Biotechnology for food, health and environment Ashwini Kumar Department of Botany University of Rajasthan Jaipur 302004. biotechnology has been increasingly applied to crop agriculture. The manipulation of whole organisms, populations of organisms and nucleic components holds much promise for improving crop productivity designing crops for specific environments Monsanto (St Louis, MO, USA), Novartis (Basel, Germany) and Archer, Daniel, Midlands (ADM) (Decatur, IL, USA) believe so Genetic engineering is applied for crop improvement. Designing crops for specific environments. The enhancement of nutrient availability, Pest and disease control, The production of herbicide resistance in crop plants and Tolerance to a variety of environmental stresses. Genes from diverse and exotic sources, are inserted in microorganisms and crop plants to confer resistance to insect pests and diseases, tolerance to herbicides, drought, soil salinity and aluminum toxicity; improved post-harvest quality; enhanced nutrient uptake and nutritional quality; increased photosynthetic rate, sugar, and starch production; increased effectiveness of biocontrol agents; improved understanding of gene action and metabolic pathways; and production of drugs and vaccines in crop plants. The United Nations have projected World Population will increase by 25% to 7.5 billion by 2020. On an average, an additional 73 million people are added annually, of which 97% will live in the developing world. leading to drought stress, excess water leading to submergence and anoxia stress, sub-optimal ambient temperature leading to low temperature stress, supra-optimal ambient temperature leading to high temperature stress, oxidative stress caused by different abiotic stresses in conjunction with high light intensity, heavy metal stress, air pollutants stress, etc.) negatively affect processes associated with biomass production and grain yield, in almost all major field-grown crops. Salinity tolerance AtHKT1 is a salt tolerance determinant that controls Na entry into plant roots. Recently, putative plasma membrane and tonoplast localized NaH transporters mediate energized transport of Na outward from the cytosol to the apoplast or into the vacuole. of cellular Na efflux, whereas the tonoplast antiporter is the AtHKT1 is a salt tolerance determinant that controls Na entry into plant roots. Recently, putative plasma membrane and tonoplast localized NaH transporters mediate energized transport of Na outward from the cytosol to the apoplast or into the vacuole. of cellular Na efflux, whereas the tonoplast antiporter is the In plants, vacuolar NaH antiporters use the proton electrochemical gradient (H) generated by the vacuolar H-translocating enzymes, H-ATPase, and H-PPiase to couple the downhill movement of H with the uphill movement of Na. Both glycophytes and halophytes use a similar strategy that involves regulation of net Na_ flux across the plasma membrane and vacuolar Compartmentalization.of the internalized cation to mediate intracellular Na_ homeostasis. UK based anti GM foods groups say we promote sustainable agriculture and food sovereignty/security, acting as a counter-balance to the models promoted by multinational biotech companies. UK based anti GM food groups say II\Environmental, health, socioeconomic, ethical and cultural questions are also taken into account, and a balance between perceptions in developed and developing countries is sought.’ UK based anti GM food groups say IIEnvironmental, health, socioeconomic, ethical and cultural questions are also taken into account, and a balance between perceptions in developed and developing countries is sought.’ In a capitalist economic system, the traits (products) developed by genetic engineers are turned into commodities that can be bought, sold and traded on the world market as property. Monitoring A high number of economically important recombinant proteins are produced in Escherichia coli based host/vector systems. Reporter green fluorescent protein (GFP) to monitorStress associated promoter regulation. GFP and its blue fluorescent variant BFP in model fermentations using E. coli indicate suitability of the fluorescent reporter proteins for the design of new strategies of on-line bioprocess monitoring.Stress associated promoter regulation. GFP and its blue fluorescent variant BFP in model fermentations using E. coli indicate suitability of the fluorescent reporter proteins for the design of new strategies of on-line bioprocess monitoring.Besides DNA sequence, chromatin organization plays a key role in determining patterns of gene expression:Less compacted euchromatin regions are the most accessible for transcription, whereas highly compacted heterochromatin regions are refractory to transcription. Biotechnology for food: Biotechnology for food: Population growth over the next 30 years will be concentrated almost exclusively in the developing countries, where more than 1 billion people currently live on less than US$ 1 per day, more than 800 million people are undernourished, and 200 million children are underweight (Smil, 