Trade-related issues in the regulation of genetically modified organisms † Second version - Comments welcome -December 2001

Olivier Cadot HEC Lausanne and CEPR

Akiko Suwa-Eisenmann INRA and DELTA

Daniel Traça Insead-Singapour Abstract This paper examines some of the trade-related aspects of the transatlantic conflict over agricultural biotechnology regulation. We first review the notion of ‘regulatory protectionism’ and the difficulties arising in its definition. We also discuss its economic rationale in the light of the strategic trade policy literature. Next, we turn to an empirical assessment of the trade effects of EU biotechnology regulations, verifying the evidence in support of US claims that those regulations have hurt US agricultural exports. Preliminary econometric results at a disaggregated level seem to indicate that there was no shock on US exports of corn seeds for planting in the years 1997-2000, but that there was a negative shock on other types of corn, suggesting that downstream traders and food retailers’ private decisions not to purchase GM products were more important than cultivation bans in explaining any drop in US corn exports to the EU. Evidence for soybeans points to a price effect. The paper then turns to a comparison of GMO regulation in the US and EU, assessing the importance of trade vs. nontrade issues. Résumé Cet article examine l’aspect commercial du conflit transatlantique portant sur la régulation de la biotechnologie agricole. Une première partie cherche à définir la notion de ‘régulation protectionniste’. La littérature sur la politique commerciale stratégique peut apporter des fondements économiques à une utilisation protectionniste de la régulation. A la lumière des arguments théoriques, nous examinons empiriquement les effets sur les exportations américaines, des directives européennes concernant les OGM, avant de comparer les approches américaine et européenne en termes de régulations.

†

Paper prepared for the workshop on European and American Perspectives on Regulating Genetically Engineered Food, Insead, 7/8 June 2001. Cadot and Suwa-Eisenmann gratefully acknowledge support from the Commissariat Général au Plan (grant 6.2000) and INRA. We are also grateful to Jose Anson for superb research assistance, to Marion Desquilbet for very useful comments on a first draft, and to Jacques Gallezot, Annie Hofstetter and Chantal LeMouel for providing us with trade data. Any remaining error remains, however, our sole responsibility.

1. Introduction In July 1999, US Agricultural Secretary Dan Glikman declared in a National Press Club speech: “Yesterday’s news was that the WTO affirmed our view that the EU is unjustifiably blocking US ranchers from selling beef produced with completely tested and safe growth hormones. […] I can assure you that trade in GMOs is looming larger over US-EU trade relations in all areas. […] We cannot let others hide behind unfounded, unwarranted scientific claims to block commerce in agriculture. […] Right now, we are fighting the battles on ensuring access to our products on many fronts. […] These are not academic problems. For 1998, 44% of our soybeans and 36% of our corn are produced from genetically modified seeds. While only a few varieties of GMO products have been approved for sale and use in Europe, many more have been put on hold by a de facto European moratorium on new GMO products. […] I am extremely concerned that failure to work out these biotech issues in a sensible way could do deep damage to our next trade round and affect both agricultural and non-agricultural issues. […] To forestall a major US-EU trade conflict, both sides of the Atlantic must tone down the rhetoric, roll up our sleeves and work toward conflict resolution based on open trade, sound science and consumer involvement. I think this can be done if the will is there. However, I should warn our friends across the Atlantic that, if these issues cannot be 1 resolved in this manner, we will vigorously fight for our legitimate rights.”

In another speech given the same year, Glickman also underscored what he viewed as the need to “ensure the continued effectiveness of the rules governing sanitary and phytosanitary measures, so that science prevails and nations cannot mask protectionism behind unvalidated, secretive studies.” 2 The USDA’s Economic Research Service made essentially the same point, if in a less adversarial tone, when noting in April 2000 that “[t]he European Union (EU) approval process for imports of bioengineered varieties has been a particular source of consternation for US exporters”, adding that “fears of having shipments delayed or halted if unapproved varieties are commingled with approved varieties has prompted some US corn exporters to forego the EU market altogether” (Ballenger et al. 2000). These quotes, a small sample from a large set of US official statements on the biotechnology issue, highlight a view widely shared in America in the late 1990s that the biotechnology dispute was essentially about market access. 1

Speech delivered before the National Press Club, Washington, DC, July 13, 1999, release # 0285.99, available at www.usda.gov/news/releases/1999/07/0285. 2 Speech given at the World Agricultural Congress, St Louis, May 24, 1999, release # 0229.99, available at www.usda.gov/news/releases/1999/05/0229. 2

American complaints about EU market-access restrictions in agricultural trade are nothing new. However, as Agriculture Secretary Glickman remarked, the biotechnology dispute had the potential to be more damaging than previous ones. There were several reasons for this. First, agricultural applications of biotechnology held the promise of new and sizable business opportunities for American companies which had leadership positions in the development of genetic engineering techniques. In this regard, the biotechnology dispute had the characteristics of a high-tech industrial dispute. Second, as its name indicates, agribiotech is an agricultural technology, and the dispute had the potential to damage other, traditional US agricultural exports, especially given the refusal of US grain exporters to segregate bio-engineered from conventional seeds. This combination of agricultural issues ? always contentious? and high-tech ones ? portent of suspicions? created a context favorable to escalation. In addition, one of the US’s key concerns was to avoid letting a series of precedents being set whereby the EU would use food-safety arguments to restrict trade in foodstuffs without adequate scientific backing, as it had in a previous dispute over hormone-treated beef. These publicly-expressed concerns could be interpreted in two ways, not mutually exclusive. On one hand, they could be taken as implying that EU regulatory restrictions, while not designed to affect international trade, had nevertheless the unintended effect of restricting it. But, on the other hand, they could also be taken as accusations of deliberate hidden protectionism designed to protect domestic farming, seed or biotechnology interests from US competition. Sorting out the trade-related issues in the biotechnology dispute is of crucial importance. Grinding down traditional forms of trade protection through multilateral negotiations, however useful and necessary, does not by itself eliminate pressures from special interests; so new forms of protectionism are likely to appear, e.g. in the form of discriminatory standards. As volatile technology-based competition unsettles stable oligopolies in traditional industries like agri-food, the temptation will be strong to use standards as surrogates for strategic trade policy. At the same time, public-opinion concerns crystallizing around idiosyncratic events such as BSE can trigger trade-restricting regulatory responses, generating friction and suspicions even in the absence of a deliberate protectionist intent. In this regard, the biotechnology dispute may well be a forerunner of many such disputes. It is therefore crucial, lest misunderstandings accumulate, to get a good grasp of the rationale (domestic or trade-related) of product standards, of their intentional or unintentional trade effects, and of causal links between the two. The present paper is a contribution to that effort. In section 2, we recall the background of the nascent conflict. In section 3, we examine, in general, the notion of regulatory protectionism and the rationale behind the strategic manipulation of product standards, arguing that a proper definition of regulatory protectionism is elusive. In section 4, we try to identify econometric evidence of a ‘GMO effect’ on US-EU bilateral agricultural trade flows in corn and soybean, two major crops potentially affected by the dispute. We find that the evidence is muddled, suggesting that the market-access issues stressed early on in the dispute by various US officials were perhaps not of such critical importance as these officials were implying. Finally, in section 5, we 3

propose an interpretation of the regulatory conflict based on domestic rather than marketaccess considerations.

2. Background

2.1 Biotechnology and US-EU Agricultural Trade The impact of biotechnology on US-EU agricultural trade concerns essentially two crops: corn and soybeans. At current trade levels, soybeans are where the stakes are highest. 42% of the US soybean crop is exported, and the EU accounts for 33% of those exports or 14% of total output, well above a billion dollars of annual exports, although these numbers vary substantially over time. US soybean exporters have been under intense price pressure in 1997-98, losing a quarter of their market share worldwide: whereas world soybean trade remained roughly constant, US exports dropped from 26 million tons in 1997 to 20 million tons in 1998.The only market where US exporters were able to maintain their market share was Mexico, where they traditionally hold a dominant position. USDA analysts explain these adverse market developments by intense price competition, 3 as conventional soybeans are a price-sensitive commodity with little product differentiation, and interpret the embrace of biotechnology by US farmers as a move to recapture eroding market shares. As noted in the previous section, genetically engineered soybeans are primarily herbicidetolerant, which means that bioengineering makes no difference to the purchaser: its only benefit, as far as soybeans are concerned, lies in its potential to reduce production costs. But as evidence of statistically significant cost reductions is so far inconclusive for herbicide-tolerant crops,4 the odds are still out as to whether biotechnology will prove a sufficient response for US exporters to fight the effects of price competition in foreign markets. What is clear, in any case, is that the breathtaking adoption rate of biotechnology had, in all likelihood, something to do with the intense competitive pressure felt by US soybean exporters and hence, indirectly, with the high value of the dollar. The situation for corn has similarities and differences with that of soybeans. Globally, US corn exporters have been under competitive pressure quite like soybean producers, with US exports dropping from 60 to 41 million tons between 1995 and 1998.5 As far as US-EU bilateral trade flows are concerned, the situation is different in that the EU is largely selfsufficient in corn and represents only a minor market for US corn producers. 18% of the US’s corn output is exported, and the EU accounted for only 4% of this at the time when genetically-modified varieties were introduced. This small niche for US exporters had been negotiated as compensation for lost agricultural markets when Spain and Portugal joined the EC in 1986. What makes the corn market an interesting one to study is that USDA researchers entirely attribute the collapse of US corn exports on European markets

