China Economic Review 20 (2009) 542–548

Contents lists available at ScienceDirect

China Economic Review

Review

The trade depressing and trade diversion effects of antidumping actions:

The case of China
Soonchan PARK ⁎
Department of Economics and International Commerce, Kongju National University, 182 Shinkwan-Dong Kongju, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: China is by far the main target of antidumping (AD) litigation for most of its major trading
Received 3 July 2007 partners. Recently, however, China itself has started using AD instrument intensively and was,
Received in revised form 25 February 2008 in fact, the third leading country in initiating AD investigations between 2002 and 2004. This
Accepted 20 March 2009
paper empirically investigates the impact of China's AD activities on trade by employing the
system GMM estimator. The empirical results show that AD protection has significant trade
JEL classification: depressing and trade diversion effects. These findings are consistent with Prusa [Prusa, T.A.,
F13
2001. On the Spread and Impact of Antidumping, Canadian Journal of Economics 34, 591–611.]
and Brenton [Brenton, P., 2001. Anti-dumping policies in the EU and trade diversion, European
Keywords:
Antidumping Journal of Political Economy 17, 593–607] that investigate the trade effects of AD actions in the
China US and EU, respectively.
Trade depressing © 2009 Elsevier Inc. All rights reserved.
Trade diversion

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
2. China's antidumping activities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543
3. Empirical model and data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544
4. Empirical results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
4.1. Trade depressing effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
4.2. Import share of named countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547

1. Introduction

China is by far the most notorious target of antidumping (AD) litigation for most of its major trading partners. China assumes
15.8% of more than 2700 antidumping investigations initiated globally during the period from 1995–2004. Recently, however,
China has started using AD instruments actively, especially after its accession to the WTO in November 2001. According to the WTO
report, China initiated 30 antidumping investigations in 2002 and 27 cases in 2004. During the 3 years between 2002 and 2004,
China was the third leading country in initiating AD investigations (79 cases) following India (148 cases) and the US (98 cases).
This increasing rate of AD filings has raised substantial concerns of China becoming an intensive user of AD (Messerlin, 2004).

⁎ Tel.: +82 41 850 8394; fax: +82 41 850 8389.

E-mail address: [email protected].

1043-951X/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.chieco.2009.03.006
 S. Park / China Economic Review 20 (2009) 542–548 543

Fig. 1. China's antidumping activities.

With the increase in China's AD activities, there have been a growing number of studies on China's AD policy. Most of them focus on
China's AD laws and regulations, investigating whether they comply with WTO rules or whether they have different characteristics
compared to those of other countries. However, little attention has been paid to the effects of China's AD activities on trade. This is for
the most part because the data necessary for analysis had not sufficiently accumulated. However, China filed 121 AD cases for the
period 1997–2004, including 10 reviews, which has now provided a sufficient accumulation of cases for statistical analysis.
The purpose of this paper is to empirically investigate the impact of China's AD activities on trade. AD is discriminatory since AD
duties are applied specifically to imports from named countries while the non-named countries supplying the same product are
not subject to such duties. AD actions therefore induce trade diversion, which in effect causes the distortion of trade patterns.
The trade effects of the AD actions filed by the United States and the European Union have been well studied. Using trade data at
the line-item level, Prusa (1997, 2001) shows that the AD actions of the United States distort trade patterns significantly. Similarly,
Brenton (2001) provides evidence to show that the AD actions pursued by the EU also cause trade diversion. On the other hand,
Konings, Vandenbussche, and Springael (2001) explore the effects of European AD actions initiated between 1985 and 1990 and
find no evidence of trade diversion. Whilst these studies focus on the trade effects of products subject to AD initiations or
measures, Vandenbussche and Zanardi (2006) examine whether AD cases restrain aggregate trade, as opposed to trade of products
subject to AD investigations, and find that AD actions lead to a reduction of annual aggregate trade by about 6.7%.
Though AD actions taken by developing countries have been proliferating, only a few studies have attempted to estimate the
trade effects of these actions.1 Ganguli (2008) investigates Indian AD actions and finds that imports from named countries fall
significantly, while imports from non-named countries increase. Since the current WTO antidumping rules are ambiguous, the
bulk of investigations are undertaken by the discretion of the authority (Blonigen, 2006). As a result the AD filings and AD tariffs
are very different from country to country, thus the impacts of AD on trade may be different.
This study may well be the first attempt to quantify the trade depressing and trade diversion effects of China's AD
investigations. In order to estimate the trade effects of China's AD activities, this paper constructs an integrated dataset by
combining AD initiations with trade data at the HS-8 digit level resulting in time-series and cross-section data. The standard
methodology used to analyze this type of data is the standard panel data technique. However, this estimator suffers from the
endogeneity problem and is inevitably biased, especially when the lagged dependent variable is included in the regression. To
supplement this problem this paper employs the Generalized Method of Moments (GMM) estimator suggested by Arellano and
Bover (1995) and Blundell and Bond (1998).
Specifically, this paper analyzes two empirical models: one that takes into account trade depressing effects and the other to
explain trade diversion effects. The empirical results show that AD investigations lead to a significant fall in imports from named
countries. Furthermore, they reduce the import share of the named countries, suggesting the existence of trade diversion effects.
The remainder of this paper is organized as follows. Section 2 provides an overview of China's AD activities. Section 3 describes
the empirical model and data. Section 4 discusses the empirical results, and Section 5 presents the conclusions.

