1. Introduction
Technical Barriers to Trade (TBTs) are on the rise at a time when the ghost of protectionism looms large across the world economy. The number of TBTs notified to the World Trade Organization (WTO) has been setting a record high each year since the trade organization was established in 1995. This stands in sharp contrast to the fact that tariff barriers have been reduced to the point at which actual tariff rates are approaching zero in advanced countries. According to the 2017 WTO annual review of the implementation and operation of the TBT Agreement, a total of 82 member economies submitted 2,585 TBT notifications compared to an average of 2,179 submissions in the previous five-year period (G/TBT/40, para. 3.1).
Technical regulations are inherently trade restrictive. They impose specific burdens – such as labelling requirements, standards on technical specifications, and quality standards – not only on domestic producers but also on foreign suppliers. On the surface, all the objectives of technical regulations look legitimate.Footnote 1 Underneath the surface, however, some regulations are more burdensome than others, particularly to foreign suppliers. It was no wonder that the 2017 TBT Committees discussed 178 Specific Trade Concerns (STCs), the largest number ever recorded. Among those cases, 151 have been raised repeatedly, indicating their unrelenting nature.
The United States (US) and the European Union (EU) stand out as the most avid players in terms of technical regulations, leading them to be ranked as the first and third, respectively, in the TBT hierarchy. They are also leading the way in which the norms of technical regulations are formed and implemented (Kelemen and Vogel, Reference Kelemen and Vogel2010; Young, Reference Young2015; Filipec, Reference Filipec2017; Michida, Reference Michida, Michida, Humphrey and Nabeshima2017).
Most notably, the EU adopted strong regulations on chemicals in 2007, commonly known as the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH). Its principle of ‘No Data, No Market’ is known to be one of the strictest regulations on chemical substances and chemical-containing products. Chemicals covered by this regulation are not allowed to enter the European market without compliance with REACH requirements (Heyvaert, Reference Heyvaert2009: 111–112; Naiki, Reference Naiki2010: 182–183).
Since the EU's notification of its implementation plan as a TBT in 2003, the REACH has caused a series of concerns for the EU's trading partners.Footnote 2 During the period from 2003 to 2017, a total of 36 countries raised 329 STCs on the REACH at the WTO TBT Committee sessions. Much ink has thus been spilled on the negative impacts of REACH on international trade. For instance, it was transcribed in the TBT Committee records of November 2004 that the member countries were concerned that ‘the REACH would be more difficult for non-EU manufacturers to comply with than for EU manufacturers’ (G/TBT/M/34, para. 23). The rest of the comments have not diverged from this point: REACH could put foreign suppliers under more trade-restrictive conditions than domestic producers (e.g. G/TBT/M/35, paras. 15–24; G/TBT/M/36, paras. 10–17; G/TBT/M/39, paras. 45–52; G/TBT/M/40, paras. 43–52).Footnote 3
Some concerns notwithstanding, new regulations encourage new technologies and the more efficient allocation of scarce resources under certain conditions (Porter, Reference Porter1990; Porter and Linde, Reference Porter and Linde1995). As a matter of fact, some foreign suppliers have better adjusted themselves to the new trade environment than others, as indicated by the growing revenues of chemical exports to the EU during the period of 2007–2017. For instance, the average annual growth rates of South Korea, China, and Taiwan were 15%, 11%, and 9%, respectively, while the world average remained at 4% during the same period.Footnote 4 This observation raises a puzzling question: who embraces technical barriers to trade better than others and why?
