Standards, trade and protection: the case of GM crops

by Kym Anderson CEPR, World Bank and University of Adelaide and Lee Ann Jackson WTO Secretariat, Geneva

Revised March 2005

We gratefully acknowledge research funding from the Australian Research Council and DfID Trust Funds though the World Bank, and helpful comments from participants at various seminars and conferences, particularly at Harvard University, 21-22 January 2005. Views expressed are our own and not necessarily those of our current employers, and we alone are responsibly for any remaining errors. Revised version of a paper prepared for the 41st Panel Meeting of Economic Policy in Luxembourg, 15-16 April 2005.

Standards, trade and protection: the case of GM crops

Over the past decade, the United States (US) and the European Union (EU) have implemented widely divergent regulatory systems to govern the production and consumption of genetically modified (GM) food and feed crops. In the US, many GM varieties have been commercially produced and marketed, while in the EU few were approved prior to a de facto moratorium being introduced in 1998 that prevented the production, import and domestic sale of most GM crops. The EUs policy altered trade flows and led in 2003 to a WTO Dispute Settlement panel to test the legality of some EU member countries barriers to GM imports. The replacement of that trade barrier in April 2004 with a strict set of labelling regulations may prove to be no less trade restricting or contentious. This paper uses a model of the global economy (GTAP) to estimate the effects of GM biotechnology adoption on economic welfare in both adopting and non-adopting countries in the absence of trade policy responses to this technology, and in their presence. Particular attention is given to the potential welfare consequences for developing countries. The poverty-reducing benefits of the technology have been reduced by the EUs response to it, but that can be redeemed if EU taxpayers are willing to compensate by boosting agricultural research spending in developing countries.

Key words: Biotechnology, trade policy, political economy, regulation of standards, CGE modeling. JEL codes: C68, D58, F13, O3, Q17, Q18

Contact author: Kym Anderson Development Research Group Mailstop MC3-303 World Bank 1818 H Street NW Washington DC 20433 USA Phone +1 202 473 3387 Fax +1 202 522 1159 kanderson@worldbank.org

Standards, trade and protection: the case of GM crops

1. Introduction Since their first commercial releases in 1996, the area sown to genetically modified (GM) crop varieties has been growing very rapidly. By 2004 GM crops accounted for 29 percent of the global land area planted to maize, soybean, canola and cotton (up from 25 percent in 2003), and one-twentieth of all cultivable cropland. But just three countries accounted for all but one-tenth of global production of GM food products in 2003, namely Argentina, Canada and the United States (US). Because of minimal regulatory impediments there, the GM shares of those three food crops now average more than 60 per cent (James 2004). By contrast, in 1998 the then-15 member countries of the European Union (hereafter just EU) imposed a de facto moratorium on approving any more GM crop varieties for production and importation. Because GM-adopting countries have lost EU market share to GM-free suppliers, notably Brazil for maize and Australia and Hungary for canola (Foster, Berry and Hogan 2003), other countries have been discouraged from adopting this new biotechnology.1 True, the EU in April 2004 replaced the de facto moratorium with labeling regulations, but the cost of conforming to those strict regulatory standards may be prohibitive for the many countries which have no such requirements domestically, making this potentially no less of a trade barrier than was the EUs de facto moratorium. In the case of soybean it may be even more of a barrier because countries that were able to provide a GM variety that was approved by the EU before 1998 now have to explicitly stigmatise it with a may contain GMOs label. These strict EU standards and their trade-diverting effects raise the empirical question of whether this new agricultural biotechnology is beneficial for GM-adopting countries and for global welfare. Using a two-country partial equilibrium theoretical model, Lapan and Moschini (2004) show there are circumstances in which the adopting countries could be worse off and the EU and even EU farmers better off via changes in international food prices resulting from adoption in the face of the EU moratorium. In this paper we go beyond the Lapan and Moschini study in the sense of making use of a multi-country, applied general equilibrium model to estimate the economic effects of GM adoption by some countries without versus with trade policy reactions by other countries. We also explore the potential global welfare gains that could result from uninhibited adoption not only of GM maize and oilseed varieties but also of GM rice and wheat. Readers can then compare these potential benefits with their perception of any costs for their region of GM food technology adoption in terms of environmental or food safety risks.

This reason evidently has been far more important to developing countries than their concerns about the environmental and food safety issues associated with GMOs, according to Cohen and Paarlberg (2002).

Of particular importance is the question of whether developing countries would gain or lose from adopting GM crop varieties if barriers to trade in GM products are to remain high in the EU. In 2004 the Copenhagen Consensus Project rated investment in agricultural technologies to alleviate malnutrition and hunger as the second-highest-payoff opportunity to improve global welfare.2 China and India are the most significant countries to consider, in the sense that they comprise almost one-third of the worlds grain production and consumption and they (especially China) have the greatest potential to rapidly apply and disseminate this new biotechnology. They also have the majority of the worlds poorest people. But SubSaharan Africa is also of crucial concern, given its extreme poverty, nutritionally lowquality diet, and strong dependence still on agriculture for employment and export earnings and, in some cases, on food aid imports (which could be problematic if food provided as aid is not GM-free, as was the case for US shipments to Southern Africa in 2002). That is, for the poor in both Asia and Africa, the welfare consequences relate not only to the present generation of GM crop varieties, which are aimed directly at boosting farmer profitability, but also to the prospective generation of GM varieties aimed at boosting the foods nutritional content. An attempt is made to evaluate the potential benefits in these developing countries from GM adoption without and with restraints on exports to the EU. Our results suggest China especially is foregoing a lot by not approving the commercial adoption of GM varieties. In response to frustration over the way the EU precautionary measures have been applied (e.g., some EU member states maintain national marketing and import bans even on GM varieties that were approved by the European Commission prior to the de facto moratorium), the United States, Canada and Argentina successfully sought the establishment of a WTO Dispute Settlement Panel on 29 August 2003 to rule on the WTO consistency of the measures. Subsequent dispute cases involving genetically modified organisms (GMOs) are likely to focus on the scientific justification for not approving for sale products that may contain GM varieties grown abroad. Since such non-tariff barriers to trade would undermine previously negotiated reductions in tariff protection, they are a direct challenge to the WTOs multilateral agricultural trade liberalization process. Clearly a great deal is at stake, the outcomes are path-dependent, and even some of the directions of the impacts on peoples welfare in various countries, let alone their magnitude, cannot be determined a priori. EU farmers, for example, may be harmed by not being able to produce most GM foods, even though a barrier to sales in the EU of GM varieties produced abroad provides an offset in the form of protection from what otherwise would be increased import competition. Developing country farmers too could lose if they do not access this new biotechnology, but would their export sales to the EU be boosted sufficiently by the reduced competition from GM-adopting countries to warrant choosing to remain GM-free for the moment? For all these reasons it is important to understand why the EU has taken such a conservative approach compared with America. The conventional explanation for the US-EU difference in GM regulations is that Europeans care more about the natural environment than do Americans, and trust their food safety regulators less. While not denying either of those as key reasons, there have been accusations by Americans that traditional agricultural protectionist forces also are at work. That raises in their minds the question of how EU farmers real incomes are affected by alternative EU policy
The highest-ranked opportunity, based on benefit-cost analysis and the combined assessments of a ten-person expert panel, is providing generic drugs to fight AIDS, while third-ranked is the removal of trade distortions. See Lomborg (2004).
2

stances. To examine this issue, we go beyond earlier empirical analyses (e.g., Nielsen and Anderson 2001; van Meijl and van Tongeren 2002; Nielsen, Robinson and Theirfelder 2003; Huang et al. 2004), by quantifying the effects within countries on real incomes of farm households. Specifically, we look at the consequences for EU farmer welfare of their adoption of GM varieties of coarse grains and oilseeds without and with an EU ban. Our CGE modeling results show that EU farmers could gain from the protection provided by the import barrier more than they would lose by being denied the opportunity to adopt GM technology. That suggests EU farmers even if they were not necessarily advocates of the moratorium of the legislation that replaced it in April 2004 may not be strong opponents of recent and current policies. Why have other agricultural-exporting countries not joined America in voicing frustration at EU import barriers, including as third parties in the Dispute Settlement case that the three early adopting countries took to the WTO in August 2003? To address this question, we examine likely impacts on the Australian and New Zealand economies and their farmers. Our model results suggest that, because of their dependence on sales in EU markets, they are slightly better off with than without the EU de facto moratorium while ever their own governments deny them the right to plant GM grains and oilseeds. The paper begins by pondering why US and EU policies over GM foods are so divergent (section 2). It then goes through the partial equilibrium economics of GM crop adoption to help provide the intuition behind the empirical results to follow (Section 3, which can be skipped by those only interested in the empirical results). The global general equilibrium economic modelling approach for quantifying possible market and welfare effects of GM technology and trade policies is then outlined (Section 4). The estimated effects of various adoption and policy response scenarios, as measured by that model, are presented for America and the EU in Section 5 and for Asia and Africa in Section 6. Section 7 focuses on what the results imply for incomes of farm and non-farm households. Caveats are discussed in Section 8, before the final section draws out some implications for GM technology and trade policies in both rich and poor countries, for the rules-based global trading system, and for the EUs development aid spending.

2. Why Do National GM Policies Diverge? Leaving aside GM cotton (which has had immediate occupational health and environmental benefits in addition to cost savings for farmers, and raises no food safety concerns), why were maize, soybean and canola crops the first ones to be targeted by biotech firms? And why have just three American countries but virtually no others seen widespread adoption of this new food technology so far? The answer to the first question may be that technologically those three crops were easiest to genetically modify. Capturing intellectual property rights, via the annual sale of seed, is also easier for those crops than for open-pollinated crops such as wheat and rice although that doesnt explain why the public agricultural research system has not bred GM varieties of key open-pollinated crops. A more likely or at least supplementary reason has to do with where those crops are grown and sold. According to FAO statistics (see www.fao.org) the US alone accounts for 30-40 per cent of global production and consumption of maize and soybean, and during 1998-2002 the US, Canada and Argentina enjoyed a combined share of global exports of 80 per cent for maize, 64 per cent for soybean, and 42 per

cent for canola. By contrast, those countries account for less than one-sixth of global wheat production and less than one-twentieth of global rice production. That concentration meant regulatory approval for soybean and maize in just three countries could potentially offer biotech and seed firms access to the lions share of those products global markets, unlike for wheat and rice. There was also likely to be less consumer resistance to GM crops that were to be processed into oil or fed to livestock prior to ending up in final products for human consumption, again in contrast to wheat and rice which can be consumed with less processing. Why was Argentina the first developing country to grow GM food crops? More than perhaps any other significant developing country, it shares with the midwest of the United States and the Canadian prairies a temperate climate and broadacre farm structure well suited to growing large fields of maize, soybean and canola. Perhaps equally important, GM varieties of those crops became available just at a time when Argentina was opening up its economy to imports after decades of being highly protective of its industrial sector. That opening up had two important effects: it reduced the implicit taxation of the export-oriented farm sector; and it lowered the farmers price of imported capital goods including tractors and other farm machinery. Hence crop agriculture was ready to boom there from the early 1990s, including at the expense of now-relatively-less-profitable cattle and sheep grazing of pastures. Farmers were also anxious to adopt new no-till cropping practises to reduce damage to their fragile soils. Herbicide-tolerant GM soybean was ideal for doing that, as it could be direct seeded immediately following the wheat harvest. That direct drilling of soybean seed cut several days off the period required to grow the crop, which effectively expanded the potential area of the country suitable for double cropping. As well, legal uncertainty in Argentina about the intellectual property rights attached to herbicide-tolerant GM soybean, and the right there for farmers to use their own crop for next years seed, meant farmers were able to obtain seed without paying the high technology fee charged to North American farmers.3 It is not surprising, therefore, that this serendipitous combination of factors led to an extraordinarily rapid adoption of GM soybean and a significant expansion of aggregate soybean area and production in Argentina. GM maize adoption naturally followed too. Furthermore, Argentinas GM varieties had already been approved in the EU prior to the imposition of the EUs de facto moratorium in 1998. That meant Argentina was able to find it much easier than the US to continue to export to the EU when its moratorium was imposed. Perhaps the only other large developing country with similarly profitable prospects for GM adoption of these crops is Brazil. That country officially resisted the temptation to approve the commercial planting of GM varieties, but farmers in southern Brazil nonetheless went ahead illegally anyway. Unhappy with this further expansion in the free use of its intellectual property, the life science firm Monsanto pressured the President of Brazil to make GM plantings legal from 2003 (initially on just a one-year basis), and to consider Monsantos proposal that a technology fee be paid by farmers as a check-off when they deliver their harvested crop to the silo (a proposal that may also be applied to Argentina). The check-off is being seriously considered because Monsanto has threatened otherwise to take legal action against importers of (in the first instance) Brazilian GM crop products.

