U.S.-China Trade: Eliminating Nonmarket Economy Methodology Would
Lower Antidumping Duties for Some Chinese Companies (10-JAN-06,  
GAO-06-231).							 
                                                                 
U.S. companies adversely affected by unfair imports may seek a	 
number of relief measures, including antidumping (AD) duties. The
Department of Commerce (Commerce) classifies China as a nonmarket
economy (NME) and uses a special methodology that is commonly	 
believed to produce AD duty rates that are higher than those	 
applied to market economies. Commerce may stop applying its NME  
methodology if it finds that China warrants designation as a	 
market economy. In light of increased concern about China's trade
practices, the conference report on fiscal year 2004		 
appropriations requested that GAO review efforts by U.S.	 
government agencies responsible for ensuring free and fair trade 
with that country. In this report, the last in a series, GAO (1) 
explains the NME methodology, (2) analyzes AD duties applied to  
China and compares them with duties applied to market economies, 
and (3) explains circumstances in which the United States would  
stop applying its NME methodology to China and evaluates the	 
potential impact of such a step. Commerce agreed with our	 
findings, commenting that our report provides timely and helpful 
information on the NME methodology and its application to China. 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-06-231 					        
    ACCNO:   A44517						        
  TITLE:     U.S.-China Trade: Eliminating Nonmarket Economy	      
Methodology Would Lower Antidumping Duties for Some Chinese	 
Companies							 
     DATE:   01/10/2006 
  SUBJECT:   Economic analysis					 
	     Evaluation methods 				 
	     Foreign governments				 
	     Foreign trade policies				 
	     Import regulation					 
	     Import restriction 				 
	     International economic relations			 
	     International trade				 
	     International trade regulation			 
	     International trade restriction			 
	     National policies					 
	     Policy evaluation					 
	     Prices and pricing 				 
	     Restrictive trade practices			 
	     Trade policies					 
	     China						 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-06-231

     

     * Report to Congressional Committees
          * January 2006
     * U.S.-CHINA TRADE
          * Eliminating Nonmarket Economy Methodology Would Lower Antidumping
            Duties for Some Chinese Companies
     * Contents
          * Results in Brief
          * Background
          * Commerce Employs a Special Methodology to Calculate China AD
            Duties
               * AD Calculations for NME Products Employ Third Country
                 Information
               * NME Companies Must Meet Certain Criteria to Be Considered
                 Eligible for Individual Duty Rates
               * Commerce Employs Different Approaches to Determine Duty
                 Rates for Eligible and Ineligible NME Companies
                    * Eligible NME Companies Receive Individually Determined
                      or Weighted Average Duty Rates
                    * Other NME Companies Receive Country-Wide Duty Rates
          * Commerce Has Applied AD Duties to China Frequently at Varied
            Rates
               * China Has Been the Most Frequent Target of U.S. Antidumping
                 Orders
               * China, Market Economy Rates Varied Widely
               * Overall, China Rates Were Higher Than Market Economy Rates
               * Difference in Average Rates Due Primarily to High China
                 Country-Wide Rates
                    * Country-Wide Duty Rates Substantially Higher Than
                      Market Economy All-Others Rates
                    * Individual Company Rates in China and Market Economy
                      Countries Not Substantially Different on Average
          * Ceasing to Apply NME Methodology Would Have Mixed Results
               * Commerce May Stop Applying NME Methodology to China in
                 Certain Circumstances
                    * Commerce Has Administrative Authority to Change China's
                      NME Status
                    * China's WTO Commitment Allowing the Use of Third-
                      Country Information Expires in 2016
               * Transition to Market Economy Methodology Would Bring
                 Significant Procedural Changes
               * Eliminating Country-Wide Rates Would Likely Reduce Duty
                 Levels for Companies Not Assigned Individually Determined
                 Rates
               * Individually Determined Rates Would Vary, Depending Upon
                 Cooperation
               * Impact of Applying Chinese Price Information Would Vary By
                 Industry and Is Likely to Decline over Time
               * Duty Rates Will Continue to Display Great Variation
               * Trade Significance of Country-Wide Rates Appears to Be
                 Declining
          * Concluding Observations
          * Agency Comments and Our Evaluation
     * Scope and Methodology
     * Additional Information on Duty Rates Applied to China and Market
       Economy Countries
          * Duty Rates Imposed on China
          * Duty Rates against China Have Fluctuated over Time
          * Average Duty Rates on 68 Orders against China and Subset of 25
            Orders Matched to Market Economy Orders
     * Regression Analysis Results
     * Comments from the Department of Commerce
          * GAO Comments
     * GAO Contact and Staff Acknowledgments
     * cov1&2.pdf
          * Report to Congressional Committees
               * January 2006
          * U.S.-CHINA TRADE
               * Eliminating Nonmarket Economy Methodology Would Lower
                 Antidumping Duties for Some Chinese Companies
     * PDF6-Ordering Information.pdf
          * Order by Mail or Phone

Contents

Tables

Figures

January 10, 2006Letter

The Honorable Richard C. Shelby Chairman The Honorable Barbara A. Mikulski
Ranking Minority Member Subcommittee on Commerce, Justice, and Science
Committee on Appropriations United States Senate

The Honorable Frank R. Wolf Chairman The Honorable Alan B. Mollohan
Ranking Minority Member Subcommittee on Science, State, Justice, and  
Commerce and Related Agencies Committee on Appropriations House of
Representatives

Imports from China have grown rapidly over the last decade, from a total
value of about $42 billion in 1995 to over $196 billion in 2004.1 While
the prices of these Chinese goods are often lower than U.S. prices and,
therefore, benefit consumers, this growth has presented a major challenge
for U.S. producers that compete with Chinese products in the U.S. market.
Some U.S. companies adversely affected by this growth have alleged that
Chinese success in the U.S. market has come partly as a result of unfair
trade practices.

U.S. companies that are adversely affected by unfair imports from China
(or other countries) may avail themselves of a number of relief measures,
including antidumping (AD) duties. The United States has classified China
as a "nonmarket economy" (NME) country since 19812 and employs a special
NME methodology to calculate AD duties on unfairly traded products from
that country. This methodology is commonly believed to result in duty
rates that are significantly higher than those applied to market economy
countries.

In light of increased concern about China's trade practices, the
conference report on fiscal year 2004 appropriations legislation requested
that GAO monitor the efforts of U.S. government agencies responsible for
ensuring free and fair trade with that country.3 In subsequent discussions
with staff from the House Appropriations Committee's Subcommittee on
Science, State, Justice, and Commerce and Related Agencies, we agreed to
provide a number of reports on import relief mechanisms and the manner in
which these mechanisms have been applied to China. To date, we have issued
three such reports, focusing on textile safeguards, safeguards applicable
to other products, and countervailing duties.4

This fourth and final report on China import relief mechanisms focuses on
AD duties. In this report, we

o explain the special methodology that the United States employs to
calculate AD duties on products from China and other NME countries,

o analyze the application of AD duties to China over the last 25 years and
compare the duty rates applied to Chinese products with the duty rates
applied to products from market economy countries, and

o explain the circumstances in which the United States would stop using
its NME methodology to calculate AD duties on Chinese products and
evaluate the potential impact of this step.

To conduct our review, we examined applicable U.S. laws and regulations
and World Trade Organization (WTO) agreements, including relevant portions
of the agreement through which China acceded to WTO membership in 2001. We
reviewed scholarly literature and consulted with trade and legal policy
experts from the U.S. government, private sector trade associations,
consulting and law firms, and academic institutions, as well as
representatives of the WTO, the government of China, and other
governments. In order to analyze U.S. application of AD duties to China
and compare the duties applied to China with those applied to market
economy countries, we used information from the Department of Commerce
(Commerce) and the U.S. International Trade Commission (ITC), including
notices of Commerce determinations appearing in the Federal Register, to
construct a database on all U.S. antidumping investigations from 1980
through 2004. We verified this database to the official sources. Our
analyses focused on the 68 AD duty orders that Commerce issued against
Chinese products during this period and especially on the 25 cases in
which Commerce also imposed duties on the same products from market
economy countries.5 We performed additional (multivariate regression)
analyses to determine the extent to which duty rate variations could be
attributed to differences between China and these other countries, or to
other factors, such as the type of product involved. Appendix I contains a
detailed description of our scope and methodology.

We conducted our work from June 2005 through December 2005 in accordance
with generally accepted government auditing standards.

Results in Brief

Commerce's special methodology for calculating the AD duties that it
applies to China and other NME countries differs from its usual (market
economy) approach in two key respects. First, since prices in NME
countries do not  reliably reflect the fair value of the merchandise,
Commerce uses price  information from surrogate countries (like India) to
construct the value of NME products-and thus provide an appropriate basis
for calculating AD duty rates-rather than relying entirely on information
from the exporting country itself. Second, Commerce requires NME companies
to demonstrate that their export activities are not subject to government
control  in order to be considered eligible for individually determined
duty rates, rather than considering all companies eligible for such rates,
as it does in market economy cases. NME companies that do not meet these
criteria, or do not participate in Commerce's investigations receive
"country-wide" rates.

Over the last 25 years, the United States has applied AD duties against
China more often than against any other country. On 25 occasions, Commerce
applied duties to the same products from both China and at least one
market economy country. The duty rates assigned in these cases varied
greatly. On average, however, the rates applied to China were over 20
percentage points higher than those applied to market economy countries.
This difference is attributable primarily to the comparatively high
country-wide duty rates applied to Chinese companies not eligible for
individual rates. These country-wide rates averaged about 98 percent-over
60 percentage points higher than the average duty rates assigned to market
economy companies not receiving individual rates. In contrast, when
Commerce calculated individual rates for Chinese companies, these rates
were not substantially different, on average, from those assigned to
individual market economy companies.

Commerce has administrative authority to declare China a market economy,
or find individual Chinese industries to be "market-oriented" in
character-provided that China overall or individual Chinese industries
meet certain criteria. Such a declaration would end application of the NME
methodology to China, in whole or in part. This would (1) eliminate
country-wide duty rates against China and (2) eliminate use of surrogate
country information to calculate AD duty rates on Chinese products. These
changes would have a mixed impact. Eliminating country-wide rates would
likely reduce duty rates applied to companies not receiving individual
rates. Individually determined rates would likely diverge into two
distinct groups, with companies that do not cooperate in Commerce
investigations receiving rates that are substantially higher than those
that do cooperate. The impact of applying Chinese price information would
likely vary by industry, and AD rates applied against China would continue
to vary widely, both within and among cases. However, it appears that the
significance of the NME country-wide rates is declining as more Chinese
companies receive individual rates, although data that would permit
quantification of the potential trade impact of these changes is not
available. This suggests that the trade significance of the NME
methodology now applied to China will likewise decline over time.

