![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Many works have examined the variables driving the formation of North-South free trade agreements (FTAs) between developed and developing countries. This study analyzes the determinants shaping their utilization in the contexts of their political economy and of Generalized System of Preferences (GSP) schemes that are unilaterally granted by developed economies to developing countries’ exports. Most of the goods liberalized through GSP are liberalized from early on in North-South FTAs; however, since FTA concessions are legally binding, goods that are excluded or only partially liberalized in GSP will be also excluded or protected in FTAs. As GSP schemes are subject to unilateral restriction/elimination by the developed country, exporters using GSP tariffs will lobby for the non-removable liberalization of their exports through an FTA and subsequently will have a high FTA utilization. These scenarios result in North-South FTAs being used to a great extent to export goods covered by and exported through GSP, thus consolidating pre-FTA trade patterns. These arguments were tested by analyzing disaggregated and rarely accessed data on Thailand’s and Malaysia’s exports through the Japanese GSP and their bilateral FTAs with Japan, as well as interviews with key actors involved in the policymaking of these FTAs. Most sectors in Thailand and Malaysia that benefited from GSP lobbied for FTA liberalization with Japan. Goods previously exported through GSP account for most of FTA utilization and the previous use of GSP preferences has a higher predictive value of subsequent FTA utilization than FTA tariff savings.
Correction statement
This article was originally published with errors, which have now been corrected in the online version. Please see Correction http://dx.doi.org/10.1080/09692290.2021.1985254
Acknowledgements
I wish to thank Emanuel Ornelas, Iain Osgood, and Kenneth Shadlen for their comments on early versions of the article. I am also grateful to the editors and three anonymous reviewers for their constructive critiques and suggestions. Lastly, I would like to deeply thank all the interviewees for their participation in the study and the Thai Ministry of Commerce and the Malaysian Ministry of International Trade and Industry for providing the preferential export data, which are used in the article.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Neoclassical trade theory associates this variation with differences in competitive advantage as well as with factor endowment and industry cleavages (Hiscox, Citation2001; Milner, Citation1999). Later contributions have identified investment and vertical intra-industry trade as major drivers in the formation of FTAs and heterogeneity in FTA preferences between firms (e.g., Chase, Citation2005; Kim & Osgood, Citation2019; Manger, Citation2009, Citation2012).
2 Most of the regional and some of the early bilateral FTAs in East Asia have been portrayed as dominated by foreign policy motivations rather than by a trade liberalization rationale; the policymaking of these FTAs was largely top-down driven and arose out the initiative of politicians and/or bureaucrats while the private sector had only a minor role (e.g., Aggarwal & Govella, Citation2013; Aggarwal & Urata, Citation2006; Higashi, Citation2008; Ravenhill, Citation2010). In contrast, pressure from Japanese businesses was key in the move by the Japanese government to embrace regionalism (Manger, Citation2005; Yoshimatsu, Citation2005).
3 Additionally, these firm-level surveys do not consider the goods for which GSP schemes and FTAs have been used or they aggregate all goods in just a few (4-12) sectors. This article analyzes GSP and FTA utilization for up to 1,300 or 6,200 categories of goods.
4 As discussed below, exporters may continue using a GSP scheme instead of switching to the FTA if the FTA offers tariff rates that are initially (or permanently) above those granted by the GSP.
5 The utilization of FTA (and GSP) preferential tariffs requires abidance by specified "rules of origin" that establish whether a good has undergone sufficient transformation within the FTA bloc (or the GSP beneficiary developing country). Compliance with these rules may impose costs on producers if they need to modify their procurement and/or production patterns (Anson et al., Citation2005). Although rules of origin’s main goal is to prevent that exports from a third country enter the FTA through the country with the lowest tariffs, FTA member countries (or GSP-granting countries) can set stricter rules of origin for protectionist purposes.
6 Article XXIV of the General Agreement on Tariffs and Trade states that FTAs should fully liberalize a significant share of trade within a reasonable period. Although these requirements are usually interpreted as applying to 90% of all existing trade within 10 years, many FTAs do not comply with these limits. The WTO Doha Round also proposed reducing higher tariffs more rapidly than reducing lower tariffs, whereas in FTAs higher tariffs can be reduced more slowly or simply excluded.