2000). This poverty is worst in rural areas where agriculture is the leading source of incomes and employment. The world’s poorest regions are typically those where agricultural investments by the public and private sectors are extremely low. Unless some mechanisms can be found to stimulate agriculture, the outlook for these poor societies is bleak China: Genetic modification techniques, it has spurred worldwide debate. The debate has been going on for decades now and has had a significantly depressing impact on the supply of biotechnology. In the meantime, the demand for the technology has continued to grow rapidly: the global area of GM crops increased from 1.7 million hectares in 1996 to 52.6 million hectares in 2001 (James, 2002). China was one of the first countries to introduce a GM crop commercially, and currently has the fourth largest GM crop area, after the USA, Argentina and Canada (James, 2002). China's agricultural biotechnology development is an interesting case and is unique in many respects. The public sector dominates the industry and the list of GM crops undergoing trials differs from those being worked on in other countries where the technologies are dominated by the private sector (Huang et al., 2002a). The Chinese government views agricultural biotechnology as a tool to help China improve the nation's food security, increase agricultural productivity and farmers' incomes, foster sustainable development and improve its competitive position in international agricultural markets (SSTC, 1990). In 2001, approximately four million small farmers in China adopted Bt cotton (Pray et al., 2002). On the other hand, there is growing concern among policy makers regarding the impact of the ongoing global debate about biotechnology on China's agricultural trade, biosafety and the potential opposition derived from public concerns about the environmental and the food safety of GM products. Because of this, although GM crops are still cultivated in public research institutes, the approval of GM crops (and particularly of food crops) for commercialization has become more difficult since late 1998 (Huang et al., 2001). This reflects the influence of the global debate about GM crops on Chinese policy makers, in particular restrictions on imports to EU countries. China also appears to take a more cautious stance. For example, in January 2002 the Ministry of Agriculture (MoA) announced three new regulations on the biosafety management, trade and labeling of GM farm products. These regulations came into effect on 20 March 2002 and require importers of GM agricultural products to apply to China's MoA for official safety verification approval, leading US producers to accuse Beijing of using the new rules to hiGM varieties in such crops as rice, maize, wheat, soybean, peanut, etc., are either in the research pipeline or are ready for commercialization (Chen, 2000; Li, 2000 and Huang et al., 2002a). A cotton variety with the Bacillus thuringiensis (Bt) gene to control the bollworm is one of the most oft-cited examples of the progress of agricultural biotechnology in China. Since the first Bt cotton variety was approved for commercialization in 1997, the total area under Bt cotton has reached nearly 1.5 million hectares (2001), accounting for 45% of China's cotton area (Table 1). In addition, other transgenic plants with resistance to insects, disease and herbicides, or which have been quality-modified, have been approved for field release and are ready for commercialization. These include transgenic varieties of cotton resistant to fungal disease, rice resistant to insect pests and diseases, wheat resistant to the barley yellow dwarf virus, maize resistant to insects and with improved quality, soybeans resistant to herbicides, transgenic potato resistant to bacterial disease, and so on (Huang et al., 2002a). Progress in plant biotechnology has also been made in recombinant microorganisms such as soybean nodule bacteria, nitrogen-fixing bacteria for rice and corn, and phytase from recombinant yeasts for feed additives. Nitrogen-fixing bacteria and phytase have been commercialized since 1999. In animals, transgenic pigs and carps have been produced since 1997 (NCBED, 2000). China was the first country to complete the shrimp genome sequencing in 2000. Rice, wheat and maize are the three most important crops in China. Each accounts for about 20% of the total area planted. The production and market stability of these three crops are a prime concern of the Chinese government as they are central to China's food security. National food security, particularly related to grains, is a central goal of China's agricultural and food policy and has been incorporated into biotechnology research priority setting (Huang et al., 2001). China's biotechnology program has also selected cotton as a targeted crop because of its large sown area, its contributions to the textile industry and trade, and the serious problems with the associated rapid increase in pesticide applications to control insects (i.e., bollworm and aphids). Pesticide expenditures in cotton