WPS6313
Policy Research Working Paper 6313
Antidumping, Retaliation Threats,
and Export Prices
Veysel Avsar
The World Bank
Development Economics Vice Presidency
Partnerships, Capacity Building Unit
January 2013
Policy Research Working Paper 6313
Abstract
Utilizing four-dimensional (firm-product-destination- duties on the unit price of the firms’ shipments. The
year) Brazilian firm-level export data, the paper shows findings suggest that AD activities in Brazil lead Brazilian
that antidumping (AD) duties result in a significant and exporting firms to increase their unit export prices for
dramatic increase in the unit values of the products that the named industries’ products to decrease the dumping
firms export to duty-imposing countries. Furthermore, margin and avoid the threat of retaliation by the target
it examines the effect of potential (retaliatory) AD countries.
This paper is a product of the Partnerships, Capacity Building Unit, Development Economics Vice Presidency. It is part
of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy
discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org.
The author may be contacted at veysel.avsar@antalya.edu.tr.
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Produced by the Research Support Team
Antidumping, Retaliation Threats, and Export Prices 1
Veysel Avsar 2
1
I am indebted to Marc Muendler for kindly inviting me to UC San Diego to work with the SECEX and RAIS data,
sharing his data preparation codes, providing computational support and answering all of my related questions. I
want to thank Lisandra Flach for helping me with the data and for many other useful suggestions. I also want to
thank the UC San Diego Department of Economics, where part of this research was conducted, for their hospitality.
I am grateful for the guidance of my advisor, Richard Chisik. This study also benefited from helpful suggestions
from Jesse Bull, Sheng Guo, Alexis Habiyaremye, Cem Karayalcin, Peter Thompson, Mehmet Ulubasoglu, three
anonymous referees, and Alain de Janvry, the editor. Financial support from Florida International University’s
Doctoral Evidence Acquisition Fellowship and Dissertation Year Fellowship is gratefully acknowledged.
2
Assistant Professor, Department of Economics, Antalya International University, Antalya, Turkey.
E-mail:veysel.avsar@antalya.edu.tr
The last two decades have witnessed rising administrative protection via antidumping
(AD) measures. When countries negotiate lower tariffs in trade agreements, domestic industries
that desire protection against imports can employ several methods to gain temporary protection.
The most popular of these methods is to claim that the trade partner is dumping or selling below
the “fair value.� This claim is often made and often generates temporary protection, even if it is
not true (Konings and Vandenbussche 2008; Aggarwal 2007). AD is an effective loophole that
has been exploited by both developed and developing countries. 1 According to the records of the
Global Antidumping Database of the World Bank (Bown 2010), roughly 4,500 AD petitions
have been filed in the last 20 years by more than 40 countries.
This study empirically examines the pricing effect of AD duties using highly
disaggregated firm-level Brazilian export data. First, we examine the effect of AD duties that are
imposed on Brazilian exporters on the export prices that they charge in their trading partners’
markets. Second, we investigate the effect of potential (retaliatory) AD measures on the export
prices of products that Brazilian firms export to markets that may file AD petitions against the
firms. When an AD petition is filed by domestic industries and exporters’ flexibility to price
discriminate between the home and target countries is restrained because of the threat of
retaliation, we expect the exporting firms to respond by increasing the prices of their shipments
to the target country to reduce the dumping margin and avoid that threat. In other words,
potential AD duties can have price-increasing effects similar to imposed duties because of
retaliatory incentives. Our findings clearly demonstrate such an impact.
In addition to the proliferation of AD duties, several studies have analyzed their pricing
and trading effects. For instance, Prusa (2001) and Ganguli (2008) estimate the effect of AD
duties on the products being imported and show that these duties have a dramatic impact on
2
import prices and import values for the United States and India, respectively. 2 Bown and
Crowley (2006) show that imposition of AD duties also affects the prices of the targeted
countries’ products in alternative destinations. Studies examining the pricing effect of AD duties
also consider the impact of events (filing, decision, and termination) during the determination
period of AD investigations into the pricing decisions of foreign firms (Staiger and Wolak 1994;
Krupp and Polard 1996). Moreover, Prusa (1992) focuses on how AD cases might lead to
collusion between domestic and foreign firms. We contribute to this literature through the
innovative use of four-dimensional customs data (firms, products, destinations, and years) to
determine the effects of AD duties and the threat of their enforcement on export prices.