About 82 percent of the producer gains from GM soybean adoption in Argentina go to the farmer, whereas for Bt cotton (for which the farmer cannot profitably withhold seed for the following years planting) only about 19 percent do, with the other 81 percent going to the input suppliers (Trigo et al. 2002).

The spectacularly rapid adoption of GM crop varieties in Americas southern cone thus had a number of unusual and possibly unique features that mean this experience will not necessarily be replicable in many other developing countries to the same extent (nor in Australia and New Zealand, where the climate makes double cropping difficult and where very little soybean and maize are grown). Given that initial adoption, why have other countries been slow to follow the American example, especially when there appears to be no hard evidence to justify the concerns reflected in the precautionary stance taken by EU member countries? History shows this is not an isolated example. It took decades before pasteurization of milk was accepted by consumers, and coffee was controversial for centuries (Juma 2005a). The consumer/food safety worries in this case have been that GM-derived food may be more toxic or carcinogenic, result in more allergies, or be nutritionally less adequate than GM-free food; and that transgenes might survive digestion and alter the genome of the person or animal consuming them. Such concerns are inconsistent with statements made by the EU scientific community (European Commission 2001), and with a major report for the UK government by eminent scientists which extensively reviewed available evidence and found no adverse effects anywhere in the world. Like previous similar reports the latter concluded that, on balance, the risks to human health are very low for GM crops currently on the market (King 2003, p. 23). Nor could the King committee find any theoretical reason or empirical evidence to suggest that GM crops would be any more invasive or persistent, or toxic to soil or wildlife outside the farmed environment than conventional crop varieties, or spread their genes to other plants.4 Taking the adoption of GM varieties in the Americas as given, several publicinterest reasons for the EU moratorium are possible. In addition to differences in environmental preferences or in consumer trust in food safety authorities in the EU versus the US, it may be in the economic interests of the EU as a bloc to react by banning GMOs because that maximizes the EUs monopoly power in international food markets: the optimal tariff argument, which has been resurrected recently by Bagwell and Staiger (1999, 2003). Such an outcome seems unlikely given the fact that the EU is foregoing the productivity gains of the new biotechnology, and that it already has a high tariff on maize, but it is proposition we can test empirically taking into account the EUs current Common Agricultural Policy which includes a zero bound tariff on oilseeds.5 More likely are private-interest explanations, consistent with Grossman and Helpman (1995, 2002) and Sykes (1999). These could take a number of different forms. One is simply that the government wishes to appease the anti-GM community. However, they are noisy on both sides of the Atlantic, with consumer surveys suggesting little difference between the US and EU in attitudes towards the desirability of GM labelling, so that is unlikely to be a major explanation for the policy difference.
The Nuffield Council on Bioethics (Thomas et al. 2004, p. 62) also concluded in a recent discussion paper: We do not take the view that there is currently enough evidence of actual or potential harm to justify a blanket moratorium on either research, field trials or the controlled release of GM crops into the environment. For a molecular biologists assessment of the benefits and risks of agricultural biotechnology, in the context of the long history of plant breeding, see Fedoroff and Brown (2004). A comprehensive survey of both the environmental and food safety literature concerning GMOs can be found in FAO (2004, Ch. 4). 5 Because of maize import restrictions and considerable domestic production of feed barley, the EU accounts for less than 4 per cent of global imports of maize, compared with more than one-third of global imports of both soybean and canola.
4

Another possible explanation is that the government is giving domestic EU biotech firms time to catch up with American competitors so that intellectual property rights are paid to domestic rather than foreign patent holders. That hypothesis is difficult to test immediately but could be examined if/when the EU and others do begin again to approve GM varieties. A third possible interest-group explanation is that farm and agribusiness groups benefit from the stance taken by the government of each country/region even if the regional economy as a whole loses. It appears that European agribusiness firms are more focused on selling farm chemicals for traditional crops and less on generating new agricultural biotechnologies than is the case in the US, but that begs the question as to whether this is a consequence of stronger anti-GM sentiment in Europe rather than a reason for anti-GM policies there. The hypothesis that EU farmers are better off with than without their GM regulations is empirically testable in principle, although some practical considerations have to be borne in mind. For example, if buffer zoning were to be required to reduce the risk that GM crops cross-pollinate with native grasses or to improve pollination of the GM variety, such regulations would be more costly (and hence more discouraging of adoption) in closely settled, small-plot, densely populated environments such as in Western Europe and Northeast Asia than in broad-acre settings such as North and South America. Also, if domestic GM production diminished the countrys profits from non-GM food sales (for example through tarnishing its generic reputation as a supplier of safe food), farmers may consider the amortised cost of that outweighs the expected benefit from the new biotechnology, bearing in mind any costs associated with co-existence requirements, any price difference that may emerge between GMand non-GM varieties because of strong opposition by environmental and consumer groups, and the fact that current GM crops are much less important to agriculture in the EU as they are in America. (The EU produces only 6 per cent of the worlds maize, 1 per cent of the soybean and 25 per cent of canola/rapeseed, while North American shares are more than 40 per cent.) Even where there is a net gain to farmers as a group, those within that group wishing to remain non-GM producers may lobby to keep it GM free so as to avoid new identity preservation and contaminationavoidance costs. For many developing countries not needing to segregate crop varieties for domestic sales, the costs of identity preservation and contamination avoidance in order to export GM varieties may be prohibitive, giving rise to what Baldwin (2001) described as a two-tier trading world. If poor countries want to continue exporting food to the EU, they may have no choice but to ban GM food production while the EU moratorium or the new strict labelling and litigation laws are in effect. The number of categories of products subject to testing in the EU is enormous, and the threshold levels of tolerance of accidental GMO contamination are very low. Moreover, feedstuffs also are now included in the EUs list and must be labelled, even though GM protein or DNA will not be present in the livestock products of feedlot operators using feed ingredients based on GM crops. Given the above, under what circumstances might it be conceivable that economies and in particular their farmers are better off by denying themselves access to GM technology, and how do current GM-adopting countries and developing countries fare in those various circumstances? In the next section we begin to answer that using the simplest theoretical model of the markets involved, before turning to a more complex empirical model of global markets.

3. Simple Economics of Adoption of GM Crop Technology Leaving aside the biotechnology research industry, this section considers the market for a single crop such as maize for which a new variety is genetically engineered and made available in the form of purchasable seed by a biotech firm or public research institution. To set the scene, we first look at the effects in a small country of adoption of this cost-reducing technology at home or abroad in the absence of any trade policy responses. This is done both without and with an existing import tariff/export subsidy in place, to show the (un)importance of current measures such as the Common Agricultural Policy on the results. The large-country case is then similarly considered, but in the case of adoption abroad we also explore the impact of it responding with either a variable levy to maintain the domestic price or a ban on imports from GM-adopting countries. The final case considers the international market, so as to be able to see the impacts on GM-adopting and third countries. 3.1 Effects of GM adoption on a small country imposing an import tariff/export subsidy Consider a small country importing this product but unable to influence its international price, Pw. The effects domestically of its GM adoption are shown in Figure 1, assuming that domestic consumers are indifferent to whether the product may contain GMOs (to be relaxed later). The cost-reducing technology causes a downward shift in the supply curve from S to S, which is assumed for convenience to be parallel. In the absence of trade policy distortions, this increases producer and national economic welfare by area abcd. If there was a tariff in place that had raised the domestic price from Pw to Pt, GM adoption would raise producer welfare by area aefd (> area abcd), but decrease government tariff revenue by area efhg (= area bcfe), so national welfare would increase by only area abcd the same as in the absence of the tariff. Were producers supported even more by an export subsidy (and an accompanying tariff to prevent imports for subsidized re-export) that held the domestic price at Ps, producers would gain and taxpayers would lose even more than with just the lower tariff: the producer surplus following adoption would be raised by area ajkd (> area aefd) while government revenue would be lowered by the increase in the export subsidy payment of area jknm (= area bckj), so net national welfare is again increased by only area abcd in Figure 1. Both cases provide the standard conclusion for a small country that an unchanged price-distorting policy alters the distribution of welfare but not the aggregate national welfare gain from a new farm technology. Presumably that protection policy thus raises the producers demand for the new technology, while lowering the taxpayers/finance ministrys incentive to subsidize such research. What about the effect on this economy of GM adoption abroad? Adoption by enough producers to lower the international price from Pw to Pw in Figure 2 would benefit this importing economy provided its tariff was not and did not remain prohibitive, but that gain is greater the smaller the tariff. With no tariff the gain in consumer welfare net of the loss to producers is area qbwx in Figure 2 the maximum gain from improved terms of international trade. If instead a nonprohibitive specific tariff of PtPw applied, the gain would be smaller but still positive, namely area yevu, following the international price fall. That is, even with an

unchanged tariff, the importing country gains less from adoption abroad the larger is that tariff. And if that country chose to raise its tariff to offset the effect of the fall in Pw on the domestic price, then it would gain even less from GM adoption abroad and nothing at all if the tariff increase fully offset the international price. The latter would also be the case if a tariff rate quota operated and the out-of-quota tariff was prohibitive and the quota itself was filled. 3.2 Effects of GM adoption on a large country imposing an import restriction What about when the importing country is large enough to influence the international market for this product? The analysis of Figure 2 in the previous paragraph applies equally to the large importing country enjoying a terms of trade improvement from GM adoption abroad (where again its consumers are assumed for the moment to be indifferent to GM food), assuming no change in this countrys tariff. The only difference is that Pw would have fallen less than in the small-country case following GM adoption abroad, because of the greater quantity demanded by this large country in the international market (to be shown later using Figure 3). When the assumption of no trade policy response is relaxed, qualifications are again needed. For example, if the importing countrys tariff is raised following the international price fall so as to keep the domestic price at Pt, or that is achieved by having a tariff rate quota that is filled and an out-of-quota tariff that is prohibitive, the national gain from GM adoption abroad again would be foregone. But unlike in the small country case, in this case the larger the importing countrys imports prior to adoption abroad, the larger would be the fall in Pw with this variable levy response. That is, the variable levy not only eliminates the gain to that country from GM adoption abroad, but it also (a) reduces the gain from this new biotechnology to the adopting countries, and (b) raises the loss to all other countries that, as net exporters of the conventional (non-GM) variant of this product, suffer a terms of trade loss (except if they enjoy frictionless preferential access to the TRQ-protected market). The same would apply with a variable export subsidy. Alternatively, if the importing country were to ban imports of this product from GM-adopting countries, that would push the domestic price above Pt to a level that could be more or less than Pt, depending on how large the GM-adopters are in international markets. As shown in the next sub-section, this reduces the gain from this new biotechnology to the adopting countries but raises the loss to other countries exporting non-GM varieties of this crop. Were this large importing country to allow GM adoption domestically, its supply curve would shift out from S to S (again assumed for convenience to be a parallel shift) in Figure 2. In the presence of a zero or fixed import tariff this would depress the international price (for diagrammatic convenience to, say, Pw) following the decline in the countrys import demand by quantity bw-rx (or ev-su in the case of a positive tariff). That would reduce the prospects of GM adopters abroad gaining and would lead to gains to domestic producers only if area daqr exceeds area qbPwPw in the case of no tariff (or if area adsy exceeds area yePtPt in the case of a positive tariff). However, if the tariff is raised after GM adoption so as to maintain the preadoption domestic price Pt, producer welfare in this country would unequivocally improve (by area daez) while that of overseas producers would be depressed by further downward pressure on Pw. 3.3 International market effects of sub-global GM adoption