This study begins with this question and aims to examine how different countries have reacted to REACH regulations and what factors have motivated some, if not all, of them to harmonize their domestic policies with REACH regulations. Of course, an individual company undertakes adjustment to a new regulatory environment in the export market through innovation. Nevertheless, the regulatory policy of export countries is equally, or even more, important than individual corporate responses. From a foreign government's perspective, a voluntary convergence of its regulatory policy is crucial for maintaining and encouraging the competitiveness of its exporting companies in the European market. Other things being equal, regulatory harmonization and convergence can promote trade and facilitate innovation (Blind, Reference Blind2001; Blind and Jungmittag, Reference Blind and Jungmittag2005; De Frahan and Vancauteren, Reference De Frahan and Vancauteren2006; Portugal-Perez et al., Reference Portugal-Perez, Reyes and Wilson2010; Vigani et al., Reference Vigani, Raimondi and Olper2012).
The remainder of this study unfolds in four sections. Section 2 discusses the theoretical background of regulatory harmonization as a source of adaptive innovation on the part of an exporting country. TBTs exist due to differences in standards, technical regulations, and conformity assessment. They can be trade restrictive unless properly harmonized. This section identifies three distinct mechanisms that induce the harmonization of technical regulations: communication-based, institution-based, and competition-based harmonization.
Section 3 identifies three causal mechanisms of policy convergence that are relevant for the EU REACH case: transnational communication, intergovernmental institutions, and competition for export markets. Based on the extant convergence literature, this section lays out a theoretical foundation for the empirical work that follows.
Section 4 conducts a random-effects ordered logistic regression analysis to test the relationship between the three causal mechanisms and the level of regulatory harmonization to the REACH. The estimation results indicate strong statistical support for two diffusion mechanisms – communication and competition – as predicted.
Section 5 summarizes key findings and draws policy implications. It highlights the fact that TBTs can promote regulatory innovation in some exporting countries when relevant factors exist, particularly the two convergence mechanisms: transnational communication, and competitive pressure for export markets.
2. Mechanisms of Regulatory Harmonization
TBTs include a wide variety of measures that countries adopt to regulate ‘product characteristics or their related processes and production methods, including the applicable administrative provisions, with which compliance is mandatory’ (technical regulation) or to provide ‘rules, guidelines or characteristics for products or related processes and production methods, with which compliance is not mandatory’ (standard) (TBT Agreement Annex 1.1). These measures are usually used for public purposes such as regulating markets, protecting consumers, and/or preserving natural resources.
In consequence, if not in intent, they can also be used to discriminate against imports in order to protect import-competing domestic industries. Country-specific technical regulations may give domestic producers advantages by imposing mandatory standards with which foreign suppliers find it more difficult to comply (Swann et al., Reference Swann, Temple and Shurmer1996: 1298–1299). As foreign suppliers are deterred from entering the market, domestic producers will increase their supplies and enjoy additional ‘producer surplus’, which is transferred from foreign suppliers.
Foreign governments have two options. They can either complain about the trade restrictiveness of importing countries’ new regulations and challenge them at the WTOFootnote 5 or adopt similar regulations on their own to mitigate the adverse effect of new barriers to trade. The latter approach – regulatory harmonization – is of interest to this study. By adopting similar measures at home, foreign governments can induce their exporting producers to comply with a new regulatory environment abroad more easily.