7 The Japanese GSP program has a two-step system for withdrawing access. Under “partial graduation”, a good cannot claim GSP preferences for at least a year when it has achieved a high level of global competitiveness. Under “entire graduation”, a developing country is excluded as a beneficiary when it has reached a certain development stage and/or its exports exceed a specified share of the total value of world exports. See United Nations Conference on Trade and Development (UNCTAD) (Citation2017:4,14–15) for current graduation criteria.
8 For instance, GSP programs may require beneficiaries to adhere to preconditions related to, inter alia, good governance, human rights, sustainable development, and labor standards.
9 FTAs include arrangements to settle disputes and most also have provisions for suspension of commitments during emergencies, as well as for a negotiated amendment or even possible termination of the agreement. However, and unlike in GSP schemes, the unilateral withdrawal of FTA commitments carries an international reputation cost.
10 In a GSP scheme, product coverage, tariff rates, product-specific ceilings, and rules of origin are all unilaterally established by the granting country. Likewise, unilateral liberalization via GSP does not address potential non-tariff barriers in the developed economy. However, all of these issues are open to potential improvement during FTA negotiations.
11 See Section 3 and the Appendix for details on how the quantitative data were processed and analyzed,
12 See the Appendix for details on how qualitative data were collected and on how the evidence of business lobbying was assessed in the semi-structured interviews.
13 Thailand has a limited scope trade agreement with India (implemented in 2004) and full-fledged FTAs with Australia (2005), New Zealand (2005), Japan (2007), Peru (2011), and Chile (2015). As a member of the Association of Southeast Asian Nations (ASEAN), Thailand is also a party of six ASEAN-centered FTAs—with China, Japan, Korea, Australia/New Zealand, India, and Hong-Kong—which are less comprehensive and liberalize goods more slowly than their corresponding bilateral agreements.
14 Malaysia has implemented bilateral FTAs with Japan (2006), Pakistan (2008), New Zealand (2010), India (2011), Chile (2012), Australia (2013), Turkey (2015). Malaysia is also a member of the six ASEAN-centered FTAs.
15 Thailand and Malaysia have not only established bilateral FTAs with Australia and New Zealand, but they have also benefited from the GSP programs offered by these two developed economies; however, there are no disaggregated data on the utilization of these GSP schemes by Thai and Malaysian exporters.
16 Countries can subdivide HS codes beyond the 6-digit level.
17 These percentages refer to goods at the 9-digit (HS9) level in the regime for developing countries not classified as “least developed countries”.
18 Author’s analyses of FTA texts (Ministry of Foreign Affairs of Japan (MOFA-J), Citation2006a, Citation2006b).
19 JTEPA and MJEPA exclude some products that are covered by the Japanese GSP; for instance, sorbitol (HS6 code 290544) and some goods at level HS6 or higher within the following categories: alcoholic preparations, lobsters and crabs, decaffeinated coffee beans, coffee and tea extracts, and fur skins (Ministry of Foreign Affairs of Japan (MOFA-J), Citation2006a, Citation2006b). Dextrin and glues, as well as certain types of plywood and laminated wood are partially liberalized by the Japanese GSP, but JTEPA imposes quotas on the former group and excludes the latter.
20 More detailed definitions of all the variables are in pages S2 and S3 of the Appendix.
21 Japan applies a zero most-favored-nation tariff to about half of all tradable goods at HS6 level, but both the strength and the level of statistical significance of the correlations in Table 2 hold when they are calculated only for goods for which Japan applies a multilateral tariff greater than zero (data not shown).
22 The exporters of a given good may use GSP or FTA preferential tariffs for most of its exports (a high utilization rate) but if export values are low, the use of the GSP or FTA for the export of that good may represent a small share of the overall GSP or FTA utilization (a low utilization share rate). Conversely, the fact that a good accounts for a large share of the overall GSP or FTA utilization (a high utilization share rate) does not necessarily mean that its traders use GSP or FTA tariffs for most of their exports. These dynamics are discussed below and illustrated in Suppl. Tables S1 and S2.