production in China increased considerably in the past decades, reaching RMB yuan 834 (approximately US$100) per hectare in 1995. In recent years, cotton production alone consumed about US$500 million annually in pesticides. Genetic traits viewed as priorities may be transferred into target crops. Priority traits include those related to insect and disease resistance, stress tolerance, and quality improvement (Huang et al., 2002a and Huang et al., 2002b). Pest resistance traits have top priority over all traits. Recently, quality improvement traits have been included as priority traits in response to increased market demand for quality foods. In addition, stress tolerance traits—particularly resistance to drought—are gaining attention with the growing concern over water shortages in northern China. Table 2. Research priority and available GM plant events in China by 1999 The BRI of CAAS recently made the other breakthrough in plant disease resistance by developing cotton resistant to fungal diseases (Table 2). Glucanase, glucoxidase and chitnase genes were introduced into major cotton varieties. Transgenic cotton lines with enhanced resistance to Verticillium and Fusarium were approved for environmental release in 1999 (BRI, 2000). Transgenic rice with Xa21, Xa7 and CpTi genes resistant to bacteria blight or rice blast were developed by the Institute of Genetics of CAS, BRI, and China Central Agricultural University. These transgenic rice plants have been approved for environmental release since 1997 (NCBED, 2000). Significant progress has also been made with transgenic plants expressing drought and salinity tolerance in rice. Transgenic rice expressing drought and salinity tolerance has been in field trials since 1998. Genetically modified nitrogen fixing bacteria for rice was approved for commercialization in 2000. Technically, various types of GM rice are ready for commercialization. However, the commercializing GM rice production has not yet been approved as the policy makers' concern about food safety, rice trade (China exports rice though the amount traded is small compared to its consumption) and its implication for the commercialization of other GM food crops such as soybean, wheat and maize. The results of their studies demonstrate that Bt cotton adopters spray 67% fewer times and reduce pesticide expenditures by 82% (Huang et al., 2002b). Because the reduction on the farmers spraying pesticide time (from an average of 20 times during one crop season to eight times), Bt cotton technology is also considered as a labor-saving tecBecause the commercialization of GM rice has not been approved yet, examination of its impacts on rice production inputs and yield are impossible from the farm level survey. However, the government has approved a number of insect, disease and herbicide resistant GM rice varieties for field trial and environmental release since the late 1990s. Interviews were conducted in the trial and environmental release areas by the authors. The results from these interviews are used to hypothesize the impacts of GM rice commercialization on rice yield and input uses (Table 3). It should be noted that Table 3 assumes the seed price difference between GM and non-GM varieties to be constant over time. This is a conservative assumption, which will tend to an underestimation of GM gains if seed prices will in fact converge to a lower level in the future. On the other hand, the hypothesized adoption rates for rice are perhaps overestimating the speed of GM rice adoption. hnology. This scenario assumes GM rice commercialization on top of the adoption of Bt cotton during 2002–2010. This mimics the current adoption process, where Bt cotton continues its rapid adoption path, but GM rice is yet to be released for commercial purposes. Consequently, the results incorporate both the Bt cotton effect and the GM rice effect, but the interaction effects between rice and cotton are negligible. This becomes evident by comparing the second and third column in Table 5. The adoption of GM rice generates cost savings due to its yield increasing, labor saving and pesticides saving impact. If the adoption will take place according to the assumed scenario the supply price of rice will be 12% lower in 2010. Almost 8%-points can be contributed to the yield increasing impact of GM rice, 4.4% to the labor saving impact, and 0.9% to pesticides saving (Table 5). The higher seed price increases the supply price with 1.1%. Despite the sharp decrease in price the output response is only 1.4%. This is due to the low income and price elasticities of domestic demand. People do not demand much more rice if the price decreases or their income increases. The increase in exports is very high (67%), but the impact on output is limited since only a small portion (1.2%) of production is exported. The estimated macro-economic welfare gains of adoption far outweigh the biotech research expenditure in China. The optimistic scenario, with high adoption rates, results in an annual income gain of roughly 5 billion USD in 2010, while the lower range estimate with lower