3

Ballenger et al. (2000). Commission, 2000a, chap 3.; Messéan, 1998. 5 Ballenger et al. (2000)

4

4

to the EU’s biotech regulations, writing that “[t]he EU represents the one documented loss of US corn exports resulting from issues related to biotech products.”6

2.2 Biotechnology and the agri-food industry As far as commercial applications are concerned, agricultural biotechnology’s début can be dated to the year 1996 when 2.6 million hectares (ha) were planted in the US. Between 1996 and 1999, areas planted with genetically modified (GM) crops increased sixteen-fold to 41.5 million ha, a growth rate of 150% per annum ? the fastest ever adoption rate of a new agricultural technology. More than two thirds of GM crops, by surface, are in the US and about 95% on the American continent (see table 1). The only country having adopted GM crops on a significant scale outside of America is China, whose government actively promotes the use of biotechnology in tobacco cultivation. Of the 41.5 million ha planted with GM crops in 1999, 69% were herbicide resistant and 21% insect resistant, with crops containing both genes accounting for another 7%. The rest (3%) was virus-resistant tobacco used only in China. However, these proportions vary across crops: whereas almost all GM soybeans are herbicide-resistant, two thirds of GM corn are insect-resistant (bt). As for GM crop types, in 1999 53% were soybeans, 27% corn, 9% cotton, 8% rapeseed, 2% tobacco and 0.1 % potatoes (Commission 2000a). Table 2 shows a breakdown of GM crops by producing country. Based on surveys carried out in March 2000, the USDA forecasts for 2000-2001 a sharp decrease in the area planted with bt corn in the US, as the profitability of bt corn relative to conventional one (which in any case varies according to infestation levels) is limited by a number of factors. These include, inter alia, a regulatory requirement imposed by the Environment Protection Agency (EPA) according to which farmers must set aside 20% of their corn land for refuges planted with conventional corn in order to slow down the genetic mutation of pests. By contrast, areas planted with bt corn are expected to grow in Argentina and South Africa. In general, market and regulatory developments in Europe and elsewhere have led to an abrupt slowdown in the growth of GM-planted areas, and growth is not forecast to resume in 2001. The emergence of biotechnology’s agricultural applications has had a major impact on the agri-food industry. First, intellectual property has become of critical importance in an industry where, traditionally, it wasn’t, as companies that had spent considerable resources developing new breeds sought to protect the value of their investments with patents. Second, seeds, which were up to the early 1990s essentially commodities with little if any product differentiation, have become technology products. This has had several implications. As research in gene technology requires sophisticated molecular biology laboratories and scarce scientific talent, the R&D cost required to put new genetically modified seeds on the market is considerable. This creates a formidable barrier to entry. Indeed, Table 3, taken from Harhoff et al., shows that the concentration of intellectual 6

Id., p. 25. 5

property in the biotechnology industry is fairly high. Second, basic research in gene technology was seen in the early 1990s to involve potential synergies between its agricultural and pharmaceutical applications. This has led to the creation of “life-science companies” whose simultaneous presence in the agricultural and pharmaceutical sectors was (at the time) considered by markets as the best way to leverage those synergies (see Magretta 1997). It turned out that a number of these companies were later forced to spin off their agri-biotech activities in order to shelter their profitable pharmaceuticals businesses from the public backlash against genetically engineered food (Cadot, Gabel and Traça, 2001), and by 2000, much of the horizontal merger activity of the 1990s had unwound. Whereas synergies between the agricultural and pharmaceutical applications of biotechnology proved elusive, more powerful complementarities appeared between cropprotection activities (insecticides, herbicides, and so on) and seed-development ones. For instance, Monsanto’s research labs have developed a trait that renders crops resistant to Roundup, one of the company’s best-selling herbicides,7 making the combination of the herbicide and so-called “Roundup-Ready” seeds an attractive marketing package. Indeed, if evidence of yield improvements has remained somewhat inconclusive for herbicidetolerant seeds, evidence of strong market-share effects is more apparent. 8 But the spillovers go both ways, as a larger market share downstream clearly raises the return on upstream product development. Recognition of these two-way vertical linkages has led biotechnology companies to integrate forward by buying seed companies or forming alliances with them, and the ones that pursued this strategy most aggressively, e.g. DuPont and Monsanto, were initially rewarded by the markets. As a result, high levels of concentration have spread downstream into the seed industry: whereas the top twelve companies controlled 20% of the worldwide seed market in 1994 (Joly and Lemarié 1998), the top ten controlled 31% of it in 2000 (RAFI 2000). Besides vertical integration, two factors have contributed to the high concentration of the emerging agri-biotech cluster. First, the crop-protection industry, dominated by large chemical companies, has always been highly concentrated, with more than 80% of the worldwide agri-chemical market controlled by the sector’s top seven companies (RAFI 2000). Through aggressive acquisition strategies, these companies have been able to reinforce their dominant positions and extend them to the new cluster. Some of the industry’s largest ones, including Monsanto and Novartis, have even carried forwardintegration strategies one step further by moving into the food processing sector, thus extending their reach and reinforcing their market power. Second, and perhaps more surprisingly, parsimonious clearances for new seeds have contributed to create monopoly positions where a less strict regulatory approach would have been more favorable to competition (although, as noted, barriers to entry were significant irrespective of regulatory issues).9

7

Roundup is a broad-spectrum herbicide that cannot be sprayed after planting unless the crop is genetically modified to tolerate it. Its US patent expired in 2000. 8 Agra Europe, July 9, 1999, EP/6, Commission 2000a, chap.3. 9 On this, see Harhoff et al, op. cit.. It is worth noting that in this regard, environmental and extraeconomic concerns have proved to work at cross-purposes with anti-trust ones, which were anyway largely out of the public debate. 6

2.3 EU regulation of GMOs In 1987, the European Council identified biotechnology as an area in which Communitylevel regulation was called for in addition to multilateral efforts towards an international regulation 10, and in 1990, it issued what was to be the key piece of European legislation in the area ––Directive 90/220 on the release of GM products into the environment (EC 1990). The Directive set out a procedure for the approval of GM products that was a highly complex mixture of ‘subsidiarity’ and centralized decision making. This procedure has been at the heart of the transatlantic dispute over market access. Under “90/220”, as the procedure came to be known, producers or importers of GMOs like GM seeds were required to notify the regulatory authority of the relevant member state (Article 11), which could either issue a favorable opinion or withhold approval. In the former case, other member states would be allowed to raise objections (Article 13). Without objection, the file would come back to the original member state for final “written consent”; with objections, member-state regulatory authorities would try to reach a consensus. If no consensus was reached, the Commission would take over and conduct a scientific review of its own at the end of which a committee of member-state representatives would rule at the qualified (two-thirds) majority. If no qualified majority emerged, the Commission itself would draft a decision and submit it to the European Council. If the Council failed to reach a decision (again at the qualified majority), the final word would go back to the Commission (Article 21) and, if the decision was positive, the file would return to the original member state for official approval.11 Then, according to the mutual-recognition principle, the product would be marketable in all member states, including those who had objected. However, directive 90/220 also included a safeguard clause whereby community approval could be suspended by a member state on the basis of “objective” health or environmental risks (art.16). The use of the safeguard clause should in theory be examined at the EU level; actually, no such ex-post evaluation ever occurred after 5 Member States invoked the safeguard clause to ban temporarily the placing on their market of 8 GMO’s varieties. Since October 1998, no further authorisations have been granted (14 applications are pending). GMO seed varieties have to be authorised in accordance with directive 90/220 before they are included in the “Common Catalogue of varieties of Agricultural Plant Species” which means that the seed can be marketed throughout the EU. Only two seeds have been so far included in the Common Catalogue (3 applications are pending).