Commerce provided written comments on a draft of this report, which are
reprinted in appendix IV. Overall, Commerce agreed with our findings. In
addition, Commerce, the Department of Homeland Security, and the Office of
the U.S. Trade Representative, provided technical comments. We took these
comments into consideration and made revisions throughout the report as
appropriate to make it more accurate and clear.

Background

Dumping refers to a type of international price discrimination wherein a
foreign company sells merchandise in a given export market (for example,
the United States) at prices that are lower than the prices that the
company charges in its home market or other export markets. When this
occurs, and when the imports have been found to materially injure, or
threaten to materially injure, U.S. producers, U.S. law permits
application of antidumping duties to offset the price advantage enjoyed by
the imported product.6

Any domestic industry that believes it is suffering material injury, or is
threatened with material injury, as a result of dumping by foreign
companies may file a petition requesting imposition of AD duties.
Interested domestic industries file petitions simultaneously with Commerce
and ITC. If Commerce determines that the petitioning parties meet certain
eligibility requirements,7 ITC determines whether the domestic industry
has suffered material injury as a result of the alleged dumping (or is
threatened with material injury).8 While ITC is completing its work,
Commerce conducts an investigation to establish the duty rates, if any,
that should be applied.

To determine the duty rates to apply in an antidumping investigation,
Commerce identifies (1) the foreign product's export price entering the
U.S. market and (2) its "normal value."  Commerce then compares these

prices to determine whether-and by how much-the product's export price is
less than its normal value. AD duty rates are based on these differences,
which are called dumping margins.9

To establish a product's export price, Commerce generally refers to the
prices charged in actual sales of that product to purchasers in the United
States.10 To establish its normal value, Commerce generally refers to the
prices charged for the product in the exporting company's home market. In
the event that the product is not sold in the exporter's home market,
Commerce may refer to prices charged for the product in another export
market or construct a normal value based on costs of production in the
exporting country, together with selling, general and administrative
expenses, and profit.11 The two agencies make preliminary and, after
additional investigation, final determinations as to whether injury has
occurred (ITC) and the size of the duty, if any, that should be imposed
(Commerce). When warranted, Commerce issues "duty orders" instructing
Customs and Border Protection to apply duties against imported products
from the countries under investigation. Both ITC and Commerce publish
their decisions in the Federal Register.

Since AD duties address unfair pricing practices, and pricing decisions
are generally made by individual companies, Commerce generally calculates
and assigns AD duty rates on an individual company basis. As a result, AD
investigations generally produce a number of individually determined,
company-specific rates, reflecting differences in the extent to which
companies have dumped their products-that is, exported them at less

than their normal value.12 In addition, AD duty orders also generally
specify a duty rate for other companies that have not been assigned an
individually determined rate.

In principle, Commerce bases its AD duty determinations on information
obtained from interested parties-including foreign producers and
exporters. Commerce obtains needed information from foreign companies by
sending them questionnaires and following up with additional questions, as
needed, and with on-site visits.13

However, both U.S. law and WTO rules recognize that, in some cases,
officials charged with completing these investigations will be unable to
obtain sufficient information. In such cases, Commerce officials apply
facts available to complete their duty determinations.14 This may include
secondary information, subject to corroboration from independent sources.

Moreover, if Commerce finds that an interested party, such as a foreign
company under investigation, "has failed to cooperate by not acting to the
best of its ability to comply with a request for information" then, in
selecting among the facts available, Commerce may apply an inference that
is adverse to the interests of that party. In applying adverse inferences,
Commerce can use (among other things) information contained in the
petition filed by the domestic industry seeking imposition of AD duties,
the results of a prior review or determination in the case, or any other
information placed on the record.15

This authority provides an incentive for foreign companies to provide the
information that Commerce needs to complete its work. For example, in a
1993 case that involved two Brazilian companies, one company attempted to
cooperate in the investigation but nonetheless was unable to provide the
information that Commerce needed, while the other declined to provide any
information at all. Commerce used facts available to determine that the
first company should be subject to a duty rate of 42 percent. For the
second company, Commerce selected adverse inferences from among the facts
available and applied these to calculate a duty rate of 109 percent.16

Commerce Employs a Special Methodology to Calculate China AD Duties

The methodology that Commerce employs in NME cases differs from Commerce's
usual (market economy) approach in two key ways. First, rather than rely
entirely on information from the exporting country itself to establish a
product's normal value, Commerce uses price information from surrogate
countries to construct these values. Second, rather than consider all
companies eligible for individually determined duty rates, Commerce
requires NME companies to meet certain criteria to be considered eligible
for such rates. Commerce generally employs different approaches to
calculate duty rates for companies that do and do not meet these criteria.

AD Calculations for NME Products Employ Third Country Information

In AD investigations involving products from NME countries, U.S. law
requires Commerce to use a special methodology to calculate duty rates in
view of the absence of meaningful home market prices and information on
production costs.  When a product from China or another NME country is the
target of an AD investigation, Commerce officials use price information
and financial data from an appropriate market economy country to construct
a normal value for the product under investigation.17 India is the most
commonly used surrogate for China.

To apply this methodology, Commerce (1) identifies and quantifies the
factors of production (e.g., various raw materials) used by the NME
producers, (2) identifies market prices for each factor in a surrogate
country; (3) multiplies volume times cost for each factor; and (4) adds
the results, together with a reasonable margin for selling, general and
administrative expenses, and profit (based on surrogate country financial
data), to produce a constructed normal value.18 The dumping margin-and
consequently the AD duty rate-is then determined by comparing this normal
value with the NME company's export price to the United States.

NME Companies Must Meet Certain Criteria to Be Considered Eligible for
Individual Duty Rates

While all companies from market economy countries are eligible for
individually determined or weighted average AD duty rates, companies from
China and other NME countries must pass a separate rates test to be
eligible for such rates. This test requires NME companies to meet two
closely related criteria: they must demonstrate that their export
activities are free from government control both in law and in fact.19 To
provide a basis for deciding whether companies meet these criteria,
Commerce requires these companies to submit information regarding

o whether there are restrictive stipulations associated with an individual
exporter's business and export licenses,

o any legislative enactments decentralizing control of companies,

o any other formal measures decentralizing government control of
companies,

o whether export prices are set by or subject to approval by the
government,

o whether the company has authority to negotiate and sign contracts,

o whether the company has autonomy in selecting its management, and

o whether the company retains the proceeds of its export sales and makes
independent decisions regarding disposition of profits or financing of
losses.20

Commerce Employs Different Approaches to Determine Duty Rates for Eligible
and Ineligible NME Companies

As shown in figure 1,  Commerce uses fundamentally different approaches to
calculate duty rates to be applied against companies that do and do not
pass the separate rates test.

Figure 1: Different Approaches to Determining AD Duty Rates

Eligible NME Companies Receive Individually Determined or Weighted Average
Duty Rates

As shown in figure 1, Commerce treats companies from China and other NME
countries that pass Commerce's separate rates test like market economy
countries when assigning duty rates. When practical, Commerce fully
investigates and establishes individually determined duty rates for each
eligible NME company, just as it does for each market economy company. To
the extent that fully investigated NME companies cooperate with Commerce,
they receive rates based on the information that they provide. As
explained in the background section of this report, Commerce uses facts
available, and may use adverse inferences, to calculate duty rates when
the companies under investigation cannot or will not provide the
information that Commerce needs.

In both NME and market economy cases, Commerce may limit the number of
companies it fully investigates when it is faced with a large number of
companies. In such situations, Commerce generally calculates individual
rates for the  companies that account for the largest volume of the
subject merchandise.21 In market economy cases, Commerce then calculates a
weighted average of these rates and applies the resulting "all others"
rate to companies that it has not fully investigated.22 Commerce does not
routinely calculate weighted average duty rates in NME cases. However,
when the number of NME companies eligible for individually determined
rates exceeds the number that Commerce can fully investigate, Commerce
calculates a weighted average rate and informs Customs of the companies
entitled to this rate.23

Other NME Companies Receive Country-Wide Duty Rates

In cases involving China or other NME countries, Commerce calculates a
country-wide duty rate for companies that could not (or did not attempt
to) pass Commerce's separate rates test. In NME cases, Commerce assumes
that all exporters and producers of a given product are subject to common
government control and that all of these companies should, therefore, be
subject to a single country-wide duty rate. Commerce begins its NME
antidumping investigations by requesting information from the government
of the country in question and from known producers and exporters. If
Commerce cannot identify all relevant producers and exporters, or if one
or more of the identified companies refuses to cooperate in the
investigation, Commerce relies on adverse inferences to calculate a
country-wide rate. Commerce then instructs Customs to apply the
country-wide rate against shipments from any company other than those
specifically listed as eligible for an individually determined or weighted
average rate.24

Commerce Has Applied AD Duties to China Frequently at Varied Rates

Over the last 25 years, the United States has applied AD duties against
Chinese products more often than against products from any other country.
While AD duty rates have varied widely, on average the rates assigned to
Chinese products have been higher than the rates assigned to the same
products from market economy countries. We found that this is attributable
primarily to the comparatively high country-wide rates applied to Chinese
companies not eligible for individually determined or weighted average
rates. When Commerce has calculated rates for individual Chinese
companies, the average rates assigned to these companies have not been
substantially different from those assigned to market economy companies.

China Has Been the Most Frequent Target of U.S. Antidumping Orders

Over the last 25 years, Commerce has both considered and actually applied
AD duties against China more often than against any other country. From
1980 through 2004, Commerce processed 1,046 AD petitions and issued 455 AD
duty orders. One hundred and ten of these petitions (11 percent) and 68 of
these orders (15 percent) focused on China-both are the largest number
against any U.S. trade partner.25

The number of orders applied to China varied from year to year. For
example, Commerce issued no AD duty orders against China in 1998 but
issued 9 in 2003. Commerce had 272 orders in place as of December 31,
2004. Fifty-five of these (20 percent) apply to China. As figure 2 shows,
this is also the highest percentage of any country. As shown in table 1,
these duty orders have targeted a wide variety of products but have been
concentrated in chemicals and plastics, metal products, and agricultural
products.