2. China's antidumping activities

China first promulgated antidumping regulations in 1997. With its accession to the WTO they were modified in February 2002
to comply with WTO rules. The new regulations significantly improved China's anti-dumping regime to the extent that they are
now almost in-line with WTO standards (Wang and Shengxing, 2002). Though China's antidumping regulations are similar to the
standard rules, Messerlin (2004) points out four striking features: lack of detail, inclusion of protectionist biases inherent in the
WTO antidumping provisions, complexity of coordination among administrative agencies, and the inclusion of Article 56, which
states that “where a country (region) discriminatorily imposes antidumping measures on the exports from the People's Republic of

1
There is also a well-established body of literature on the rise of AD actions by developing countries. These studies hold two common hypotheses on the
proliferation of AD: AD as a retaliation and AD as a quid pro quo for tariff liberalization (Feinberg and Reynolds, 2007). Feinberg and Reynolds (2006), Prusa and
Skeath (2004), and Francois and Niels (2004) present evidence to show that the global spread of the AD actions can be largely attributed to ‘tit-for-tat’ retaliatory
policies. On the other hand, Blonigen and Bown (2003) find that AD activity both at the industry and government levels in the United States dampen when
retaliation threats exist.
544 S. Park / China Economic Review 20 (2009) 542–548

Table 1
Regional distribution of AD cases initiated by China (1997–2004).

Country Korea Japan USA EU Row Total

No. of AD initiations 25 21 18 16 31 111
(Share) (22.5%) (18.9%) (16.2%) (14.4%) (27.9%) (100%)

Source: annual report of AD (WTO).

China, China may, on the basis of the actual situations, take corresponding measures against that country (region).” This provision
opens the possibility of China using antidumping regulations as a retaliatory instrument. To coordinate the administrative
functions of antidumping agencies, the Chinese government revised its regulations in 2004. Therefore, the Ministry of Commerce
(MOC) is currently the sole competent authority with regard to antidumping in China.
The first AD case pursued by Chinese authorities was filed in December 1997 against imports of newsprints from the US,
Canada, and the Republic of Korea. Since 2001 the number of cases initiated has increased remarkably: China filed 30 cases in 2002
and 27 in 2004. Fig. 1 presents the trends in China's use of antidumping investigations.
The imposition of AD duties usually expires after 5 years from the date of entry into force. However, many countries have
extended that period through a review clause that can be initiated upon request by any party interested. China's antidumping
regulations also provide for this type of review and it has begun to carry out antidumping reviews from 2001 as original AD duties
expired. Hence, it is expected the total number of AD measures in force will further increase in the future.
As is shown in Table 1, the main countries targeted by China are Korea, Japan, the US, and the EU: the share of these four
countries in the total number of AD initiations by Chinese authorities is over 70%. This is likely due to the rapid increase of imports
from these countries. As Blonigen (2005) reveals, higher import penetration leads to greater AD activity. Imports from Korea and
Japan, for example, increased about 527% and 120%, respectively, during the period 1998–2004.
Table 2 shows the AD duties levied on firms. Average provisional duties amount to 33.8%, whereas the maximum duty is 144%.
These are conspicuously high levels, considering that the Chinese most-favored-nation (MFN) tariff on industrial products is on
average only 9.7%. Given the prohibitively high level of AD duties, AD activities are effective in significantly hindering imports from
targeted countries.