10

UN Biosafety protocol (Cartagena protocol) was adopted in Montreal in January 2000. It is based on the Precautionary principle and aims to ensure the safe transboundary transfers of living modified organisms that may affect biodiversity. In June 1999, the G8 requested OECD to carry out an analysis on Food Safety and Biotechnology ( 3 working groups : on Food Safety / on Novel Food and Feeds / on the Harmonisation of Regulatory Oversight in Biotechnology). Last, the Codex Alimentarius has also established a Task Force on Foods derived from Biotechnology with the objective to develop standards, guidelines or recommendations for GMO foods (full report by 2003). The Codex Committee on Food Labelling is working on a standard for the labelling of GMO food. 11

A very good description of EU biotechnology regulation can be found in Princen (2000). 7

The other piece of EU legislation is Regulation 258/97 on novel foods, adopted in January 1997 (EC 1997), which was concerned essentially with foodstuffs containing GMOs and their risks for food safety. Regulation 258/97 set up a procedure for foodstuffs that was fairly similar to the “90/220” procedure. So far, no GM food has successfully completed all of the procedure’s steps (11 cases are pending). Two genetically modified plants (soya and maize) have been previously authorised before the entry into force of the Novel Foods Regulation. However, foodstuffs produced from GM plants but containing no biologically active DNA and “substantially equivalent” (in terms of sanitary risk and nutritional properties) to their conventional counterpart have been authorized, e.g. oil made from GM rapeseed or corn. The authorization procedure for the latter group requires only notification by the firm to the Commission, who in turn informs Member States and publishes a list of authorized foods. Since 1997, labelling to indicate the presence of GMOs is mandatory for food and seeds (currently, there is no specific legislation on feed labelling). The labelling requirement for foods no longer containing GMO is based on the concept of equivalence. The characteristic which is not equivalent to an existing counterpart must be labelled if it was obtained by a genetic modification. The threshold for adventitious contamination of GM material on conventional food is 1% de minimis (regulation 49/2000) per ingredient. The one-percent threshold, which applies per ingredient, has been widely perceived as a very stringent one. First, for some crops like rapeseed, natural dissemination can produce GMO concentrations above 1%. Second, for a foodstuff to qualify as “GMO free” none of its ingredients taken individually, however small its share, must contain more than 1% of GMO12. With only eighteen GM products cleared under Regulation 90/220 and a moratorium in force on new clearances since June 1999, the EU regulatory process had virtually ground to a halt, and was widely perceived to be a political mess. The Financial Times, for instance, wrote in 1999 that “European Union rules for approving genetically modified products are absurdly cumbersome and, in some cases, offend basic rules of democracy.”13 A new updated directive (2001/18/EC) will enter into force on October 2002. It strengthens the existing rules by introducing mandatory information to the public and mandatory labelling and traceability at all stages. It also foresees mandatory monitoring requirements of long term effects on environment. First approvals will be limited to a maximum of ten years and consultation of the Scientific Committee (set directly at the European level) becomes obligatory. The new directive also introduces an obligation to consult the European Parliament and the possibility for Council of Ministers to adopt or reject a Commission proposal for the authorisation of a GMO by a qualified majority. As announced in the White Paper on Food Safety, the Commission currently works on a Regulation proposal on the labelling and traceability and a new regulation on genetically modified food and feed that would harmonise all existing regulations on all GM products, seeds excepted. The regulatory short-cut for ‘substantially equivalent’ GM food (a simple 12 13

Regulations 1139/98 and 49/2000 Financial Times, 13 June 1999. 8

notification) has been very controversial and will not be included in the new regulation. The proposal extends the current labelling provisions to all genetically modified food, irrespective of the detectability of DNA or protein (eg. oils made from GM seeds will be labelled). Thus, all products which are subject to authorisation would also be subject to mandatory labelling (ex.GM Feeds will be all labelled ; currently, 4 out of 8 types of feeds are authorised without labelling requirement). In return, products which are not subject to authorisation would also not be subject to mandatory labelling (ex. Products obtained from an animal fed with GM feed would not be subject to the labelling requirement). In order to qualify for GMO-free status under the reinforced traceability criterium ( or Identity Preservation), a product would have to be guaranteed as GMO-free at all stages of production, not just at the first (seeds) and final stages (food) of the production process. Unsurprisingly, such a requirement would be of serious concern to US producers (see Bullock and Nisi, 2001, Bullock and Desquilbet, 2001). As Europe is almost entirely GMO-free, European products would qualify at no cost; however, the cost of tracking down and labeling GMO content at each step of the value chain could be substantial to US producers. In sum, EU labeling could have discriminatory effects and, if so, could be interpreted as technical barriers to trade.

3. Trade and standards: a review of the issues 3.1 “Regulatory protectionism” in search of a definition Unlike import tariffs, product standards are not inherently discriminatory against imports inasmuch as they apply equally to imported and domestically-produced goods. However, they can have the intentional or unintentional effect of deterring imports, in which case they become, in the jargon of international trade law, Technical Barriers to Trade (TBT). In general, product standards have the effect of raising the cost of producing the goods to which they apply. Variable costs can be raised by compulsory devices (safety or other) that are expensive to manufacture and to integrate in existing designs. Development costs nd lead times can be raised by long or uncertain certification processes in the importing country, and sometimes also by the need for multiple designs if the foreign manufacturer finds it unprofitable to adopt the importing country’s standard on its whole production line. These cost-raising effects can affect domestic and foreign producers asymmetrically. Unless the exporting and importing countries have mutual-recognition agreements, which typically comes only with close trade ties and similar levels of economic development, certification must be sought in the importing country whose bureaucracy may be unfriendly to foreign producers. Such was often reported to be the case in Japan, for instance. Of course, importing-country certification needs not be designed as a trade barrier: it can be justified, when there is a consumer hazard, if the exporting country does not consume the exported good and has consequently little incentive to set up rigorous certification procedures. But in practice, the importing country’s domestic producers are likely to be,

9

for various reasons, in a better position than their foreign competitors to minimize conformity-assessment costs. Given this likely asymmetry in compliance costs, the question is: when is a standard ‘protectionist’? A legal definition can be found, stated negatively, in the Uruguay Round version of the GATT’s TBT Agreement (the first version of the Agreement was adopted at the Tokyo Round), which expanded and specified GATT Article III. Namely, Members shall ensure that technical regulations are not prepared, adopted or applied with a view to or with the effect of creating unnecessary obstacles to international trade. For this purpose, technical regulations shall not be more trade-restrictive than necessary to fulfill a legitimate objective, taking account of the risks non-fulfillment would create. Such legitimate objectives are, inter alia: national security requirements; the prevention of deceptive practices; protection of human health or safety, animal or plant life or health, or the environment. In assessing such risks, relevant elements of consideration are, inter alia: available scientific and technical information, related processing technology or intended end-uses of products.14

Thus, the mere fact that a standard restricts trade is not by itself sufficient to make it a TBT; in order to be so, it must restrict trade “more than necessary”. The TBT Agreement gives some indication about what kind of concern can provide legitimate ground to impose traderestricting standards; but its wording leaves substantial ambiguity both on the actual scope of ‘legitimate’ concerns and on the procedures to assess their reality. As far as foodstuffs are concerned, it is complemented by the somewhat more constraining Agreement on Sanitary and Phytosanitary Standards (SPS). The SPS Agreement is similar in spirit to the TBT Agreement, but stresses the obligation for the country imposing trade-restricting measures to base those measures on a scientific risk assessment (Jackson 1997), an issue which was crucial in the transatlantic dispute over hormone-treated beef and is equally relevant to the GMO debate. If considerable progress has undoubtedly been achieved in the quest for an operational definition of regulatory protectionism, from GATT Article III to the Tokyo Round version of the TBT Agreement to its Uruguay Round version and to the SPS Agreement, much remains to be made precise. Indeed, the transatlantic regulatory conflict over GMOs has exposed some of the ambiguities of the legal framework for product standards, be they about the norms applicable to scientific risk assessment, the treatment of labeling requirements, the measure of trade effects, or the respective domains of the SPS and TBT agreements (i.e. which one should apply). Economists have also struggled somewhat with the proper definition of ‘regulatory protectionism’ ? in fact, the very notion of ‘protectionism’ is itself surprisingly ambiguous. Chambers and Pick (1994), for instance, cite three different definitions. The first, due to Walter (1972; see also Baldwin, 2000, for a discussion) classifies as protectionist any measure that distorts the volume, composition or direction of trade. By that yardstick, any standard that restricts trade, whether it does so “more […] than 14

Agreement on Technical Barriers to Trade, Article 2.2, Final Act of the Uruguay Round Agreement, p. 118. 10

necessary to fulfill a legitimate objective” or not, is protectionist. This criterion has the merit of being simple and easily measurable. It is also, by its very simplicity, likely to appeal to politicians and commentators and therefore to be an implicit, if not explicit driver of trade disputes. In other words, a regulation that does not affect international trade at all is unlikely to ever lead to an international dispute. Thus, it provides a good starting point to analyze the GMO dispute, which has been portrayed in the US as a market-access issue. We will accordingly start our empirical investigation of the dispute, in the following section, by an assessment of the reality of damage done to US exports by EU GMO regulations. The problem with this definition is that it does not have anything to say about trade-offs between domestic and trade-related objectives, even though they are acknowledged in the TBT Agreement’s wording. Is a standard that slightly restricts international trade in order to mitigate a strong negative externality at home ‘protectionist’?15 By the yardstick of this definition, it is. But then, the case for forcing a country to abandon that standard would obviously be so weak that it would merely encourage non-compliance. Thus, any operational definition of protectionism needs to be less sweeping. An alternative proposed by Hillman (1991) classifies as protectionist any measure that impedes or discriminates against imports. This definition is slightly better conceptually than the first one, being grounded in the national treatment principle. However, as we noted earlier, standards typically apply to domestic as well as foreign producers, and it is only through indirect manipulation that they end up discriminating against foreign producers. The question is then, where should the red line be drawn? The question is particularly relevant for agricultural biotechnology, where only foreign producers have adopted a technology (see Table 1 below) that is viewed by EU regulators and the public as potentially hazardous. Any regulatory requirement, labeling or other, imposed on GM seeds and products ends up applying primarily to imports. Is this ‘discriminatory’? Is it ‘protectionist’? In the absence of an established jurisprudence, there is no clear-cut answer to these questions, although WTO experts tend to view it as negative. Moreover, absent a counterfactual history, one cannot be sure that the absence of domestic cultivation of GM seeds in the EU is not the result of the regulations, in which case the fact that the regulation applies only to imports would be itself endogenous. Thus, although this second definition appears, prima facie, as less crude conceptually than the first one, it raises more questions (both with respect to measurement and definition) than it answers. Moreover, it is also vulnerable to the criticism levied against the first definition ? namely, that it fails to acknowledge trade-offs between trade-related and domestic objectives. The third, and most analytic one, proposed by Baldwin (1970), takes as protectionist any measure that reallocates resources in traded sectors in such a way as to reduce world welfare. According to this definition, a standard is not protectionist, even if it distorts or 15