23 See the Appendix for details. The adjusted GSP (or FTA) utilization rate does not take into account the export value of goods for which the preferential tariff margin offered by the GSP (or FTA) is zero; that is, it not only excludes the export value of the around 50% of goods at HS6 that can enter Japan tariff-free under the multilateral regime, but also that of goods for which the tariff rate in the GSP (or FTA) is the same as the multilateral most-favored-nation tariff rate.
24 The relatively low unadjusted average overall utilization indicates that most Thai exports enter Japan under zero most-favored-nation tariffs (e.g., mechanical and electrical machinery), whereas the higher adjusted utilization rate indicates that Thai exporters have a high use of the tariff savings offered by JTEPA.
25 Japan’s most-favored-nation tariff on petroleum products (HS6 codes 200900 to 271390) averages only 1.1%. JTEPA and MJEPA eliminated tariffs on most of these products.
26 For instance, the overall utilization by Mexican exporters of the North-American Free Trade Agreement stood at 64.0% in 2000, with 82.7% for adjusted utilization (Anson et al., Citation2005).
27 In the more than 12 years under study, only 39 and 54 goods at HS4 were represented among the top 20 exports through JTEPA and MJEPA, respectively (data not shown).
28 Of the top 20 Thai and Malaysian exports at HS4 through JTEPA and MJEPA in the first four complete years after implementation, 16 and 18 goods, respectively, overlapped with the top 20 Thai and Malaysian exports that used the Japanese GSP program during the previous four years (Tables 3 and 4). The cumulative utilization share of the non-overlapping top 20 Thai exports (not included in Table 3) between GSP (14 goods) and JTEPA (10 goods) were 15.6% and 22.8%, respectively. In turn, the cumulative utilization share of the non-overlapping top 20 Malaysian exports (not included in Table 4) between GSP (9 goods) and MJEPA (7 goods) were 11.4% and 6.8%, respectively. The description of goods at HS4 in Tables 3 and 4 has been shortened because of space limitations.
29 It should be noted that because a high utilization rate is not necessarily accompanied by a high utilization share rate, the cumulative share of the goods with the highest utilization rate in Suppl. Tables S1 and S2 is different from that of the top 20 exports in Suppl. Figures S1 and S2. Thus, both FTAs have also been used almost entirely (utilization ≥ 80%) to export some goods that account for a small share of total FTA utilization (utilization share ≤ 0.1%) (Suppl. Tables S1 and S2).
30 Interviews with government officials (mainly in the Thai Ministries of Foreign Affairs, of Commerce, and of Industry) as well as representatives from the peak and sectoral business associations and individual firms in Thailand. See the Appendix for details. Goods marked with a “Yes” in Table 3 and Suppl. Table S1 refer to those for which there is evidence that their Thai exporters pressed in favor of FTA liberalization. Businesses pressured the Thai government for or against JTEPA liberalization, not only at invited consultations but also through direct connections within government agencies and/or even the Cabinet. On the import side, Japanese automakers also lobbied for the removal of Thai tariffs on vehicles and intermediate inputs imported from Japan, a liberalization opposed by Western automakers (Manger, Citation2005, Citation2009; Postigo, Citation2014, Citation2016).
31 The Japanese GSP offers lower or no tariff on certain types of yarns, textiles and fabrics, but excludes footwear (HS2 code 64) and most garments (HS2 61 and 62). Japan’s most-favored-nation tariffs on footwear and garments average a high 16.2% and 9.2%, respectively. Thai exporters of organic chemicals as well as copper and articles thereof benefited from GSP preferential tariffs and now have high utilization and utilization share rates of JTEPA; however, except for Thai dextrins producers, there was no evidence in this study that these sectors had lobbied for liberalization in JTEPA.
32 Interviews with the heads of the committees on rules of origin in the Federation of Thai Industries and the Chamber of Commerce.
33 See footnote 19 for goods covered by the GSP but excluded by JTEPA. During 2007-2019, just three processed food goods at HS4 level (1602,1604,1605) accounted on average for a quarter of total JTEPA utilization. The share that the exports of garments in HS2 code 61 represents in total exports to Japan has doubled since JTEPA implementation. However, although JTEPA eliminated tariffs on footwear and on garments included in HS2 code 62, their exports have barely changed probably due to the trade diversion of these products from the FTA between Japan and Vietnam (Author’s calculations from data in www.trademap.org).