3. Empirical model and data

Combining the AD cases with import data at the detailed product and origin country levels, this paper investigates the trade
effects of China's AD investigations initiated between 1997 and 2004. The first issue addresses to what extent the imports of named
countries change due to AD investigations. The second question is whether AD activities generate trade diversion effects. This
paper formulates two different methods to this end. The first is to examine how trade volumes of named countries are affected by
AD actions. Second, by investigating how the share of named countries changes at the product level, this paper directly measures
the trade diversion effects of AD policies.
Consistent with Prusa (1997, 2001), Brenton (2001), and Konings et al. (2001), the basic empirical model employed in this
paper is as follows:

lnðimportit Þ = α + γ1 lnðimportit − 1 Þ + β1 ðln dutyit 4Dt Þ + δi + uit ð1Þ

where importit denotes imports from country i at time t. Following Prusa (2001), time is normalized so that t = 0 presents the year
the investigation was initiated. ln duty is specified as ln(1 + AD duties) to avoid missing values when taking the log. Dt is the
dummy variable equal to 1 if the AD duty is imposed at time t, and δi denotes the case-specific effects.
In the presence of case-specific effects, ordinary least squares (OLS) may induce a bias as a result of the correlation of
unobserved case-specific factors and the variables of interest. In addition, there exists a close relationship between imports in t and
t-1. In such a framework, OLS might result in inconsistent estimates since imports in t-1 and δi are unavoidably correlated. The
standard solution to these problems is to use a fixed effects estimator that transforms the equation to eliminate the term δi.
However, the fixed effects estimator fails to consistently estimate the parameters of interest as the transformation only leads to
consistent estimates when regressors are strictly exogenous. In particular, for panels with short time periods, this transformation

Table 2
AD duties imposed on firms (1997–2004).

Min. Average Std. D. Max.

Provisional 0 33.8 28.2 144.0
Affirmative 0 24.5 25.1 144.0

Source: annual report of AD (WTO).

 S. Park / China Economic Review 20 (2009) 542–548 545

Table 3
Summary statistics.

Mean Std D. Min. Max.

ln (import) 9.90 2.15 0.00 15.82
ln (total import) 12.20 1.49 8.74 16.45
ln (share) 2.33 1.31 − 4.44 4.30
ln (duty,t + 1) 0.54 1.30 0.00 4.92
ln (duty,t + 2) 0.39 1.13 0.00 4.49
ln (duty,t + 3) 0.19 0.80 0.00 4.38

induces a correlation between the transformed lagged dependent variable and the transformed error term (Bond, 2002; Nickel,
1981).
To resolve these problems, this paper employs the Generalized Method of Moments (GMM) estimator that allows control of the
lagged endogeneity by differencing and incorporating lagged values as instruments following Arellano and Bond (1991).
The dynamic panel model is represented as

Δlnðimportit Þ = γ1 Δlnðimportit − 1 Þ + β1 ðΔln dutyit 4Dt Þ + Δuit ; ð2Þ

where Δ is the difference operator.

This first-differenced GMM estimator has been widely used in the estimation of dynamic panel data models. However, Blundell
and Bond (1998) have shown that the first-differenced GMM estimator can generate substantially biased results, when the time
series is persistent, but the number of time series is small. As an alternative, Arellano and Bover (1995) and Blundell and Bond
(1998) suggest a system GMM estimator that combines a levels equation with the first-differenced equation and employs both
lagged levels and differences as instruments, which allows for dramatic reduction of biases.
The AD data are collected from the WTO and the Ministry of Commerce (MOC) of China, which archive product codes at the HS
8-digit level and AD duties. The source of annual import data for each investigated product of a named country is China's Customs
Statistics provided by the General Administration of Customs in the People's Republic of China and Korea International Trade
Association. Import data for each named country were constructed based upon AD cases. For the cases that involved more than one
product code, all product codes were summed up to match the import data in each AD case.
The dataset includes AD duties and imports from named countries for the period from 1995 to 2004. Since the number of AD
cases initiated is different from year to year, the number of observations varies per year. Table 3 presents summary statistics for the
data used in the estimation.

4. Empirical results

Three different estimation models are applied to investigate the trade effects of China's AD activities: the fixed-effects
estimation, the first-differenced GMM, and the system GMM. In addition, an estimation of random effects, which assumes that
individual case-specific effects are random variables, is included. A Hausman test is accordingly applied comparing the fixed effects
to random effects under the null hypothesis to asses whether individual effects are in fact uncorrelated with other regressors in the
model (Hausman, 1978). The Hausman specification test reveals that the null hypothesis is rejected indicating that the fixed and
random estimates are different from one another and, thus, only the fixed estimates are reported here.