As an example, consider bt corn seeds for sowing. A Cornell University study published in Nature suggested that bt corn could create a hazard for certain butterflies (J.E. Losey, L.S. Rayor and M.E. Carter, “Transgenic pollen harms Monarch larvae”, Nature 399, 1999, 214-215). If the problem was serious enough, banning the cultivation of bt corn seeds might be a rational response, even though this would hurt US exports. Given that US exports of corn seeds to the EU are in any case small, would it be reasonable to look at the regulation only from a trade angle? 11

restricts trade, provided that it is the first-best instrument to deal with a market failure (externality or other), or if not first best, provided that it is the most efficient instrument given existing constraints on other ones. Protectionism is thus, according to this definition, a departure from global optimality due to the fact that the policy-making authority is accountable to a local (national) constituency instead of a global one, leading to beggarthy-neighbour policies (Engel, 2000). This way of looking at things has the merit of highlighting the importance and usefulness of harmonization, mutual recognition, and delegation of decision-making power to supranational bodies. However, it is unlikely to be easily verifiable. Moreover, it assumes (as most of the literature on environmental economics does) that the form and extent of the externality to be dealt with are common knowledge. But this is rarely the case: indeed, a large part of the GMO debate revolved around the issue of risk assessment, with the EU arguing for the application of the precautionary principle and the US voicing doubts about the reality of the risks. Thus, instead of a single, broadly accepted definition of what is a ‘protectionist standard’, we are left with a menu of definitions in which sophistication must be traded off against measurability. Fortunately, the more sophisticated definitions are also stricter, so that the most primitive one ? alteration of trade flows? must be verified in any case.16 This is what we do in section 4 below, but before doing so we discuss the economic rationale, if any, for the manipulation of standards in a way that affects trade flows. 3.2 Strategic manipulation of standards: Racing to the Top or to the Bottom? The difficulty of identifying the crime (protectionism) underscores the need to understand its motives. Why would a government adopt a standard or set its level in a way that departs from the one that a global social planner would choose? First, standards can act as surrogates for industrial policy and shift rents to domestic firms at the expense of foreign ones. Moreover, governments and regulatory agencies can be captured, at home, by special interests tilting policies away even from the ‘local optimum’. The use of standards as a surrogate for strategic industrial policy has been extensively analyzed, directly or indirectly, in the environmental-economics literature (see e.g. Conrad, 1993a and 1993b; Ulph, 1993; Barrett, 1994; Hung, 1994; Kennedy, 1994, Long and Soubeyran, 1998). The literature’s basic thrust is that under imperfect competition, governments will trade off the treatment of domestic production externalities (pollution) by environmental taxes or standards against a desire to favor domestic firms at the expense of foreign ones. The result will be a relaxation of national production standards below the level that would be globally welfare-maximizing. As the same argument applies to all trading partners, the ultimate outcome will be a race to the bottom. More recently, however, Fisher and Serra (2000) have considered a different setting where a government sets a strict product (consumption) standard so as to take advantage of the fact that if the home market is small enough, the foreign firm may simply choose to stay out 16

Baldwin (2000) cites a case where a standard could raise world welfare while simultaneously reducing trade, but this is likely to be a theoretical curiosum more than anything else. 12

in order to avoid the additional costs created by the standard. In other words, instead of using the relaxation of a production standard as a hidden subsidy, the home government uses the tightening of a consumption standard as a hidden import tax. The argument can be carried one step further. Suppose that production generates a local externality ? pollution? for which the first-best response is, for the domestic government, to impose a tax on local producers only, since the welfare of foreign residents polluted by their own exporters is of no concern. But a production standard puts domestic firms at a competitive disadvantage vis-à-vis foreign ones; recognizing this, the domestic government may choose to impose a consumption standard instead of a production one (thus applying it to imported goods as well) so as to level the playing field. In this logic, suppose that the environmental hazards of GMOs were all that really matters; a government might nevertheless claim that otherwise nonexistent food-safety issues warrant cost-raising consumption regulations (say, labeling) in order to avoid putting domestic farmers at a disadvantage. The result of all countries doing this would then be a mixture of instrument substitution (consumption standards for production ones) and of a race to the top (over-strict standards). It is worth noting, however, that the argument applies only if the foreign government does nothing to control the production externality in its own jurisdiction; if such is the case, global welfare may be improved if the home government, in choosing a consumption standard, acts as a surrogate global planner. In the same vein, Mattoo (2001) considers the issue of “voluntary” standards (with which companies can decide to comply or not) and whether they can be set strategically. The model is a three-stage game where a government sets the standard for its area of jurisdiction (stage one) and two duopolists independently decide to comply or not (stage two), after which they compete à la Cournot (stage three). He shows that even if the standard is not inherently discriminatory in the sense that it applies equally to the domestic and foreign firms, it may have a de facto discriminatory effect. The reason is that if the market in which the standard applies is small enough and firms cannot differentiate their product (so as to comply only in the affected market) there is an equibrium in which the firm that has the lowest marginal cost decides not to comply whereas the one with the highest cost does. In that equilibrium, the high-cost firm enjoys a de facto monopoly in the market affected by the standard and its profits increase with the standard’s tightness. If that firm happens to be the home firm, it makes sense for the home government to set a “strategically tight” standard. The evolution of the agri-food industry clearly fits some if not all the assumptions of the strategic-trade literature. By the late 1990s, what used to be a commodity, low-tech industry with a relatively stable oligopoly structure had become a high-tech one in which first-mover advantages were crucial. In agri-biotech as elsewhere, product innovation tends to be a winner-takes-all game; perhaps even more so given that, as noted earlier, few products end up being approved, especially in Europe. Moreover, the shining stock-market performance of the industry’s leaders suggested that the rate of return on capital invested in agri-biotech seemed very high right up to the critical years 1998-99, at least ex post (i.e. for those companies that had made successful product introductions). Finally, the industry’s growth prospects were virtually unlimited as second- and third-generation products were expected to generate business opportunities in both industrial and emerging markets. In sum, the agri-biotech industry was taking on all the characteristics of a “strategic industry” 13

? i.e. of an industry where strategic trade policy of the type discussed here could have dramatic profit-shifting and welfare effects. If enthusiasm for strategic trade and industrial policy had, by the 1990s, considerably abated on both sides of the Atlantic, the potentially high stakes and scope for strategic policy moves undoubtedly generated uneasiness, as evidenced by US official quotes given in the introduction. However, one common critique against the strategic-trade literature, which applies equally well to the models surveyed above, is that it is fairly sensitive to changes in modeling assumptions. The strategic manipulation of standards makes sense only if there are rents to be shifted, which in turn requires entry barriers and a certain type of competition ? say, Cournot oligopoly. Should firms play a Bertrand oligopoly game, for instance, the rationale for strategic standard manipulation would become s a lot less compelling; as a matter of fact, in most cases, it would be nonexistent. As far as the upstream part of the agri-biotech industry is concerned, the necessary assumptions may not be too far-fetched. But as one travels downstream the industry, these assumptions are no longer met; in particular, as European agriculture is already heavily subsidized, the rationale for ‘shifting rents’ away from US farmers is unclear. 3.3 Conflicts of jurisdiction Another case where governments can set standards that will restrict market access is the problem of partial juridictions. National decisions that have cross-border spillovers will not be globally optimal if the spillovers are not taken into account by national authorities ? and they won’t in the absence of a coordinating mechanism. In that case, governments do not use standards strategically in order to hurt foreigners. They simply disregard foreigners in the determination of domestic standards. The issue of “partial jurisdictions” has a parallel in competition policy, where a national decision to approve a merger or not that properly balances domestic consumer and producer surpluses may nevertheless fail to be globally welfare-maximizing if it disregards sufficiently large spillover effects on foreign consumers or firms (see Cadot, Grether and de Melo, 2000). Indeed, as in competition policy, two national authorities applying the same criteria may take different decisions, ending up in regulatory conflict. Suppose that both US and EU authorities set standards on GM products, say labeling requirements, so as to balance the technology’s risks for consumers against its profits for producers. Suppose further that the sole owners of intellectual property over the technology are American companies (on this, see Harhoff, Regibeau and Rockett, 2001). Ceteris paribus, US authorities will set the standard in a way that is more favorable to producers than EU authorities will do, even if both apply identical risk-assessment procedures and welfare criteria. Thus, there will be a conflict over outcomes even if the rules have been harmonized. Only mutual recognition or delegation of decision-making authority to a supranational agency would solve this type of conflict, which may have trade implications without having trade motivations.