Our retaliation analysis broadly fits into the literature that investigates retaliatory
incentives in AD activity. Many existing studies propose a number of political and strategic
considerations to explain the proliferation of AD duties even though the ideal AD case aims to
hinder unfair competition in the international market. One proposed explanation is that countries
tend to retaliate against countries by which they were previously targeted in AD investigations
(Miranda et al. 1998; Prusa and Skeath 2002; Francois and Niels 2004; Feinberg and Reynolds
2006). These studies suggest that this tit-for-tat strategy in AD actions has created an “AD club,�
which consists of countries that have implemented and actively utilize AD law. Blonigen and
Bown (2003), however, provide another perspective on the debate. They suggest that the
worldwide spread of AD regimes may dampen AD activity because countries become able to
retaliate when they begin enforcing these laws. Concentrating on AD activity in the United
States, these authors show that the risk of a potential reaction from the target country decreases
the probability of an AD petition from U.S. industries as well as decreasing the likelihood of an
affirmative AD decision from the U.S. AD agency. Whether retaliation encourages or dampens
3
the spread of AD activity worldwide has been widely examined, but nothing is known about
domestic exporters’ responses to the threat of AD retaliation. To our knowledge, this is the first
paper to address this issue by documenting exporters’ price adjustments due to this threat.
Empirical works using firm-level data are scarce in the AD literature. One strand of the
literature focuses on the productivity effects of AD duties on domestic firms (Konings and
Vandenbussche 2008; Pierce 2011). In contrast, Belderbos (1997) investigates the relationship
between AD enforcement and foreign direct investment. Konings and Vandenbuscche’s (2009)
study is the only one to focus on the effect of AD duties on export sales using firm-level data. To
our knowledge, this study is the first attempt to use disaggregated firm-level data to analyze the
pricing effect of AD duties.
In terms of the effects of a country's own AD activity on its domestic exports, our
analysis of the pricing effects of retaliation threats most closely complements the study by
Konings and Vandenbussche (2009), which examines the impact of France’s AD duties on the
export sales of protected firms. Using French exporters’ firm-level customs data, they show that
foreign export sales decreased dramatically among the protected firms. They also note that this
decrease might be a consequence of the limitations of price setting in the international market,
stemming from AD protection in the home market.
Like Konings and Vandenbussche's (2009) study, our AD threat analysis focuses on the
effect of a country's AD activity on its export markets rather than its domestic market. However,
the two works differ in two key respects. First, Konings and Vandenbussche (2009) use a
difference-in-differences approach to determine the effect of AD on the exports of protected
firms compared with a control group of unprotected firms. In contrast, we focus on the threat of
retaliation resulting from ongoing AD investigations, regardless of their tentative outcomes,
4
rather than the impact of AD protection in the home market. Prusa (2001) shows that exporters
react to a tentative AD duty imposition immediately after an AD investigation is filed and that
AD filings reduce the imports of the countries named in a filing even if the complaints are not
upheld. From this point of view, beginning an investigation triggers retaliatory incentives for the
target country. 3 Additionally, our analysis is based on the trading effects of AD activity in a
particular export destination, namely, the country targeted by the domestic industries rather than
the entire international market. In this regard, the four-dimensional information on firms,
products, and export destinations in the Brazilian firm-level export data fits our research question
perfectly.
In addition to its rich firm-level export data, Brazil is a suitable country for our analysis
in terms of AD. Brazil implemented an AD law in 1987, and it ranks in the top 10 of all World
Trade Organization (WTO) members in terms of both the number of AD petitions filed and the
number of AD duties targeted against it in the post-WTO period. Brazil was the defendant
country in 26 AD cases during our sample period, which corresponds to 208 unique six-digit
Harmonized System (HS) products from 11 different industries. 4 Furthermore, Brazil filed 71
AD cases for 96 six-digit HS products, which is far above the mean and median of worldwide
AD activity.
As shown in table 2, Brazil named its major trading partners, which are also very active
AD users. Brazil seems to be a member of the “AD club�; its major trading partners also targeted
Brazil in several cases after the creation of the WTO. Therefore, the threat of AD retaliation is an
important consideration for exporters, not only because of the dramatic impact of AD duties on
trade flows, but also because of the target countries’ past AD activities against Brazil. In tables 3
and 4, we document the AD cases filed by and against Brazil by three-digit ISIC industry codes.
5
In line with worldwide AD activities, the chemical and iron-steel industries dominate the filings
in terms of the number of cases. 5
The remainder of the paper is organized as follows. The second section describes the
data, the third section presents the empirical methodology and results, and the fourth section
concludes.