Adoption of that new GM maize variety in some maize-exporting countries provides importing countries with the option of continuing to buy from those exporters a crop that now may contain GMOs, or buying a GM-free product from non-adopting countries. The former international market is shown in Figure 3(a), the latter in Figure 3(b), in both cases assuming no price-distorting policies are introduced. In Figure 3(a), ESg and EDg are the excess supply and excess demand curves for the GM-adopting countries surplus prior to adoption; and in Figure 3(b), ESn and EDn are the excess supply and excess demand curves for the exporting countries choosing not to adopt GM varieties yet. Adoption of a GM variety by some producers in the first group of exporting countries is assumed to lower production costs there such that ESg shifts down to ESg. If consumers in importing countries (as is assumed throughout to be the case in exporting countries) were indifferent about whether the crop may contain GMOs, EDg would remain unchanged and the unit value of that bilateral trade would fall from b to e. Net economic welfare would increase in the importing countries by area bcfe (the gain to their consumers would exceed the loss to their domestic producers), while net economic welfare in the GMadopting countries would change by the area def minus area abc. A sufficient condition for that to be positive is that ESg is parallel to ESg.6 What if some consumers in some of the importing countries consider the GMfree variety to be superior and therefore prefer it over supplies that may contain GMOs? That would cause the EDg curve in Figure 3(a) to shift leftwards to EDg. It would depress the unit value of that bilateral trade even more, to g, and may even lead to the export volume being smaller than before adoption (if h is to the left of c). Consumers of the GM variety benefit even more from its low price, but producers in the adopting countries (who are assumed not to segregate GM and non-GM varieties) are more likely to be worse off. How does all this affect the GM-free exporting countries? On the one hand, if consumers consider the GM and non-GM varieties as perfect substitutes, then the increased volume and lower price of exports from GM-adopting countries would shift EDn to EDn in Figure 3(b), lowering export revenue and net economic welfare in non-adopting exporting countries where producer losses would exceed gains to consumers in those countries by area vwzy. On the other hand, if some consumers in some importing countries prefer the GM-free product, the leftward shift in EDn would be smaller or that curve may even end up to the right rather than the left of its original position. If that substitution effect is strong enough to place EDn to the right of EDn, producers in the GM-free exporting countries would be better rather than worse off following GM adoption abroad, and conversely for consumers in those exporting countries. Should some of the importing countries choose to ban imports of this product from countries adopting the GM variety (in the assumed absence of segregation and an identity preservation system), the leftward shift of EDg in Figure 3(a) would be
If, however, that supply shift is wedge-shaped such that points a and d are closer together or coincide, we know from the theory of immizerizing growth (Bhagwati 1958) that the GM-adopting countries could be worse off if the excess demand curve they face is sufficiently inelastic. Lindner and Jarrett (1978) show that, even with a completely inelastic demand curve, a parallel shift (but not a pivotal shift) downwards in the supply curve will not reduce the exporting countries surplus. Were the supply curve to shift downwards only to the left of point c in Figure 3(a) -- that is, if only lower-cost inframarginal producers were able to adopt the new biotechnology -- all the benefits would stay with the GM-adopting producers (but be shared with the biotech firm that engineered that new GM variety, not modeled above) and none would be shared with consumers abroad or at home.
6

greater and that of EDn in Figure 3(b) would be less (or be more likely to be a rightward shift). In this case the governments import ban is effectively forcing all of its consumers to buy only GM-free varieties, which makes them higher-priced, so consumer welfare in such importing countries and producer welfare in GM-adopting countries will be less than if consumers were free to choose whether to avoid imports that may contain GMOs.7 However, producers in the countries imposing the ban, and in GM-free exporting countries, will be better off with such a ban in place than under free trade while ever they are denied the right to adopt the new technology. Hence it is an empirical question as to whether they would be better off in that situation or being able to adopt the technically more productive GM variety if the latter required (to avoid violating WTO national treatment rules) the removal of the GM import ban. Were that import ban to be replaced by strict labelling and associated liability laws, as happened in the EU in April 2004, importers would be even less inclined to buy from GM-adopting countries -- including in cases where the GM crop varieties grown there had been approved by the EU prior to 1998 (as with some GM soybean varieties) because of the now-higher costs of compliance with the new labelling laws that require segregation and identity preservation back down the value chain to the farmer. The more GM-adopting countries there are that cannot provide credible certification to importers, the larger would be the further leftward shift in the EDg curve in Figure 3(a). It is clear from the above sample of situations that not even some of the signs, let alone the sizes, of the welfare effects of GM adoption by a sub-set of countries can be determined a priori when there are trade policy responses by other countries. Hence the need for quantitative analysis of global markets for the relevant products. Such analysis needs to go beyond the above one-crop model so as also to take account of products that are close substitutes or complements in production and/or consumption or are inputs into other activities. Of particular importance in this case is the livestock sector, since maize and soybean (the first two GM food crops developed) are major inputs into the intensive segment of livestock production but not the extensive segment that still relies on grazing pastures. The most comprehensive way to meet these needs is to use a global economy-wide model of trading nations.

4. The GTAP Model Modifications and Scenarios We use a well-received computable model of the global economy (the Global Trade Analysis Project, or GTAP, model) to examine the effects of some countries adopting the new GMO technology without and then with government and consumer responses in other countries. Being a general equilibrium model, GTAP describes both the vertical and horizontal linkages between all product markets both within the model's individual countries and regions as well as between countries and regions via their bilateral trade flows. The Version 5.4 database used for these applications draws on the global economic structures, trade flows and trade policies of 1997, around the time of the take-off in adoption of GM crop varieties and the year before the EUs de
The presumption here is that the deadweight welfare cost of segregation and identity preservation through the value chain is modest, as it seems to be for segregating different qualities of each traditional grain at present. For a model in which such costs are non-trivial and so the welfare costs are less clear, see Moschini and Lapan (2005).
7

10

facto moratorium was introduced. To make the results easier to digest, the GTAP model has been aggregated to depict the global economy as having 16 regions (to highlight the main participants in the GMO debate), and 14 sectors (with the focus on the primary agricultural sectors affected by the GMO debate and their related processing industries).8 The scenarios analysed here assume that GM-driven farm productivity growth has occurred only in a subset of countries and only for a few of the GTAP sectors. Specifically, coarse grain (primarily maize in the key countries considered) and oilseeds (primarily soybean and canola in the key countries considered) are included in all scenarios, but to illustrate what might happen soon we also look at adoption of GM rice and wheat in some countries. We have modified the GTAP model so it can capture the effects of productivity increases of GM crops, some consumer aversion to products containing GMOs, and substitutability between GM and non-GM crop varieties as intermediate inputs into final consumable food. There are five types of productive factors in the version used here: skilled labour, unskilled labour, agricultural land, other natural resources, and other (non-human) capital. All factors except natural resources (used only in primary production) are assumed to be perfectly mobile throughout the economy. 4.1 Production The Unites States, Argentina and Canada are assumed to be the major adopters of GM crops. Not all other countries are assumed to adopt GM crops in every scenario but, in simulations where we explore what would happen if they did adopt, we assume they would do so to a lesser extent than the first GM-adopting countries. None of these countries is as intensive in the use of maize, soybean and canola as the first GM adopters, and few have the same degree of broad-acre agriculture. Hence they are more likely to be constrained by government regulations in how they plant GM varieties. In addition, unlike the first GM-adopters, some may have segregation and identity preservation costs imposed on them, which further reduces the profitability for them of GM adoption. In all GTAP simulations we assume 45 per cent of US and Canadian coarse grain production is GM. When they adopt, all Latin American countries and Australia are assumed to adopt GM coarse grains at two-thirds the level of the US (i.e., 30 per cent of coarse grain production is GM) while all other countries are assumed to adopt GM coarse grains at one-third the level of US adoption (i.e., 15 per cent of coarse grain production is GM). For oilseeds, we assume that 75 per cent of oilseed production in the US, Canada and Argentina (and Brazil when we allow it) is GM. Again Other Latin American countries and Australia are assumed to adopt at twothirds the extent of the major adopters and the remaining regions adopt at one-third the extent of the major adopters. For the prospective rice scenarios, major assumed adopters, including the US, Canada, China, India, and all other Asian countries are assumed to produce 45 per cent of their crop using GM varieties. All other regions adopt at two-thirds this rate (i.e., 30 per cent of their rice crop is GM). Prospective
The GTAP (Global Trade Analysis Project) model is a multi-regional, static, applied general equilibrium model based on neo-classical microeconomic theory with international trade described by an Armington (1969) specification (which means that products are differentiated by country of origin). See Hertel (1997) for comprehensive model documentation and Dimaranan and McDougall (2002) for the GTAP 5.4 database used here.
8

11

GM wheat adoption is assumed to occur to the same extent as coarse grain adoption for all regions. In the constant-elasticity-of-substitution production nest, producers choose first between imported and domestic inputs according to the models Armington (1969) elasticities, and then choose whether or not to use GM or non-GM varieties as intermediate inputs in their production of final goods. 4.2 Productivity shocks The adopting sectors are each sub-divided into GM and non-GM varieties. Based on available evidence in GM-adopting countries and, in the case of wheat and rice, in countries undertaking field trials (see references in Table 1), the model simulations of GM technology assume technical change in coarse grain, oilseed and (prospectively) wheat production is Hicks-neutral, involving an output-augmenting productivity shock of 7.5 per cent for coarse grain, 6 per cent for oilseeds and 5 per cent for wheat; and for rice, where more-precise information is available from Huang et al. (2004), differing partial productivity improvements are assumed that are of a similar average magnitude to the productivity gains for oilseeds and wheat (Table 1).9 4.3 Consumption In order to capture consumer aversion to GM products, elasticities of substitution between otherwise-like GM and non-GM products are introduced. For the European Union, Australia and New Zealand, where consumers are most GM-averse, those elasticities are set at relatively low levels to capture the perceived low substitutability of these products, whereas they are set at higher levels elsewhere. 4.4 Factor ownership GTAP provides a comprehensive decomposition of changes in national economic welfare as measured by the equivalent variation in income. National and world measures of welfare changes hide the distributional implications within countries of GM policies, however, and so fail to provide insights into the political economy of GM policy choices. While the total economic benefits from trade typically decrease when inefficient policies such as import bans are implemented, some groups within national economies will be beneficiaries. Hence post-simulation analysis is needed. We examine the effects on intra-regional distribution of income by dividing the economy into three groups of households: farmers, unskilled labourers, and owners of human and other capital. Income of each group comes from a combination of factors. Farm households earn income from farm and non-farm activities. The
Because of the variance across GM varieties of any crop in the inputs they save, and because it makes little difference to the trade and welfare results being analysed here, we simply follow most previous analysts in assuming that the productivity effects of genetic modification do not differ across inputs except for rice. Changing the size of the shock (e.g. doubling it) generates near-linear changes (i.e. roughly a doubling) in the effects on prices and quantities. The lower supply price of GM crops is assumed to be net of the technology fee paid to the seed supplier (which is assumed to be a payment for past sunk costs of research) and of any mandatory may contain GMOs labeling and identity preservation costs. These costs are otherwise ignored in the computable general equilibrium analysis to follow, but further research might explicitly include them and, to fine-tune the welfare calculations, even keep track of which country is the home of the (typically multinational) firm receiving the biotechnology fee.
9