Figure 2: U.S. AD Orders in Place by Country as of December 31, 2004

Note: From 1980 through 2004 Commerce applied 68 orders against China.
Thirteen of these were revoked, leaving 55 in place as of the end of 2004.

Table 1: Products Affected by AD Duty Orders against China, 1980-2004

                                        

          Type of product            Examples of affected           Number of 
                                           products                    orders 
Chemicals, plastics,          Barium chloride                           26 
pharmaceuticals                                            
                                 Polyethylene retail carrier  
                                 bags                         
                                                              
                                 Bulk aspirin                 
Steel, other metals           Carbon steel butt-weld pipe               20 
                                 fittings                     
                                                              
                                 Chrome-plated lug nuts       
                                                              
                                 Pure magnesium               
Agricultural products         Crawfish                                   5 
                                                              
                                 Garlic                       
                                                              
                                 Honey                        
Other products                Brake rotors                              17 
                                                              
                                 Hand tools                   
                                                              
                                 Cotton shop towels           
                                                              
                                 Automotive replacement glass 
                                 windshields                  
                                                              
                                 Folding gift boxes           

Source: GAO AD database.

Note: Product categories based on the Harmonized Tariff Schedule of the
United States, Annotated. See appendix I for more information.

Moreover, petitions for AD duties against China have resulted in
application of duties more often than those against other countries. As
shown in figure 3, 62 percent of the petitions filed against China over
the last 25 years resulted in duty orders, while the equivalent figure for
all countries was about 43 percent.26

Figure 3: Results of AD Petitions, 1980-2004

Note: Petitions may not result in an AD order for several reasons. For
example, if ITC does not find the domestic industry to be materially
injured, or threatened with material injury, or if Commerce does not find
that dumping has occurred, then the case is terminated. Also, Commerce may
suspend a case if the United States reaches an agreement with the foreign
government that would eliminate the impact of dumping (e.g., restrictions
on exports). In addition, the petitioners may decide to withdraw their
petition before ITC and Commerce have completed their work.

China, Market Economy Rates Varied Widely

Over this 25 year period, Commerce issued duty orders against the same
products from China and at least one market economy country on 25
occasions.27 In 18 of these cases, Commerce calculated individual rates
for companies from China and at least one market economy country. Fifteen
of these cases involved more than one market economy country. In all, the
orders applying to these 25 products contained a total of 243 individual,
weighted average, and country-wide duty rates. Appendix II provides
detailed information on the rates applied in each of these cases, as well
as another 43 cases that we identified wherein Commerce applied duty rates
to China but not to any market economy country.

These rates varied a great deal-both among the orders applied to different
products and within the orders applied to the same products. Overall,
these duty rates varied from zero to 218 percent for China and from zero
to about 244 percent for market economy countries. Figure 4 shows the
extent to which duty rates applied to a single product can vary.28

Figure 4: AD Duty Rates Applied against Imported Polyethylene Retail
Carrier Bags (2004)

Note: In addition to the five individual Chinese companies shown above,
two Chinese companies received de minimus rates and were excluded from the
order.

Overall, China Rates Were Higher Than Market Economy Rates

The average AD duty rates imposed on Chinese (NME) exporters over the last
25 years have been significantly higher than those imposed on market
economy exporters of the same products. Taking all rates into
consideration (including those calculated for individual companies,
weighted averages of these rates, and country-wide rates applied to China)
the average rate applied to Chinese companies in the 25 cases we examined
was about 67 percent-over 20 percentage points higher than the average
rate of 44 percent applied to market economy companies. As figure 5 shows,
the overall average rates applied against China were higher for 18 of the
25 products in which there were AD orders against both China and at least
one market economy.

Figure 5: Differences between Overall Average Duty Rates-China and Market
Economies, 1980-2004

Difference in Average Rates Due Primarily to High China Country-Wide Rates

The difference between average China and average market economy duty rates
was due primarily to the fact that the NME country-wide duty rates applied
to China were substantially higher than the comparable all-others duty
rates applied against market economy countries. In contrast, the
individually determined duty rates assigned to Chinese companies in these
cases were not substantially different, on average, from the individually
determined rates assigned to market economy companies.

Country-Wide Duty Rates Substantially Higher Than Market Economy
All-Others Rates

On average, the country-wide rates applied to China in these 25 cases were
substantially higher than the comparable all-others rates applied to
market economy countries. The country-wide duty rates applied against
China averaged about 98 percent-over 60 percentage points higher than the
average 37 percent all-others duty rate applied to market economy
exporters of the same products. Figure 6 shows that the China country-wide
rate was higher than the market economy all-others rate in 21 of 25 cases.
As explained below, this difference was due largely to the use of
different methodologies to calculate country-wide and all-others rates.

Figure 6: Differences between China Country-Wide and Market Economy
All-Others Duty Rates, 1980-2004

Note: In several cases, Commerce issued orders against several market
economy countries, and we calculated an average all-others rate for all of
the affected market economy countries.

Country-wide rates were nearly always equal to or higher than the highest
individually determined rate applied to any Chinese company, due to
application of adverse inferences.29 According to Commerce, NME country
governments themselves have never provided the information that Commerce
needs to establish an appropriate country-wide duty rate. In addition,
Commerce officials stated that, in most cases, participating NME companies
have accounted for only a portion of known exports to the U.S. market from
their country, indicating that others had not come forward. In most cases,
therefore, Commerce has used adverse inferences to determine country-wide
rates. For example, in its investigation of carbazole violet pigment,
Commerce assigned three fully investigated Chinese companies individually
determined rates of about 6, 27, and 45 percent. However, since other
known Chinese producers did not respond to Commerce's request for
information, Commerce used adverse inferences to determine that all other
Chinese producers should be subject to an NME country-wide rate of about
218 percent.30

In contrast, the comparable market economy all-others rates were lower
than the highest individual company rates assigned in any given case (if
more than one other individual rate was assigned).31 This is because, as 
discussed earlier, Commerce generally calculates all-others rates by
averaging individually determined rates-excluding those derived entirely
through application of facts available and those that are de minimis or 
zero. With regard to carbazole violet pigment, for example, Commerce
investigated not only China but also India. Commerce assigned two fully
investigated Indian companies rates of about 10 and 50 percent and
weight-averaged these rates to determine that shipments from all other
Indian producers should be subject to a duty rate of about 27 percent.

Individual Company Rates in China and Market Economy Countries Not
Substantially Different on Average

On average, there was little difference between the individually
determined rates applied to companies from China and those applied to
market economy companies. The average individually determined rate applied
to Chinese companies in these cases was 53 percent-a little less than the
average rate of 55 percent applied to market economy companies.32 The
median rate for Chinese companies was 42 percent-the same as the median
rate for market economy companies. Figure 7 displays the average
individual company rates assigned to Chinese and market economy companies
in the 18 cases in which Commerce assigned individual rates to both. As
the figure shows, the rates assigned to Chinese companies were higher than
the market economy rates in ten of these cases and lower in the other
eight.

Figure 7: Differences between Average Individual Company Duty Rates-China
and Market Economy Countries, 1980-2004

Our statistical  analyses provided additional support for the importance
of the country-wide rates in accounting for the overall difference between
the duty rates applied to China and to market economy countries. Using
multivariate regression analysis, we found that a number of variables,
such as the type of product involved, accounted for some of the overall
variation in duty rates. However, after accounting for the China
country-wide rates there was no statistically significant difference
between the duty rates applied to China and those applied to market
economy countries. As explained in more detail in appendix III, we found
essentially the same results when we expanded our analyses to include data
on AD actions against NMEs other than China.33

Ceasing to Apply NME Methodology Would Have Mixed Results

In certain circumstances, Commerce may stop using its NME methodology in
China cases-and thus begin applying its market economy methodology to
determine AD duty rates against that country. Such a step would lead to
important changes in the methods that Commerce employs to determine China
AD duty rates and in the duty orders resulting from these proceedings.
These changes would have mixed results. Eliminating country-wide duty
rates would likely reduce duty levels for Chinese companies that are not
assigned individually determined rates. Individually determined rates
would likely diverge into two distinct groups, with companies that do not
cooperate in Commerce investigations receiving rates that are
substantially higher than those assigned to companies that do cooperate.
The impact of applying Chinese price information to calculate the normal
value of Chinese products would likely vary by industry. In any case,
rates would continue to vary widely based on the circumstances of each
case. While trade data that would permit analysis of the potential trade
impact of these changes is not available, it appears that the trade
significance of country-wide duty rates is declining.

Commerce May Stop Applying NME Methodology to China in Certain
Circumstances

Commerce has administrative authority to reclassify China and other NME
countries as market economies or individual NME country industries as
market-oriented in character. Such reclassifications  would end Commerce's
authority to apply its NME methodology to such countries or industries.
Also, China's WTO accession agreement specifies that members may apply
third-country information to calculate AD duty rates against that country,
but this provision expires in 2016.

Commerce Has Administrative Authority to Change China's NME Status

Commerce has the authority to reclassify China as a market economy
country, in whole or in part. As we explained in more detail in a prior
report,34 U.S. trade law authorizes Commerce to determine whether
countries should be accorded NME or market economy status and specifies a
number of criteria for Commerce to apply in making such determinations.35
Countries classified as NMEs may ask for a review of their status at any
time.36 China has actively sought market economy status among its trading
partners, and a number of them have designated China as a market economy.
However, Commerce informed us that Chinese officials have not yet
officially requested a determination as to whether their country merits
reclassification under the criteria specified in U.S. law. In April 2004,
the United States and China established a Structural Issues Working Group
under the auspices of the U.S.-China Joint Commission on Commerce and
Trade. This group is examining structural and operational issues related
to China's economy that may give rise to bilateral trade frictions,
including issues related to China's desire to be classified as a market
economy.37

Commerce also has the authority to designate individual NME industries as
market oriented in character, but has denied all such requests to date.
Commerce determined in a 1992 case against China that, short of finding
that an entire country merits designation as a market economy, it could
find specific industries within such countries to be market oriented in
character.38 Commerce officials noted that on several occasions Chinese
industries responding to antidumping duty petitions have requested
designation as market-oriented industries. To date, Commerce has denied
such requests-primarily on the grounds that the Chinese companies in
question submitted information that was insufficient or was provided too
late in Commerce's process to allow an informed decision.