<> DATA
We employed detailed export data from the Brazilian customs office, SECEX (Secretaria
de Comércio Exterior), which gathers export reports using product codes at the plant, month, and
NCM (Nomenclatura Comum do Mercosul) levels. The NCM codes are eight-digit numbers, the
first six digits of which coincide with the first six digits in the HS. The destination information is
transformed from the Brazilian country codes into the international ISO system. The product
codes at the six-digit level in the Brazilian data, for which no corresponding HS entries exist,
were removed from the data. All export values in the SECEX data are reported in current U.S.
dollars (USD), free on board (fob). We used observations on exporting plants, declared export
values, export destinations and export quantities, and we aggregated monthly plant-level export
information to years and firms. We focused solely on manufactured products. The main variable
of interest in our analysis is the unit price of the exported good p, shipped by firm f to destination
d, which is modeled as follows:
í µí°¸í µí±¥í µí±?í µí±œí µí±Ÿí µí±¡ í µí±£í µí±Ží µí±™í µí±¢í µí±’í µí±“í µí±?í µí±‘
í µí±¢í µí±“í µí±?í µí±‘ = .
í µí±„í µí±¢í µí±Ží µí±›í µí±¡í µí±–í µí±¡í µí±¦í µí±“í µí±?í µí±‘
6
Our sample covered the years from 1997 to 2000 because export quantities were only
available for this period. We also removed all observations for which the export value was zero,
and there was no information on the quantity of shipments.
The Brazilian firms’ employment data were obtained from the RAIS (Relacao Anual de
Informacoes Sociais), a collection of annual reports with individual information on workers and
employees. Similarly, we aggregated the monthly worker-plant information to years and firms. 6
The data on AD activity were obtained from the Global Antidumping Database of the
World Bank. 7 This database provides detailed case-level information on AD filings and
outcomes, such as the dates of the filing and the decision, the target country, the final decision of
the AD authority, and the HS codes of the products subject to filings. To match the six-digit HS
product codes with three-digit ISIC codes, we used the concordances in the World Bank’s Trade
Production and Protection Database. Finally, for the product-level import data, we used the UN
COMTRADE database.
<> EMPIRICAL ANALYSIS
As discussed above, we are primarily interested in the variations in Brazilian firms’ unit
export values within products across the export destinations. To quantify the effect of AD duties
and the retaliation threats on the prices of exports being shipped, we begin by estimating the
following baseline equation:
logï¿½í µí±¢í µí±“í µí±?í µí±‘í µí±¡ � = í µí»¾0 + í µí»¾1 (í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘í µí±¡ + í µí»¾2 (í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘,í µí±¡âˆ’1 + í µí¼€í µí±“í µí±?í µí±‘í µí±¡ ,
where f denotes the firm, p denotes the six-digit HS product code, and d denotes the
export destination. As noted earlier, the variable í µí±¢í µí±“í µí±?í µí±‘í µí±¡ denotes the unit values of the exported
products; (í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘í µí±¡ is a binary indicator and is unity if there is an AD duty in force in year t
7
by country d against Brazil for product p. (í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘,í µí±¡âˆ’1 is a binary indicator that takes a value
of one if Brazil filed an AD investigation against country d in the three-digit ISIC industry
category k in the previous year. We consider it plausible that firms in the same industry may be
connected via unions, trade associations, and other business organizations. Thus, retaliation may
be reflected at the industry level, although AD filings target a narrow product line. 8,9 In addition,
this variable is zero for countries that have not implemented an AD law because they have no
capacity to retaliate. We treat the European Union countries as a single country because AD is a
European Union-wide policy. 10
<> Other Considerations
We include firm-product-destination fixed effects in our specifications to control for
firm-level unobservables, such as managerial ability, to account for the heterogeneity in the
quantities’ units of measurement, which are used to create the unit export values and to control
for time-invariant, country-specific factors, such as distance. 11 Moreover, we use the total export
revenues of the firm, which measure not only the size but also the market power in the
international market. 12
In addition, as stated in Verhoogen (2008) and Hallak and Sivadasan (2008), larger and
more productive firms pay higher wages and employ better inputs, resulting in higher quality in
exports, which are sold for higher prices. Although we do not have information on domestic
sales, we use the “number of workers� obtained from the linked employment data to represent
the size of the firm. 13
Recent literature (including, but not limited to, Hallak 2006, Kneller and Yu 2008,
Harrigan and Deng 2008, Manova and Zhang 2009, Bastos and Silva 2010), on the other hand,
has documented that the unit value of shipments increases with distance, size and the income of
8
the destination country. Therefore, part of the variation in unit export values can be explained by
these country-specific covariates. In this regard, we use country-year dummies to control for all
time-varying destination market aggregates, such as GDP and exchange rate. Further, Manova
and Zhang (2009) demonstrated a positive relationship between the revenue exporters earn and