12

existing GTAP database provides information about the availability and use of land, unskilled labour, skilled labour, other natural resources and other capital in the agricultural sector, and likewise in other sectors. Non-farm activities of farm households are assumed to earn income from factors in the same proportion as activities conducted by the typical urban capital-owning household. Hence factor shares for farm households are a weighted sum of factor shares used in agricultural production and the factor income shares of capital owners.10 The shares of farm household income from non-farm activities are assumed to be 85 per cent for Japan and Korea, 75 per cent for US and Canada, 55 per cent for EU-15, 40 per cent for Australia, China, Eastern Europe and New Zealand, and 25 per cent for all other regions.11 Unskilled labourers are assumed to receive all their income from unskilled non-farm labour. The expenditure shares are assumed to be the same for all households, so real household incomes are calculated simply by deflating by the national consumer price index. 4.5 Simulations Several sets of simulations are considered below to address questions posed in the introduction and issues raised by the theory in Section 3. We begin with GM adoption for just coarse grains and oilseeds but then add rice and wheat, to get a feel for the relative economic importance to different regions and the world as a whole of current versus prospective GM crop technologies. We look at the impacts of GM adoption by just the US, Canada and Argentina first, without and then with policy reactions by EU and Northeast Asian countries. Then we add the EU to the list of adopters to explore the tradeoffs for the EU between productivity growth via GM adoption and the benefits of remaining GM-free given the prior move to adopt in the Americas. A change of heart in the EU would reduce the reticence of the rest of the world to adopt GM crop varieties, so we examine too the effects of all countries adopting. Some additional scenarios focus on adoption by Asia and Africa.12 Specifically, the base case in the GTAP model, which is calibrated to 1997, is compared with three sets of simulations. The first set examines the effects of adoption
This measure of impact on farmer income is different from the partial equilibrium measure of producer surplus used by, for example, Lindner and Jarrett (1978) who show that even with a completely inelastic demand curve a parallel shift (but not a pivotal shift) downwards in the supply curve will not reduce producer surplus. The measure of farm household income change used here can generate a loss for producers partly because it is a general equilibrium measure that also captures offfarm earnings of farm households, but also because the technology shock only applies to the GM varieties which then have to compete with the (sometimes preferred) non-GM varieties of that crop. Hence the price-depressing impact can more than offset the effect of the productivity improvement on profits of GM adopters relative to traditional producer profits. 11 These estimates for OECD countries are from OECD (2003), and those from developing countries are guestimates. Precise numbers are not possible because of the varying definitions across countries of what is a farm household, particularly over inclusion or otherwise of hobby farms and large corporate farms. In the US, for example, this number is for commercial farmers and those known as intermediates. If just commercial farmers are considered (they account for only one-third of the total number of farmers, another one-quarter being rural residents earning most of their income off farm, and the rest are intermediates), then the share of farm household income earned from non-farm sources falls from 75 to 35 per cent (USDA 2001). 12 In this paper we ignore nutritionally enhanced GM varieties such as golden rice, which aims not so much to boost farm productivity but rather to boost the health of rice consumers through enhancing it with pro-vitamin A (Beyer et al. 2002). For an economic analysis of its possible benefits as compared with those from GM rice that simply boosts farm productivity, see Anderson, Jackson and Nielsen (2005).
10

13

of currently available GM varieties of maize, soybean and canola13 by the current adopters (Argentina, Canada and the US) without and then with an EU ban on imports of GMOs. In doing this we model an EU import ban as a way of representing the current (April 2004) EU policy as it even more effectively limits imports from GMadopting countries than did the de facto moratorium, given that the adopters typically do not segregate GM from non-GM varieties throughout the supply chain and so cannot credibly comply with the new EU regulations even if the varieties they seek to deliver have been approved in the EU.14 Sim 1a: The US, Canada and Argentina adopt first-generation GM varieties of coarse grain and oilseeds that raise farm productivity Sim 1b: As for Sim 1a + the EU bans imports of those crops from GM-adopting countries Sim 1c: As for Sim 1a + EU adopts GM varieties of coarse grain and oilseeds Sim 1d: All countries adopt GM varieties of coarse grain and oilseeds (no moratoria). The second set of simulations involves a repeat of the first set except that China and India are assumed to join America in adopting GM crops, and GM rice and wheat are assumed to be made available to the GM adopting countries farmers, again without and then with import moratoria: Sim 2a: The US, Canada, Argentina, China and India adopt GM varieties of coarse grain, oilseeds, rice and wheat that raise farm productivity Sim 2b: As for Sim 2a + the EU, Japan and Korea impose a ban on imports of directly affected crops from GM-adopting countries The third set of simulations adds to the second set the adoption of GM coarse grain, oilseeds rice and wheat by the South African Customs Union (SACU), and focuses on what difference it would make if Southern African Development Community (SADC) countries other than SACU members either banned imports of GM varieties or allowed their farmers and consumers access to them, again with and without the EU import moratorium: Sim 3a: As for Sim 2a + SACU joins those adopters of all four GM crops

This has to be done in a slightly inflating way in that the GTAP model is not disaggregated below coarse grains and oilseeds. However, in the current adopting countries (Argentina, Canada and the US), maize, soybean and canola are the dominant coarse grains and oilseed crops. 14 The counterfactual in our model in the absence of a ban is the existing CAP policy package as represented in the standard GTAP model for the EU15 countries. This includes a zero tariff on oilseeds and a positive tariff on maize. The WTO bound out-of-quota tariff on maize is slightly above the MFN applied rate (in 2001 they were 69 and 51 per cent, respectively), which means even if the applied rate was not prohibitive the EU could effectively operate a variable levy to insulate the domestic price from an international price fall. And if that tariff is prohibitive and the quota is filled at the in-quota tariff rate, the domestic price again would be unaffected. So with respect to maize the GTAP model estimates the potential benefits to the EU from an improvement in its terms of trade, rather than what would actually happen in the absence of a GM policy response. A recent GTAP modeling exercise by van Meijl and van Tongeren (2004) estimates that correcting for this may as much as halve the estimated welfare gain that the EU would otherwise enjoy from a terms of trade improvement following GM adoption abroad, but it has very little impact on welfare results for other countries. It only halves, rather than eliminates the gain (as happens in Section 3 and Figure 2), largely because oilseed tariffs are already zero.

13

14

Sim 3b: As for Sim 3a + the EU imposes a ban on imports of directly affected crops from GM-adopting countries Sim 3c: As for Sim 3b + the rest of SADC also imposes a ban on imports of directly affected crops from GM-adopting countries including SACU Sim 3d: As for Sim 3b + Rest of SADC adopts GM varieties of coarse grain, oilseeds, rice and wheat Sim 3e: All countries adopt GM varieties of coarse grain, oilseeds, rice and wheat (no moratoria) These simulations, which are summarized in Table 2, are clearly only a small subset of possible simulations, but they are chosen to illustrate key current trade tensions and choices facing developing countries.

5. Effects of American and European Policy Choices In the absence of any adverse reactions abroad, in the first set of scenarios the GM-adopting countries expand their output and net exports of coarse grains and oilseeds (and meat) while the opposite happens in the rest of the world. Consumption of these products expands in all regions because they are now cheaper, but especially in the GM-adopting regions since in this model the Armington assumption ensures that imported products are an imperfect substitute for domestically produced products.15 However, when the EU moratorium is imposed on imports from GMadopting countries (Sim 1b), the international prices of coarse grains and oilseeds fall more so much so as to cause GM-adopting countries to reduce their output of these crops slightly. In Europe, the opposite occurs because the import ban drives up domestic prices (Table 3).16 If instead the EU were to also adopt GM varieties (Sim 1c), EU production and net exports are higher instead of lower, the increases in production and exports by the first GM-adopters are slightly less, and the decreases in production and net exports by the rest of the world are slightly more because international prices of coarse grains and oilseeds fall more. A comparison of the welfare effects of Sim 1c and those from the EU moratorium scenario (Sim 1b) provides a conservative estimate of the cost of the EUs current policy compared with following the American strategy of embracing GM technology for coarse grain and oilseeds -- but only if the EU stance has no effect on other countries GM policies. In so far as the rest of the world is delaying adoption solely because of the EU stance, then an upper-bound estimate of the cost of the EUs current policy can be found by comparing the EU moratorium case with a scenario in which all countries adopt GM varieties of coarse grain and oilseeds (Sim 1d). The aggregate economic welfare effects of these various cases are summarized in Table 4 for all of the first set of scenarios (Sims 1a to 1d), and those effects are disaggregated for scenarios 1a and 1b in Table 5 into three parts: that due to changes in resource allocative efficiency (given the trade and subsidy policies in place in each

The price falls are less than in Nielsen and Anderson (2001) because the present study distinguishes GM from non-GM varieties and applies the productivity shocks only to the former, whereas the earlier study applied it to all production in GM-adopting countries. 16 To reduce the number of numbers in tables, price and quantity results are shown for the sum of GM and non-GM varieties.