China's WTO Commitment Allowing the Use of Third-Country Information
Expires in 2016

When joining the WTO, China agreed that other WTO members could use
third-country information to calculate normal values in antidumping
actions against Chinese companies. Specifically, China's WTO accession
agreement provides that in determining price comparability in antidumping
investigations WTO members may use "a methodology that is not based on a
strict comparison with domestic prices or costs in China."39 However, the
accession agreement also specifies that this provision will expire 15
years after the date of the agreement-that is, by the end of 2016.40

After 2016, the ability of WTO members to continue using third-country
information in AD calculations involving China would be governed by
generally applicable WTO rules, according to officials at the Office of
the U.S. Trade Representative. These rules recognize that when dumping
investigations involve products from a country that "has a complete or
substantially complete monopoly of its trade and where all domestic prices
are fixed by the state," importing country authorities may have difficulty
making the price comparisons through which AD duty rates are normally
established. In such situations, importing countries may "find it
necessary to take into account the possibility that a strict comparison
with domestic prices in such a country may not always be appropriate."41
WTO rules do not provide any specific guidance about how this provision
should be implemented; such decisions appear to be left up to individual
members.

Transition to Market Economy Methodology Would Bring Significant
Procedural Changes

Ending application of the NME methodology to China would bring two
significant procedural changes in AD duty investigations against Chinese
products. First, such a step would eliminate NME country-wide duty rates
from China AD orders. Commerce would instead assign an individually
determined rate to every relevant Chinese producer or exporter. If the
number of companies involved were too great to allow full investigation of
all relevant companies, Commerce would apply an all-others rate-a weighted
average of the individually determined rates to all other Chinese
companies (excluding those rates based entirely on facts available or that
are de minimis or zero). However, Commerce would retain its authority to
use facts available to determine the rates that it applies to individual
Chinese companies. Second, transition to the market economy methodology
would end Commerce's use of surrogate country information to calculate the
normal value of Chinese products. Application of the market economy
methodology would generally require Commerce to set the normal value of
Chinese products equal to their sales price in China. If the product were
not sold in China, Commerce could refer to prices charged for the product
in another export market or construct the product's normal value, or it
could continue to construct the product's normal value-using factor prices
from the Chinese companies under investigation rather than from surrogate
countries.

Eliminating Country-Wide Rates Would Likely Reduce Duty Levels for
Companies Not Assigned Individually Determined Rates

The elimination of country-wide duty rates against China would likely
reduce the duty rates applied to some Chinese companies. If Commerce
applied its market economy approach to China, duty rates for companies not
receiving individually determined rates would, in most cases, no longer be
determined by applying facts available. Rather, Commerce would, for the
most part, determine these rates by averaging the rates applied to fully
investigated Chinese companies, with some exclusions. The default rate for
uninvestigated Chinese companies would move, in most cases, from being the
highest rate found to the average rate found among companies that
cooperate in Commerce investigations.

Though not predictive, available evidence suggests that the all-others
rates that Commerce would apply to China under the market economy
methodology would be significantly lower than the country-wide rates
currently applied to that country. As already shown, China country-wide
rates have generally been significantly higher than the all-others rates
that Commerce has assigned to market economy sources of the same products.
As shown in table 2, the average country-wide rate for the 25 cases in
which Commerce assigned duties to both China and one or more market
economies was 98 percent, while the average market economy all-others rate
was 37 percent. The average rate assigned to individual Chinese companies
was 53 percent, and Commerce calculates all-others rates by weight
averaging individually determined rates, excluding those that are derived
entirely through application of facts available and those that are de
minimis or zero.

Table 2: Comparison of China, Market Economy AD Duty Rates, Methodological
Changes, and Potential Effects

                                        

                                           Individual 
                                            rates b   
               Overall a   Group a         All        Cooperative Adverse       
                                                                  inferences    
Average     67%         98%             53%        51%         (Rarely       
China (NME)                                                    applied)c     
rates                   (Country-wide)                         
Average     44%         37%             55%        16%         77%           
market                                                         
economy                 (All others)                           
rates                                                          
Change from Chinese     Country-wide    Separate               
NME to      price       rates           rates test             
market      information eliminated,     eliminated             
economy     replaces    uninvestigated                         
methodology surrogate   companies                              
for China   price       receive all                            
companies   information others rate                            
Potential   Effect      Rates likely    Rates      Rates       Rates likely  
effect on   unknown but to be           likely to  likely to   to be         
average     likely to   significantly   fall into  be          relatively    
China rates vary by     lower           two        relatively  high for      
               industry                    distinct   low for     uncooperative 
                                           groups     cooperative companies     
                                                      companies   

Source: GAO analysis of Commerce data.

aAverages based on 25 products with comparable China and market economy
cases (1985-2004).

bAverages based on subset of 18 products with comparable China and market
economy cases in which individual rates were applied. However, averages
are nearly identical for full group of 25 products.

cCommerce applied adverse inferences only three times, for an average rate
of 78 percent.

Individually Determined Rates Would Vary, Depending Upon Cooperation

A simple comparison of the average individually determined duty rates
calculated under the NME and market economy methodologies suggests that a
change in methodology would not result in any significant overall change
in duty rates applied to individual Chinese companies. For the comparable
cases, individual AD duty rates for Chinese companies averaged 53 percent
and were not substantially different from individual market economy
company rates, which averaged 55 percent.

However, a more detailed examination of the data indicates that the impact
of a change in methodology on individual Chinese company duty rates would
depend on the extent to which Commerce applies adverse inferences to
calculate these rates. The rates assigned to individual companies under
the market economy methodology fell into two distinct groups, depending on
whether the companies cooperated with Commerce investigations. In the 25
cases that we examined in detail, about half of the fully investigated
market economy companies cooperated with Commerce. On average, Commerce
assigned a duty rate of about 17 percent to these companies.42 Commerce
found the other half of the fully investigated companies uncooperative
and, therefore, applied adverse inferences to determine the duty rates to
be applied to these companies. On average, Commerce assigned a duty rate
of about 77 percent to these uncooperative market economy companies.43

Though not predictive, this suggests that a change from the NME
methodology for China would result in a significant number of
(cooperative) companies receiving relatively low rates, while another
significant group of (uncooperative) companies would receive relatively
high rates.44 Our regression analysis confirmed the importance of adverse
inferences as a determinant of variation in duty rates. As explained in
appendix III, we found that application of adverse inferences tends to
increase duty rates by a large margin.

Impact of Applying Chinese Price Information Would Vary By Industry and Is
Likely to Decline over Time

The impact of using Chinese price information on China AD duty rates would
likely vary from one industry to another under the market economy
methodology. Chinese prices are widely viewed as distorted to varying
degrees. Where prices for key inputs are artificially low, relying on
Chinese price information would produce an artificially low normal value.
The result would be an AD duty that is lower than would be obtained by
applying surrogate country input prices. Conversely, where Chinese prices
are artificially high, AD duty rates may be higher if based on Chinese
prices. To the extent that Chinese economic reforms bring Chinese prices
more into line with world markets, the impact of abandoning the use of
surrogate country information can be expected to decline. At any point in
time, however, the probable effect of such a methodological change in an
individual industry investigation would depend on the particular facts
applying to that industry. The net impact of changing the source of price
information on overall China duty rates cannot be estimated with
confidence.

Duty Rates Will Continue to Display Great Variation

Our multivariate regression analyses suggest that, regardless of changes
in methodology, there will continue to be a great deal of variation among
the AD duty rates applied to products from China and other countries. Our
analyses showed that application of country-wide duty rates to China
largely explained the difference between the overall average duty rates
applied to China and to market economy countries. Eliminating these rates
would likely have a substantial overall reducing effect on China rates.
However, a number of other factors, such as the type of product involved,
also helped to account for differences among rates overall, and these
factors will continue to have an impact on duty rates, regardless of
whether Commerce applies country-wide rates to China. Furthermore, even
after taking these factors into account, our analyses still explained only
about half of the total variation in duty rates.45 This means that about
half of the

variation in duty rates is attributable either to idiosyncratic factors or
to systematic factors that we did not capture in any of our variables.46

Trade Significance of Country-Wide Rates Appears to Be Declining

Available evidence suggests that the volume of trade affected by
country-wide rates is declining and that, consequently, the trade impact
of China duty orders will in the future depend increasingly on the
magnitude of the individually determined rates. Commerce officials
observed that in the early 1980s it was not unusual for China AD duty
investigations to produce only a country-wide rate. However, as the
Chinese economy has evolved, individual Chinese companies have become more
likely to request-and receive-individually determined or weighted average
rates. Since 1980, Commerce has applied country-wide rates alone in only
15 of 68 Chinese AD orders, and the last of these occasions was in 1995.
The majority of all Chinese AD orders (about 78 percent), and all such
orders issued over the last 10 years, have included at least one
individual company rate.

Neither Commerce nor Customs and Border Protection maintain trade data
that would permit analysis of changes over time in the relative volume or
value of products imported into the United States under the country-wide
or various individual duty rates listed in AD duty orders. However, as
figure 8 shows, the average number of Chinese companies assigned
individually determined rates (or assigned a weighted average rate) has
been growing, though there continues to be variation from year to year.
For example, in 2004 Commerce placed duties on six Chinese products and in
doing so assigned individually determined or weighted average rates to 53
Chinese companies. Anecdotal evidence suggests that along with this rise
in company interest in obtaining individual rates has come an increase in
the volume of trade covered by these rates. For example, in one recent
case Commerce fully investigated and assigned individually determined
rates to four companies accounting for more than 90 percent of Chinese
exports to the U.S. market. Commerce then assigned a weighted average of
these rates to 9 additional companies, leaving only a very small portion
of all Chinese exports to be covered by the country-wide rate.47

Figure 8: Average Number of Individual Rates per Case per Year, All Orders
against China, 1980-2004

Notes: This table includes only the companies assigned individually
determined or weighted average rates in initial AD duty orders. Other
companies may be assigned such rates in subsequent reviews of these
orders. The annual figures are the number of individual companies granted
such rates divided by the number of AD duty orders. Commerce did not issue
any individual company duty orders against China in 1980, 1981, 1982,
1985, 1987, 1988, 1989, or 1998. In the tapered roller bearings cases,
Commerce originally put an order in place in 1987, but amended it in 1990.
We use the information from the 1990 amendment in the above graphic.