Then why do some exporting countries facilitate regulatory harmonization and market-conforming policies instead of settling the differences through international arbitration? Why do some countries embrace otherwise trade-restrictive regulations imposed by other countries? Based on the existing regulatory harmonization and convergence literature,Footnote 6 this study identifies three distinct but related mechanisms through which regulatory harmonization and policy convergence take place vis-à-vis REACH regulations.Footnote 7
First, transnational communication can facilitate policy convergence by disseminating best policy practices. Regular meetings and information sharing among governments can motivate them to emulate policy measures that work well in other countries (DiMaggio and Powell, Reference DiMaggio and Powell1991; Simmons and Elkins, Reference Simmons and Elkins2004: 175). Most notably, Allport's (Reference Allport1954) contact hypothesis, also known as intergroup contact theory, posits that under certain conditions interpersonal contact has a positive impact on relations between individuals and groups, and that close contact effectively reduces prejudices between them. This can also be extended to interstate relations. The simple act of intergovernmental communication can facilitate international cooperation (Haas, Reference Haas1964). The key mechanism at work is a better understanding and appreciation of others and their ways, customs, practices, and concerns. Policy diffusion can be further accelerated if intergovernmental communication channels are formalized rather than remaining ad hoc or temporary features that may lapse over time (Strang and Meyer, Reference Strang and Meyer1993; Kern et al., Reference Kern, Kissling-Näf, Landmann, Mauch and Löffelsend2001). They can also take the form of loose networks of policy experts (Haas, Reference Haas1992).Footnote 8
Second, regime-based harmonization can emerge when like-minded countries work together to address common global or regional problems such as climate change, biodiversity, and epidemic diseases. Due to the cross-border nature of such problems, national efforts alone fall short of resolving the negative externalities of certain issues (Holzinger et al., Reference Holzinger, Knill and Sommerer2008: 557). Increased communication and connectivity between government officials and non- governmental actors can help develop an international regime, defined as ‘sets of implicit or explicit principles, norms, rules and decision–making procedures around which actors’ expectations converge in a given area of international relations’ (Krasner, Reference Krasner1982). If there is an international regime or institution around which countries’ expectations converge, cooperation and collaboration among like-minded countries would become much easier. In particular, dense networks of intergovernmental organization (IGO) membership are known to contribute to the peaceful settlement of mutual conflicts because commonly acquired and shared knowledge through participation in IGOs can reduce uncertainty and the transaction cost of international interactions (Drezner, Reference Drezner2001: 60).Footnote 9 In fact, the pace of network building, official and unofficial, formal and informal, and bilateral and multilateral, has grown substantially in the area of technical regulations over the past decades.Footnote 10
Third, regulatory harmonization can also result from fierce competition for export markets. This competitive pressure can work in two opposite directions. It can facilitate the ‘regulatory race to the bottom’ in which countries keep lowering regulatory standards to reduce production costs (Dobbin et al., Reference Dobbin, Simmons and Garrett2007: 457–460).Footnote 11 Alternatively, the ‘regulatory politics of vortex’ can induce rivals in the export market to adopt higher technical standards to better respond to foreign consumer preferences for eco-friendly and high-quality products (Vogel, Reference Vogel1997; Holzinger and Knill, Reference Holzinger and Knill2004; Holzinger et al., Reference Holzinger, Knill and Sommerer2011; Perkins and Neumayer, Reference Perkins and Neumayer2012). This is particularly true of large export markets such as the EU. As Drezner (Reference Drezner2005: 843) notes, ‘the larger the economy, the stronger the pull for producers to secure and exploit market excess. As demand increases, firms will have greater incentives to mirror that market's preferences’. Well-informed governments vying for a fixed quantity of trade have little choice but to comply with new technical regulations to keep their exports competitive, especially when other countries competing in the same export market have already done so (Dobbin et al., Reference Dobbin, Simmons and Garrett2007: 457). Prakash and Potoski (Reference Prakash and Potoski2006) present empirical evidence connecting the market size of an importing country to the regulatory harmonization of the exporting country. They found that trade encourages exporting countries to adopt ISO 14001, the high-level environmental standard, when their major export markets have already adopted that standard.
3. Data and Methods
3.1 Estimation Model
This study uses a panel data regression method for 78 WTO member countries with consistent records of exporting chemical products to the EU during the period of 1995–2017. Twenty-seven EU countries are subject to examination, except for Croatia, which joined the EU in 2013 after the EU REACH regulation came into force in 2007. Meanwhile, Iceland and Norway are included in the analysis as EEA (European Economic Area) countries that are subject to the REACH. The UNCOMTRADE database is used as it classifies total exports by product categories according to the harmonized system (HS) code. Export data on chemical products numbered 28–38 are extracted, with the exception of No. 30 because pharmaceutical products are exempted from the application of the REACH.