15

15

economy), to changes in the regions terms of trade, and to technological change. Several points can be drawn from Tables 4 and 5: The global benefits of the first groups GM adoption is a substantial US$2.3 billion per year (net of the gains to the biotech firms) if there are no adverse reactions elsewhere (Sim 1a), and about one-quarter of it is shared with the major importing regions of the EU and Northeast Asia, while Brazil, Australia, New Zealand and Sub-Saharan Africa other than SACU losing very slightly (because of an adverse change in their terms of trade and, in the case of Brazil, a reduction in allocative efficiency as resources move into more-protected activities); When the EU imposes its moratorium (Sim 1b), however, this is similar to an increase in farm protection there and causes the EU to be worse off by $3.1 billion per year (less whatever value EU consumers place on having avoided consuming GMOs), as well as reducing by one-third the gain to GM-adopting North America, while improving welfare for Brazil considerably but for foodimporting regions of the rest of the world only very slightly; If the EU were to take the opposite view and allow GM adoption (Sim 1c), it would gain more because of its own productivity gains and so too would net importers of these products elsewhere in the world, while net exporters of coarse grains and oilseeds (both GM adopters and non-adopters) would be slightly worse off, so the net global gains would be just 7 per cent more than in Sim 1a because coarse grains and oilseeds are minor crops in the EU compared with North America, assuming the EU moratorium has no impact on the GM policies of other countries; However, if by adopting that opposite stance in the EU the rest of the world became uninhibited about adopting GM varieties of these crops (Sim 1d), global welfare would be increased by nearly twice as much as it would when just North America and Argentina adopt, the EU too would gain more in this scenario as compared with just the EU alone joining the GM adopters (because of improved terms of trade), and almost all of the extra global gains would be enjoyed by developing countries (final column of Table 4). The cost of the EUs moratorium (not counting any domestic consumer benefit from knowing the EU is GM-free) can be thought of as in the range of the difference between columns 4 and 2 and the difference between columns 5 and 2 of Table 4, depending on how much one believes the EUs policy stance is determining the rest of the worlds reluctance to adopt GM varieties of these crops. For the EU that cost range is (406 + 3145 =) $3551 million to (595 + 3145 =) $3740 million per year, while for the world as a whole the range is 2.43 + 1.24 =) $3.67 billion to (4.05 + 1.24 =) $5.29 billion per year (less whatever value EU consumers place on having avoided consuming GMOs). That is, the cost to people outside the EU15 has been up to $1.55 billion per year. But even that $5.29 billion number understates the global welfare cost of the EUs policy in at least three respects. First, we have not included in Sim 1b the fact that the EUs stance has already induced some other countries to also impose similar moratoria. Sri Lanka was perhaps the first developing country to ban the production and importation of GM foods. In 2001 China did the same (with some relaxation in 2002), having been denied access to the EU for some soy sauce exports because they may have been produced using GM soybeans imported by China from the US. If that China moratorium was included along with the EU moratorium, the global welfare

16

loss in Sim 1b would become -$2548 million instead of -$1243 million; and the effect on Chinas economic welfare would be -$1105 million instead of +$111 million, or an extra $1.2 billion per year. Second, these are comparative static simulations that ignore the fact that GM food R&D is on-going and that investment in this area has been reduced considerably because of the EUs extreme policy stance, as biotech firms redirect their investments towards pharmaceuticals and industrial crops instead of food crops. And third, the above results refer to GM adoption just of coarse grains and oilseeds. We turn now to consider possible effects of GM varieties of other grains.

6. Effects of Asian and African Policy Choices The worlds other two major food crops are rice and wheat, for which GM varieties have been developed and are close to being ready for commercial release. How much impact would they have, should governments choose to approve them? Indications to date are that North American farmers are disinterested in adopting GM rice and wheat until they see clearer signs that consumers in Europe and elsewhere would accept them. But if China were to decide to approve the release of GM rice and wheat varieties, India would probably follow soon after. China and India account for 55 per cent of the worlds rice market and 30 per cent of the wheat market Being close to self sufficient in both, they need not worry greatly about market access abroad. If that led to enough other non-EU countries accepting GM varieties of these grains, this could well lead North American and Argentina also to adopt them. Allowing China and India to join the GM-adopters group, and adding rice and wheat to coarse grain and oilseeds, almost doubles the potential global gains from this biotechnology. The global economic welfare gain if there is no policy response by the EU or others is $3.9 billion with just rice added or $4.3 billion if wheat is also added Sim 2a), instead of $2.3 billion per year (compare column 1 of Tables 4 and 6). North America gains only a little more from the addition of GM rice and wheat, which might seem surprising given the importance to it of wheat, but it is because its productivity gain is almost offset by a worsening of its terms of trade as a consequence of their and the other adopters additional productivity. Two-thirds of the extra $2.0 billion per year from adding rice and wheat accrues to China and India, with other developing countries, as a net grain-importing group, enjoying most of the residual via lower-priced imports. What about when the EU moratorium is in place? A comparison of the differences between columns 1 and 2 in Tables 4 and 6 reveals an increase in the cost to the EU of its moratorium, from $3.4 to $5.1 billion per year (again not counting the benefit to EU consumers of knowing they are not consuming GMOs), while for the rest of the world the difference is small. But again a more appropriate comparison if the EU policy is discouraging GM adoption elsewhere is between the EU moratorium case and the case where all countries including the EU adopt (Sim 3e), shown in the penultimate column of Table 6. That difference is $5.5 billion for the EU and $8.4 billion for the world as a whole. Those numbers compare with $3.7 and $5.3 billion, respectively, in the earlier situation that excluded rice and wheat an increase of 1.6 times the estimated global cost of the EUs policy. And the adding of further crops to the GM family would continue to multiply that estimate. Recently South Africa approved the release of GM maize and the initial take up has been very impressive, including by smallholders (Gouse et al. 2004), and

17

Kenya and Egypt are engaging in GM field trials (Thomson 2004). Africa lagged far behind Asia and Latin America following the Green Revolution with dwarf cereal varieties that began in the 1960s, contributing importantly to that continents relatively slow growth in per capita food production particularly up to the 1990s (Evenson and Gollin 2003). Partly as a result, Africa now accounts for one-third of the worlds people living on less than $1 a day (up from one-tenth two decades ago Chen and Ravallion 2004). Since the vast majority of those poor people in SubSaharan Africa are dependent on agriculture for their livelihood and much of their food, the new biotech revolution provides a potential opportunity for raising the welfare of hundreds of millions of Africans. Yet African countries exporting food products fear that they will find food-importing countries discounting or denying access to their products if their farmers adopt GM technology or even if they import GM food (because of the risk of contamination of domestically produced non-GM food). A critical question is: would African food exporters gain more from reduced American competition in that market than from trying to develop and adopt new GM crop varieties? If that improved competitiveness required in turn a ban on imports of all food and feed from GM-adopting countries by those African countries so as to avoid contamination (as ostensibly feared by Mozambique, Zambia and Zimbabwe when they were offered food aid from the US in 2002), would the domestic economic loss to net buyers of food outweigh the gains to farmers in those countries? How would welfare in the rest of Sub-Saharan Africa be affected if the South African Customs Union (SACU) and perhaps other members of the Southern African Development Community (SADC) choose to adopt GM varieties of food crops? Assuming no adverse reaction by consumers or trade policy responses by governments, a comparison of Sims 2a and 3a in Table 6 shows that the adoption of GM varieties of coarse grains and oilseeds by SACU in the absence of the EU moratorium would benefit SACU an extra $6 million p.a. while helping the rest of SSA by $1 million. However, in the presence of the EU ban, SACU would be $1 million worse off and the rest of SSA $1 million better off (compare Sims 2b and 3b). In Sim 3c, SADC members other than SACU place a ban on imports of products that may contain GMOs, while in Sim 3d they embrace the technology. In the first of those cases SACU is made slightly worse off relative to Sim 3b (by $4 million p.a.), while the rest of SADC is hurt even more (by $14 million p.a.) assuming consumers there are indifferent to consuming food that may contain GMOs; and other SSA welfare remains virtually the same. By contrast, if the rest of SADC were to adopt GM varieties along with SACU, as in Sim 3d, its welfare would be boosted by $26 million instead of reduced by $10 million, and SACUs would be up by a further $4 million annually despite the assumed continuance of the EU moratorium.17 It is instructive to focus also on the impacts on domestic food prices and quantities in the Rest of SADC (other than SACU). Table 7 reveals the extent to which domestic food production would be greater -- but by more in Sim 3c where Other SADC chooses not to adopt, and to ban GM imports, than in Sim 3d where Other SADC embraces the new biotechnology. This is the standard consequence of increasing agricultural protectionism, reflected also in the greater decline in net

These values might seem small in absolute terms, but they are more significant when thought of as a percentage of household income in Africas poor economies. Moreover, these GM varieties considered apply just to major grains, and not to many other staple foods that might be genetically modified. Anderson and Jackson (2005) show that the gains could be especially large if new GM varieties developed for Africa were nutritionally enhanced.

17

18

imports of food in Sim 3c than 3d, and in the increase in domestic food prices in Sim 3c compared with their decline in Sim 3d.

7. Impacts on terms of trade and farm household incomes Two political economy questions were raised earlier in the paper. Is the EU policy response to GM adoption in the Americas consistent with a regional publicinterest strategy aimed at capturing terms of trade benefits for the EU; and are farmers in the EU and other non-adopting countries benefiting from the EU policy even if the national economy has a measured welfare loss? 7.1 EU terms of trade The terms of trade for the EU improve when the EU implements a moratorium in response to North America and Argentina adoption of GM coarse grains and oilseeds (Table 5). However, the extent of farm protection provided by the moratorium in the presence of the Common Agricultural Policy is far more than is optimal in terms of boosting EU economic welfare through improved terms of trade. The comparison shows that EU welfare is $3.4 billion per year lower because of the moratorium: the loss in allocative efficiency of $3.6 billion, due to resources moving to more-protected sectors, greatly outweighs the gain from the terms of trade change of just $0.2 billion (compare the upper and lower parts of Table 5, Sim 1b versus Sim 1a). If when the EU abandoned its moratorium and allowed uninhibited domestic production and imports of GM coarse grains and oilseeds, the EU would have gained from its own actions (again, not counting the value EU citizens place on knowing they are not consuming GMOs). If this policy change induced other countries to also allow GM production of those crops, the EU welfare would improve by an additional (595 406 =) $189 million per year (compare Sims 1c and 1d in Table 4), due mainly to a further improvement in the EUs terms of trade. Clearly these results do not support the view that an improvement in the terms of trade is the primary reason for the EUs protectionist response to GM adoption by others. 7.2 Farmers incomes Do EU farmers gain from the current policy regime even though it means foregoing the opportunity to adopt a lower-cost technology? What about farmers and unskilled non-farm labourers in developing countries? The effects on real farm household incomes are summarized in Table 8. The first three rows of that table show Argentinean farmers are slightly better off and farmers in the US and Canada are only slightly worse off as a result of their adoption of GM varieties. Even though the productivity gains are more than offset by the price declines for North American farmers (since they are such a dominant part of the global market for maize and soybean), if any one sub-group of them did not adopt they would be even worse off by suffering the price decline but not enjoying the productivity growth. Note from columns 2 and 3 of Table 8 that their welfare is worsened greatly by the EU moratorium. Farmers in the EU, on the other hand, while only slightly worse off if there is GM adoption in the Americas, are made better off if the EU moratorium on American imports is imposed, but that advantage disappears if