Concluding Observations

On average, Commerce's application of its NME methodology has produced AD
duties on Chinese products that are substantially higher than those
applied to the same products from market economy countries. Changing
China's NME status-and thus eliminating the application of this
methodology-would have a variety of impacts. The duty rates applied to
companies that do not receive individual rates would likely decline.
Chinese companies that cooperate in Commerce investigations may also
receive comparatively low rates. However, the impact of these lower rates
on overall China averages may be offset, to some extent, by application of
adverse inferences to assign relatively high rates to individual Chinese
companies that do not cooperate in Commerce investigations.

The net effect of these changes cannot be predicted. Such a prediction
would require knowledge of price distortions in diverse Chinese
industries, changes in these distortions over time, pricing decisions by
Chinese companies in reaction to these changes, and decisions by U.S.
companies about whether they should seek relief. Nonetheless, while the
NME methodology is applied, it appears that the actual trade impact of
using this methodology will decline as the portion of total export trade
conducted by Chinese companies assigned individual rates increases, and as
the country-wide rates that largely account for the comparatively high
average rates applied to China decline in importance.

Agency Comments and Our Evaluation

The Department of Commerce provided written comments on a draft of this
report. These comments are reprinted in appendix IV. Overall, Commerce
agreed with our findings, observing that the report provided timely and
helpful information on the NME methodology and its application to China.

Commerce identified a small number of apparent errors in our database. We
re-examined our data, making corrections when necessary, and updated our
analyses; these corrections did not have any significant impact on our
findings. Commerce also made a number of technical comments, focusing
primarily on our description of its NME methodology. We took these
comments into consideration and made changes throughout the report to
insure its clarity and accuracy. We also made a number of technical
corrections suggested by the Department of Homeland Security and the
Office of the U.S. Trade Representative.

We are sending copies of this report to the Secretaries of Commerce and
Homeland Security, the International Trade Commission, the U. S. Trade
Representative, appropriate congressional committees, and other interested
parties. We also will make copies available to others upon request. In
addition, the report will be available at no charge on the GAO Web site at
http://www.gao.gov .

If you or any of your staff have any questions about this report, please
contact me at 202-512-4347 or [email protected] . Contact points for our
Offices of Congressional Relations and Public Affairs may be found on the
last page of this report. GAO staff who made major contributions to this
report are listed in appendix V.

Loren Yager Director, International Affairs and Trade

Scope and Methodology Appendix I

To address our objectives, we examined and summarized applicable U.S. laws
and regulations, as well as relevant World Trade Organization (WTO)
agreements. These included the Agreement on Implementation of Article VI
of the General Agreement on Tariffs and Trade 1994-commonly known as the
"antidumping agreement"-and China's WTO accession agreement. We conducted
a literature search and reviewed relevant scholarly and legal analyses and
Department of Commerce (Commerce) determinations.1 In order to corroborate
and broaden our understanding, we consulted with trade and legal policy
experts from the U.S. government, private sector trade associations,
consulting firms, academic institutions, law firms with broad experience
in trade actions involving China, as well as representatives of the WTO,
the government of China, and other governments concerned about Chinese
trade practices, including the European Union and Canada.

In order to analyze the application of antidumping (AD) duties to China
and compare duty rates applied to China with those applied to market
economy countries (our second objective) and to evaluate the potential
impact of ceasing to apply the nonmarket economy (NME) methodology to
China (our third objective), we collected information from the Department
of Commerce and the International Trade Commission, including notices of
Commerce determinations appearing in the Federal Register. We used this
information to construct a database on all U.S. AD investigations from
1980 through 2004. In addition to information on the countries and
products involved and the status of each investigation, our database
included the duty rates applied upon completion of each new antidumping
investigation against China during this period, as well as the duty rates
applied against any producers of the same products from other countries.
This database is accessible on-line at www.gao.gov/cgi-bin/getrpt?
GAO-06-652SP . We verified this database to the official sources and found
the data to be sufficiently reliable for the purposes of this report.

Our analyses focused on the 68 cases during this time period wherein
Commerce imposed AD duties on Chinese products, and especially on the
subset of 25 cases in which Commerce imposed duties against a similar

product from one or more market economy countries.2 Specifically, the 25
cases included all market economy cases that had the same product name and
were initiated within 1 year of an AD investigation against China. In all,
we assembled data on 303 company-specific, weighted average, and
country-wide duty rate determinations on Chinese products, and an
additional 168 duty rate determinations on market economy products.
Appendix II provides additional analyses of this data.

As part of our examination, we also performed multivariate regression
analyses to determine the extent to which duty rate variations could be
attributed to differences between China and these other countries, or to
other factors, such as the type of product involved. Appendix III provides
more information on these analyses and their results.

In addition to comparing China and market economies, we also collected
information on duty rates that Commerce applied to products from other NME
countries at the same time as it applied them to similar products from
either China or a market economy. Appendix III provides information on the
results of our analyses of this data.

We did not collect or analyze information on duty rates applied against
market economy countries in cases where no parallel action was taken
against China or any other NME country. Therefore, our analyses of market
economy duty rates are specific to the sample of market economy orders in
which a corresponding NME order was also in effect. Inclusion of other
market economy product duty rates may have produced different results.
However, we determined that the appropriate comparison between China and
market economy countries was between the 25 similar products. We found
through our regression analyses (discussed in app. III) that the product
being investigated does help explain the variation among rates and it is,
therefore, important to make an appropriate comparison. In addition, duty
rates for the 43 remaining orders against China alone followed a similar
pattern as those contained in the 25 cases where we drew comparisons with
market economy duty rates. The average country-wide rate for these 43
orders against China was higher than the country-wide rate for the 25
orders (118 percent compared to 98 percent), and the average individual
rate was lower (41 percent compared to 53 percent) for the 18 orders with
individual rates. These results were consistent with our findings that the
country-wide rates tend to be significantly higher than individual rates.

In order to group specific products subject to AD orders into groups of
similar products, we used the Harmonized Tariff Schedule (HTS)
classifications for each product, as reported in the Federal Register
announcement of the order. The HTS is the official U.S. classification of
goods imported into the United States and includes 99 chapters covering
all goods imports. In addition, the HTS chapters are grouped into larger
sections that cover broad types of related products. The categories we
used in this report are based on these HTS sections and chapters.
Specifically, the category "Chemicals, plastics, pharmaceuticals"
comprises HTS chapters 28 through 40 (which includes all chapters under
the section "Chemical or allied industries"). The category "Steel, other
metals" comprises HTS chapters 72 through 81 (which includes most chapters
under the section "Base metals and articles of base metals" except those
chapters covering articles of base metals). The category "Agricultural
products" comprises HTS chapters 1 through 24 (which includes all chapters
under the sections "Live animals; animal products," "Vegetable products,"
"Animal or vegetable fats, etc.," and "Prepared foodstuffs, beverages,
spirits, and vinegar; tobacco and manufactured tobacco substitutes"). The
category "Other products" comprises all other HTS chapters.

We conducted our work from June 2005 through December 2005 in accordance
with generally accepted government auditing standards.

Additional Information on Duty Rates Applied to China and Market Economy
Countries Appendix II

This appendix provides additional information on the antidumping (AD) duty
rate data that we assembled for this report and provides some additional
analytical information, including brief discussions of variation in the
duty rates applied to China over time, Department of Commerce (Commerce)
determinations on whether Chinese companies should be considered eligible
for individual rates, and duty rates applied to selected market economy
countries.

Duty Rates Imposed on China

The overall average duty rate for all 68 orders against China from 1980
through 2004 was 65 percent. This was the result of 72 country-wide rates
(on 68 products) with an average duty of 111 percent and 158 individual
company rates with an average duty of 44 percent.1 These rates ranged from
zero to about 384 percent (see table 3). In our analysis, we identified 25
orders against China in which there was also an order against a market
economy country on the same product put in place within 1 year from the
order against China. Table 3 shows overall average duty rates from the 25
orders against China that were matched to market economy cases and the 43
orders in which no market economy order was identified. Table 4 at the end
of this appendix provides information on each of the 68 orders against
China, and table 5 provides comparative information for each of the 25
cases in which duties were also applied against market economy producers.

Table 3: Summary Data on China AD Duty Orders, 1980-2004

                                        

                               Type of rate    Number of Mean Minimum Maximum 
                                                   rates                      
                                                          (%)     (%)     (%) 
25 orders matched to     Individual company        50   52       0     162 
similar market economy   rates                                     
orders                   Country-wide rate         25   98       3     218 
                            Total                     75   67       0     218 
43 orders not matched to Individual company       108   41       0     292 
similar market economy   rates                                     
orders                   Country-wide rate         47  118       1     384 
                            Total                    155   64       0     384 
All 68 orders            Individual company       158   44       0     292 
                            rates                                     
                            Country-wide rate         72  111       1     384 
                            Grand total              230   65       0     384 

Source: GAO AD database.

Note: The overall average individual duty rates listed in this table for
the "matching" orders are for the 25 products with both China and market
economy orders. However, only 18 of these products had individual rates
imposed on both China and at least one other market economy. We use these
18 products for our comparison of individual company duty rates imposed on
China and market economies, rather than the 25 products, in the rest of
the report. The average individual company duty rate imposed on China for
these 18 products is 53 percent, rather than 52 percent.

