When talking about protectionist trade policy, one quickly thinks of ‘America First’ or the Sino-American trade war (e.g. Bown 2018, 2019). However, the EU is also a frequent user of trade defence instruments, in particular against China. In June 2021, the European Council agreed on a mandate to launch negotiations on an international procurement instrument, designed to protect European procurement markets. In 2019, 51% (61 cases) of all EU antidumping (AD) duties – the most frequently used trade defence instrument – in force were directed against China (Garcia-Herrero et al. 2020).
China has long been treated differently from other countries in the EU’s AD regulation because of its non-market economy status (NMES). In a nutshell, market economy status (MES) is assigned by the importing country to the exporting country and determines the way AD duties are calculated. Exporters situated in MES countries receive firm specific duties, while AD duties imposed against NMES countries are often the same across all exporting firms and on average higher than those imposed against MES exporters. In 2017, the EU officially abandoned the NMES concept, whereas the US continues to treat China as a non-market economy. The role played by MES for the effectiveness of AD duties has so far been almost completely ignored by the literature.
While correctly estimating the trade effects of AD duties is essential for deriving the welfare consequences that may result from trade destruction and higher import prices, it is also notoriously difficult. AD measures are often imposed in response to an increase in imports, resulting in biased estimates of the treatment effect (Bown and Crowley 2012, 2013). In this column, we present two recent approaches to tackling endogeneity inherent in trade policy. The first takes the importer’s perspective and relies on the 2004 EU enlargement as a natural experiment and exogenous source of variation in AD policy from the perspective of the new member states (Sandkamp 2020). The second approach uses Chinese firm-level export data, exploiting differences across firms exporting the same product to estimate treatment effects (Felbermayr and Sandkamp 2020).
Evidence from the 2004 EU enlargement
The first approach uses data on import prices and quantities at the CN8 digit product level for the ten countries that joined the EU in 2004 (Eurostat 2017) as well as information on AD duties (Bown 2015). When joining the EU in 2004, the new member states adopted the Union’s AD duties imposed against third countries. Assuming that the decision to join the EU was not driven by its AD policy, the enlargement lead to an implementation of AD duties in the accession countries that was independent of their existing trade flows. The effect of AD duties on import prices and quantities can hence be estimated using a simple difference-in-differences regression with fixed effects by comparing the change over time in exporter-product combinations that become subject to AD duties relative to those that do not.
Using data for the years 2003 and 2005, the estimation shows that AD duties on average raise pre-duty import prices by 25%, implying a negative terms of trade effect for the importer (usually positive for normal tariffs). Prices of imports from non-targeted exporters also increase, which constitutes empirical evidence for tariff scares (Crowley et al. 2017). These results are highly relevant against the background of the recent surge in producer prices, experienced all over the EU. In Germany, the EU’s largest economy, producer prices increased by 10.4% in July 2021 compared to the same month in the previous year (Destatis 2021).
As AD duties are often imposed on intermediate products, reducing or eliminating them could take some pressure off prices. AD duties are most prominent in the metals and chemicals sector – 48% (23%) of EU AD cases in force in 2019 were imposed in the metals (chemicals) sector (Bown 2020). For EU AD duties against China in particular, average ad-valorem duties in these sectors amount to more than 30% (Yalcin et al. 2016). In July 2021, metals and chemicals belonged to the sectors experiencing some of the strongest surges in producer prices (33.9% in the manufacture of basic metals and 19.3% in the chemical sector in the EU27 compared to July 2020; see Eurostat 2021), indicating that removing duties could contribute to stabilising prices in these sectors.