19

EU farmers are allowed to adopt these GM varieties (in which case the price decline evidently fully offsets the productivity gain for them see columns 1 to 3 of row 8 of Table 8). In short, American farmers are made worse off,18 and EU farmers better off, by the EU ban on production and imports of products that may contain GMOs, compared with the alternatives of embracing the new technology as in America. The cost of that moratorium to non-farm households in the EU, however, is less than $20 per capita even when wheat and rice are considered, so it is not surprising that EU consumers appear willing to bear that. The right-hand half of Table 8 refers to GM adoption also of rice and wheat, with China and India joining the GM adopters. This depresses international grain prices even more but the gain to Argentinean farmers from higher wheat productivity more than compensates, while North American farmers are slightly worse off than if GM approval is restricted to coarse grains and oilseeds. Farmers in China and India in this case gain, and those gains to Chinese and Indian farm households are only slightly diminished by the EU moratorium (since they exports very little food to the EU). Since there are also large national economic welfare gains from adoption for China, how can these results be reconciled with Chinas decision to not yet approve GM food production and to ban imports of GM products in 2001 (subsequently weakened in 2002 but only after strong protests from the US)? Import-competing coarse grain19 and oilseed farmers there, as in the EU, might be expected to support tough GM standards, as might the many traditional livestock producers who do not use industrially prepared feedstuffs, if they feel their farm management systems would not allow them to take much advantage of GM technology. But traditionally those farmer interests have had very little political clout in China, as in so many developing countries; and the modern feed-intensive livestock sector would be a counterforce there as in Japan and Korea. Chinas policy is all the more puzzling given that China (a) has the technology and could release numerous GM crop varieties including rice almost immediately (Huang, Wang and Zhang 2001; Huang and Wang 2002; Huang, Hu, Roxelle and Pray 2004), (b) exports very few food products and then mostly to East Asia and so is not likely to suffer serious problems of market access, particularly in the years ahead as industrialization causes Chinas agricultural export competitiveness to diminish, and (c) as a poor country would gain from GM adoption of those crops about 20 times as much as North America when the gain is express as a share of GDP (and India would gain nearly 40 times as much). Officially the reason is that food safety tests are still under way, but an alternative or additional possibility is that China is stalling until it has its own GM varieties ready for release so as to avoid paying foreign firms for intellectual property rights. What about Australia and New Zealand (ANZ), which have chosen so far not to approve GM food production? ANZ would lose slightly less in net economic welfare terms from joining with all others in adopting GM varieties of these four crops (Table 4), but the difference is less than $1 per capita per year. Even if that was
As part of our sensitivity analysis, we recalculated the first row of Table 5 assuming the share of farm household income in the US earned off the farm was 0.75 instead of 0.35 (to allow for smaller farms, even though they typically would not be very influential in lobbying). Even though the magnitude of all the effects on real US farmer incomes was reduced (by about 60 per cent), the signs remained the same. 19 China has been a net exporter of maize in recent years but those exports represent only a small share of production and were helped by export subsidies that have since been phased out.
18

20

sufficient to offset the negative value ANZ consumers place on not knowing if they may be consuming GM products, rows 6 and 7 of Table 8 show the average ANZ farm household income would not improve from GM adoption by it and others even with rice and wheat included regardless of whether the EU moratorium remains. Hence one should not expect ANZ farmers to be pushing hard for rapid approval of GM production until consumer concerns fade and GM varieties of crops of more economic importance to those countries are developed. Nor are there any huge ANZowned biotech firms developing the technology and hence lobbying for its adoption in those relatively small markets in the same way as there are in the US.

8. Qualifications As with all CGE modelling results, the above are subject to a number of qualifications. One has to do with the way consumer preferences are handled. The estimated market and welfare effects vary with the elasticities of substitution assumed between GM and non-GM varieties of a product. Anderson, Nielsen and Robinson (2002) examine this issue and show that this is unlikely to be an important issue because results do not vary much as those elasticities (which are set very low for Europe, Northeast Asia and ANZ and moderate elsewhere) are altered. Of more importance is that we have no satisfactory way of valuing any loss of welfare for consumers who would like to avoid consuming foods containing GMOs but cannot if such foods are introduced into their marketplace without credible labelling. Assuming that loss to be zero is to overstate the net gains from adopting this technology. Also not included explicitly in the analysis is a cost of segregation and identity preservation. We included it implicitly by choosing conservative cost savings due to the new technology, saying they were net of any fees charged for segregation and identity preservation. According to Burton et al. (2002) such fees may be as high as 15 per cent of the farm gate price, which would make it unprofitable to market many GM varieties if that was a required condition of sale. Others suggest those costs could be miniscule at least in developed economies on the grounds that such segregation is increasingly being demanded by consumers of many conventional foods anyway (e.g., different grades or varieties or attributes of each crop) so the marginal cost of expanding such systems to handle GM-ness would not be great, at least in countries that have already shown a willingness to pay for product differentiation. Whatever their cost, if it was sufficiently small to enable profitable segregation and identity preservation, then the global cost of the EUs new regime could be less than what we have modelled as its moratorium, because the non-GM part of GM adopting countries crop output may then be able to access EU markets. The version of the GTAP database used in the above modelling does not include tariff preferences enjoyed by Africans exporting to the EU. In so far as they enjoy preferences on the products considered above, then African and other leastdeveloped and ACP countries exporters are currently receiving the domestic EU price minus trading costs (including the share of the tariff rent enjoyed by the importing firms). That price would be raised by the EU moratorium. In practice this issue is likely to be of minor importance though, for two reasons. One is that the EUs delivered rate of protection on coarse grains is low and on oilseeds is zero, and hence so is the margin of preference. The other is that many exporters find the rules of

21

origin so complicated that it is cheaper for them just to pay the MFN import duty rather than try to take advantage of preferences. Nor does the version of the GTAP database used above represent tariff rate quotas ideally, so our estimate of the welfare gain the EU would receive from an improvement in its terms of trade following GM adoption abroad is more potential than actual, since the counterfactual in practice for TRQ-protected maize (though not for unprotected oilseeds) could be no increase in imports. Unfortunately it would be a major job to install all the 1400-plus current TRQs more appropriately in the GTAP model. In all these simulations we assume for simplicity that there are no negative environmental risks net of positive environmental benefits associated with producing GM crops, and that there is no discounting and/or loss of market access abroad for other food products because of what GM adoption does for a countrys generic reputation as a producer of clean, green, safe food. We also assume all productive factors are saved equally. This is not just for simplicity but also because it is the most neutral assumption when the savings vary within each crop depending on the nature of new varieties (e.g., Bt versus Roundupready: the first uses less pesticide whereas the second may involve more chemical spraying). In any case, sensitivity analysis suggests varying this assumption makes little difference to the mostly macro results presented above. Perhaps more importantly, we have assumed that the available farm land is fixed whereas future generations of biotechnologies may be able to bring unused land back into production (as with drought-tolerant or salt-tolerant varieties, for example), thereby multiplying the gains. We have ignored the owners of intellectual property in GM varieties, and simply assumed the productivity advantage of GM varieties is net of the higher cost of GM seeds. Even if only half the producer gains are captured by the input suppliers, that would suggest our global gains are much larger than those shown above, minus the amortized cost of the enabling R&D. The high profits recorded by the worlds big 7 biotech firms (three based in the US, four based in Europe) in recent years suggest their gains may be at least as large as the above estimates of gains from farmer and consumer adoption of the technology. Even bearing in mind that developing countries are being offered some of this new biotechnology at no cost by its private sector developers, the cost of adapting the off-the-shelf technology to local conditions in Africa especially may well be nontrivial, and may require a better-functioning agricultural research system than has operated in the past four decades (as evidenced by Africas relatively poor take-up of the previous green revolution see Evenson and Gollin 2003). Credit markets may have to improve to help farmers finance the up-front technology fee in cases where seed needs to be purchased from the private sector too. Monsanto has recognised that in the case of Bt cotton and its response has been to foster microcredit schemes a small example of what Juma (2005b) calls the coevolution of biotechnology and social institutions. Finally, and perhaps most importantly, the above comparative static modelling assumes GM technology delivers just a one-off increase in total factor productivity (TFP) for that portion of a crops area planted to the GM varieties. But what is more likely is that, if/when the principle of GM crop production is accepted, there would be an increase in the rate of agricultural TFP growth into the future. Similarly, secondgeneration GM varieties with additional health attributes such as those associated with golden rice (see e.g. Bouis 2002, Welch 2002, Zimmerman and Qaim 2002) would be

22

quicker in coming on stream the more countries embraced the technology. Hence the present value of future returns from GM adoption may be many times the numbers shown above. For that reason, care is needed in interpreting cases where our results suggest that when rich countries introduce trade barriers against GM products, foodimporting developing countries benefit. This is because our analysis does not take into account that moratoria have slowed the investment in agricultural biotechnology, and so reduced future market and technological spillovers to developing countries from that prospective R&D. A further possibility not accounted for is that investments are being increasingly directed to GM varieties of industrial crops, where regulatory hurdles are much less than is the case for food crops. That diversion of R&D effort will be eventually reflected in increased relative profitability for industrial relative to food cropping, which will raise the relative price of food, ceteris paribus.

9. Conclusions From the viewpoint of developing countries, the above results in some senses provide good news. The GM crop technology promises much to the countries willing to adopt these new varieties. The first-generation, farm-productivity enhancing GM varieties alone will boost welfare in the adopting countries (and those welfare gains could be multiplied if second-generation GM varieties such as golden rice were also to be embraced see Anderson, Jackson and Nielsen 2005 and Anderson and Jackson 2005). Those estimated gains are only slightly lower if the EUs policies continue to effectively restrict imports of affected crop products from adopting countries. More importantly, developing countries do not gain if they impose bans on GM crop imports even in the presence of policies restricting imports from GM-adopting countries: the measured consumer loss net of that protectionism boost to Asian and Sub-Saharan African farmers (assuming no consumer aversion to GMOs) is more than the small gain in terms of greater market access to the EU. The stakes in this issue are thus high, with welfare gains that could alleviate poverty directly and substantially in those countries willing and able to adopt this new GM food crop technology. Developing countries thus need to assess whether they share the food safety and environmental concerns of Europeans regarding GMOs. If not, their citizens in general, and their poor in particular, have much to gain from adopting GM crop varieties. Unlike for North America and Argentina, who are heavily dependent on exports of maize and oilseeds, the welfare gains from GM crop adoption by Asian and Sub-Saharan African countries would not be greatly jeopardised by rich countries banning imports of those crop products from the adopting countries. Rather, they would be shared between domestic producers and consumers, net of whatever is received by the biotech and seed firms. Labelling policies potentially provide a more efficient mechanism than trade moratoria for accommodating consumers preferences for non-GM food, provided their cost to the global economy in terms of necessary segregation and identity preservation systems is sufficiently modest. More economic modelling research is required to include the costs of segregating GM-inclusive from GM-free products and to explore the incidence of the identity preservation cost between GM and non-GM farmers, between farmers as a group and others, and between rich and poor countries. If the reason for Chinas reluctance to approve GM varieties for domestic production is because it wants to restrict approval to indigenously developed GM varieties so as to capture the intellectual property earnings domestically, then one can