About 78 percent (53 AD orders) of the 68 AD orders included not only
country-wide rates but also individually calculated rates for at least one
Chinese company. Of these, about 54 percent (37 orders) included
company-specific rates that were lower than the country-wide rates imposed
in the same cases. With regard to nonmalleable cast iron pipe fittings,
for example, two Chinese companies submitted detailed information and met
Commerce's criteria for assignment of individually determined rates. Other
Chinese pipe fitting companies, however, did not provide any information.
Commerce assigned the two cooperating companies duty rates of between 6
and 8 percent-a fraction of the 76 percent country-wide duty rate applied
in this case.2

Only 15 orders issued against China during this period included just a
country-wide rate. Most of these orders date from the period before 1991
when Commerce had not yet begun applying its separate rates test. However,
from 1991 through 1995 Commerce issued six orders that contained only a
country-wide rate. In most of these cases, Chinese companies failed to
respond to Commerce requests for information. For example, in one case
Commerce solicited information through both the Chinese government and the
relevant Chinese industry association. However, the industry association
responded that no Chinese producer or exporter wanted to participate in
Commerce's investigation. Commerce, therefore, used facts available to
establish a country-wide duty rate of about 156 percent.3

In 12 of the 68 orders, all the individual rates issued were equal to the
country-wide rate. In some cases, Commerce specified an individual rate
for one company and then used this rate as "facts available" to establish
a country-wide duty rate at the same level.4 For example, in its
investigation of refined brown aluminum oxide from China, Commerce
requested information from the government of China and more than 20
Chinese companies. Only one of these companies responded. Commerce found
that this company qualified for its own duty rate and determined that this
rate should be about 135 percent. Commerce determined that the failure of
the other companies to provide requested information justified application
of an adverse inference to determine the country-wide rate. Since the rate
established for the lone cooperating company was higher than any of the
rates suggested in the petition requesting imposition of duties on this
product, Commerce set the country-wide rate equal to the rate applied to
the one cooperating company-135 percent.5

Duty Rates against China Have Fluctuated over Time

We found that there was a slight tendency for duty rates applied against
Chinese products to rise over the period of our analysis, as well as to
fluctuate over time. As figure 9 shows, individual company and
country-wide duty rates tended to be larger from 1992-2004 than from
1980-1991. In addition, the individual company rates demonstrate a
cyclical pattern over time. In our regression analysis, we found that
there was a small positive trend in AD duty rates against China over time
that was statistically significant. This result is consistent with
research that has shown that overall U.S. AD margins have increased over
time.6

Figure 9: Average Country-Wide and Individual Rates, All Orders against
China, 1980-2004

Note: Commerce did not issue any new AD duty orders against China in 1980,
1981, 1982, 1985, 1987, 1988, 1989, or 1998. In the tapered roller
bearings cases, Commerce put an order in place in 1987 but amended it in
1990. We use 1990 information in the above graphic.

Average Duty Rates on 68 Orders against China and Subset of 25 Orders
Matched to Market Economy Orders

Table 4 shows the duty rates on the 68 orders imposed on China between
1980 and 2004. Table 5 then shows the duty rates on the 25 orders imposed
on China in which we also found matching orders imposed on market
economies.

Table 4: Average Duty Rates for 68 Orders against China, 1980-2004

                                        

                                          Individual 
                                           company   
                                            rates    
 Year and product   Average      Average   Number of     Average Minimum Maximum 
                       (all country-wide       rates         (%)     (%)     (%) 
                     rates)     rate (%)                                 
                        (%)                                              
(1983) Cotton shop       35           36           2  34      30      37 
towels                                                                   
(1983) Greig             22           22                                 
polyester cotton                                                         
print cloth                                                              
(1984) Barium            15           15           1  15      15      15 
chloride                                                                 
(1984)                   58           58           1  58      58      58 
Chloropicrin                                                             
(1984) Potassium         40           40           1  40      40      40 
permanganate                                                             
(1986) Iron              12           12                                 
construction                                                             
castings                                                                 
(1986) Paint            127          127                                 
brushes                                                                  
(1986)                   67           67                                 
Porcelain-on-steel                                                       
cookware                                                                 
(1986) Steel wire         6            6                                 
nails                                                                    
(1986) Wax candles       54           54           1  54      54      54 
(1990) Industrial        78           78           1  78      78      78 
nitrocellulose                                                           
(1990) Tapered            3            3           2   3       1       5 
roller bearings                                                          
(1991)                   42           42           1  42      42      42 
Chrome-plated lug                                                        
nuts                                                                     
(1991) Hand tools        36           36                                 
(1991) Oscillating        1            1           8   1       0       2 
fans and ceiling                                                         
fans                                                                     
(1991) Silicon          139          139                                 
metal                                                                    
(1991) Sparklers         59           76           3  54       2      94 
(1991) Sulfur            28           28                                 
chemicals                                                                
(1991) Tungsten         151          151                                 
ore concentrates                                                         
(1992) Carbon           114          183           6 102      35     155 
steel butt-weld                                                          
pipe fittings                                                            
(1992) Sulfanilic        52           85           1  19      19      19 
acid                                                                     
(1993) Compact          127          127                                 
ductile iron                                                             
waterworks                                                               
(1993)                  138          138                                 
Ferrosilicon                                                             
(1993) Helical           89          129           2  70      70      70 
spring lock                                                              
washers                                                                  
(1994) Cased             14           45           4   6       0      17 
pencils                                                                  
(1994) Garlic           377          377                                 
(1994) Paper clips       73          127           3  55      46      61 
(1994) Sebacic           98          243           4  61      44      85 
acid                                                                     
(1994)                  150          150                                 
Silicomanganese                                                          
(1995) Furfuryl          46           45           2  47      44      50 
alcohol                                                                  
(1995) Glycine          156          156                                 
(1995) Pure             108          108                                 
magnesium 1                                                              
(1996) Manganese         33          143           4   6       1      12 
metal                                                                    
(1996) Polyvinyl         78          117           2  58       0     117 
alcohol 1                                                                
(1997) Brake             10           43           6   5       0      16 
rotors                                                                   
(1997) Collated          39          118           2   0       0       0 
roofing nails                                                            
(1997) Coumarin          87          161           2  51      31      70 
(1997) Crawfish         133          202           5 120      92     157 
(1997) Melamine           2            7           4   1       0       3 
dinnerware                                                               
(1997) Persulfates       55          119           3  34      32      35 
(1999) Mushrooms        155          199           4 144     121     162 
(2000) Apple juice       19           52           7  14       0      28 
(2000) Bulk              57          144           2  14      11      17 
aspirin                                                                  
(2000) Creatine          47          154           6  30       0      58 
(2000) Synthetic         96          130           2  80      80      80 
indigo                                                                   
(2001) Certain           76           91           4  73      64      91 
hot-rolled carbon                                                        
steel flat                                                               
products                                                                 
(2001) Foundry          109          215           4  83      49     106 
coke products                                                            
(2001) Honey             72          184           4  45      26      57 
(2001) Pure             165          306           1  25      25      25 
magnesium 2                                                              
(2001) Steel            133          133           1 133     133     133 
concrete                                                                 
reinforcing bars                                                         
(2002) Automotive        37          125           3   8       4      12 
replacement glass                                                        
windshields                                                              
(2002) Certain           58          165           2   5       2       9 
folding gift boxes                                                       
(2002) Folding           28           71           2   7       0      14 
metal tables and                                                         
folding metal                                                            
chairs                                                                   
(2003) Barium            58           81           1  34      34      34 
carbonate                                                                
(2003) Certain           31          111           4  11       7      16 
malleable iron                                                           
pipe fittings                                                            
(2003) Cut to            62          129           5  49      17     129 
length carbon                                                            
steel plate                                                              
(2003)                   40           67           1  13      13      13 
Ferrovanadium                                                            
(2003) Lawn and           6           16           3   2       0       7 
garden steel fence                                                       
posts                                                                    
(2003)                   30           76           2   7       6       7 
Non-malleable cast                                                       
iron pipe fittings                                                       
(2003) Polyvinyl         52           98           1   7       7       7 
alcohol 2                                                                
(2003) Refined          135          135           1 135     135     135 
brown aluminum                                                           
oxide                                                                    
(2003) Saccharin        288          330           3 274     249     292 
(2004) Carbazole         74          218           3  26       6      45 
violet pigment 23                                                        
(2004) Certain           27           78           5  17       5      26 
color television                                                         
receivers                                                                
(2004) Hand trucks      105          384           4  35      26      46 
(2004) Ironing           99          158           3  80       9     158 
tables                                                                   
(2004)                   27           78           8  21       0      41 
Polyethylene                                                             
retail carrier                                                           
bags                                                                     
(2004)                  137          137           1 137     137     137 
Tetrahydrofurfuryl                                                       
alcohol                                                                  
Total                    65          111         158  44       0     292 

Source: GAO AD database.

Note: The average (all rates) is calculated as the average of the
country-wide rate and each of the individual company rates that Commerce
issued in its order.

Table 5: Average Duty Rates for 25 Products with Orders against both China
and Market Economies, 1980-2004

                                        

                                                   Individual 
                                                    company   
                                                     rates    
 Product and year   Country  Average      Average   Number of Average Minimum Maximum 
                      and            country-wide       rates                         
                   number of    (all      or all-                 (%)     (%)     (%) 
                    orders    rates)  others rate                             
                                                                              
