974, the Administering Authority has applied a test commonly referred to as the "10/90/10 test" in determining whether substantial quantities of sales at less than the cost of production should be disregarded in determining FMV. An exception has been made for perishable agricultural products where the test has been 50/90/10 (i.e. sales will not be disregarded unless more than 50 percent of home market sales are found to be made at below cost). The question is, however, 10 or 90 percent of what? Over the years three major variations have arisen creating an inconsistent, unpredictable practice. Sometimes, the Department applies the 10/90/10 cost test over an entire such or similar category (all merchandise that can be reasonably compared under Section 771(16)). Other times, the Department has applied the test on a "model specific" basis (model being defined as the such or similar merchandise as defined under 771(16)). Finally, a hybrid approach where the test is first conducted on a "macro" or such or similar basis and then if more than 10 percent but less than 90 percent are below cost, the analyst would conduct a "micro" or model specific basis. The dumping margin can be significantly affected by the choice of methodology. This has been an issue for several years and was recently raised in the Remand of the less than fair value investigation of Mexican cement. In the Mexican cement investigation, the Petitioner alleged sales below cost of one type of cement (type []). The Department did not initiate the COP investigation stating: Because such sales of type [ ] cement comprise less than [ ] percent of the relevant 'such or similar' category, we would not disregard any of these sales for purposes of calculating foreign market value (FMV) even if we were to find all of these sales to be sold below cost. This [ ] percent volume falls far short of the quantitative threshold that we evaluate, among other factors, to determine whether to disregard below cost sales. See Timken Co. v. United States, 11 CIT 786 806-077, 673 F. Supp. 495, 514 (CIT 1987). Because we could never achieve the intended purpose of the COP statutory provisions in this case, i.e. the exclusion of below cost sales for FMV purposes, initiating a cost investigation under these circumstances would be a futile exercise. Letter from the Department to Respondent March 3, 1990 In the cement investigation, the analysis was conducted using the such or similar methodology. Sales of type [ ] cement were less than 10 percent of total home market sales. It would not have been logical, therefore, to initiate the investigation. However, had the analysis been done using the model specific methodology, the question would have been, has petitioner made a reasonable claim that a significant portion of type [ ] cement sales in the home market been made at less than the COP? The fact that type [ ] cement represented a negligible amount of total home market sales would have been irrelevant. In that case, the Department was able to avoid addressing the issue because the number of sales in question were minuscule such that either methodology would have had a negligible affect on the margin. The Judge, however, did comment on the methodology used: The question is which group of sales is at issue -- the full 'such or similar merchandise' that is 'the sales on which the Secretary could base the calculation of foreign market value.' In this case, a separate FMV for Type [ ] sales was calculated by Commerce. Only the Type [ ] FMV was used in making the fair value comparisons between Cemex's Type [ ] cement sales in the United States and those in the home market. Thus, the sales used by Commerce for calculating FMV for Type [ ] cement were the Type [ ] home market sales which were alleged to be below COP. ITA's statement of reasons in its determination causes the court some concern. Because it has sub-divided the 'such or similar merchandise' for purposes of the price comparisons, it could be basing a part of the margin on a priced based FMV that should have been replaced by a cost-based FMV. The ten percent rule used in the past has been approved in the context of a unitary 'such or similar merchandise' category. See Timken 11 CIT at 806-807, 673 F. Supp. at 514-15. Thus, the court has not previously approved ITA's reasoning that even when it sub-divides 'such or similar merchandise' for FMV and U.S. price comparison purposes, it may nonetheless re-aggregate the 'such or similar' category for purposes of its ten percent rule. The court leaves this issue for another day because, whatever the correct approach, it would not order ITA to undertake a COP investigation in this fair value investigation. (due to the small value of the sales involved). CIT Slip op. 92-24 at 13-14 Though the Judge did not rule on this issue, this case clearly shows the differences in the two methodologies and how they could have an effect on the margin. This case also shows the difficulties which might arise defending one methodology while several methodologies are in practice within IA. Analysis: If the purpose behind the COP provision of the law is to evaluate the rationality of exporters pricing practices, the such or similar approach is appropriate. The price of any model may be a function, among other things, of the price of other models. It is reasonable to believe that from time to time a producer will price certain models at less than COP, but be profitable overall. As an example, an automobile maker may offer low end cars below COP in order to build owner loyalty. In this case, the pricing decision is rational. Is it fair to look at the price of only one model when the prices are set on a product line basis? If the purpose of the COP provision is to avoid basing FMV on prices below cost (with certain exceptions), the model specific methodology is appropriate, since it focuses on the prices actually used for FMV. FMV itself is based on a model specific comparison, that of the most similar model. In the price-to- price comparison, the prices of models that are not used in the comparison are irrelevant to the determination of FMV. Similarly, in the cost test, the fact that models not used for comparison are priced above or below cost is irrelevant to determining if the prices to be used for FMV are above or below cost. We consider that the second method accords more with the departmental practice of calculating model specific FMVs than does the first, and 773(b) directs us to disregard below cost sales in calculating FMV. Statement of Policy We compare the home market sales price of each model to the corresponding cost of production. Then we apply the 10/90/10 test to the home market sales of each model sold in the home market. If we find that less than 10 percent (50 percent for perishable agricultural products) of the sales of a model are made at less than cost, we include all home market sales of that model in the calculation of FMV. If between 10 and 90 percent (50 and 90 percent for perishable agricultural products) of home market sales of a specific model are made at less than cost, (assuming below cost sales occur over an extended period of time at less than cost recovery prices) we disregard those sales made at less than cost and use the above cost sales in the price comparisons. If more than 90 percent of home market sales of a model are made at less than cost (with the previous assumption), we disregard all sales of that model. Implementation The above practice should be followed in all investigations and administrative reviews for which a preliminary determination has not been reached by the issue date of this bulletin, and in all final determinations in which the issue has been raised in comments from interested parties. If this method is challenged in comments after a preliminary determination, you should use the reasoning expressed in the bulletin, as well as citing to any appropriate cases. In responding to the comment, the bulletin itself, standing alone, cannot be cited as authority for an action, as it is not a regulation.