23

only hope for the sake of their and other developing countries consumers and farmers that such varieties will be ready soon (and that India and subsequent potential GM adopters will be willing to use Chinese or other GM varieties rather than cause further delays while their biotech researchers catch up).20 What are the implications for the WTO rules-based global trading system? If it is in the interests of farmers in food-importing countries of Europe and elsewhere to forego adopting this new biotechnology in order to reduce their competitive disadvantage or get preferential market access vis a vis GM-adopting export-oriented producers in America and elsewhere, then those producers have incentives to support rather than oppose consumer and environmental groups lobbying for tough GMO standards and our results suggest it may not even be in the interests of Cairns Group farmers such as in Australia and New Zealand to oppose that stance. The risk as perceived by some Americans is that such standards could provide a replacement for the traditional forms of government assistance to agriculture that are under pressure to be dismantled in agricultural-protectionist countries, following the Uruguay round of trade negotiations. Not only could that negate the benefits of negotiating lower farm support programs in the current Doha round of WTO negotiations,21 but it promises to raise the level of friction in the WTOs Dispute Settlement Body. One final point. The EU evidently has been willing to forego substantial economic benefits from GM food technology. The numbers in Tables 4 and 6 (the difference between sims 3e and 1b) suggest that in static terms this has been almost $4 billion per year just for the main grains and oilseeds. But that is not the full cost of that policy choice. If EU policies are indeed responsible for the reticence of developing countries to adopt this technology, as reported in Cohen and Paarlberg (2002), then the EU also should compensate developing countries for their benefits foregone. From the difference between sims 3e and 1b in Tables 4 and 6 again, in static terms this is around $4.8 billion per year, or several times that if other crops were taken into account -- not to mention the dynamic boost to the rate of growth of farmer productivity that potentially could come from this technology. An increase in EU payments of, say, just $20 per capita would add around $10 billion per year to the budgets of the key international agricultural research centres and the national public research systems they support in developing countries. Such a boost in the EUs development aid, rather than being an act of charity, would mean simply EU taxpayers paying a bit more of the full global cost of them choosing to avoid consuming GMOs. How effective could such an aid boost be? The CGIARs international agricultural research centres were extraordinarily effective in catalysing researchers who generated the first Green Revolution in the 1960s and 1970s, but they have been starved of new funding over the past two decades. If donors were to reverse their
In the case of minor and open-pollinated crops, public research agencies are more likely to be the providers of GM varieties in low-income countries than private biotech firms (Sithole-Niang, Cohan and Zambrano 2004; Cohen 2005). 21 The prospective gains from multilateral agricultural reform under the WTOs Doha Development Agenda has a wide range. With complete removal of all farm protection and subsidies globally, the world is estimated to be up to $170 billion per year better off by 2015; but with only the modest cuts to bound rates of support and protection implied by the July 2004 Framework Agreement among WTO members (WTO 2004), and given the large gap between bound and applied rates, those gains could be as little as $16 billion by 2015 (Anderson and Martin 2005). It would not take too many more advances in agricultural biotechnology for the potential gains shown in Table 6 (which apply to a 1997 depiction of the world economy for just key grains and oilseeds) plus the associated gains to biotech firms to be of a similar order of magnitude to that lower estimate of the possible gains from Doha.
20

24

current position and instead look favourably on agricultural biotechnology investments by the CGIAR, there would be great scope to boost collaboration with national agricultural researchers in developing countries. The latter are starting to make rapid progress in this area: a survey of fifteen countries at end-2003 identified public research pipelines for GM crops containing no less than 201 genetic transformation events for 45 different crops (Cohen 2004). Just under 60 percent of those transformations were aimed at disease and insect resistance; the rest were focused on drought and salt tolerance, on enhancing the shelf life of products, and on raising nutritional content. A funding injection at this time could make a major contribution -- as clearly identified in the Copenhagen Consensus Project (Lomborg 2004) -- toward meeting the key UN Millennium Development Goals of reducing poverty and hunger, as well as yielding a high rate of return to that investment.

References Anderson, K. and L.A. Jackson (2005), Some Implications of GM Food Technology Policies for Sub-Saharan Africa, Journal of African Economies 14(1), March. Anderson, K., L.A. Jackson and C.P. Nielsen (2005), GM Rice Adoption: Implications for Welfare and Poverty Alleviation, Journal of Economic Integration 20 (forthcoming). Anderson, K. and W, Martin (2005), Agricultural Trade Reform and the Doha Development Agenda, The World Economy 28(9), September (forthcoming). Armington, P.A. (1969), A Theory of Demand for Products Distinguished by Place of Production, IMF Staff Papers 16: 159-178. Bagwell, K. and R.W. Staiger (1999), An Economic Theory of GATT, American Economic Review 89(1): 215-48, March. Baldwin, R. (2001), Regulatory Protection, Developing Nations, and a Two-tier World Trade System, Brookings Trade Forum 3: 237-280. Bernauer, T. (2003), Genes, Trade and Regulation: The Seeds of Conflict in Food Biotechnology, Princeton: Princeton University Press. Beyer, P., S. Al-Babili, X. Ye, P. Lucca, P. Schaub, R. Welsch and I. Potrykus (2002), Golden Rice: Introducing the Beta-Carotene Biosynthesis Pathway into Rice Endosperm by Genetic Engineering to Defeat Vitamin A Deficiency, Journal of Nutrition 132: 506-10. Bhagwati, J.N. (1958), Immizerizing Growth: A Geometric Note, Review of Economic Studies 25: 201-05, June. Bouis, H.E. (2002), Plant Breeding: A New Tool to Fight Micronutrient Malnutrition, Journal of Nutrition 132: 491S-494S. Burton, M., S. James, R. Lindner and J. Pluske (2002), A Way Forward for Frankenstein Foods, Ch. 1 in Market Development for Genetically Modified Foods, edited by V. Santaniello, R.E. Evenson and D. Zilberman, London: CAB International. Chen, S. and M. Ravallion (2004), How Have the Worlds Poorest Fared Since the Early 1980s? mimeo, World Bank, Washington DC, April. Downloadable at www.worldbank.org/research/povmonitor/ Cohen, J.I. (2005), Poorer Nations Turn to Publicly Developed GM Crops, Nature Biotechnology 23(1): 27-33, January. Cohen, J.I. and R. Paalberg (2002), Explaining Restricted Approval and Availability of GM Crops in Developing Countries, AgBiotechNet 4: 1-6, October.

25

Dimaranan, B.V. and R.A. McDougall (eds.) (2002), Global Trade, Assistance, and Production: The GTAP 5 Data Base, West Lafayette: Center for Global Trade Analysis, Purdue University. Evenson, R.E. and D. Gollin (2003), Assessing the Impact of the Green Revolution, 1960-2000, Science 300: 758-62, May 2. European Commission (2001), GMOs: Are There Any Risks? Press Release by Research Directorate-General European Commission, October 8, 2001 available at <http://europa.eu.int/comm./research/press/2001/pr0901en.html> FAO (2004), The State of Food and Agriculture 2003-04: Agricultural Biotechnology, Rome: UN Food and Agriculture Organization. Federoff, N.V. and N.M. Brown (2004), Mendel in the Kitchen, Washington DC: National Academies Press. Foster, M., Berry, P. and Hogan, J. 2003, Market Access for GM Products: Implications for Australia, ABARE eReport 03.13, Canberra: ABARE, July. Fulton, M. and K. Giannakis (2004), Inserting GM Products into the Food Chain: The Market and Welfare Effects of Different Labeling and Regulatory Regimes, American Journal of Agricultural Economics 86(1): 42-60, February. Gouse, M., C.E. Pray, J. Kirsten and D. Schimmelpfennig (2004), A GM Subsistance Crop in Africa: The Case of Bt White Maize in South Africa, Journal of International Biotechnology. Grossman, G.M. and E. Helpman (1995), Trade Wars and Trade Talks, Journal of Political Economy 103(4): 675-708, August. Harhoff, D., P. Regibeau and K. Rockett (2001), Genetically Modified Food: Evaluating the Economic Risks, Economic Policy 16(33): 264-99, October. Hertel, T.W. (ed.) (1997), Global Trade Analysis: Modeling and Applications, Cambridge and New York: Cambridge University Press. Huang, J., H. Hu, S. Rozelle and C. Pray (2004), GM Rice in Farmer Fields: Assessing Productivity and Health Effects in China, mimeo, Center for Chinese Agricultural Policy, Beijing, October. Huang, J., R. Hu, H. van Meijl and F. van Tongeren (2004), Biotechnology Boosts to Crop Productivity in China: Trade and Welfare Implications, Journal of Development Economics 75(1): 27-54, October. James, C. (2004), Global Status of Commercialized Biotech/GM Crops: 2004, International Service for the Acquisition of Agri-biotech Applications, Ithaca NY. Downloadable at http://www.isaaa.org/ Johnson, H.G. (1967), The Possibility of Income Losses from Increased Efficiency or Factor Accumulation in the Presence of Tariffs, Economic Journal 77: 15154, March. Juma, C. (2005a), Innovation, Discontinuity and Controversy, mimeo, Kennedy School, Harvard University, Cambridge MA, January. Juma, C. (2005b), Biotechnology in a Globalizing World: The Coevolution of Technology and Social Institutions, Bioscience 55(3): 2-9, March. King, D.K. (2003), GM Science Review: First Report, Prepared by the GM Science Review Panel under the chairmanship of Sir David King for the UK Government, July. Lapan, H.E. and G.C. Moschini (2004), Innovation and Trade with Endogenous Market Failure: The Case of Genetically Modified Products, American Journal of Agricultural Economics 86(3): 634-48, August.

26

Lindner, R.J. and F.G. Jarrett (1978), Supply Shifts and the Size of Research Benefits, American Journal of Agricultural Economics 60(1): 48-58, February. Lomborg, B. (ed.) (2004), Global Crises, Global Solutions, Cambridge and New York: Cambridge University Press. Marra, M., P. Pardey and J. Alston (2002), The Payoffs to Agricultural Biotechnology: An Assessment of the Evidence, AgBioForum 5(2): 43-50. Downloadable at http://www.agbioforum.org/v5n2/v5n2a02-marra.pdf van Meijl, H. and F. van Tongeren (2004), International Diffusion of Gains from Biotechnology and the European Unions Common Agricultural Policy, Agricultural Economics 31(2): 307-16. Moschini, G. and H. Lapan (2005), Labeling Regulations and Segregation of Firstand Second-Generation GM Products: Innovation Incentives and Welfare Effects, paper presented at the Conference on Economics of Regulation of Agricultural Biotechnologies, Arlington VA, 10-12 March. Nielsen, C.P. and K. Anderson (2001), Global Market Effects of European Responses to Genetically Modified Organisms, Weltwertschaftliches Archiv 137(2): 320-46, June. Nielsen, C., S. Robinson and K. Theirfelder (2003), Consumer Preferences and Trade in Genetically Modified Foods, Journal of Policy Modeling 25(8): 777-94, November. Qaim, M. and D. Zilberman (2003), Yield Effects of Genetically Modified Crops in Developing Countries, Science 299: 900-02. Sithole-Niang, I., J.I. Cohen and P. Zambrano (2004), Putting GM Technologies to Work: Public Research Pipelines in Selected African Countries, African Journal of Biotechnology 3(11): 564-71, November. Thomas, S. et al (2004), The Use of Genetically Modified Crops in Developing Countries, London: Nuffield Council on Bioethics. Thomson, J.A. (2004), The Status of Plant Biotechnology in Africa, AgBioForum 7(1&2): 9-12. Trigo, E., D. Chudnovsky, E. Cap and A. Lopez (2002), Los Transgenicos en la Agricultura Argentina: Una Historia con Final Abierto, Buenos Aires: Libros del Zorzal (in Spanish, but available in English in mimeo form from the authors as Genetically Modified Crops in Argentine Agriculture: An Open Ended Story). USDA (2001), Agricultural Income and Finance Outlook, ERS Report AIS-77, Washington, DC: US Department of Agriculture, 25 September. Welch, R.M. (2002), Breeding Strategies for Biofortified Staple Plant Foods to Reduce Micronutrient Malnutrition Globally, Journal of Nutrition 132: 495S499S. WTO (2004), Doha Work Programme: Decision Adopted by the General Council on 1 August 2004, WT/L/579, Geneva: World Trade Organization (The July Framework Agreement). Zimmermann, R. and M. Qaim (2002), Projecting the Benefits of Golden Rice in the Philippines, ZEF Discussion Paper No. 51, Bonn: Center for Development Research, September.

27

Table 1: Assumed impact of adoption of first-generation GM crop technology on factor productivity for GM varieties relative to current non-GM varieties, by sector (per cent difference)

Land Skilled labour Unskilled labour Capital Chemical inputs

GM coarse grains 7.5 7.5 7.5 7.5 7.5

GM oilseeds 6 6 6 6 6

GM wheat 5 5 5 5 5

GM rice (nongolden) 6 8 8 0 5

Sources: Authors assumptions, based on literature reviews by Marra, Pardey and Alston (2002), Qaim and Zilberman (2003), Huang, Hu, van Meijl and van Tongeren (2004), Huang, Hu, Rozelle and Pray (2004) and FAO (2004).