14

4. Interpreting the biotech dispute

Official US comments cited in the introduction suggest at least a suspicion that EU biotechnology regulations are “technical barriers to trade”. Moreover, we showed in the previous section that the literature suggests plenty of reasons why countries may use safety and other standards as hidden barriers to trade, and as many ways of doing so. We now turn to the empirical evidence in order to assess whether EU regulations have had a statistically traceable impact on US-EU trade flows in sectors that stand to be affected, namely corn and soybean.

4.1 Explaining market-share changes: empirical evidence In order to check for a possible ‘GMO effect’ on US-EU agricultural trade flows we have estimated import-demand equations for the EU using data from Eurostat’s Comext database at the tariff-line level (9-digit, the highest disaggregation level) and re-aggregating them into four product categories: corn seeds, other corn, corn gluten feeds and soy products (soybeans and oilcakes used as animal feed). The import-demand equations control, directly or indirectly, for the usual explanatory variables (own price, price of substitutes, and income). The price variables are the series of unit values obtained by dividing import values by tonnage, both for the US and for other sources (total EU imports minus imports from the US). For each product category, the relative-price variable (LNREL), expressed in logs, is the ratio of US unit values divided by the unit values of other import sources. Other determinants of demand are approximated by the EU’s total imports of the product in question (LNTOT), which pick up the effect not only of income, but also of shifts in preferences and food technology. We also report market-share estimates, which avoid having an endogenous variable on the right-hand side of the equations. For corn gluten, we use import data for a broader aggregate of protein feed (LNPROT) as a ‘consumption’ control, as the US accounts for virtually all of the EU’s imports. We also control for exchange-rate variations with the ECU-per-dollar rate (LNEXR). We use quarterly data in order to have more degrees of freedom but use quarterly dummies (Q1 to Q4) to control for strong seasonal variations (some of the US’s competitors, like Argentina, are in the Southern hemisphere and have consequently different seasonal output patterns). We also include a time trend (TREND) and, whenever appropriate, use a Prais-Winsten transformation to control for AR-1 error processes without losing one observation in the estimation. Durbin-Watson statistics are reported for all regressions. Finally, in each equation we add a dummy variable (GMO) for the years 1997-2000 where US crops contained GMOs. We take the significance of this variable as measuring the statistical trace of a ‘GMO effect’ on US-EU trade. Tables 4-7 report both OLS estimates and Instrumental Variable (IV) estimates in which prices are replaced by vectors of instruments (prices cannot be treated as exogenous, since they are determined simultaneously by the interplay of demand and supply). We have also estimated simultaneously demand-supply systems in which the US export equation included US domestic output as a control for weather and other supply determinants. However, demand parameter estimates were close to those of OLS and IV estimation whereas supply 15

elasticity estimates were unconvincing, the data in hand being apparently not sufficient to identify the supply equation correctly. Accordingly, simultaneous estimation results are not reported. It is worth noticing that, depending on the equations, IV parameter estimates are not very different from those obtained by OLS estimation (indeed, Hausman tests for endogeneity of the price variables proved insignificant), but IV estimation reduces the price variable’s level of significance, partly because the available instruments are not very good ones. We therefore report both OLS and IV results. Reported R2 statistics suggest that our formulation explains a substantial part of the observed variation in the dependent variable. As far as soybeans and derived products are concerned, the numbers tell a story that is consistent with the USDA’s analysis reported in section 2.1 above. EU imports of soybeans and oilcakes from the US went down substantially during 1998 (figure 7), whereas imports from other sources (primarily Argentina) simultaneously increased. This could be attributable to a shift away from American GM products, but once price effects are controlled for, the ‘GMO dummy’ picking up the effect of idiosyncratic events linked to the stiffening of GMO regulation in Europe is statistically insignificant (Table 4). The reason can be seen by looking at Figure 8: the drop in US market shares after 1998 is essentially due to a rise in the relative price of US soybeans.17 Thus, with all the caveats that a rough-and-ready econometric exercise such as the present one calls for, we are left with the conclusion that there is, so far at least, no market-access issue in soybeans and derived products. Results for corn seeds, reported in Table 5, show the ‘GMO dummy’ to be insignificant at any level. This is not terribly surprising given the pattern of imports shown in Figure 1: most of the time series’ variation is seasonal, and there is little evidence of a collapse in US sales after 1997. This negative result is, by itself, interesting, especially when compared to the result for “other corn” reported in Table 6. Given that EU regulations restricted the cultivation of GM seeds more than their use in processed foods, if they had caused US market-share losses, the effect would have been stronger for upstream products such as corn seeds than for downstream ones such as “other corn”. Yet, the opposite pattern is apparent. The “other corn” category shows a sharp drop in the EU’s overall imports from 1998 onwards; imports from the US, in particular, collapse to near zero. Figure 4 shows that the prices charged by US producers have increased substantially relative to those charged by their competitors (unit values for “other corn” imports originating from the US went up after 1997 whereas unit values for all other partners went down). Nevertheless, even after controlling for price and other effects, the GMO dummy remains significant at the 5% level in the IV regression reported in Table 6. By contrast, corn gluten imports from the US (Europe’s quasi-exclusive supplier of the stuff) do show a sizable decrease between a high point of 1.695 million tons in the first quarter of 1995 and a low point of 749,000 tons in the fourth quarter of 1997 (see Figure 5). If the timing of the decrease does not match exactly the introduction of the EU’s GMO regulations, econometric estimates suggest that, after controlling for price effects (the 17

This price effect may be related to exchange -rate fluctuations as no parallel price increase is observable on the US domestic soybean market, although the parameter estimate on the dollar-ECU exchange rate is insignificant. 16

landed price of US corn gluten in the EU dropped by 38% between the end of 1996 and the middle of 1999), the “GMO effect” is statistically significant. It is also quantitatively important, at least relatively to export levels that are themselves of modest size, since the estimated decrease in US corn gluten exports to the EU attributable to the “GMO dummy” is equal to 23% of base-period exports, or 27 million dollars.18 In sum, with the notable exception of corn gluten, it is fair to say that, so far, evidence of a strong market-access effect attributable to EU GMO regulations is less than conclusive. However, this is by itself not sufficient to establish that those regulations were not intended as technical barriers to trade, since they may adversely affect future US export opportunities that have not yet materialized and therefore cannot be measured statistically. We now turn to a comparison of EU and US regulatory approaches in biotechnology highlighting the relative unimportance of trade-related motivations in Europe and providing a “non-trade” interpretation of the regulatory conflict. 4.2 Comparing regulatory approaches to GMOs At the outset, EU biotechnology regulation, unlike its US equivalent, was based on the premise that GM products require specific legislation. But even with different premises, how restrictive EU regulation has become over time is a puzzle, especially in a continent where risky technologies like nuclear electricity or high-speed trains have been embraced without much public debate. A number of factors have contributed to the evolution of the EU’s regulatory climate. First, food regulation is a sensitive issue in Europe because it lies at the intersection of three types of concerns, each one of which is of special importance in some of the EU’s member states. The first has to do with ‘food as culture’. Nowhere have the ideas of free trade and economic efficiency less legitimacy than in the agri-food sector, as food is considered in countries like France as part of culture and something that ought to escape economic logic ? at least to some extent. The second type of concerns have to do with food safety. The outbreak of mad-cow disease, first in England and later in the Continent, raised the level of public anxiety while simultaneously undermining confidence in ‘official science’, even in the UK where food culture is not an issue.19 Finally, environmental concerns (the risks that GM crops can pose for ecosystems) have special prominence in Northern European countries with strong environmentalist traditions. Thus, biotechnology regulation is one of a few areas in which public opinion weighs on the same side, albeit for different reasons, throughout the continent. This rare and fortuitous convergence helps explain the strength of the anti-GMO movement in the EU.

18

This is somewhat puzzling given that corn gluten produced with Monsanto’s bt seeds has had, at least in principle, no clearance problem. Under Secretary of State Alan Larson noted in a March 2000 speech to the Commodity Classic in Orlando that “only 4.6% of the US 1999 planted corn acreage is unapproved in the EU.” 19 Thomas Barlow, “A Bug’s Life Goes Global”, Financial Times, 11 September 1999. 17