the prices that they charge in the export market. For this consideration, we also include the total
revenue that the firm earns in the export destination. 14
Another concern in estimating (2) is that export revenues and quantities may suffer from
measurement error. To remedy this concern, we removed outliers, firm-product-destination
combinations with export values below the 1st percentile or above the 99th percentile of the
distribution, to reduce the bias associated with measurement error. The extensive fixed effects in
our specifications also attenuate measurement error concerns. 15
<> RESULTS
After removing all observations lacking quantity or export value information, roughly
450,000 observations remain. 16 All of the regression results are derived from a panel-data fixed
effect estimation in which the panel is a firm-product-destination triplet. All of the specifications
include country-year dummies to control for aggregate variation in the export market. Table 5
contains the main results derived from estimating equation (2). In column (1), we report our
estimates without controls. We include our controls in the remaining specifications. The variable
log(exportrevenue)í µí±“í µí±‘í µí±¡ denotes the total revenue that firm f earns in market d,
log(exportrevenue)í µí±“í µí±¡ denotes the total export revenue of the firm, log(m)í µí±?í µí±‘í µí±¡ is the average
unit import value of the six-digit HS product p in a particular destination, and
log(employment)í µí±“í µí±¡ is the number of workers that firm f employs. The availability of
employment data determines the sample size in the last specification.
9
Table 5 lists the regression estimates. First, regarding the effect of the imposed AD
duties, we see that the coefficient of the AD duty variable in the first three columns suggests a
large and significant pricing effect. AD duties lead to a 23 percent increase in the price charged
by Brazilian exporting firms in the duty-imposing country for the products named in an AD
filing. 17
Regarding the effect of potential AD duties, the positive coefficient of the retaliation
threat variable suggests that Brazilian firms react to potential AD investigations resulting from
Brazil’s AD activity by increasing the prices of their shipments. In terms of economic
magnitude, when Brazil files an AD complaint against the importers in an industry from a
particular country, Brazilian exporters increase the price of their exported products within the
same industry to the countries named in the complaint by 10 percent because of the risk of
retaliation. Although this effect is not comparable to the actual effect of an imposed AD duty, it
is commensurate with the pricing effect of the firms’ total export sales. In addition,
complementing previous studies, the coefficient estimates of the size and revenue variables
suggest that larger firms and those that generate more revenue in a destination charge higher
prices for product varieties in the same destination. We also find a positive and significant effect
of the average product import prices at the export destinations (compared with the rest of the
world) on the unit value of Brazilian firms’ exports.
In the fourth column of table 5, we report results for the estimates in which we used two
additional lags for our policy variables. 18 In terms of the AD duty, the pricing effect is slightly
higher for the first and second years after an affirmative AD decision. In contrast, the pricing
effect of the potential AD risk decreases in the years after the AD activity, indicating that
10
exporters expect retaliation shortly after their own country’s AD activity and decrease the
dumping margin more in the first year compared with the following years.
Some readers may question whether our results are driven by the products of the steel and
chemical industries because they constitute a majority of the AD cases. To address this concern,
we interact our policy variables with steel/chemical and non-steel/chemical dummies. The last
column in table 5 reports the estimates for the regressions that include these interaction terms.
All of the interaction terms are significant. This result confirms that our results are not driven by
particular industry categories. Interestingly, firms in the steel and chemical industries are more
reactive to retaliation threats because of the political sensitivity of these industries in worldwide
AD activity.
<> FURTHER DISCUSSION ON RETALIATION
<> The Role of Export Share
In the previous section, we showed that retaliation threats affect the unit price of the
products exported by firms to countries that are subject to AD investigation in the firm’s home
country. In this section, we will address the argument that potential AD claims in an export
destination are not equally important for all firms. Exporting firms for which products in a
particular industry in the target country constitute a significant share of exports are more
responsive to the risk of retaliation. Therefore, to reduce exposure to retaliation, these firms
would be expected to decrease the dumping margin more than firms with smaller export shares
for the same product group in the target country. To test this hypothesis, we first create a variable
to measure the export share, which is calculated as the ratio of the firm’s (f) export sales of
products within the particular three-digit ISIC industry (k) in a particular country (d) to the firm’s
11
(f) total export sales. Second, we include an interaction term between the dummy variable that
denotes the risk of retaliation and the measure of the export share in our specifications.