adoption rates still delivers 3 billion USD. These gains are recurring annually and may be compared to R&D expenditures reported in Huang and Wang (2003). They estimate biotech research expenditures in 2000 at about 40 million USD. The accumulated expenditure between 1986 and 2000 amount to about 450 million USD (in real 2000 prices). The implied social rates of return to research are certainly very high. 1. Introduction Because biotechnology—one of this century's most promising and innovative technologies—employs genetic modification techniques, it has spurred worldwide debate. The debate has been going on for decades now and has had a significantly depressing impact on the supply of biotechnology. In the meantime, the demand for the technology has continued to grow rapidly: the global area of GM crops increased from 1.7 million hectares in 1996 to 52.6 million hectares in 2001 (James, 2002). China was one of the first countries to introduce a GM crop commercially, and currently has the fourth largest GM crop area, after the USA, Argentina and Canada (James, 2002). China's agricultural biotechnology development is an interesting case and is unique in many respects. The public sector dominates the industry and the list of GM crops undergoing trials differs from those being worked on in other countries where the technologies are dominated by the private sector (Huang et al., 2002a). The Chinese government views agricultural biotechnology as a tool to help China improve the nation's food security, increase agricultural productivity and farmers' incomes, foster sustainable development and improve its competitive position in international agricultural markets (SSTC, 1990). In 2001, approximately four million small farmers in China adopted Bt cotton (Pray et al., 2002). On the other hand, there is growing concern among policy makers regarding the impact of the ongoing global debate about biotechnology on China's agricultural trade, biosafety and the potential opposition derived from public concerns about the environmental and the food safety of GM products. Because of this, although GM crops are still cultivated in public research institutes, the approval of GM crops (and particularly of food crops) for commercialization has become more difficult since late 1998 (Huang et al., 2001). This reflects the influence of the global debate about GM crops on Chinese policy makers, in particular restrictions on imports to EU countries. China also appears to take a more cautious stance. For example, in January 2002 the Ministry of Agriculture (MoA) announced three new regulations on the biosafety management, trade and labeling of GM farm products. These regulations came into effect on 20 March 2002 and require importers of GM agricultural products to apply to China's MoA for official safety verification approval, leading US producers to accuse Beijing of using the new rules to hinder imports and protect Chinese soybean farmers. China, like many other developing countries, now has to decide how to proceed on the further commercialization of GM crops. Policy makers have raised several issues. Should China continue to promote its agricultural biotechnology and commercialize its GM food crops (i.e. rice and soybean)? How important are the trade restrictions imposed on GM products, particularly those imposed by the EU and by other countries in East Asia? What will be the impact of alternative biotechnology policies (in both China and the rest of world) on China's agricultural economy and trade? Answers to these questions are of critical importance for policy makers and the agricultural industry. The central theme of this paper is to provide a cost–benefit analysis of research and development of GM crops in China in the face of likely international policy developments. To achieve this, the paper is organized as follows. In Section 2, a general review of agricultural biotechnology development in China is provided. The impacts of Bt cotton adoption in China are presented in Section 3. The results from the empirical studies on Bt cotton and the hypothesized results of GM rice commercialization are the data used for the later simulation analyses with a tailored version of the multi-country general equilibrium GTAP model. Section 4 presents the model and scenarios that are used in the impact assessments. The results of the impacts of alternative biotechnology development strategies are discussed in Section 5. The final section provides concluding remarks and areas for policy actions. 