28

Table 2: Simulation scenarios considered

US, CAN + ARG adopt GM coarse grain and oilseeds EU bans imports of affected crops from GM adopters EU adopts GM coarse grain and oilseeds US, CAN, ARG, CHINA + INDIA adopt GM coarse grain, oilseeds, rice and wheat Japan + Korea ban imports of affected crops from GM adopters SACU adopts GM coarse grain, oilseeds, rice and wheat SADC SACU bans imports of affected crops from GM adopters All SADC adopts GM coarse grain, oilseeds, rice and wheat All countries adopt GM coarse grain and oilseeds All countries adopt GM coarse grain, oilseeds, rice and wheat

Scenario:

1a 1b 1c 1d 2a 2b 3a 3b 3c 3d 3e

X X X

X X X X X X X X X

X X X X X

X X X

X X X

Source: Authors choices (see text for reasoning).

Table 3: Production, consumption, price and trade impacts of GM coarse grains and oilseeds adoption by the US, Canada and Argentina
(percentage changes)

(a) Sim 1a (with no moratoria responses)

Australia Production Coarse grains Oilseeds Meat products Domestic market prices Coarse grains Oilseeds Meat products Imports Coarse grains Oilseeds Meat products Exports Coarse grains Oilseeds Meat products New Zealand -0.16 -0.59 -0.24 Argentina Canada United States 0.80 2.93 0.31 China India EU-15

-0.20 -3.17 -0.14

3.01 1.36 0.09

0.24 1.13 0.01

-0.16 -0.70 -0.02

-0.00 -0.03 -0.00

-0.42 -1.13 -0.02

-0.05 -0.06 -0.04

-0.04 -0.04 -0.03

-1.98 -4.03 -0.16

-1.61 -2.00 -0.30

-1.94 -2.90 -0.42

-0.02 -0.03 -0.03

-0.02 -0.02 -0.02

-0.08 -0.11 -0.03

2.40 6.71 0.38

2.24 1.87 0.73

-1.35 -7.88 -0.35

0.93 0.50 0.16

-2.31 -3.63 -1.00

0.76 6.25 0.83

7.92 1.19 0.20

0.67 2.20 0.06

-0.97 -4.60 -0.34

-0.93 -4.06 -0.43

6.62 12.16 0.20

0.89 2.78 -0.03

3.67 6.37 1.42

-1.58 -2.38 -0.29

-1.41 -1.68 -0.18

-0.65 -1.36 -0.06

(b) Sim 1b (with the EU moratorium on GM products)

Australia Production Coarse grains Oilseeds New Zealand 0.00 0.21 Argentina Canada United States 0.28 -8.87 China India EU-15

-0.11 -2.55

-0.46 -1.58

-0.59 -8.20

-0.16 -0.46

0.01 0.15

7.30 22.31

Meat products

-0.11

0.01

0.33

0.32

0.50

-0.01

0.00

-0.38

Average domestic market prices Coarse grains -0.06 Oilseeds -0.07 Meat products -0.05 Imports Coarse grains Oilseeds Meat products Exports Coarse grains Oilseeds Meat products

-0.01 -0.01 -0.01

-2.36 -4.42 -0.37

-1.96 -2.45 -0.57

-2.37 -3.39 -0.65

-0.03 -0.03 -0.03

0.01 0.01 0.01

1.42 2.06 0.46

2.91 8.04 0.34

2.76 0.89 1.34

-4.86 -8.24 -2.17

1.02 -0.33 -0.20

-3.44 -6.18 -1.57

0.79 7.17 0.91

9.60 1.00 -0.22

-13.11 -24.28 0.49

-0.75 -3.32 -0.26

1.92 8.46 0.06

-2.04 7.55 2.09

-2.05 -21.54 0.62

-0.82 -21.86 2.51

-1.64 8.56 0.02

4.93 9.58 -0.12

10.36 25.68 -0.72

Source: Authors GTAP model simulation results.

Table 4: Economic welfare effects of adoption of first-generation GM coarse grain and oilseed varieties by various regions, without and with EU reactions (equivalent variation in income, US$ million)

US, Canada and Argentina adopt GM coarse grain and oilseed varieties

United States Canada Argentina Brazil Other Latin America Australia New Zealand EU-15 Eastern Europe China India Japan + Korea Other Developing Asia Sth African Customs Union Other SADC

with no moratoria responses Sim 1a 939 72 312 -36 125 -9 -5 267 7 107 0 322 36 3 0

with EU moratorium Sim 1b 628 7 247 256 184 -4 2 -3145 -10 111 3 341 44 7 2

US, Canada, Argentina and EU adopt with no moratoria Sim 1c 928 70 307 -53 128 -10 -5 406 8 110 0 335 37 4 0

All countries adopt

with no moratoria Sim 1d 897 65 287 317 356 2 -6 595 35 235 252 430 134 9 18

Other Sub-Saharan Africa Rest of World WORLD

-2 152 2290

12 75 -1243

-2 167 2429

42 380 4047

Source: Authors GTAP model simulation results.

Table 5: Decomposition of economic welfare effects of adoption of first-generation GM coarse grain and oilseed varieties by various regions, without and with EU moratorium (per cent changes)
(a) With no policy responses (Sim 1a) Allocative efficiency impacts United States 70 Canada 17 Argentina 19 Brazil -18 Other Latin America 70 Australia 2 New Zealand 0 EU-15 181 Eastern Europe 5 China 85 India 3 Japan + Korea 98 Other Developing Asia 17 South African Customs Union 3 Other SADC 0 Other Sub-Saharan Africa 0 Rest of World 98 WORLD 647 Terms of trade impacts -368 -43 -50 -14 55 -11 -4 102 -1 27 -3 239 19 1 0 -2 54 0 New GM technology impacts 1204 101 338 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1643 Total

939 72 312 -36 125 -9 -5 267 7 107 0 322 36 3 0 -2 152 2290

(b) With EU moratorium response (Sim 1b) Allocative efficiency Terms of trade New GM technology Total

United States Canada Argentina Brazil Other Latin America Australia New Zealand EU-15 Eastern Europe China India Japan + Korea Other Developing Asia South African Customs Union Other SADC Other Sub-Saharan Africa Rest of World WORLD

impacts 192 17 5 100 79 3 0 -3431 -15 85 -3 98 13 2 0 1 33 -2821

impacts -690 -111 -89 125 106 -7 2 288 1 28 6 250 33 5 3 9 42 -1

impacts 1153 96 330 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1579

628 7 247 256 184 -4 2 -3145 -10 111 3 341 44 7 2 13 75 -1243

Source: Authors GTAP model simulation results.

Table 6: Estimated economic welfare effects of GM coarse grain, oilseed, rice and wheat adoption
US, CAN, ARG, CHN and IND adopt
without policy response with EU and NE Asia moratoria

US, CAN, ARG, CHN, IND + SACU adopt

without policy response with EU moratorium with EU + Other SADC moratoria with adoption by Other SADC, and EU moratorium

All countries adopt

with no moratoria

Change in economic welfare (equivalent variation in income, $million) SACU Rest of SADC Rest of SSA

Sim 2a 7 0 5 350 83 1045 841 669 355 964 4308

Sim 2b 11 4 23 285 -23 754 833 654 -4717 1322 -892

Sim 3a 13 0 6 350 83 1045 841 669 358 953 4319

Sim 3b 10 4 24 285 -23 754 832 654 -4754 1285 -928

Sim 3c 6 -10 25 284 -24 756 833 654 -4760 1290 -946

Sim 3d 10 26 25 285 -23 754 833 654 -4750 1286 -900

Sim 3e 15 22 165 312 63 1041 899 669 810 3509 7506

EV as % of GDP (sim 3e)

0.01 0.05 0.12 0.12 0.01 0.01 0.25 0.14 0.01 0.03 0.024

Argentina Canada
US China India EU-15

Rest of World WORLD

Source: Authors GTAP model simulation results

Table 7: Trade and domestic production and price impacts in Other SADC (excluding SACU) of GM adoption, with Other SADC either banning or also adopting GM varieties of coarse grains, oilseeds, rice and wheat (percentage changes) US, CAN, ARG, CHN, IND, SACU adopt GM varieties with EU + Rest of SADC Moratoria Sim 3c Production Coarse grains Oilseeds Rice Wheat Meat Domestic market prices Coarse grains Oilseeds Rice Wheat Meat Imports Coarse grains Oilseeds Rice Wheat Meat Exports Coarse grains Oilseeds Rice Wheat Meat 1.0 5.8 1.5 15.6 0.0 0.3 0.3 0.3 0.4 0.2 -25.7 -52.5 -6.9 -21.2 -0.4 1.2 9.2 7.0 1.7 0.6 US, CAN, ARG, CHN, IND, SACU adopt GM varieties + Rest of SADC adopts with EU Moratorium Sim 3d 0.4 1.8 0.9 0.7 0.3 -0.8 -1.2 -1.0 -0.3 0.0 -2.2 -0.6 -4.1 -0.2 -0.9 4.2 12.0 11.7 -0.3 1.4

Source: Authors GTAP model simulation results.

Table 8: Percentage change in farm household real income in selected regions, various GM adoption and policy response scenarios

<GM coarse grains and oilseeds> US, Canada, and Argentina adopt US, Canada, Argentina and EU adopt with no moratorium responses
Sim 1c -0.19 -0.27 0.00 -0.02 -0.07 -0.04 -0.03 -0.05 -0.03 -0.02 -0.03 -0.01 -0.04 -0.04 -0.01 -0.04

with no moratoria responses

United States Canada Argentina Brazil Other Latin America Australia New Zealand EU-15 Eastern Europe China India Japan + Korea Other Developing Asia South African Customs Union Rest of Sub-Saharan Africa Rest of World Sim 1a -0.18 -0.26 0.01 -0.00 -0.06 -0.04 -0.03 -0.03 -0.03 -0.02 0.00 -0.01 -0.04 -0.03 -0.01 -0.04

with EU moratorium
Sim 1b -0.36 -0.57 -0.10 0.15 -0.06 -0.03 0.00 0.74 0.08 -0.02 0.00 -0.01 -0.03 0.02 0.04 0.03

<GM coarse grains, oilseeds, rice and wheat> US, Canada, Argentina, All China and India adopt countries including EU adopt with no with EU moratoria moratorium responses
Sim 2a -0.20 -0.29 0.07 -0.01 -0.07 -0.07 -0.02 -0.04 -0.03 0.06 0.01 0.00 -0.03 -0.03 -0.01 -0.04 Sim 2b -0.43 -0.63 -0.15 0.12 -0.06 -0.04 0.00 0.86 0.11 0.05 -0.02 -0.02 -0.02 0.03 0.04 0.03 Sim 3e -0.29 -0.36 -0.07 -0.03 -0.14 -0.17 -0.18 -0.07 -0.08 0.12 -0.01 0.00 -0.04 -0.07 0.01 -0.09

Source: Authors GTAP model simulation results.

Figure 1: Effects of GM crop adoption at home by a small distorted economy

Price

S S

Ps

Pt

e g b a c

f h m n

Pw

Quantity

Figure 2: Effects of GM crop adoption abroad or at home for a distorted economy

Price

S S

Pt Pt Pw Pw
a q b r y

e s

v u w x

Quantity

Figure 3: Effects of GM crop adoption by some exporting countries