As shown in the first column of table 6, we obtain a significant estimate for the
interaction term. The retaliation threat variable also remains significant. This result suggests that
the pricing effect of the retaliation threats is significantly increasing in the share of the exports of
the named industries’ products in the target country. To gauge the economic significance,
consider a one standard deviation increase from the mean in the export share of the firms’
products that are exposed to retaliation. Such an increase would be associated with a 3% increase
in the unit price of the exported products to the same country.
<> Exporters Serving Richer Countries
Do firms that export to developed countries have more exposure to retaliation? The
abstract in Hansen and Nielsen (2009) notes:
The GATT/WTO rules allow for antidumping measures if domestic
producers, exposed to price discrimination, also demonstrate injury
where price-undercutting is an important indicator of the latter. The
paper shows that the procedure for calculating injury is flawed due
to negligence of quality differences in the calculation of the margin
price-undercutting. This gives countries with high quality producers
an option to practice protectionism. This asymmetry between
countries in ability to implement antidumping measures
predominantly favors the developed countries which are specialized
in producing high quality products.
12
This critical view of AD investigations has important implications for our research
question. Firms that export to developed countries experience stronger competition in these
markets because firms there produce higher-quality products. Because quality differentials are
often ignored in the calculation of “fair value,� the risk of being subject to an AD investigation is
higher in these countries when exporters price discriminate to survive. 19 Following this line of
argument, when there is a focus on the threat of retaliation, exporters that serve countries that are
richer than Brazil will bear more of the risk of retaliation compared with their counterparts that
serve the less-developed world. Consequently, we expect a greater increase in the unit price of
exports for firms serving developed markets as a result of retaliation threats.
To examine this issue, we include an interaction term between our policy variable and a
binary indicator, “Rich,� which takes a value of one if the destination country is classified by the
World Bank as a “developed country.� The coefficient estimates for this specification are
reported in the second column of table 6. Both the interaction term and the variable of interest
are statistically significant. This result demonstrates a greater reaction to a threat of retaliation
from exporters that serve developed countries.
In terms of the magnitude of the effect, there is a 14 percent increase in the unit price of
firms’ exported products to rich countries because of the threat of retaliation within an industry
that is subject to AD filing in Brazil.
<> Exporters Serving Heavy AD Users
Although AD policy is a global phenomenon, some countries' AD activity levels are
significantly higher than others. For instance, Japan, which filed the first AD investigation in
1982, only filed six AD petitions before 2000. Conversely, India, one of the most recent adopters
of AD law, has become the heaviest AD user in the world. Obviously, the likelihood of
13
retaliation is higher for firms serving heavy AD users. Given this higher risk, firms might be
more responsive to the threat of retaliation in these countries.
To investigate whether the destination country’s AD activity affects the export prices of
the products of firms that are exposed to retaliation, we interact our variable of interest with
another dummy variable, “Heavy,� which takes a value of one if the destination country’s total
number of AD filings is higher than the world average over the previous five years. The
regression estimates obtained using this variable are reported in the last column of table 6.
Similar to the earlier estimates, both the retaliation-threat variable and the interaction term are
statistically significant. This result suggests that the increase in the unit price of products from
exporting firms that are susceptible to retaliation is higher in export destinations that are heavy
AD users. In terms of economic significance, there is a 13 percent increase in the price of
products that firms export to heavy AD users when these countries are defendants in Brazil’s AD
investigations within the same industry.
<> CONCLUSION
AD has become the most frequently implemented nontariff trade barrier. After the
inception of the WTO in 1995 and the dramatic tariff cuts that countries experienced, the
importance of AD duties increased for firms that seek alternative instruments of protection.
Because AD duties are very large compared with most favored nation tariffs, they have a
dramatic impact on both dumped imported varieties and domestic production. In this study, we
analyze one dimension of this impact and contribute to the literature by investigating the pricing
effect of AD duties in two ways. First, using rich, four-dimensional trade data, we show that AD
duties imposed on Brazilian exporters significantly increase the unit value of firms’ exported
products. Second, we document a significant price increase for the exported products stemming
14
from potential AD duties. This finding demonstrates that exporters are willing to avoid the threat
of retaliation by decreasing their dumping margins in a particular country with importers that are
named in AD petitions in the exporters’ home country. We also show that the pricing effect of
retaliation threats is increasing in firms' shares of each industry's product in their total exports.