2. Agricultural biotechnology development in China 2.1. An overview Biotechnology in China has a long history. Several research institutes within the CAAS (the Chinese Academy of Agricultural Sciences), the CAS (the Chinese Academy of Sciences) and various universities initiated their first agricultural biotechnology research programs in the early 1970s.1 However, the most significant progress in agricultural biotechnology has been made since China initiated a national high-tech program (the ‘863’ program) in March 1986. Since then, agricultural biotechnology laboratories have been established in almost every agricultural academy and major university. There are now over 100 laboratories in China involved in transgenic plant research (Chen, 2000). By 2000, eighteen GM crops had been generated by Chinese research institutes; four of these crops have been approved for commercialization since 1997. , 2 GM varieties in such crops as rice, maize, wheat, soybean, peanut, etc., are either in the research pipeline or are ready for commercialization (Chen, 2000; Li, 2000 and Huang et al., 2002a). A cotton variety with the Bacillus thuringiensis (Bt) gene to control the bollworm is one of the most oft-cited examples of the progress of agricultural biotechnology in China. Since the first Bt cotton variety was approved for commercialization in 1997, the total area under Bt cotton has reached nearly 1.5 million hectares (2001), accounting for 45% of China's cotton area (Table 1). In addition, other transgenic plants with resistance to insects, disease and herbicides, or which have been quality-modified, have been approved for field release and are ready for commercialization. These include transgenic varieties of cotton resistant to fungal disease, rice resistant to insect pests and diseases, wheat resistant to the barley yellow dwarf virus, maize resistant to insects and with improved quality, soybeans resistant to herbicides, transgenic potato resistant to bacterial disease, and so on (Huang et al., 2002a). Table 1. Bt cotton adoption in China Region I includes Hebei, Shangdong and Henan, regional II includes Anhui, Jiangsu and Hubei, and all rest of China are in region III. Source: Author's surveys. Progress in plant biotechnology has also been made in recombinant microorganisms such as soybean nodule bacteria, nitrogen-fixing bacteria for rice and corn, and phytase from recombinant yeasts for feed additives. Nitrogen-fixing bacteria and phytase have been commercialized since 1999. In animals, transgenic pigs and carps have been produced since 1997 (NCBED, 2000). China was the first country to complete the shrimp genome sequencing in 2000. 2.2. Research priorities Rice, wheat and maize are the three most important crops in China. Each accounts for about 20% of the total area planted. The production and market stability of these three crops are a prime concern of the Chinese government as they are central to China's food security. National food security, particularly related to grains, is a central goal of China's agricultural and food policy and has been incorporated into biotechnology research priority setting (Huang et al., 2001). China's biotechnology program has also selected cotton as a targeted crop because of its large sown area, its contributions to the textile industry and trade, and the serious problems with the associated rapid increase in pesticide applications to control insects (i.e., bollworm and aphids). Pesticide expenditures in cotton production in China increased considerably in the past decades, reaching RMB yuan 834 (approximately US$100) per hectare in 1995. In recent years, cotton production alone consumed about US$500 million annually in pesticides. Genetic traits viewed as priorities may be transferred into target crops. Priority traits include those related to insect and disease resistance, stress tolerance, and quality improvement (Huang et al., 2002a and Huang et al., 2002b). Pest resistance traits have top priority over all traits. Recently, quality improvement traits have been included as priority traits in response to increased market demand for quality foods. In addition, stress tolerance traits—particularly resistance to drought—are gaining attention with the growing concern over water shortages in northern China. 2.3. GM cotton and rice China is one of the world's leading countries in the production of GM cotton and rice and the related technology (Table 2). The Biotechnology Research Institute (BRI) of CAAS developed insect-resistant Bt cotton. The Bt gene's modification and plant vector construction technique was granted a patent in China in 1998. The Bt gene was introduced into major cotton varieties using the pollen tube pathway developed in China (Guo and Cui, 1998 and Guo and Cui, 2000). By early 2002, sixteen Bt cotton varieties with resistance to bollworms generated by China's public institutions and five Bt cotton varieties from Monsanto had been approved for commercialization in nine provinces. Table 2. Research priority and available GM plant events in China by 1999 Source: Authors' surveys. The BRI of CAAS recently made the other breakthrough in plant disease resistance by developing cotton resistant to fungal diseases (Table 2). Glucanase, glucoxidase and chitnase genes were introduced into major cotton varieties. Transgenic cotton lines with enhanced resistance to Verticillium and Fusarium were approved for environmental release in 1999 (BRI, 2000). More efforts have been put on the GM rice sector. Numerous research institutes and universities have been working on transgenic rice resistant to insects since the early 1990s. Transgenic hybrid and conventional Bt rice