34 Interviews with officials from different government agencies (mainly in the Ministry of International Trade and Industry) and representatives in business associations and individual firms in Malaysia. See the Appendix for details. Goods marked with a “Yes” in Table 4 and Suppl. Table S2 refer to those for which there is evidence that their Malaysian exporters pressed in favor of FTA liberalization. In contrast to Thailand, where sectoral business associations and individual firms carried much of the lobbying in favor or against JTEPA, lobbying in Malaysia was mainly channeled through the peak business association, the Malaysian Manufacturers Federation; the most notable exception was the Malaysian Textile Manufacturers Association which, as in Thailand and for similar reasons (footnote 31), played a proactive role in favor of MJEPA. Also as in Thailand, on the import side, Japanese automakers also pushed for the liberalization of Malaysia’s protected automotive sector, but Malaysian and Western automotive firms assembling in Malaysia opposed opening up to imports from Japan (Manger, Citation2009; Postigo, Citation2014, Citation2016).
35 See footnote 19 for goods covered by the GSP but excluded by MJEPA. Just four goods at HS4 level, namely, plastic items (3907,3920,3923) and palm oil (1511) account for around a third of the overall utilization of MJEPA. Since 2012, both total palm oil exports to Japan and the use of MJEPA preferences for its exports have been halved.
36 FTA and GSP utilization rates and the FTA preferential tariff margin are expressed as percentages. The value of total exports and exports through the FTA are in current United States dollars as “free-on-board” (see Appendix).
37 The utilization of the Japanese GSP was averaged over the four years prior to the implementation of each FTA, while FTA utilization, preferential tariff margin, and trade value were averaged for periods of 2 or 3, 6, and 4 years, as indicated in Tables 6 and 7. The results obtained when calculating individual years were similar to calculating the mean of several years (not shown).
38 The utilization of an FTA should not influence the previous use of the GSP, the FTA tariff concessions that both countries agreed to, or trade values.
39 Colinearity analyses in the most complete specification (model 4) showed that for both FTAs and in all periods examined, tolerance coefficients were all > 0.9 and variance inflation factors (VIF) were all <1.1.
40 The robust standard errors were calculated with the HC3 estimator (Davidson & Mackinnon, Citation1993).
41 The declining capability of variation in GSP utilization to explain variation in FTA utilization over time is due not only to the expansion in FTA coverage to goods not included in the GSP but also to changes in the overall composition of exports to Japan (data not shown).
42 In that line, during the 2006–2011 period, 62.1% of the value of exports from Honduras—the country with the greatest dependence on the United States’ GSP in Manger and Shadlen (Citation2014:89)—to the United States via the Dominican Republic-Central America FTA corresponded to goods at HS4 level that Honduras had previously exported using preferential tariffs granted unilaterally by the United States (GSP and the Caribbean Basin Initiative schemes); this figure is similar to that of JTEPA with respect to the Japanese GSP (Author’s calculations from data in https://www.usitc.gov).
43 Most Asian countries have moved to a self-certification system of compliance with the rules of origin in FTAs. The FTA between the European Union and Vietnam entered into force in August 2020 and there is no historical data for analysis.
44 Author’s calculations using data from Trade Map (https://www.trademap.org/).
45 The restrictiveness of rules of origin in most Japanese FTAs is lower than those in the Japanese GSP (Kniahin et al., Citation2020).
46 See footnote 22.
47 Most of these goods are also excluded from the Japanese GSP (footnote 19).
48 See footnotes 31, 33, and 34.
49 See footnote 31.
50 The WTO framework still offers developing countries room to implement some country-specific development strategies. However, developing governments often surrender this policy space in exchange for market access in their bilateral FTAs with developed economies (Shadlen, Citation2005).
Additional information
Notes on contributors
Antonio Postigo
Antonio Postigo holds a PhD from the London School of Economics and Political Science. He has lectured at universities in the United States, Europe, and Asia and has worked for United Nations, The World Bank, and several think tanks. His main area of research is the comparative and international political economy of trade and economic integration, with a focus on East and Southeast Asia.