This price increase is higher for exporters serving the developed world and heavy AD users.
1
See Blonigen and Prusa (2003) for an extensive survey.
2
India and the United Stated are the top two AD users in the world.
3
Studies on retaliation and AD also use filing behavior as the dependent variable.
4
See table 1 for the AD duties of importing countries.
5
Most of the AD cases are filed for multiple products. Therefore, the number of cases does not always represent the
number of six-digit HS products investigated.
6
See Molina and Muendler (2009) and Hirakawa, Muendler and Rauch (2010) for more information about SECEX
and RAIS data.
7
http://econ.worldbank.org/ttbd/gad.
8
See Feinberg and Reynolds (2006).
9
Given the large costs associated with initiating AD investigations, firms often come together via lobbies to file
petitions and share the cost of filing.
10
Our results are insensitive to treating each European Union country as a different market and to removing
European Union countries from the sample.
11
The linked employment data also report the wage bill paid by the firm. However, the availability of this variable is
so limited in our sample that it is unusable.
12
Although we do not report the results here, the alternative measure of the use of total revenue that a firm earns
from the same product in the destination does not alter our findings.
13
The linked employment data also report the wage bill paid by the firm. However, the availability of this variable is
limited in our sample, which does not allow us to utilize the data.
14
While we do not report the results here, the use of total revenue that the firm earns from the same product in the
destination as an alternative measure does not alter our findings.
15
See Manova and Zhang (2009) for the same argument.
16
Individual exporting activities are unit observations where there is no firm code.
17
We use the formula in Kennedy (1981) to convert the coefficient of the dummy variable to its true marginal effect.
18
Instead of using the AD duty in force, which is one until the duty is revoked, the AD variables are now defined as
follows: AD duties for one year or two years after the affirmative AD decision.
19
See Vandenbussche and Wauthy (2001) for the same argument.
<> REFERENCES
Aggarwal, A. 2007. “The Antidumping Agreement and Developing Countries.� Oxford
University Press.
Bastos, P. and J. Silva. 2010. “The Quality of a Firm’s Exports: Where You Export to Matters.�
Journal of International Economics 82 (2): 99-11.
15
Belderbos, R. 1997. “Antidumping and Tariff Jumping: Japanese Firms DFI in the European
Union and the United States.� Review of World Economics (Weltwirtschaftliches Archiv) 133
(3): 419-45.
Blonigen, B. and C. Bown. 2003. “Antidumping and Retaliation Threats.� Journal of
International Economics 60 (2): 249-73.
Blonigen, B. and T.J. Prusa. 2003. “Antidumping.� In Handbook of International Trade, ed. E.K.
Choi and J. Harrigan, 258-84. Oxford, U.K. and Cambridge, MA: Blackwell Publishers.
Bown, C. 2010. Global Antidumping Database. Washington, DC:
http://econ.worldbank.org/ttbd/gad.
Bown, C. and Crowley M. 2006. “Policy Externalities: How US Antidumping Affects Japanese
Exports to the EU.� European Journal of Political Economy 22 (3): 696-714.
Feinberg, R. and K. Reynolds. 2006. “The Spread of Antidumping Regimes and the Role of
Retaliation in Filings.� Southern Economic Journal 72 (4): 877-90.
Francois, J. F. and G. Niels. 2004. Political Influence in a New Antidumping Regime: Evidence
from Mexico. Timbergen Institute Working Paper No. TI 2004-011/2.
Ganguli, B. 2008. “The Trade Effects of Indian Antidumping Actions.� Review of International
Economics 16 (5): 930-941.
Hallak, J. C. and J. Sivadasan. 2008. “Productivity, Quality and Exporting Behavior under
Minimum Quality Constraints." MPRA Paper 24146, University Library of Munich, Germany.
Hallak, J. C. 2006. “Product Quality and the Direction of Trade.� Journal of International
Economics 68 (1) :238-65.
Hansen, J. and J. Nielsen. 2009. “Dumping and Injury Margins in Markets with Horizontal as
well as Vertical Product Differentiation.� Journal of Industry, Competition and Trade 9 (3): 233-
50.
Harrigan J. and H. Deng. 2008. China’s Local Compataive Advantage. NBER Working Paper
No. 13963.