This study modifies the gravity model to estimate the degree to which exporting countries harmonize their regulatory policies with the REACH at the bilateral level. There are a number of studies that take advantage of the extensive applicability of the gravity model when measuring the trade policy effects of tariff and/or nontariff measures on a pair of trading partners (Otsuki et al., Reference Otsuki, Wilson and Sewadeh2001; Anders and Caswell, Reference Anders and Caswell2009; Liu and Yue, Reference Liu and Yue2009; Bao and Qiu, Reference Bao and Qiu2010; Ferro et al., Reference Ferro, Otsuki and Wilson2014; Crivelli and Groeschl, Reference Crivelli and Groeschl2016; Koo and Kim, Reference Koo and Kim2018). Even in the WTO-UNCTAD guidance book, the gravity equation is introduced as a proper analytical tool for trade policy analysis (WTO and UNCTAD, 2012: 103–105).
The dependent variable of this study is measured on a nominal scale ranging from 0 to 4. As equal distances between the levels cannot be reasonably assumed for the ordinal dependent variables, we can use an ordered logistic regression model. With a random-effects ordered logistic regression model, this study estimates parameters that best describe the extent of regulatory harmonization adopted by different countries in accordance with the REACH. Unlike fixed-effect estimators, the parameters of the random-effect model can explain the variance of the dependent variable between countries as well as within a country. To address the heteroskedasticity of error terms, this study utilizes robust standard errors clustered by country. Five different regression models have been estimated in a hierarchical order from the base (only control variables included) to the full model (all explanatory variables included) to show that the estimates do not change significantly.
In this model, it is posited that a five-point scale ordered dependent variable Y it is derived from the latent continuous variable 
$Y_{it}^\ast $, such that
$$Y_{it}\;= \left\{{\matrix{ {\matrix{ {0}&{\rm if}&\quad\quad\,\,\,{Y_{it}^\ast \le T_0} \cr {1}&{\rm if}&{T_0 \lt Y_{it}^\ast \le T_1} \cr {\vdots } \cr {4}&{\rm if}&{T_3 \lt \;Y_{it}^\ast .}\hfill $$where T 0.1.2.3 represents a set of four thresholds or cut points to classify the country's regulatory harmonization status into five different levels. In regard to the latent variable 
$Y_{it}^\ast $, the whole latent regression model can be defined as
$$Y_{it}^\ast= X_{it}\beta + \upsilon _i + \;\varepsilon _{it}$$or specified as:
$$\eqalign{Y_{it}^\ast & = \beta _1communication_{it} + \beta _2institution_{it} + \beta _3competition_{it} + \beta _4REACH\,\,impact_t \cr& \quad + \beta _5\ln \lpar {GDP\,\,per\,\,capita_{it}\,\times \,EU\,GDP\,\,per\,\,capita_t} \rpar \cr & \quad + \beta _6\ln \lpar {\,population_{it}\,\times \,EU\;population_t} \rpar + \beta _7EU\,GSP_{it} + \beta _8EU\;FTA_{it}\;+ \;\upsilon _i + \varepsilon _{it}} $$where υi is independent and identically distributed 
$N\lpar {0\comma \;\;\sigma_\upsilon^2 } \rpar $, and explains unobservable individual specific effects which do not change by time (random effects). The error term ɛit is posited to follow logistic distribution with mean zero and variance π 2/3, which is independent of υi (Green and Hensher, Reference Green and Hensher2010). The parameter β k can be obtained using maximum likelihood estimation (MLE) and assuming conditional distribution for the dependent variable with probability of observing outcome m (where m = 0, 1, 2, 3, 4).