A second contributing factor is the rising risk-aversion of regulators and politicians in Europe. This is particularly visible in France,20 where, in the last few years, the judiciary has been very proactive ––both by its own past standards and in the absolute–– in going after unscrupulous politicians. The contaminated-blood scandal in which medical authorities were accused of knowingly putting HIV-infected blood on the market for commercial reasons was a watershed in this regard, the reach of criminal investigations and severity of the sentences having no precedent in this type of scandal. Rising exposure of politicians to penal responsibility may be related to idiosyncratic events in France, but it is also part of a broader evolution of the continent’s disparate political systems toward increased accountability, and is unlikely to reverse itself. Its effect is to make regulatory decisions more conservative than they used to, and this effect is reinforced in the context of biotechnology by the fact that joint decision-making between Brussels and member states makes it very hard to know who would be responsible in a penal sense should things turn sour (interview). However unfavorable is the overall context, political decisions are always the result of a cost-benefit analysis. The problem with biotechnology was that if the potential costs loomed large for reasons explained above, the benefits looked thin. On the consumer side, the industry’s decision to prioritize herbicide-resistant seeds (see supra) instead of products that would bring direct benefits to final consumers or the environment meant that not much could be said in favour of taking risks that would convince public opinion. In the US, the lack of consumer benefits was compensated by the potential profit opportunities to companies that were significant employers and, perhaps more importantly, carried weight in Washington. But this wasn’t the case in Europe, where the biotechnology industry, which was growing comparatively slowly (Europabio 1997), carried much less weight economically and politically. Thus, regulatory barriers that risked slowing down the development of GM products were not seen in Europe as particularly costly in terms of foregone economic opportunities, because these opportunities were in any case likely to be seized by American rather than domestic firms (interview). However, for precisely that reason, they were perceived across the Atlantic as discriminatory, de facto if not de jure (Pollack and Shaffer 2000). In the terminology of section 3.3, there is little doubt that the GMO conflict was, if nothing else, a case of ‘partial jurisdictions’. If the peculiarities of the European political context have created favorable conditions for a regulatory conflict with the United States, there is nevertheless little evidence that, apart from uncorrelated public-opinion cycles, European and American regulatory philosophies regarding the environment are seriously diverging. In fact, as noted by Vogel (2001) and Cadot and Vogel (2001), the over-reaction of Europe’s regulatory and political system to the biotechnology crisis illustrates how similar it has become to America’s.21 However, as we will see, the political context in which American biotechnology regulations were set up was equally idiosyncratic. 20

On this, see also Cadot and Vogel 2001, and Vogel 2001. The creation in 2001 of a European Food Authority is part of that evolution; ho wever it should be noted that the EFA is unlikely to have the decision-making power of the FDA, and that its mandate includes objectives (e.g. social, ethical) that would clearly fall outside of the scope of food safety regulation in the US (Commission 2000b). Thus, food regulatory approaches still do differ; on this, see Pollack and Shaffer (2000). 21

18

The first key piece of US regulation in the biotechnology area was the “Coordinated Framework” issued in 1986 by President Reagan’s Office of Science and Technology Policy22 and whose objective was to define the responsibilities of a number of government agencies sharing the biotechnology industry’s regulatory oversight. Under the resulting architecture, the Department of Agriculture (USDA) was in charge of crop safety (ensuring that GM crops would not damage other crops), the Environmental Protection Agency (EPA) was in charge of wider environmental risks, and the Food and Drug Administration (FDA), part of the Department of Health and Human Services, was in charge of food safety. One of the key features of the Coordinated Framework was that, rather than developing a new regulatory approach dealing specifically with genetically modified (GM) products, it chose instead to rely on existing legislation. The legal basis of the FDA’s regulatory oversight was the Federal Food, Drug and Cosmetic Act under which the FDA issued clearance for food additives and had the power to prohibit the sale of unsafe food products. The clearance process was waived for additives that were ‘Generally Recognized as Safe’, either because they had already been cleared by the FDA or because they had a history of safe use before 1958. Moreover, whether a new additive needed clearance or not was left to the judgment of the firm that had developed it. Additives and foodstuffs produced through genetic manipulations were subject to those same rules; in particular, formal review of GM additives by the FDA was optional, and the clearance process for GM products was “based on the intended use of each product on a case-bycase basis” (OSTP 1986, p. 23310). The Coordinated Framework was updated in 1992 by the FDA’s ‘Statement of Policy: Foods Derived from New Plant Varieties’ (FDA 1992), which upheld the Coordinated Framework’s basic premise that GM products did not differ in any fundamental way from their conventional parents. It also confirmed the FDA’s ‘voluntary’ approach to regulatory oversight, leaving to the developer the responsibility of identifying risks and deciding whether or not formal review was needed. Thus, from the outset the United States adopted a light-handed approach to the regulation of biotechnology. Yet, until the 1990s, the US was, in many areas, leading the world in the development of a comprehensive and fairly tight body of environmental legislation. Moreover, a large part of that body was based on what came to be known as the “precautionary principle” (Cameron 1999). For instance, the US Congress stated in 1977 that the EPA’s duty was “to assess risks rather than wait for proof of actual harm.”23 A 1976 Court of Appeals ruling was even more explicit: “A statute allowing for regulation in the face of danger is, necessarily, a precautionary statute. Regulatory action may be taken before the threatened harm occurs. […] the statutes and common sense demand regulatory action to prevent harm, even if the regulator is less than certain that harm is otherwise inevitable.”24

22

The following description of the US regulatory system draws on Princen (2000). HR Report 294, 1977, p. 49; quoted in Cameron (1999), p. 251. 24 Quoted in Vogel (1995), p. 182. 23

19

In sum, if the EU’s over-cautious approach to biotechnology regulation was surprising, the relaxed attitude of US regulators was as much of a puzzle, for which a number of explanations have been proposed. First, Vogel (2001) noted that the policy process that led to the determination of US biotechnology regulation was a fairly closed one, with little involvement of NGOs or the public, in contrast to much of the country’s environmental legislation whose genesis was often fairly political. Successful containment of the biotechnology debate within industry and expert circles was undoubtedly facilitated by the low level of public anxiety about GM food in the US. Moreover, until the late 1990s consumer awareness of the rising use of GMOs in foodstuffs was, in the US, very low (Pollack and Shaffer 1999; Lynch and Vogel 2000). Second, in the US as elsewhere, government agencies were somewhat divided. The USDA, the FDA and the White House stood, alongside with industry, in favor of light regulatory oversight and rapid clearance of new GM products. By contrast, the EPA favored relatively strict regulatory oversight. This line-up was not special to the US, as Agricultural and Environment Ministries held similarly conflicting views in most European countries. However, notwithstanding the EPA’s position, the overall political context of the 1990s was, in the US, unfavorable to ‘environmentalism’. As noted by Lynch and Vogel (2000), NGOs and environmentalist groups were on the defensive throughout the 1980s, trying to prevent the Reagan Administration from watering down existing environmental legislation. They were thus in no position to pick fights in new areas like biotechnology. Perhaps more importantly, observers have also noted that the US regulatory process was to a large extent captured by the biotechnology industry, which exerted considerable influence in Washington: “What Monsanto wished for from Washington, Monsanto — and, by extension, the biotechnology industry — got. If the company's strategy demanded regulations, rules favored by the industry were adopted. And when the company abruptly decided that it needed to throw off the regulations and speed its foods to market, the White House quickly ushered through an unusually generous policy of self-policing. Even longtime Washington hands said that the control this nascent industry exerted over its own regulatory destiny — through the Environmental Protection Agency, the Agriculture Department and ultimately the Food and Drug Administration — was astonishing. ‘In this area, the U.S. government agencies have done exactly what big agribusiness has asked them to do and told them to do,’ said Dr. Henry Miller, a senior research fellow at the Hoover Institution, who was responsible for biotechnology issues at the Food and Drug Administration from 1979 to 1994.”25

Washington’s friendly approach was, of course, part of an overall pro-business philosophy that was common to the Reagan and Bush-I administrations; there is little doubt that it was also partly attributable to the revolving door between industry and various government

25

Kurt Eichenwald, Gina Kolata and Melody Petersen, “Biotechnology Food: From the Lab to a Debacle”; The New York Times online, 25 January 2001. 20

bodies.26 But whatever its causes, its effect was to set US biotechnology policy, in an unforeseen way, on a collision course with the European Union which was about to travel the opposite way. 5. Concluding remarks We noted in the introduction that one of the peculiarities of the biotechnology dispute was its combination of agricultural market-access issues and strategic trade policy ones. This combination provided an ideal setting for a trade conflict between the US and the EU, especially given that transatlantic trade relations had already been put under stress by previous conflicts over hormone-treated beef and bananas. Washington’s concern, expressed on a number of occasions, not to let the EU establish a pattern of noncompliance with WTO rulings, contributed to raise the stakes. Yet, the disaster has, so far, not happened. Why? Although many questions remain, our analysis provides, we believe, some elements toward an answer. As traditional EU agricultural lobbies gradually lose steam under internal and external pressures for change, a scenario in which they turn to the strategic manipulation of product standards to maintain high levels of protection is certainly not an implausible one. However, this scenario does not seem to be the relevant one. Agricultural market-access issues for biotechnology concern primarily two commodities: corn and soybeans. US corn exports to the EU are marginal (4% of US corn exports in 1996, 1% in 2000) so the stakes were minimal to begin with. Nevertheless, studying the corn market’s reaction to EU biotechnology regulations was a worthwhile exercise, as corn has been claimed by the USDA to be the one documented case of a US agricultural export collapse related to the biotechnology issue, and the relative unimportance of US beef exports to the EU did not prevent a trade conflict from erupting over hormone regulations. As it turns out, we found very little statistically traceable effect of the EU’s anti-GMO reaction, whether market- or government-driven. The soybean issue was potentially more relevant economically, as US soybean exports to the EU are worth more than a billion dollars a year. But again we found no statistically significant ‘GMO effect’ after controlling for price effects: consistent with the USDA’s market analysis, we found that the US’s share of the EU soybean market eroded essentially because US producers priced themselves out of the market. If the numbers are to be believed, we are left with the conclusion that the ‘agricultural protectionism’ aspect of the dispute is largely irrelevant. What remains, then, is the ‘strategic trade policy’ issue discussed theoretically at the beginning of this paper. This is primarily an issue of future business opportunities lost by US market leaders. It is possible that EU biotechnology regulations have been designed to protect EU-based players in the industry (Aventis, BASF) at the expense of American ones (Monsanto, DuPont, or Dow). The evidence in favor of such a ‘conspiracy theory’ is, however, hard to pin down. There is little doubt that business opportunities lost to erratic or over-strict regulations weigh little in the calculus of politicians if those opportunities were anyway to be taken by 26