4.1. Trade depressing effects

Table 4 presents the results from both the fixed-effects estimates, the first-differenced GMM, and the system GMM.2 The
empirical model fits the data well, explaining a major part of changes in import volume of named country. The GMM estimator
validity and consistency depends on two assumptions: i) the lagged value of the dependent variable and other explanatory
variables are valid instruments, and ii) the error terms do not show serial correlation. The Hansen test is satisfied and the test for
second order correlation is rejected, suggesting the equation has been correctly specified.
Most of the estimated coefficients are statistically significant and show the expected signs. Column (1) of Table 4 reports the
results using the fixed-effects estimator. The estimated coefficient on ln (duty,t + 1) is negative and significant, indicating that an
imposition of a 10% AD duty will lead to about a 10.7% decrease in imports in the first year following an AD investigation.
Column (3) and (5) of Table 4 present the estimation results of the trade depressing effects of AD duties in the system GMM and
first-differenced GMM, respectively. The first-differenced GMM estimates are very close to the fixed-effects results. However, the
estimated coefficients on AD duties in the system GMM are higher than those in the first-differenced GMM and the fixed-effects
estimation, raising suspicions of a serious finite sample bias in the dataset.

2
For convenience, the notation of the difference operator Δ in the first-differenced GMM estimates is abused.
546 S. Park / China Economic Review 20 (2009) 542–548

Table 4
Trade depressing effects of China's AD.

Dep. Variable Fixed-effects System GMM Diff-GMM

ln (import) (1) (2) (3) (4) (5) (6)

Constant 7.563 7.617 0.858 0.784 − 1.449 − 1.495
(0.639)⁎⁎⁎ (0.641)⁎⁎⁎ (1.083) (1.133) (0.290)⁎⁎⁎ (0.306)⁎⁎⁎
ln (import, t-1) 0.615 0.615 0.774 0.796 0.394 0.383
(0.043)⁎⁎⁎ (0.044)⁎⁎⁎ (0.057)⁎⁎⁎ (0.066)⁎⁎⁎ (0.121)⁎⁎⁎ (0.127)⁎⁎⁎
ln (import, t-2) 0.151 0.131 0.048 0.057
(0.085)⁎ (0.091) (0.074) (0.076)
ln (duty, t + 1) − 0.107 − 0.147 − 0.110
(0.038)⁎⁎⁎ (0.036)⁎⁎⁎ (0.033)⁎⁎⁎
ln (duty, t + 2) − 0.111 − 0.174 − 0.107
(0.038)⁎⁎⁎ (0.043)⁎⁎⁎ (0.047)⁎⁎
ln (duty, t + 3) − 0.065 0.009 − 0.069
(0.052) (0.070) (0.079)
t+1 − 0.351 − 0.505 − 0.347
(0.115)⁎⁎⁎ (0.123)⁎⁎⁎ (0.108)⁎⁎⁎
t+2 − 0.366 −0.490 − 0.363
(0.138)⁎⁎⁎ (0.165)⁎⁎⁎ (0.140)⁎⁎⁎
t+3 − 0.224 0.097 − 0.275
(0.184) (0.255) (0.345)
N. obs. 453 453 287 287 202 202
R-sq 0.83 0.83
Hansen test 0.99 0.97
AR(1) (p-value) − 3.17 (0.00) − 3.36 (0.00) − 1.74 (0.08) − 1.83 (0.07)
AR(2) (p-value) 0.90 (0.37) 0.91 (0.37) 0.73 (0.46) 0.71 (0.48)

Note: ⁎ denotes statistical significance at the 10%, ⁎⁎ at the 5%, ⁎⁎⁎ at the 1% level. Robust standard errors are in parenthesis. Calendar year dummies are included in
the estimation. The Hansen test shows a p-value. In the first-differenced GMM, standard errors are robust to both heteroskedasticity and serial correlation, thus the
Sargan test statistics are not reported.