Hirakawa, O., M. Muendler, and J. Rauch. 2010.Employee Spinoffs and Other Entrants: Stylized
Facts from Brazil. Working paper, 10/0879. International Growth Centre, London, UK.
Kennedy, P.E. 1981. “Estimation with Correctly Interpreted Dummy Variables in
Semilogarithmic Equations.� American Economic Review (71) :801.
16
Kneller, R. A. and Z. Yu. 2008. Quality Selection, Chinese Exports and Theories of
Heterogeneous Firm Trade. University of Nottingham, GEP, Discussion Papers.
http://EconPapers.repec.org/RePEc:not:notgep:08/44.
Konings, J. and H. Vandenbussche. 2008. “Heterogeneous responses of firms to trade
protection.� Journal of International Economics, 76 (2) :371-83.
Konings, J. and H. Vandenbussche. 2009. Antidumping Protection Hurts Exporters: Firm-level
Evidence from France. CEPR Discussion Paper No. DP7330.
Krupp C. and P. Polard. 1996. "Market Responses to Antidumping Laws: Some Evidence from
the U.S. Chemical Industry.� Canadian Journal of Economics 29 (1): 199-227.
Manova, K. and Z. Zhang. 2009. Quality Heterogeneity across Firms and Export Destinations.
NBER Working Paper No. 15342.
Miranda J., Torres R. and M. Ruiz. 1998. “The International Use of Antidumping: 1987-1997.�
Journal of World Trade 32 (5): 5-71.
Molina D. and M. Muendler, 2009, Preparing to Export. Working paper, University of California
San Diego (http://www.econ.ucsd.edu/muendler/papers/brazexp.pdf)
Pierce J. 2011. “Plant-Level Responses to Antidumping Duties: Evidence from U.S.
Manufacturers.� Journal of International Economics 85 (2) :222-33.
Prusa T. 1992. “Why are so Many Antidumping Petitions Withdrawn?� Journal of International
Economics 33 (1-2) :1-20.
Prusa T. 2001. “On the Spread and Impact of Antidumping.� Canadian Journal of Economics 34
(3): 591-611.
Prusa, T. and S. Skeath. 2002. “The Economic and Strategic Motives for Antidumping Filings.�
Weltwirtschaftliches Archiv 138 (3): 389-413.
Staiger, R. and F. Wolak. 1994. “Measuring Industry Specific Protection: Antidumping in the
United States.� Microeconomics 1994: 51-103.
Vandenbussche, H. and X. Wauthy. 2001. “Inflicting Injury through Product Quality: How
European Antidumping Policy Disadvantages European Producers.� European Journal of
Political Economy, 17: 101-116
Verhoogen E. 2008. “Trade, Quality Upgrading and Wage Inequality in the Mexican
Manufacturing Sector.� Quarterly Journal of Economics 123: 489-530.
17
Table 1. Antidumping Duties in Force against Brazilian Exporters
(by Importing Country) (1997–2000)
Number of
Country cases
Argentina 14
European Union 4
Mexico 3
Canada 2
United States of America 1
South Africa 1
Source: Bown(2010)
18
Table 2. Defendant Countries in Brazil’s Antidumping Investigations (1996–1999)
Number of
Country cases
USA 10
China 9
Germany 4
South Africa 4
France 3
India 3
Mexico 3
Spain 3
Romania 3
United Kingdom 3
Chile 2
Japan 2
Kazakhstan 2
Russia 2
Venezuela 2
Argentina 1
Australia 1
Bangladesh 1
Bulgaria 1
Cuba 1
Denmark 1
Hong Kong 1
Italy 1
Netherlands 1
New Zealand 1
Poland 1
Thailand 1
Taiwan 1
Ukraine 1
Uruguay 1
Source: Bown(2010)
19
Table 3. Brazilian Industries Facing AD Duties
Number of
Three-digit ISIC Industry cases
311 Food products 1
313 Beverages 0
314 Tobacco 0
321 Textiles 1
322 Apparel except footwear 0
323 Leather products 0
324 Footwear except rubber or plastic 0
331 Wood products except furniture 0
332 Furniture except metal 0
341 Paper and products 2
342 Printing and publishing 0
351 Industrial chemicals 3
352 Other chemicals 0
353 Petroleum refineries 0
354 Miscellaneous petroleum and coal products 0
355 Rubber products 0