On this, see Levy (2000). See also BioDemocracy and Organic Consumers Association (www.purefood.org/Monsanto/revolvedoor.cfm). 21

foreign companies rather than domestic ones. But using the terminology of Section 3, this is an issue of conflict of jurisdictions rather than of strategic trade policy. Indeed, our historical account of the genesis of EU biotechnology regulations highlights the importance of domestic considerations (in particular, the legacy of a series of health and food alarms) rather than trade-related or even industrial-policy ones. By contrast, the regulation of biotechnology in the US seems to have been influenced since the early stages by a desire to encourage the development of biotechnology and of the business opportunities that it was expected to generate. This, according to industry observers, reflected not only a general pro-business orientation of successive US Administrations in the 1980s and early 1990s, but also the political influence of the industry itself. Whether this influence also colored early US positions on the biotechnology dispute with the EU is a question that, at this stage, we can only leave open.

22

References Bagwell K. and R. Staiger (1999) “Domestic Policies, National Sovereignty and International Economic Institutions”, NBER working paper 7293. Baldwin (1970) “Non-tariff Distortions of International Trade” Washington, DC: The Brookings Institution. Baldwin R. (2000), “Regulatory protectionism, Developing Nations and a Two-Tier World Trade System”, CEPR discussion paper 2574. Ballenger, Nicole, M. Bohman and M. Gehlhar (2000), “Biotechnology: Implications for US Corn & Soybean Trade”, USDA, Economic Research Service, Agricultural Outlook/April 2000. Barrett, S. (1994), “Strategic Environmental Policy and International Trade”; Journal of Public Economics 54, 325-338 Bullock, D and M. Desquilbet (2001), “Who pays the costs of non-GMO segration and identity preservation ?”, mimeo. Bullock, D. and E. Nitsi (2001), “Roundup Ready Soybean technology and farm production costs : measuring the incentives to adopt”, American Behavioural Scientist, 44, 12831301. Cadot, O. , J-M. Grether and J. de Melo (2000) “Trade and Competition Policy: Where Do We Stand?”, Journal of World Trade 13, 1-24. ----- and D.Vogel (2001), “France’s role in the transatlantic dispute over biotechnology: Public concern or protectionist conspiracy?”, Brookings Briefings, Brookings Institution, 2001. -----, L. Gabel and D. Traca, Monsanto and Genetically Modified Organisms; Insead case, 2001. Cameron, James (1999), “The Precautionary Principle”; in Gary Sampson and Bradnee Chambers and, eds., Trade, Environment, and the Millenium, UNU Press. Chambers and Pick (1994) “Marketing Orders as Non-tariff Barriers”; American Journal of Agricultural Economics 76.

23

Commission of the European Communities (2000a), Economic Impacts of Genetically Modified Crops on the Agri-Food Sector: A First Review; Directorate General for Agriculture, working document Rev. 2. Commission of the European Communities (2000b), White Paper on Food Safety, COM(1999)710 Final, 12 January 2000. Conrad, K. (1993a), “Taxes and Subsidies for Pollution-Intensive Industries as Trade Policy”; Journal of Environmental Economics and Management 25, 121-135 --- (1993b), ``Optimal Environmental Policy for Oligopolistic Industries in an Open Economy''; University of Mannheim, Department of Economics working paper 476-93 Engel E. (2000) “Poisoned grapes, mad cows, and protectionism”, Journal of Policy Reform 4 (2). European Council (1990), Council Directive 90/220/EEC of April 23, 1990, on the Deliberate Release into the Environment of Genetically Modified Organisms; OJ L 117, 8 May 1990, pp. 15-27. European Council (1997), Regulation 258/97 of the European Parliament and of the Council of 27 January 1997 Concerning Novel Foods and Novel Food Ingredients, OJ L 43, 14 February 1997, pp. 1-7. Europabio (1997), Benchmarking the competitiveness of Biotechnology in Europe, Brussels. FDA (1992), “Statement of Policy: Foods Derived from New Plant Varieties”, Federal Register 57, May 1992, pp. 22984-23005. Fischer R. and P. Serra (2000), “Standards and Protection”, Journal of International Economics 52. Harhoff, D., P. Régibeau and K. Rockett (2001), “Some Simple Economics of GM Food”, forthcoming, Economic Policy. Hillman, J. (1991), Technical Barriers to Agricultural Trade, Westview Press. Hung, N. M. (1994), “Environmental Taxation in an International Duopoly Game”; Economics Letters 44, 339-343. Jackson, John H. (1997), The World Trading System; MIT Press, 2nd edition.

24

Joly, Pierre-Benoît, and S. Lemarié (1998), “Industry Consolidation, Public Attitude and theFuture of Plant Biotechnology in Europe”; AgbioForum 11, Fall. Kennedy, P. W. (1994), “Equilibrium Pollution Taxes in Open Economies with Imperfect Competition”; Journal of Environmental Economics and Management 27, 49-63. Levy, David (2000), “Oceans Apart? Comparing Business Responses to the Environment in Europe and North America”, mimeo. Long, N.G. and A. Soubeyran (1998), ``Entwining Environmental and Trade Policies'', Université d'Aix Marseilles II, GREQAM working paper 98A38, 1998. Lynch, Diahanna, and David Vogel (2000): “Apples and Oranges: Comparing the Regulation of Genetically Modified Food in Europe and the United States”, mimeo, Berkeley, 2000. Magretta, Joan (1997) “Growth Through Global Sustainability: An Interview with Monsanto’s CEO, Robert B. Shapiro”, Harvard Business Review, January-February, 8189. Maskus, Keith, J.S. Wilson and T. Otsuki (2001), “Quantifying the Impact of Technical Barriers to Trade”, mimeo, The World Bank. Mattoo, Aaditya (2001), “Discriminatory Consequences of Non-discriminatory Standards”, Journal of Economic Integration 16, 78-105. Messéan, A. (1998), “Comment apprécier l’intérêt des OGM pour la compétitivité de l’agriculture? », Organismes génétiquement modifiés à l’INRA : environnement, agriculture et alimentation. Office of Science and Technology Policy (1986), “Coordinated Framework for Regulation of Biotechnology: Announcement of Policy and Notice for Public Comment”, Federal Register 51, 26 June 1986, pp. 23302-23350. Pollack, Mark, and Gregory Shaffer (1999), “The Challenge of Reconciling Regulatory Differences: Food Safety and Genetically Modified Organisms in the Transatlantic Relationship”; mimeo, University of Wisconsin-Maddison. Princen, Sebastian (2000), “Genetically Modified Foods and Food Products”, Ph.D dissertation, unpublished. Rural Advancement Fund International (2000), RAFI’s Seed Company Chart. Ulph, A. (1993), “Environmental Policy and International Trade When Governments And Producers Act Strategically”; University of Southampton, Discussion paper in Economics and Econometrics 9318.

25

Vogel, David (1995), Trading Up: Consumer and Environmental Regulation in A Global Economy, Harvard University Press. ––– (2001), “Ships Passing in the Night: GMOs and the Politics of Risk Regulation in Europe and the United States”, mimeo. Walter, I. (1972) “Non-tariff Protection Among Industrial Countries: Some Preliminary Empirical Evidence,” Economia Internazionale 25.