The logistic model estimates the odds ratio (OR) of an outcome m (e.g., the adoption of a REACH-like integrated chemical control system) occurring relative to the other category (e.g., minimal or no chemical regulations). The OR can be obtained by exponentiating the ordered logit coefficients (
$exp\lpar {\widehat{{\beta_k}}} \rpar \rpar $. The log-odds (logit) of cumulative probabilities for categories of m are modelled as linear functions of independent variables. If we view the change in the levels of the dependent variable in a cumulative sense and interpret the coefficients in odds, we are comparing the outcome in categories greater than m versus the one in categories less than or equal to m. In other words, for a one unit change in the independent variable, the odds for cases in a category that is greater than m are the OR times larger than cases in a category that is less than or equal to m.Footnote 12
3.2 Dependent Variable
The dependent variable, convergence, is a five-point scale ordered variable to measure the degree to which a country adopts chemical regulations in accordance with the EU REACH in a given dyad year. Each scale is derived from a stepwise hierarchy in terms of chemical regulatory harmonization with the REACH. Individual conditions must be met to move up to the next level of scale.
0 = Minimal or no chemical regulations notified to the WTO TBT Committee are present. The exporting country makes no effort at harmonizing its regulatory policy with the EU REACH.
1 = Partial control or standardization of certain chemical substances and products such as fertilizer and detergent are present. The exporting country makes a moderate and partial adjustment without adopting a REACH-like integrated chemical control system.Footnote 13
2 = Relevant regulations used as a basis for a REACH-like integrated chemical control system are present, implying that the exporting country manages and controls chemicals in a comprehensive way. Examples include identification, classification, and labelling requirements for all chemicals. These requirements are a prerequisite for implementing the entire REACH-like procedures.Footnote 14
3 = Regulations requiring registrants to submit and register safety data for all new chemicals to a national regulatory authority are present. The exporting country explicitly applies a REACH-like integrated chemical control system.
4 = Regulations requiring registrants to submit and register safety data for all existing chemicals to a national regulatory authority year after year are present. This level of chemical regulation is fully harmonized with the REACH.Footnote 15
These indicators show the level of stringency of chemical management, which is one of the most important requirements introduced by the REACH. As the regulatory race-to-the-top or race-to-the-bottom literature suggests, any (new) regulations are either upward (more stringent) or downward (less stringent) in terms of the strength of their requirements. Using the TBT notification archive (WTO TBT IMS), all the relevant chemical regulation policies notified by 78 countries as TBTs have been coded accordingly.Footnote 16
3.3 Independent Variables
Following the regulatory harmonization and convergence literature, this study hypothesizes the relationship between the degree of regulatory convergence to the REACH and the regulatory harmonization mechanisms.
Hypothesis 1: If an exporting country communicates more with the EU vis-à-vis the REACH than other countries and/or than before, the exporting country is more likely to harmonize its chemical regulations with the REACH than other countries and/or than before.
The communication hypothesis emphasizes information exchange through regular channels. There might be many different paths to transfer and obtain information about potentially better forms of technical regulation. Among them is the TBT Committee that allows WTO members to discuss any STCs that they consider necessary for further information and consultation during the WTO TBT Committee sessions. It is a privilege available for all WTO members pursuant to Article 10 of the WTO TBT Agreement, which demands that a regulation-adopting country answer all inquiries from other members. In terms of regular channels, the WTO holds TBT Committee meetings three times a year.
During the period 1995–2017, the most frequently cited reason for STCs was to ‘seek further information and clarification’ from TBT-adopting countries, which means that concerned parties are more likely to ‘work towards mutually acceptable solutions’ (G/TBT/40, para. 3.32) as a result of communication mechanism.Footnote 17 In fact, studies have found that many STCs disappear before growing into formal WTO litigations, even though the number of STCs raised each year is on the rise (Horn et al., Reference Horn, Mavroidis and Wijkström2013; Holzer, Reference Holzer2019).Footnote 18
In the estimation model, communication cumulatively measures how many times a country has communicated with the EU by raising STCs at STC discussion sessions until a given year during the period 2003–2017. Once an STC is raised, bilateral or multilateral forums for consultation must follow between STC-raising countries and respondent countries. As noted earlier, Australia has submitted the largest number of STCs (32), followed by the US (31), China (27), Japan (24), and Canada (23). A total of 30 out of 78 countries have raised STCs on the REACH and thus communicated with the EU either bilaterally or multilaterally.Footnote 19
Hypothesis 2: If a country has a joint membership with the EU in international IGOs, it is more likely to harmonize its chemical regulations with the REACH than other countries without a joint membership and/or than before when a joint member is absent.