Interestingly, these positive price effects are only observed for imports from MES countries; pre-duty prices of imports originating from NMES countries remain constant. While this still implies increasing post-duty prices as duties are passed on fully to importers, the overall effect is smaller. As explained in detail in Sandkamp (2020), exporters in MES countries can benefit from raising prices in the form of lower duties. In addition, low-price exporters receive larger duties and are thus more likely to exit. Both channels are absent in NMES countries. The results align the seemingly contradicting findings of previous research by demonstrating that price effects are driven by the MES of the exporter investigated. Somehow paradoxically, the abandonment of NMES in the EU’s AD legislation in 2017 might thus have contributed to stronger increases in import prices following the imposition of AD duties. On the other hand, this effect is counteracted by lower average duties imposed against MES exporters.
Import quantities on average fall by 74%. The estimated coefficient is larger than those in previous studies, suggesting that those may indeed be subject to several sources of omitted variable bias. Driven by the higher average duties imposed by the EU against NMES exporters, imports from these countries fall by more (85% on average) compared to imports from countries with MES (68%). The recent abandonment of NMES in EU AD regulation might thus be expected to reduce the trade dampening effects of AD duties. However, the new methodology recognizes the concept of price and cost distortions that might ultimately result in the same treatment of China as before (Garcia-Herrero et al. 2020).
Figure 1 illustrates the trade effects of EU AD duties over time by showing estimated treatment effects before and after the enlargement took place in 2004. Both estimated price and quantity coefficients are insignificant before the beginning of the treatment. At the time of the accession, positive (negative) price (quantity) effects can be observed. These increase in magnitude in 2005 and stay relatively constant until the end of the sample period in 2007. This indicates persistent effects of AD duties, even beyond their revocation (several duties were lifted before the end of the sample period so that one would expect a decline in treatment effects over time). While revoking AD duties may thus not immediately lower pre-duty import prices, it should certainly lower post-duty prices.
Figure 1 Effect of AD duties on import prices and quantities
Note: EU accession (beginning of treatment) in May 2004. Percentage change in import prices and quantities of treated products on vertical axis, year on horizontal axis. The plot shows point estimates with 95% confidence intervals, indicating large and significant price and quantity effects of AD duties since their introduction in 2004.
Source: Sandkamp (2020).
Firm-level evidence from China
As China was treated as a NMES country by the EU and the US in the period of investigation, most duties are imposed at the product level. However, several Chinese firms successfully applied for special treatment in US and EU AD investigations and are hence subject to a firm-specific duty. An alternative approach to reduce endogeneity inherent in trade policy is hence to use firm-level data to exploit the difference in duties across different firms exporting the same product. To do so, annual export data at the destination country-product-firm level from the Chinese customs office was merged with firm specific AD duties (Bown 2015).
Motivated by a simple firm-level gravity equation, we control for a plethora of unobserved variables using a battery of fixed effects. Unobserved supply-side variables, including firm-product specific subsidies, are controlled for using product-firm-time fixed effects. Importing country-product-time fixed effects control for unobserved demand-side variables and lobbying power of import competing industries. Importing country-product-firm fixed effects control for the fact that large exporters often apply for special treatment and thus on average receive smaller duties. We show empirically that these theoretically motivated fixed effects eliminate several sources of omitted variable bias, overall resulting in a larger estimated treatment effect.
At the firm level, AD duties reduce exports. In particular, a 1% increase in EU (US) AD duties reduces firm-level export value by 7.3% (4.7%). Pre-duty price effects are not significantly different from zero, indicating 100% pass-through. This provides additional evidence that removing AD duties (or at least not imposing new ones) would take some pressure off prices. Along the extensive margin, a one percent increase in EU (US) duties reduces the number of exporters by 0.8% (0.4%).
Chinese exports to the EU react differently compared to those to the US. Panel (a) of Figure 2 shows that the US on average impose larger AD duties on Chinese exporters than the EU (for a detailed discussion see Felbermayr and Sandkamp 2020). This is true for product-specific duties (no special treatment of individual firms) as well as for firm-specific duties (the distribution of AD duties across exporting firms is illustrated in Panel (b) of Figure 2). Consequently, considering both duty size and elasticities, the overall trade dampening effect of US AD duties is stronger.