                                 (%)          (%)                             
Carbazole violet   China          74          218           3      26       6      45 
pigment 23 (2004)  Market         29           27           2      30      10      50 
                   economies                                                  
                   (1)                                                        
Carbon steel       China         114          183           6     102      35     155 
butt-weld pipe     Market         25           39           3      21       0      51 
fittings (1992)    economies                                                  
                   (1)                                                        
Certain hot-rolled China          76           91           4      73      64      91 
carbon steel flat  Market         24           23          11      25       3      48 
products (2001)    economies                                                  
                   (7)                                                        
Chrome-plated lug  China          42           42           1      42      42      42 
nuts (1991)        Market          8            7           2       9       6      11 
                   economies                                                  
                   (1)                                                        
Collated roofing   China          39          118           2       0       0       0 
nails (1997)       Market         22            3           3      28       3      40 
                   economies                                                  
                   (1)                                                        
Ferrosilicon       China         138          138           0                 
(1993)             Market         27           23           4      29       3      89 
                   economies                                                  
                   (2)                                                        
Ferrovanadium      China          40           67           1      13      13      13 
(2003)             Market        116          116           2     116     116     116 
                   economies                                                  
                   (1)                                                        
Furfuryl alcohol   China          46           45           2      47      44      50 
(1995)             Market         10           10           2      10       8      12 
                   economies                                                  
                   (2)                                                        
Helical spring     China          89          129           2      70      70      70 
lock washers       Market         32           32           3      32      32      32 
(1993)             economies                                                  
                   (1)                                                        
Honey (2001)       China          72          184           4      45      26      57 
                   Market         36           30           3      38      27      55 
                   economies                                                  
                   (1)                                                        
Industrial         China          78           78           1      78      78      78 
nitrocellulose     Market         37           37           6      37       4      66 
(1990)             economies                                                  
                   (6)                                                        
Iron construction  China          12           12           0                 
castings (1986)    Market         11           12          10      11       0      59 
                   economies                                                  
                   (3)                                                        
Melamine           China           2            7           4       1       0       3 
institutional      Market         18            6           6      22       0      53 
dinnerware (1997)  economies                                                  
                   (2)                                                        
Polyethylene       China          27           78           8      21       0      41 
retail carrier     Market         81           44          10      88       2     123 
bags (2004)        economies                                                  
                   (2)                                                        
Polyvinyl alcohol  China          78          117           2      58       0     117 
1 (1996)           Market         63           48           6      68      19      77 
                   economies                                                  
                   (2)                                                        
Polyvinyl alcohol  China          52           98           1       7       7       7 
2 (2003)           Market        103           54           5     123      39     144 
                   economies                                                  
                   (2)                                                        
Porcelain-on-steel China          67           67           0                 
cooking ware       Market         15           17           8      15       2      57 
(1986)             economies                                                  
                   (2)                                                        
Potassium          China          40           40           1      40      40      40 
permanganate       Market          5            5           1       5       5       5 
(1984)             economies                                                  
                   (1)                                                        
Preserved          China         155          199           4     144     121     162 
mushrooms (1999)   Market         86           57           7      98       6     244 
                   economies                                                  
                   (3)                                                        
Silicomanganese 1  China         150          150           0                 
(1994)             Market         41           18           1      65      65      65 
                   economies                                                  
                   (1)                                                        
Silicon metal      China         139          139           0                 
(1991)             Market         58           50           3      63       9      93 
                   economies                                                  
                   (2)                                                        
Steel concrete     China         133          133           1     133     133     133 
reinforcing bars   Market         57           37          12      64      17     102 
(2001)             economies                                                  
                   (4)                                                        
Sulfanilic acid    China          52           85           1      19      19      19 
(1992)             Market        115          115           0                 
                   economies                                                  
                   (1)                                                        
Sulfur chemicals   China          28           28           0                 
(1991)             Market         75           75           2      75      50     100 
                   economies                                                  
                   (2)                                                        
Tapered roller     China           3            3           2       3       1       5 
bearings (1987)    Market         54           65           2      36      36      37 
                   economies                                                  
                   (3)                                                        
Total              China          67           98          50      52       0     162 
                   Market         44           37         114      48       0     244 
                   economies                                                  
                   (54)                                                       

Source: GAO AD database.

Notes:

The average (all rates) is calculated as the average of the country-wide
rate and each of the individual company rates that Commerce issued in its
order.

The overall average individual duty rates listed in this table are for the
25 products with both China and market economy orders. However, only 18 of
these products had individual rates imposed on both China and at least one
other market economy. We use these 18 products for our comparison of
individual company duty rates imposed on China and market economies,
rather than the 25 products, in the rest of the report. The average
individual company duty rate imposed on China for these 18 products is 53
percent and for market economies it is 55 percent.

Regression Analysis Results Appendix III

In order to examine the difference between duty rates applied to China and
those applied to market economy countries, we performed multivariate
regression analyses on the cases in which the Department of Commerce
(Commerce) applied duties to both China and at least one market economy
country. These involved 25 different products, affected by 25 duty orders
against China, and 54 duty orders against market economies. Multivariate
regression analysis makes it possible to examine the simultaneous effect
of several different factors on the duty rates and to determine the extent
to which these factors, taken together, explain variation in these rates.
To determine whether our analytical results for China held true for all
nonmarket economy (NME) countries, we also identified six instances in
which Commerce applied duties to a nonmarket economy other than China, and
at least one market economy country, and reran our analyses using data for
all 31 products.

Table 6 shows the results of our multivariate regression analysis of
variation in the dependent variable (the antidumping [AD] duty rate)
attributable to the following independent variables:

o China (a variable indicating whether the AD duty rate is for China or
not)

o the country-wide rate (a variable indicating whether the AD duty rate is
a country-wide rate), and

o year (a variable indicating the year in which the duty went into
affect).

We also included a constant term. The regression involved 25 products
covered by 25 orders against China and 54 orders against market economies
and included a total of 243 duty rates (the dependent variable) from these
79 orders.

The results show that the variable for China as the target country had a
coefficient of 3.002 percent, indicating that duty rates against China 
tended to be about 3  percentage points higher than those against market
economies, on average. However, this coefficient is not statistically
significant, meaning that there was no statistically significant
difference between the rates assigned to China and market economy
countries, when the other factors in the regression are included. The
coefficient for the country-wide rate, on the other hand, shows that there
is a 52 percentage point difference between country-wide rates against
China and other rates. This result is statistically significant at above
the 99 percent level. The variable for the year of the order is also
statistically significant, but it has a small coefficient.1 The adjusted
R-square measure shows that about 15 percent of the overall variation in
duty rates is explained by the independent variables included here.

Table 6: Results of Multivariate Regression Analysis of Duty Rates on
Explanatory Variables for China, Country-Wide Rates, and Year

                                        

    Dependent   
    variable =  
AD duty rate 
Independent  Unstandardized Standard Standardized t-statistic Significance 
    variables     coefficients    error coefficients                       of 
                                              (Beta)              t-statistic 
                           (B)                                   
China                 3.002    7.511         .028        .400         .690 
Year                  2.095     .497         .252       4.218         .000 
Country-wide         52.050   11.398         .316       4.567         .000 
rate                                                          

Source: GAO analysis of Commerce data.

Note: R-square = 0.164; Adjusted R-square = 0.154; Observations = 243.

We then included additional variables for product groups, such as
agriculture and steel, and, in separate regressions, individual product
variables for each type of product. The additional variables generally
improved the overall "fit" of the regression; the adjusted R-square
measure with the individual product variables included showed that the
regression explained between 24 and 31 percent of the overall variation in
duty rates across the sample compared with 15 percent in the regression
above. Also, certain types of products, such as agriculture products,
tended to have higher duty rates relative to other types.

Table 7 shows the regression results when individual product variables are
included. Once again the coefficient for China is insignificant, while the
coefficient for the country-wide rate is significant at the 99 percent
level. Some coefficients for individual products are significant (e.g.,
carbon steel butt-weld pipe fittings), but many are not. The overall
adjusted R-square measure shows that this regression model explains about
31 percent of total variation in the duty rates.

Table 7: Results of Multivariate Regression Analysis of Duty Rates on
Explanatory Variables for China, Country-Wide Rates, and Individual
Products

                                        

Dependent variable 
  = AD duty rate   
Independent     Unstandardized Standard Standardized t-statistic Significance 
    variables        coefficients    error coefficients                       of 
                                                 (Beta)              t-statistic 
                              (B)                                                
China                       2.055    7.302         .019        .281         .779 
Year                       -6.022   20.636        -.726       -.292         .771 
Country-wide rate          50.928   10.609         .309       4.801         .000 
Carbazole violet          113.617  269.338         .380        .422         .674 
pigment 23 (2004)                                                                
Carbon steel               70.893   30.579         .294       2.318         .021 
butt-weld pipe                                                                   
fittings (1992)                                                                  
Certain hot-rolled         82.479  207.275         .482        .398         .691 
carbon steel flat                                                                
products (2001)                                                                  
Collated roofing           46.480  126.162         .155        .368         .713 
nails (1997)                                                                     
Ferrosilicon               40.090   58.403         .134        .686         .493 
(1993)                                                                           
Ferrovanadium             136.023  249.034         .386        .546         .585 
(2003)                                                                           
Furfuryl alcohol           30.585   86.044         .102        .355         .723 
(1995)                                                                           
Helical spring             49.701   47.927         .166       1.037         .301 
lock washers                                                                     
(1993)                                                                           
Honey (2001)               98.896  207.608         .373        .476         .634 
Industrial                 21.805   30.079         .101        .725         .469 
nitrocellulose                                                                   
(1990)                                                                           
Iron construction         -33.870  105.554        -.158       -.321         .749 
castings (1986)                                                                  
Melamine                   32.449  125.663         .146        .258         .796 
institutional                                                                    
dinnerware (1997)                                                                
Polyethylene              122.291  268.962         .686        .455         .650 
retail carrier                                                                   
bags (2004)                                                                      
Polyvinyl alcohol          81.135  105.532         .337        .769         .443 
1 (1996)                                                                         
Polyvinyl alcohol         147.492  248.661         .556        .593         .554 
2 (2003)                                                                         
Porcelain-on-steel        -25.756  105.698        -.107       -.244         .808 
cooking ware                                                                     
(1986)                                                                           
Potassium                 -44.131  147.104        -.112       -.300         .764 
permanganate                                                                     
(1984)                                                                           
Preserved                 141.530  163.666         .680        .865         .388 
mushrooms (1999)                                                                 
Silicomanganese 1          67.142   69.110         .148        .972         .332 
(1994)                                                                           
Silicon metal              51.868   25.308         .161       2.049         .042 
(1991)                                                                           
Steel concrete            111.738  207.339         .584        .539         .591 
reinforcing bars                                                                 
(2001)                                                                           
Sulfanilic acid            51.956   41.151         .115       1.263         .208 
(1992)                                                                           
Sulfur chemicals           44.355   26.424         .126       1.679         .095 
(1991)                                                                           
Tapered roller             -7.471   85.958        -.027       -.087         .931 
bearings (1987)                                                                  

Source: GAO analysis of Commerce data.

Note: R-square = 0.386; Adjusted R-square = 0.309; Observations = 243.

In order to examine the effect of applying adverse inferences and facts
available (other than adverse inferences) on the duty rates, we added
additional variables indicating when Commerce used these approaches. The
results show that application of adverse inferences is a significant
variable and has a large effect on the duty rates, but that application of
facts available (other than adverse inferences) is not. When adverse
inferences are introduced, this results in the country-wide rate variable
becoming insignificant (see table 8). However, this is likely due to the
fact that the adverse inferences variable is highly correlated with the
country-wide rate. Therefore, it is not surprising that the country-wide
rate is no longer significant since the adverse inferences variable is
already accounting for much of the variation. In addition, the variable
for China once again becomes significant. As we discuss in the body of
this report, Commerce uses adverse inferences in very few determinations
for Chinese companies granted their own rates. Adverse inferences were
applied in making only 3 out of the 50 individual determinations used in
this analysis. However, Commerce used adverse inferences in nearly half of
its determinations against individual market economy companies.