There are two main conditions that stimulate the institutional mechanism to work: one is the pan-regional nature of problems and the other is the presence of an international institution or regime to coordinate matters. The case considered here satisfies the first condition in that the objective of the REACH regulation is to protect people and the environment from the misuse of chemical substances. The case satisfies the second condition as well in that the United Nations (UN) has proactively promoted chemical safety. The United Nations Environmental Program (UNEP) has endorsed a chemical policy framework called the Strategic Approach to International Chemicals Management (SAICM) and led member countries to make a commitment to follow in 2006 (Lee, Reference Lee2015: 400). Moreover, a multilateral consensus to create globally harmonized chemical classification standards was first formed in 1992. In 2002, it was formally institutionalized as the Globally Harmonized System of Classification and Labelling of Chemicals (GHS) (Winder et al., Reference Winder, Azzi and Wagner2005; United Nations, Reference Nations2017). According to the United Nations Economic Commission for Europe (UNECE), it is confirmed that 72 countries have implemented the GHS standard so far in various ways.Footnote 20 If a country takes part in these regimes and institutions, it is more likely to adopt chemical regulations in accordance with the REACH.
Institution is a binary variable to measure the participation in the UN institution: a value of 1 is given from the year when a country implements the GHS for the first time; it is otherwise 0. A total of 37 out of 78 countries are confirmed to have implemented the GHS standard in their own ways.Footnote 21
Hypothesis 3: If a country has more commercial interests to protect in the European chemical market than other countries or than before, the country is more likely to harmonize its chemical regulations with the REACH than other countries and/or than before.
It is obvious that the European market is attractive enough for many exporters to make the changes necessary to survive in that market. Notwithstanding the strict level of regulation, many countries will be induced to make their regulations harmonized with the REACH. It is no coincidence that major exporters of chemicals such as South Korea, Taiwan, and China have eventually adopted new chemical regulations in line with the REACH (EU SME Center, Reference Centre2011).
Finally, competition measures the share of chemical exports to the EU in a country's total export revenues during 1995–2017. This variable is a proxy to capture the competitive pressure that an exporting country feels from the EU chemical market: if a country depends on the chemical export to the EU market for its national accounts more than others do and/or more than before, it will feel more pressure to maintain its EU market share. It is log transformed to obtain normality as follows:
$$Competition{\rm \;} = {\rm log}\left({\displaystyle{{Total\;annual\;export\;revenue\;of\;chemical\;products\;to\;the\;EU} \over {Total\;annual\;export\;revenue\;of\;all\;goods\;\;to\;the\;world}}} \right)$$3.4 Control Variables
The estimation model includes a set of control variables that can influence regulatory harmonization. REACH impact is a binary variable to measure the effect of policy implementation per se. It gives a value of 1 from 2007 when the regulation was implemented (2007–2017); otherwise, it is 9 (1995–2006). This study sets a time lag of one year.
In accordance with the gravity model, two multiplicative variables – GDP per capitait × EU GDP per capitat and populationit × EU populationt – are included to capture the joint effect of economic size on the likelihood of stringent chemical regulations to be adopted by an exporting country. Log normalization is used to squeeze the values of GDP and population.