Figure 2 AD duty rate against China in the EU and the US
Note: Panel (a): "Product-specific Duty'' is the unweighted mean of country-wide duty levels over all affected HS6 products; "Weighted Duty" is the mean of firm-specific duty levels weighted by export value in USD; ``Firm Duty (IT and MET only)'' is the unweighted mean of firm-specific duties over all firm-product combinations receiving individual or market economy treatment; Panel (b): Boxplots show the distribution of ad valorem AD duties across all Chinese exporters to the EU and the US.
Source: Felbermayr and Sandkamp (2020).
Smaller, potentially less productive exporters on average receive higher duties and are thus more likely to exit. This within-industry reallocation may increase the overall competitiveness of Chinese exporters, reducing the protective effects of AD-duties for import competing firms in the imposing countries. Finally, we find evidence that trade deflection observed in the literature is driven by the extensive margin, as Chinese exporters enter new markets following the imposition of EU or US AD duties.
Overall, the two papers presented in this column shed new light on the true trade-dampening effects of AD duties. The empirical analyses show that both EU and US AD duties constitute significant trade barriers that strongly reduce imports. Estimated coefficients are larger than those found in the literature so far, suggesting stronger negative welfare implications than previously thought. Driven by higher average duties, the trade-dampening effect is, however, stronger for the US than for the EU.
The natural experiment revealed that trade destruction is particularly strong for imports from China and other exporters with non-market economy status. On the other hand, price effects – positive for MES exporters – are absent for these countries. Our results imply that removing AD duties could ease the pressure on both input as well as consumer prices that are currently plaguing EU economies, in particular in the metals and chemical industries.
Finally, the recent abandonment of NMES by the EU can – if fully implemented – be expected to have implications for the effectiveness of its AD policy. In particular, the MES concept may lead to an increase in pre-duty import prices. Afterall, this is the official goal of the instrument, although a further increase in prices is certainly not desirable given the current environment. Both methodologies certainly reduce import quantities, so that the switch will not render the EU defenceless against unfair trade practices.
Bown, C P (2015), Global Antidumping Database, The World Bank, June.
Bown, C P (2018), “Trump's steel and aluminium tariffs: How WTO retaliation typically works”, VoxEU.org, 8 March.
Bown, C P (2019), “The 2018 trade war and the end of dispute settlement as we knew it”, VoxEU.org, 13 June.
Bown, C P (2020), Global Antidumping Database, The World Bank
Bown, C P and M A Crowley (2012), “Antidumping as cooperation”, VoxEU.org, 14 July.
Bown, C P and M A Crowley (2013), “Self-enforcing trade agreements: evidence from time-varying trade policy”, American Economic Review 103(2): 1071–1090.
Crowley, M A, H Song and N Meng (2017), “Protectionist threats jeopardise international trade: Chinese evidence for Trump’s policies”, VoxEU.org, 10 February.
Destatis (2021), “Producer prices of industrial products” (accessed on 29 August 2021).
Eurostat (2017), Eurostat COMEXT database.
Eurostat (2021), Producer prices in industry, domestic market - monthly data.
Felbermayr, G and A Sandkamp (2020), “The trade effects of anti-dumping duties: Firm-level evidence from China”, European Economic Review 122.
Garcia-Herrero, A, G Wolff, J Xu, N Poitier, G Felbermayr, R Langhammer, W-H Liu and A Sandkamp (2020),” EU-China trade and investment relations in challenging times”, European Parliament, May.
Sandkamp, A (2020), “The trade effects of antidumping duties: Evidence from the 2004 EU enlargement”, Journal of International Economics 123.
Yalcin, E, A Sandkamp and P Lang (2016), Reformen der EU-Handelsschutz-Instrumente - Unter Berücksichtigung des Marktwirtschaftsstatus der Volksrepublik China, Bertelsmann Stiftung.