Prusa (1997, 2001) argues that AD duties do not fully capture potential trade depressing effects, since they are only levied when
a case is affirmatively determined as injurious. Furthermore, many theoretical studies contend that the mere threat of protection is
likely to lead to a reduction in trade (Pauwels, Vandenbussche, & Weverbergh, 2001). That is, there exists a so-called chilling effect
of AD in which the initiation of AD itself depresses trade, even if the case does not receive an affirmative decision. These
implications are empirically identified by Staiger and Wolak (1989) and Krupp and Pollard (1996). In particular, Vandenbussche
and Zanardi (2006) show evidence that AD laws and their implementation cause a substantial depressing effect on aggregate
imports worldwide.
To capture the potential trade depressing effects of AD, dummy variables t + 1, t + 2, and t + 3 are introduced. For example, the
dummy variable t + 1 takes the value of 1, if t = 1 (following the first year of AD investigations); otherwise, it is zero. The results
are shown in Columns (2), (4), and (6). Column (2) of Table 4 is the fixed-effects estimation of log imports on the AD dummy
variables. The results show a statistically significant and negative relationship between imports and AD. The estimate on AD t + 1
implies that AD investigations reduce imports of named countries by about 29.6% in the immediate following year. Furthermore,
the trade depressing effects in the system GMM results are larger than those in fixed-effects estimations and the first-differenced
GMM. However, the coefficients for t + 3 are not statistically significant in any of the specifications. This may be due to the small
number of observations. Although China filed 79 AD cases during the period from 2002–2004, the dataset covers only the years
leading up to 2004; thus, these cases are actually excluded in the third year estimations.

4.2. Import share of named countries

Considering that only a limited number of import suppliers are subject to AD investigations, AD is discriminatory in the sense
that the AD duties are applied specifically against imports from named countries. Countries that are not listed, but supply the same
product are not subject to AD duties. AD investigations, thus, lead to an increase in the demand for domestically produced products
and identical products from unlisted countries and, consequently, induce trade diversion.
Most previous studies identify the trade diversion effects of AD by examining the changes in imports of non-named countries.
However, Brenton (2001) takes a different approach that examines the share of named countries of products subject to AD
investigations. Adopting a similar methodology, this paper also analyzes the changes in the trade share of named countries. The
empirical model for estimating the impacts of AD on the import share of a named country is specified as follows

lnðshareit Þ = α + γ2 lnðshareit − 1 Þ + β2 ð ADit Þ + δi + uit ; ð3Þ

where shareit denotes the import share of the case i at time t and ADit presents AD duties or AD dummies.
Table 5 shows the trade diversion effects of China's AD activities: the dependent variable is the log value of the named country's
share of products under AD investigation. The results from the first-differenced GMM are not reported, since the estimates are
similar to those from the fixed-effects estimator. Except for the third year (t + 3), the coefficients on the variables that represent AD
 S. Park / China Economic Review 20 (2009) 542–548 547

Table 5
Trade diversion effects of China's AD.

Fixed-effects System GMM

(1) (2) (3) (4)

Constant 3.701 2.113 0.170 0.145
(0.392)⁎⁎⁎ (0.405)⁎⁎⁎ (0.183) (0.171)
ln share(t-1) 0.509 0.510 0.931 0.940
(0.039)⁎⁎⁎ (0.039)⁎⁎⁎ (0.026)⁎⁎⁎ (0.023)⁎⁎⁎
ln share(t-2)
ln (duty, t + 1) − 0.093 − 0.153
(0.024)⁎⁎⁎ (0.035)⁎⁎⁎
ln (duty, t + 2) − 0.078 − 0.074
(0.029)⁎⁎⁎ (0.038)⁎
ln (duty, t + 3) − 0.058 − 0.034
(0.040) (0.054)
t+1 − 0.313 − 0.547
(0.089)⁎⁎⁎ (0.122)⁎⁎⁎
t+2 − 0.254 − 0.242
(0.105)⁎⁎ (0.140)⁎
t+3 − 0.146 − 0.119
(0.143) (0.192)
N. obs. 450 450 371 371
R-sq 0.65 0.63
Hansen Test (p-value) 0.97 0.96
AR(1) (p-value) − 2.40 (0.02) − 2.41 (0.02)
AR(2) (p-value) 1.51 (0.13) 1.49 (0.14)

Note: ⁎ denotes statistical significance at the 10%, ⁎⁎ at the 5%, ⁎⁎⁎ at the 1% level. Robust standard errors are in parenthesis and calendar year dummies are
included in the estimation.

investigations have an expected sign and are significant. This means that the import share of a named country decreases when an
AD investigation is initiated. Therefore, substantial trade diversion effects are noticeable, although they become weaker in the
second year.