26

Figure 1 EU imports of corn seeds, 1988:1 to 2000:4 60000.00 50000.00 40000.00 All extra-EU

30000.00

United States

20000.00 10000.00

94 -1 95 -3 97 -1 98 -3 20 00 -1

91 -1 92 -3

88 -1 89 -3

.00

Metric tons Source: Eurostat, COMEXT

Figure 2 Relative price, US vs. other sources, 1988:1 to 2000:4 5 4.5 4 3.5 3 2.5

Relative US/others

2 1.5 1 0.5

88 -1 89 -3 91 -1 92 -3 94 -1 95 -3 97 -1 98 -3 20 00 -1

0

27

Figure 3 EU imports of other corn, 1988:1 to 2000:4 4000000 3500000 3000000 2500000 All extra-EU

2000000

United States

1500000 1000000 500000

95 -3 97 -1 98 -3 20 00 -1

91 -1 92 -3 94 -1

88 -1 89 -3

0

Metric tons Source: Eurostat, COMEXT

Figure 4 Relative price of other corn, US vs. other sources, 1988:1 to 2000:4 5.00 4.50 4.00 3.50 3.00 2.50

Relative US/others

2.00 1.50 1.00 0.50

88 -1 89 -2 90 -3 91 -4 93 -1 94 -2 95 -3 96 -4 98 -1 99 -2 20 00 -3

0.00

28

Figure 5 EU imports of corn gluten, 1988:1 to 2000:4

United States

97 -1 98 -3 20 00 -1

All extra-EU

91 -1 92 -3 94 -1 95 -3

88 -1 89 -3

2000000 1800000 1600000 1400000 1200000 1000000 800000 600000 400000 200000 0

Metric tons Source: Eurostat, COMEXT

Figure 6 Import price (unit value) of corn gluten imported from US, 1988:1 to 2000:4

1.4 1.2 1 0.8 Unit value 0.6 0.4 0.2

88 -1 89 -2 90 -3 91 -4 93 -1 94 -2 95 -3 96 -4 98 -1 99 -2 20 00 -3

0

29

Figure 7 EU imports of soybeans and oilcakes, 1988:1 to 2000:4 10000000 9000000 8000000 7000000 6000000

All extra-EU

5000000 4000000 3000000 2000000 1000000 0 2000-1

98-3

97-1

95-3

94-1

92-3

91-1

89-3

88-1

United States

Metric tons Source: Eurostat COMEXT

Figure 8 Relative price of soybeans and oilcakes, US vs. other sources, 1988:1 to 2000:4 2.50

2.00

1.50 Relative US/others 1.00

0.50

90 -3 91 -4 93 -1 94 -2 95 -3 96 -4 98 -1 99 -2 20 00 -3

88 -1 89 -2

0.00

30

Table 1 Areas planted with GM crops, 1996-99 million ha

1996

1997

1998

1999

1999 in %

USA Argentina Canada China Brazil Australia South Africa Mexico Europe Total

1.45 0.05 0.11 1.00 0.00 0.00 0.00 0.00 0.00 2.61

7.16 1.47 1.68 1.00 0.00 0.20 0.00 0.00 0.00 11.51

20.83 3.53 2.75 1.10 0.00 0.30 0.06 0.05 0.00 28.62

28.64 5.81 4.01 1.30 1.18 0.30 0.18 0.05 0.01 41.48

69.10 14.00 9.70 3.10 2.80 0.70 0.40 0.12 0.03 100

Source: Commission (2000a), Table 1.2

Table 2 Major GM crops by country million ha Soybean USA Argentina Canada Brazil Corn USA Argentina Canada South Africa Cotton USA China Australia South Africa Mexico Rapeseed USA Canada

1996

1997

1998

1999

% GM, 1999

0.40 0.05 0.00 0.00

3.64 1.40 0.00 0.00

10.12 3.43 0.04 0.00

15.00 5.50 0.10 0.00

51 75 10 10

0.30 0.00 0.00 0.00

2.27 0.07 0.27 0.00

8.66 0.09 0.30 0.05

10.30 0.31 0.50 0.16

36 11 44 5

0.73 0.00 0.00 0.00 0.00

1.23 0.00 0.00 0.00 0.00

2.00 0.10 0.30 0.01 0.05

3.25 0.30 0.30 0.02 0.05

55 8 79 13 25

0.01 0.10

0.02 1.40

0.03 2.40

0.06 3.40

15 61

Source: Co mmission (2000a), Tables 1.3-1.6

31

Table 3 Largest companies in the agri-biotech sector Mergers and Acquisitions Estimated Number of (date) share in GM food total Agri- patents, 2000 biotech (US, EPO) market (%) Monsanto, Monsanto + Pharmacia 80 65, 36* USA (2000); agri-business unit divested. Acquired Cargill's international seed business, PlantGenetics, Calgene, Dekalb (1998). Aventis, Hoechst + Rhöne-Poulenc 7 10, 7 Germany (1999). Agribusiness unit (seeds and agrochemicals, comprising Agrevo, formerly owned by Hoechst and Shering) to be divested. Syngenta, Astra-Zeneca + Novartis 5 26, 15 Switzerland (1999). Novartis itself from Sandoz + Ciba-Geigy (1996) BASF, Acquired 40% of Svalöf 5 n.a. Germany Weilbull, Sweden (1998); spun off Metanomics and SunGene, plant biotech breeding companies, Germany DuPont, USA Acquired Pioneer (1999) 3 37, 20

Share of Seed sales Rank in approved 1999 world GM crops, ($ million) seed sales 1999 (%)

55

1700

2

19

288

n.a.

13

947

3

n.a.

n.a.

3

8

1850

1

Limagrain, France

n.a.

n.a.

n.a.

700

4

Grupo Pulsar, Seminis,world largest for Mexico fruit and vegetable and Savia

n.a.

n.a.

n.a.

531

5

Advanta , UK Subsidiary of AstraZeneca, that was not part of the deal with Novartis KWS AG, KWS and Limagrain merge Germany their corn and soybean activities in North America in 2000. Dow, USA Mycogen, Cargill North America

n.a.

n.a.

n.a.

416

6

n.a.

n.a.

n.a.

355

8

n.a.

n.a.

n.a.

350

9

(source)

Wood (Harhoff et (Harhoff et (RAFI) Mackenzie al.) al.)

(RAFI)

* Obtained by adding up total number of patents for Monsanto (24,17), Calgene (28,7) and Plant Genetics (13,12).

Appendix : estimation of demand equations The estimated equation is either LNUSt = a0 + a1 LNRELt + a2 LNTOTt + a3 LNEXRATEt + St at DUM t + a4 GMOt + a5 TREND + ut

(1)

or LNUSt/LNTOT t = a0 + a1 LNRELt + a2 LNEXRATEt + St at DUMt + a3 GMOt + a4 TREND + ut

(2)

The relative-price variable (LNREL), expressed in logs, are the ratio of US unit values divided by the unit values of other import sources. Other determinants of demand are approximated by the EU’s total imports of the product in question (LNTOT), which pick up the effect not only of income, but also of shifts in preferences and food technology. For corn gluten, we have used import data for a broader aggregate of protein feed (LNPROT) as a ‘consumption’ control, as the US accounts for virtually all of the EU’s imports. We also control for exchange-rate variations with the ECU-per-dollar rate (LNEXR). Finally, we added in each equation dummy variables for the years in which US crops contained GMOs (GMO), namely 1997-2000. Estimates from market-share equation (2) were similar and are not reported.

Table 4 Regression results, corn seeds Dependent variable: EU imports (corn seeds) OLS Estimate LNREL 0.040 LNEXR -1.597 LNTOT 0.971 Q1 0.033 Q2 -0.801 Q3 -0.631 Q4 GMO 0.146 TREND -0.007 CONST -0.523 # obs. R2 DW

52 0.91 2.016

from US, tons, cat. 10510

t-stat 0.217 -1.393 4.985 0.109 -2.953 -0.997 0.423 -0.798 -0.294

IV Estimate t-stat 0.066 0.157 -2.000 -1.721 0.910 6.988 0.983 2.895

0.917 3.464 0.148 0.418 -0.003 -0.293 -1.001 -1.061 48 0.92

33

Table 5 Regression results, other corn Dependent variable: EU imports from US, tons, cat. 10590 (other corn) OLS IV Estimate t-stat Estimate t-stat LNREL -2.037 -13.771 -2.209 -4.918 LNEXR -0.434 -0.594 -0.564 -0.763 LNTOT 0.966 7.654 0.903 8.197 Q1 0.809 0.437 -0.027 -0.198 Q2 0.158 0.757 Q3 0.204 1.292 Q4 -0.220 -1.407 GMO -0.920 -3.670 -0.817 -1.999 TREND -0.005 -0.933 -0.001 -0.195 CONST -0.170 -0.105 0.706 0.482 # obs. R2 DW

52 0.95 1.88

48 0.95

Table 6 Regression results, corn gluten Dependent variable: EU imports from US (corn gluten meal and corn gluten feed) 27 Prais-Winsten IV Estimate t-stat Estimate t-stat LNUSP -0.812 -2.849 -0.987 3.161 LNPROT 0.287 0.379 -0.416 -0.577 LNEXR 0.000 -0.001 0.088 0.383 Q1 0.103 2.178 0.146 2.947 Q2 0.074 1.571 0.077 1.787 Q3 -0.024 -0.564 -0.031 -0.748 Q4 GMO -0.256 -2.981 -0.239 -3.648 TREND -0.003 -0.652 -0.002 -0.582 CONST 9.341 1.224 16.365 2.307 # obs. R2 DW

52 0.90 1.72

51 0.60

27

The tariff-line code was changed during our sample period. Raw data provided by Eurostat must be verified carefully, no indication being given as to sources of error of this type, even upon request. In this case, multiple clarification requests proved necessary.

34

Table 7 Regression results, soybeans Dependent variable: EU imports oilcakes) OLS Estimate LNREL -2.844 LNEXR -2.222 LNTOT 0.926 Q1 0.709 Q2 -0.034 Q3 -1.452 Q4 GMO -0.343 TREND 0.002 CONST -0.253 # obs. R2 DW

52 0.83 1.99

from the US, tons (soybeans and

t-stat -4.591 -2.262 1.141 3.634 -0.176 -7.633 -1.183 0.261 -0.020

IV Estimate t-stat -2.879 -1.012 -2.520 -1.183 0.342 0.196 1.487 5.652

0.817 2.842 -0.266 -0.576 0.007 0.469 7.905 0.297 48 0.61

35