EU GSP is a dummy variable, which stands for the generalized system of preferences offered to developing countries by the EU. The concept of GSP is to help the least developed and developing countries stand on their own feet: if a country is designated to be in the GSP group by the EU, it can receive preferential treatments including lower tariff rates and favourable quota. Other things being equal, it is assumed that GSP countries are more likely to adjust their chemical regulations to the REACH than non-GSP countries.Footnote 22
Lastly, EU FTA measures the effect of a free trade agreement between an exporting country and the EU on policy convergence. It is coded as 1 if there is a concluded FTA in a given year; otherwise, it is 0. Other things being equal, FTA member countries are more likely to harmonize their regulations with the REACH. Table 1 shows the descriptive statistics of all the variables.
Table 1. Descriptive statistics
Note: a. The UNCOMTRADE omits some pieces of data and thus this study considers them as missing values and regresses the model without including them.
4. Estimation Results
The estimation results show strong statistical support for two diffusion mechanisms – communication and competition – as predicted. These results show who embraces technical barriers to trade better than others and why. Table 2 reports the estimation results.
Table 2. Estimation results
Note: *, **, and *** denote the level of significance at 10%, 5%, and 1% respectively.
First of all, the results support Hypothesis 1 that the intergovernmental communication mechanism will facilitate regulatory convergence. In the full model (5), other things being equal, a one-unit increase in communication measures is expected to increase the odds of obtaining convergence = n + 1 to 1.097 times greater than the odds of obtaining convergence ⩽ n. This can be interpreted in two different but related terms: (1) if a country has more chances of communication with the EU than other countries, the likelihood of making its regulation more harmonious with the REACH increases more than with other countries (‘between’ effect); and (2) if a country communicates more with the EU than before, the likelihood of its regulations becoming harmonious with the REACH increases more than before (‘within’ effect).
Secondly, Hypothesis 2 is not statistically supported by any models, with a p value larger than the significance level of 10%. We did not find evidence that the joint participation of countries in a credible international institution generates strong commitment or peer pressure to comply with the institution's guidelines.
Thirdly, the findings suggest that competition pressure in the European chemical market encourages exporting countries to harmonize their regulations with the REACH. The full model (5) indicates that a one-unit increase in the competition score, either between countries or within a country, will make the odds of convergence = n + 1 as compared to convergence ⩽ n 1.205 times larger. As predicted, strong commercial ties have a strong influence on regulatory harmonization: if the total export revenue of an exporting country depends on the chemical export to the EU market more than other countries and/or more than before, that country is more likely to harmonize its regulatory policies with the REACH in order to maintain the market share in the EU. This finding is of particular interest because it supports the theory of the regulatory politics of vortex or the regulatory race to the top, as opposed to the theory of the regulatory race to the bottom.
With regard to the control variables, the results of all the five models consistently show that implementation of the REACH has a statistically significant positive impact on the level of regulatory harmonization. With other variables held at their means, for a one unit increase in the enforcement of the REACH, the odds of obtaining convergence = n + 1 versus convergence ⩽ n are 2.051 times larger.
The classical gravity model variables – the multiplication of GDP per capita and population – have a statistically significant positive impact on the level of regulatory harmonization with p values lower than 1% in all models. The greater the log-transformed multiplication of GDP per capita incomes of an exporting country and the EU, the higher the expected level of regulatory harmonization (OR = 7.996), indicating that rich countries can afford stronger protection to their citizens from chemical hazards. In a similar vein, the log-transformed multiplication of populations of an exporting country and the EU has a statistically significant positive impact on the level of regulatory harmonization (OR = 3.403). Other things being equal, populous countries are more conscious about hazardous chemicals and thus have a higher likelihood of stringent measures to regulate them.
Finally, the estimation results for institutional proximity variables – EU GSP and EU FTA – are mixed. None of the five models obtained statistically significant results for EU GSP to reject the null hypothesis that the favourable export conditions made available by the preferential agreement are not exploited by developing countries. The impact of EU FTA on the regulatory harmonization is found to be statistically significant in models 2 and 5. This indicates that, unlike GSP partners, EU's FTA partners take advantage of their preferential market access to the EU.