Since adverse inferences are already factored into this model, as is the
country-wide rate, the remaining differences accounted for by the China
variable in table 3 are between individual (noncountry-wide) Chinese rates
and individual market economy rates in which adverse inferences are not
used. Table 8 shows that there is a statistically significant 27
percentage point difference between these rates. However, because there
are methodological differences between the NME and market economy
methodologies for individual companies, it is not clear how often adverse
inferences would be used against individual Chinese companies should they
move to a market economy methodology. In other words, we cannot predict
the extent to which, under a market economy methodology, individual
Chinese companies would cooperate with Commerce or Commerce would find it
necessary to use adverse inferences in its determinations against Chinese
companies. It is possible that some Chinese companies that currently have
an individually determined rate under the NME methodology would face
adverse inferences under a market economy methodology, whereas others
would not. This could produce a result similar to the market economy cases
we have examined in which the overall average (for example, 55 percent) is
the result of some companies receiving comparatively high duty rates
(e.g., 77 percent) when adverse inferences are used and others receiving
comparatively low rates (e.g., 16 percent) when adverse inferences are not
used (see table 2). In any case, these results show that there is a
remaining difference between these two groups after accounting for the use
of adverse inferences and the country-wide rate.

Table 8: Results of Multivariate Regression Analysis of Duty Rates on
Explanatory Variables for China, Country-Wide Rates, Adverse Inferences,
Other Facts Available, and Individual Products

                                        

Dependent variable 
  = AD duty rate   
Independent     Unstandardized Standard Standardized t-statistic Significance 
    variables        coefficients    error coefficients                       of 
                                                 (Beta)              t-statistic 
                              (B)                                   
China                      27.342    6.856         .252       3.988         .000 
Year                        4.831   17.572         .582        .275         .784 
Country-wide rate           9.910   10.064         .060        .985         .326 
Adverse inferences         60.196    6.845         .587       8.794         .000 
Facts available            10.135    8.495         .069       1.193         .234 
Carbazole violet          -40.728  229.493        -.136       -.177         .859 
pigment 23 (2004)                                                   
Carbon steel               39.749   26.189         .165       1.518         .131 
butt-weld pipe                                                      
fittings (1992)                                                     
Certain hot-rolled        -57.704  176.678        -.337       -.327         .744 
carbon steel flat                                                   
products (2001)                                                     
Collated roofing          -45.478  107.679        -.152       -.422         .673 
nails (1997)                                                        
Ferrosilicon               -3.176   49.788        -.011       -.064         .949 
(1993)                                                              
Ferrovanadium             -40.348  212.436        -.114       -.190         .850 
(2003)                                                              
Furfuryl alcohol          -13.868   73.344        -.046       -.189         .850 
(1995)                                                              
Helical spring            -18.938   41.354        -.063       -.458         .647 
lock washers                                                        
(1993)                                                              
Honey (2001)              -31.946  176.875        -.121       -.181         .857 
Industrial                  8.668   25.589         .040        .339         .735 
nitrocellulose                                                      
(1990)                                                              
Iron construction          25.459   89.838         .118        .283         .777 
castings (1986)                                                     
Melamine                  -53.825  107.215        -.242       -.502         .616 
institutional                                                       
dinnerware (1997)                                                   
Polyethylene              -55.280  229.336        -.310       -.241         .810 
retail carrier                                                      
bags (2004)                                                         
Polyvinyl alcohol          -5.196   90.169        -.022       -.058         .954 
1 (1996)                                                            
Polyvinyl alcohol         -22.256  211.996        -.084       -.105         .916 
2 (2003)                                                            
Porcelain-on-steel         32.037   89.998         .133        .356         .722 
cooking ware                                                        
(1986)                                                              
Potassium                  33.392  125.194         .085        .267         .790 
permanganate                                                        
(1984)                                                              
Preserved                  32.301  139.519         .155        .232         .817 
mushrooms (1999)                                                    
Silicomanganese 1           3.633   59.124         .008        .061         .951 
(1994)                                                              
Silicon metal              39.950   22.201         .124       1.799         .073 
(1991)                                                              
Steel concrete            -41.334  176.900        -.216       -.234         .815 
reinforcing bars                                                    
(2001)                                                              
Sulfanilic acid            -1.893   35.427        -.004       -.053         .957 
(1992)                                                              
Sulfur chemicals           -8.757   23.151        -.025       -.378         .706 
(1991)                                                              
Tapered roller             12.952   73.025         .046        .177         .859 
bearings (1987)                                                     

Source: GAO analysis of Commerce data.

Note: R-square = 0.562; Adjusted R-square = 0.502; Observations = 243.

In order to examine whether the above results hold for all NMEs, we ran
the same regressions for a larger set of 31 products (compared with the 25
products above) in which we found matching cases between nonmarket
economies other than China and market economies. The data set on these 31
products included rates from 128 orders (26 on China, 82 on market
economies, and 20 on NMEs other than China) that contained 355 duty rates
(dependent variable).

These analyses confirmed our China-market economy only analyses but also
showed that other NME countries tend to have duty rates that are
statistically higher than market economy rates for this sample of matching
cases. (Note that the number of additional products-six-is relatively
small.) Controlling for both the NME designation and the country-wide
rate, the NME designation itself is a significant variable at the 97
percent level of confidence with a coefficient of 23 percent (the
coefficient for China is not statistically significant). The country-wide
variable is also significant (99 percent level) and larger with a
coefficient of 48 percent. As additional variables are added for
individual products, the NME designation continued to be significant along
with the country-wide rate variable.

There may be other systematic factors not accounted for in this regression
model that would explain some of the variability not accounted for by the
variables we included. As shown in table 7, our model accounted for about
50 percent (half) of the variation in rates. Some of this variation may be
idiosyncratic and related to differences in individual companies'
practices, other may relate to how Commerce has implemented its analysis.
However, these unexplained factors do not appear to be systematically
related to whether the case involved China or a market economy since the
regression analysis already controls for that difference.

Comments from the Department of Commerce Appendix IV

The following are GAO's comments on the Department of Commerce's letter
dated December 8, 2005.

GAO Comments

1.We re-examined our data, making corrections as appropriate, and updated
our analyses. The report reflects these corrections, though they did not
have a significant impact on any of our findings.

2.As discussed in the report, the overall difference between the duty
rates applied to China and those applied to market economy countries is
largely explained by the application of comparatively high country-wide
rates to China. Therefore, the model allows us to conclude that
elimination of the NME methodology-and thus these country-wide rates-would
result in lower duties for some Chinese companies. Nevertheless, there
would still be variation in duty rates among companies and products due to
a range of other factors.

GAO Contact and Staff Acknowledgments Appendix V

Loren Yager (202) 512-4347

In addition to the individual named above, Adam R. Cowles, Monica Ghosh,
R. Gifford Howland, Michael McAtee, Richard Seldin, Ross Tuttleman,
Roberto Walton, and Timothy Wedding made significant contributions to this
report.

(320327)

www.gao.gov/cgi-bin/getrpt? GAO-06-231 .

To view the full product, including the scope

and methodology, click on the link above.

To view the database on antidumping case information against Chinese
companies, click on the following link: www.gao.gov/cgi-bin/getrpt?
GAO-06-652SP . For more information, contact Loren Yager at (202) 512-4347
or [email protected].

Highlights of GAO-06-231 , a report to congressional committees

January 2006

U.S.-CHINA TRADE

Eliminating Nonmarket Economy Methodology Would Lower Antidumping Duties
for Some Chinese Companies

U.S. companies adversely affected by unfair imports may seek a number of
relief measures, including antidumping (AD) duties. The Department of
Commerce (Commerce) classifies China as a nonmarket economy (NME) and uses
a special methodology that is commonly believed to produce AD duty rates
that are higher than those applied to market economies. Commerce may stop
applying its NME methodology if it finds that China warrants designation
as a market economy.

In light of increased concern about China's trade practices, the
conference report on fiscal year 2004 appropriations requested that GAO
review efforts by U.S. government agencies responsible for ensuring free
and fair trade with that country. In this report, the last in a series,
GAO (1) explains the NME methodology, (2) analyzes AD duties applied to
China and compares them with duties applied to market economies, and (3)
explains circumstances in which the United States would stop applying its
NME methodology to China and evaluates the potential impact of such a
step.

Commerce agreed with our findings, commenting that our report provides
timely and helpful information on the NME methodology and its application
to China.

Commerce's methodology for calculating AD duties on nonmarket economy
products differs from its market economy approach in that (1) since NME
prices are unreliable, it uses price information from surrogate countries,
like India, to construct the value of the imported products and (2) it
limits eligibility for individual rates to companies that show their
export activities are not subject to government control. Companies that do
not meet the criteria or do not participate in Commerce investigations
receive "country-wide" rates.

China has been the most frequent target of U.S. AD actions. On 25
occasions, Commerce has applied duties to the same product from both China
and one or more market economy. China (NME) duties were over 20 percentage
points higher than those applied to market economies, on average. This is
because average China country-wide rates were over 60 points higher than
comparable market economy rates. Individual China company rates were
similar to those assigned to market economy companies, on average.

Comparison of China and Market Economy Antidumping Rates for 25 Products
(1985-2004)

Commerce can declare China a market economy if the country meets certain
criteria, thus ending the use of surrogate price information and
country-wide rates in China AD actions. These changes would have a mixed
impact. Duties would likely decline for Chinese companies not assigned
individual rates. Individual company rates would likely diverge, with
those that do not cooperate with Commerce receiving rates that are
substantially higher than those that do cooperate. In any case, it appears
that the actual trade impact of the NME methodology will decline as the
portion of total export trade conducted by Chinese companies assigned
individual rates increases and as the country-wide rates that largely
account for the comparatively high average rates applied to China decline
in importance.

Report to Congressional Committees

January 2006

U.S.-CHINA TRADE

Eliminating Nonmarket Economy Methodology Would Lower Antidumping Duties
for Some Chinese Companies

GAO's Mission

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( www.gao.gov ). Each weekday, GAO posts newly
released reports, testimony, and correspondence on its Web site. To have
GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone

The first copy of each printed report is free. Additional copies are $2
each. A check or money order should be made out to the Superintendent of
Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more
copies mailed to a single address are discounted 25 percent. Orders should
be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Waste, and Abuse in Federal Programs

Contact:

Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: [email protected]
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional Relations

Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548

Public Affairs

Paul Anderson, Managing Director, [email protected] (202) 512-4800 U.S.
Government Accountability Office, 441 G Street NW, Room 7149 Washington,
D.C. 20548
*** End of document. ***