356 Plastic products 0
361 Pottery, china, earthenware 0
362 Glass and products 1
369 Other non-metallic min. products 1
371 Iron and steel 12
372 Non-ferrous metals 0
381 Fabricated metal products 1
382 Machinery except electrical 0
383 Machinery electric 3
384 Transport equipment 0
385 Professional and scientific equipment 1
390 Other manufactured products 1
Source: Bown (2010)
20
Table 4. Brazil’s Antidumping Filings by Industry
(1996–1999)
Number of
Three-digit ISIC industry
cases
311 Food products 5
313 Beverages 0
314 Tobacco 0
321 Textiles 3
322 Apparel except footwear 0
323 Leather products 0
324 Footwear except rubber or plastic 0
331 Wood products except furniture 0
332 Furniture except metal 0
341 Paper and products 0
342 Printing and publishing 0
351 Industrial chemicals 27
352 Other chemicals 3
353 Petroleum refineries 0
354 Miscellaneous petroleum and coal products 0
355 Rubber products 6
356 Plastic products 0
361 Pottery, china, earthenware 0
362 Glass and products 1
369 Other non-metallic min. products 4
371 Iron and steel 15
372 Non-ferrous metals 0
381 Fabricated metal products 2
382 Machinery except electrical 1
383 Machinery electric 0
384 Transport equipment 0
385 Professional and scientific equipment 2
390 Other manufactured products 1
Source: Bown (2010)
21
Table 5. Estimation Results
Dependent Variable: logï¿½í µí±¢í µí±“í µí±?í µí±‘í µí±¡ �
1 2 3 4 5
(í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘í µí±¡ 0.236 0.232 0.228 0.196
(0.07)*** (0.07)*** (0.06)*** (0.08)***
(í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘í µí±¡ í µí±¥ í µí°¶â„Ží µí±’í µí±ší µí±–í µí±?í µí±Ží µí±™/í µí±†í µí±¡í µí±’í µí±’í µí±™ 0.278
(0.05)***
(í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘í µí±¡ í µí±¥ í µí±?í µí±œí µí±› í µí°¶â„Ží µí±’í µí±ší µí±–í µí±?í µí±Ží µí±™/í µí±†í µí±¡í µí±’í µí±’í µí±™ 0.178
(0.08)**
(í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘ í µí±¡âˆ’1 0.199
(0.09)**
(í µí°´í µí°·í µí±‘í µí±¢í µí±¡í µí±¦)í µí±?í µí±‘ í µí±¡âˆ’2 0.206
(0.10)**
(í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘ í µí±¡âˆ’1 0.102 0.102 0.098 0.088
(0.04)** (0.04)** (0.05)* (0.02)**
(í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘ í µí±¡âˆ’1 í µí±¥ í µí°¶â„Ží µí±’í µí±ší µí±–í µí±?í µí±Ží µí±™/í µí±†í µí±¡í µí±’í µí±’í µí±™ 0.145
(0.03)***
(í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘ í µí±¡âˆ’1 í µí±¥ í µí±?í µí±œí µí±› í µí°¶â„Ží µí±’í µí±ší µí±–í µí±?í µí±Ží µí±™/í µí±†í µí±¡í µí±’í µí±’í µí±™ 0.079
(0.04)*
(í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘ í µí±¡âˆ’2 0.061
(0.02)**
(í µí±‡â„Ží µí±Ÿí µí±’í µí±Ží µí±¡)í µí±˜í µí±‘ í µí±¡âˆ’3 0.016
(0.00)*
log(í µí±ší µí±?í µí±‘í µí±¡ ) 0.045 0.056 0.054 0.054
(0.02)** (0.02)** (0.02)* (0.02)*
log (revenue)fdt 0.028 0.021 0.020 0.020
(0.00)*** (0.00)*** (0.00)*** (0.00)***
log (revenue)ft 0.108 0.096 0.114 0.114
(0.00)*** (0.00)*** (0.05)*** (0.05)***
log (employment)ft 0.035 0.039 0.039
(0.01)*** (0.01)*** (0.00)**
Country-year dummies Yes Yes Yes Yes Yes
Firm-product-destination fixed effects Yes Yes Yes Yes Yes
R2 0.17 0.23 0.29 0.29 0.29
Observations 443253 443253 343252 343252 343252
Source: Authors’ analysis based on data described in the text.
Notes: p denotes six-digit HS product, k denotes three-digit ISIC industry, f denotes firm, d denotes destination,
and t denotes time in years. Standard errors are in parentheses (clustered by firm-product combinations). ***, **,
and * indicate the level of significance at 1%, 5% and 10%, respectively. All specifications include a constant
term that is suppressed.
22