5. Conclusion

As China continues to emerge as one of the major actors in the world, much literature has focused on the dynamic of the
Chinese economy. However, little attention has been paid to estimating the trade effects of China's AD activities. China itself has
started using AD instrument intensively and was, in fact, the third leading country in initiating AD investigations between 2002
and 2004, following India and the US.
This paper has empirically investigated the impacts of China's AD activities on trade and trade patterns. In order to achieve this,
this paper has integrated data pertaining to China's AD investigations with import data at the tariff-line level. Based on previous
studies regarding this topic (Brenton, 2001; Konings et al., 2001; Prusa, 2001), this paper has also examined the evolution of
imports and changes in the import share of named countries. The empirical model employed is autoregressive, meaning that the
OLS estimation is likely to be affected by a problem of endogeneity. To adjust for this, the system GMM estimator was applied.
The empirical results show that AD protection has significant trade depressing and trade diversion effects. These findings are
consistent with previous studies of AD actions in the US and the EU. In particular, the existence of trade diversion effects suggests
that a decrease in imports from named countries is to a large extent offset by an increase from non-named countries, thus the
benefits of import competing domestic industries may be smaller than expected.

References

Arellano, M., & Bond, S. R. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic
Studies, 58, 277−297.
Arellano, M., & Bover, O. (1995). Another look at the instrumental-variable estimation of error components models. Journal of Econometrics, 68, 29−52.
Blonigen, B. A. (2005). The effects of NAFTA on antidumping and countervailing duty activity. World Bank Economic Review, 19, 407−424.
Blonigen, B. A. (2006). Evolving discretionary practices of U.S. antidumping activity. Canadian Journal of Economics, 39, 874−900.
Blonigen, B. A., & Bown, C. P. (2003). Antidumping and retaliation threats. Journal of International Economics, 60, 249−273.
Blundell, R. W., & Bond, S. R. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87, 115−143.
Bond, S. (2002). Dynamic panel data models: A guide to micro data methods and practice. CEMMAP working paper CWP 09/02, The Institute for Fiscal Studies: UCL.
Brenton, P. (2001). Anti-dumping policies in the EU and trade diversion. European Journal of Political Economy, 17, 593−607.
Feinberg, R. M., & Reynolds, K. M. (2006). The spread of antidumping regimes and the role of retaliation in filings. Southern Economic Journal, 72, 877−890.
Feinberg, R. M., & Reynolds, K. M. (2007). Tariff liberalization and increased administrative protection: Is there a quid pro quo? The World Economy, 30, 948−961.
Francois, J. F., & Niels, G. (2004). Political Influence in a New Anti-Dumping Regime: Evidence from Mexico, CEPR Discussion Paper No. 4297.
Ganguli, B. (2008). The Trade Effects of Indian Antidumping Actions. Review of International Economics, 16, 930−941.
Hausman, J. (1978). Specification tests in econometrics. Econometrica, 46, 1251−1271.
Konings, J., Vandenbussche, H., & Springael, L. (2001). Import diversion under European antidumping policy. Journal of Industry, Competition and Trade, 1, 283−299.
548 S. Park / China Economic Review 20 (2009) 542–548

Krupp, C., & Pollard, P. S. (1996). Market responses to antidumping laws: Some evidence from the US chemical industry. Canadian Journal of Economics, 29,
199−227.
Messerlin, K. A. (2004). China in the World Trade Organization: Antidumping and safeguards. The World Bank Economic Review, 18, 105−130.
Nickel, S. J. (1981). Biases in dynamic models with fixed effects. Econometrica, 49, 1417−1426.
Pauwels, W., Vandenbussche, H., & Weverbergh, M. (2001). Strategic behaviour under European antidumping duties. International Journal of the Economics and
Business, 8, 75−99.
Prusa, T. A. (1997). The trade effects of U.S. antidumping actions. In Robert C. Feenstra (Ed.), Effects of U.S. Trade Protection and Promotion Policies Chicago: University
of Chicago Press.
Prusa, T. A. (2001). On the spread and impact of antidumping. Canadian Journal of Economics, 34, 591−611.
Prusa, T. A., & Skeath, S. (2004). Modern commercial policy: Managed trade or retaliation? In E. K. Choi & J. Hartigan (Eds.), The Handbook of International Trade
Volume II: Economic and Legal Analyses of Trade Policy and Institutions (pp. 358−382). Oxford, UK: Blackwell Publishers.
Staiger, R. W., & Wolak, F. A. (1989). Strategic use of antidumping law to enforce tacit international collusion, NBER Working Paper, No. 3016: NBER.
Vandenbussche, H., & Zanardi, M. (2006). The Global Chilling Effects of Antidumping Proliferation, CEPR Discussion paper, No. 5597: CEPR.
Wang, L., & Shengxing, Y. (2002). China's new antidumping regulations: Improvements to comply with the World Trade Organization rules. Journal of World Trade,
36, 903−920.