5. Conclusion and Policy Implications
This study began with an observation that technical barriers to trade are on the rise across the world. For some, this is a worrisome development because TBTs are inherently trade restrictive. For others, however, TBTs are not necessarily counter-productive because under certain conditions they can provide incentives for innovation and a race to the top. Why then do some countries do better than others in terms of addressing new regulatory challenges imposed by an importing country?
The European Union stands out as one of the most active advocates of technical regulations for a variety of products and purposes, Using the case of the European REACH as one of the most stringent regulations on chemicals, this study attempted to unravel the underlying causal mechanisms for regulatory harmonization adopted by different exporting countries. The study first measured how exporting countries have reacted to the REACH regulation and then established three causal mechanisms for regulatory harmonization: transnational communication, intergovernmental institutions, and competitive pressure for export markets.
The random-effects ordered logistic regression analyses yielded strong statistical support for two causal mechanisms: (1) if an exporting country communicates more with the EU vis-à-vis the REACH than others and/or than before at the STC discussion sessions, it is more likely to harmonize its chemical regulations with the EU's REACH than others and/or than before due to greater understanding and clarity (Hypothesis 1 on communication); and (2) if the total export revenue of a country depends relatively more on the chemical export to the EU market than other countries do and/or than before, the country is more likely to harmonize its chemical regulations with the EU's REACH than others and/or than before due to greater competitive pressure (Hypothesis 3 on competition). Hypothesis 2 on intergovernmental institutions found no statistical support for its causality.
The key implications of these findings are as follows. Given the fact that two different harmonization mechanisms work for the REACH, the technical regulations adopted by the REACH can be market-friendly rather than market-distorting. This in turn implies that technical regulations, if understood correctly through communication and/or motivated by strong commercial incentives, can create upward pressure for global regulatory harmonization. South Korea, Taiwan, and China, among others, illustrate this point. It is seemingly ironic that these export-oriented Asian countries have gained more than the others (and more than before) in the European chemical market since the implementation of the REACH. The key is in the regulatory harmonization that these countries have adopted vis-à-vis the REACH.
These findings also support the regulation-driven innovation or innovation-friendly regulation argument. Many product innovations can come from the rigorous process of complying with more stringent regulatory requirements. The positive outcomes that the two seemingly contradictory concepts – regulation and innovation – can produce are worth further research. In this regard, more attention will have to be paid to the regulatory politics mechanism of exporting countries, which was treated as a black box in this study. For instance, Wilson (Reference Wilson1980) explains that we can better understand how business and government influence each other in a regulatory context by examining how each stakeholder affected by a particular regulation perceives its costs and benefits.
Finally, the findings of this study will contribute to the broad literature on regulatory convergence and the quantitative and qualitative effects of technical regulations and standards in various areas ranging from agriculture to food safety (McDonald, Reference McDonald2005; Mangelsdorf et al., Reference Mangelsdorf, Portugal-Perez and Wilson2012; Vigani et al., Reference Vigani, Raimondi and Olper2012; Lazo and Sauve, Reference Lazo and Sauve2018). It is particularly important during a time when a growing number of governments are being asked to do more to better protect their citizens with or without an intention to discriminate against imports and to protect import-competing domestic industries. This study can also branch out to non-state actors who can affect the way in which their formal counterparts manage such global issues in a country-specific manner (Auer, Reference Auer2000).
Supplementary Materials
To view supplementary materials for this article, please visit https://doi.org/10.1017/S1474745620000130.
Acknowledgement
We would like to thank the two anonymous reviewers and the Editor, L. Alan Winters CB, for their constructive comments. In addition, we thank Professor Kilkon Ko and Dr. Seo Young Kim for their helpful comments on earlier versions of this article. Finally, we are grateful to the Ministry of Education and the National Research Foundation of the Republic of Korea (NRF- 2017S1A3A2067636) for their support.