varieties, resistant to rice stem borer and leaf roller were approved for environmental release in 1997 and 1998 (Zhang, 1999). The transgenic rice variety that expressed resistance to rice plant hopper has been tested in field trials. Through the anther culture, the CpTi gene and the Bar gene were successfully introduced into rice, which expressed resistance to rice stem borer and herbicide (NCBED, 2000 and Zhu, 2000). Transgenic rice with Xa21, Xa7 and CpTi genes resistant to bacteria blight or rice blast were developed by the Institute of Genetics of CAS, BRI, and China Central Agricultural University. These transgenic rice plants have been approved for environmental release since 1997 (NCBED, 2000). Significant progress has also been made with transgenic plants expressing drought and salinity tolerance in rice. Transgenic rice expressing drought and salinity tolerance has been in field trials since 1998. Genetically modified nitrogen fixing bacteria for rice was approved for commercialization in 2000. Technically, various types of GM rice are ready for commercialization. However, the commercializing GM rice production has not yet been approved as the policy makers' concern about food safety, rice trade (China exports rice though the amount traded is small compared to its consumption) and its implication for the commercialization of other GM food crops such as soybean, wheat and maize. 3. Impact of Bt cotton in China: factor biased technical change One cannot simply assume that the GM technologies imply a Hicks-neutral productivity boost., 3 The productivity impact of GM technologies in crops is typically factor-biased., 4 That is, cost reductions on some of the production factors can be achieved in varying degrees. See for example European Commission (2001) for a survey and Van Meijl and van Tongeren (2002) for an application to Bt maize and Ht soybean technology. To examine the impact of biotechnology on various input uses and crop yield (after control for input uses) in the cotton production, Pray et al. (2001) and Huang et al. (2002b) used both farm budget analysis and damage control production function approach based on the production practices of 282 cotton farmers (including Bt and non-Bt farmers) in 1999 in Hebei and Shandong provinces, where the bollworm has seriously damaged the local cotton production (Region I in Table 1). A budget analysis by Pray et al. (2001) shows that while there is no significant difference in fertilizer and machinery uses between Bt and non-Bt cotton production, significant reductions were recorded in pesticide and labor use (labor used for spray pesticide). More sophisticated measures based on the same data that applied multivariate regression to estimate the pesticide use and cotton production functions show similar results for the effect of Bt cotton on input uses. The results of their studies demonstrate that Bt cotton adopters spray 67% fewer times and reduce pesticide expenditures by 82% (Huang et al., 2002b). Because the reduction on the farmers spraying pesticide time (from an average of 20 times during one crop season to eight times), Bt cotton technology is also considered as a labor-saving technology. While costs of pesticides and labor inputs are reduced, seed costs of Bt varieties are higher than those of non-Bt cotton by about 100–250% (based on author's survey in 1999, 2000 and 2001 in five provinces where Bt cotton is adopted, the price difference between Bt and non-Bt cotton declined over time). But this is much lower than the market price ratio of Bt cotton seed (40–50 yuan/kg) and non-Bt conventional cotton seed (4–8 yuan/kg) in our sampled areas. The lower seed use per hectare in Bt cotton production and farmers' saved Bt cotton seed partly offset the seed price difference. After controlling for all input differences and geographical location, Huang et al. (2002b) found that adoption of Bt cotton also impacts on cotton yield. Bt cotton contributed to about 7–15% (with an average of about 10%) of yield increase in the Hebei and Shangdong (cotton region I) in 1999. , 5 These results are re-confirmed by two similar surveys conducted in 2000 (which also covered Henan province) and in 2001 (which also covered Anhui and Jiangsu provinces, cotton region II). However, new surveys in 2000–2001 also revealed that the extent of the impacts (pesticide and labor inputs and yield) decline with moving Bt cotton from the region I to region II (authors' survey). We derive productivity effects of Bt cotton based on our 3 years (1999–2001) surveys of primary cotton farmers (1052 farms) in five provinces, including the two major cotton producing regions (regions I and II). We compute the average inputs of pesticides, seed and labor and yield of cotton per hectare for both Bt cotton and non-Bt cotton. The productivity impacts are measured as the difference of input use and yield between Bt and non-Bt cotton. These differences or impacts for regions I and II are reported in the first row (2001) of Table 3. Impacts of Bt cotton in region III in 2001 was estimated by interviewing provincial agricultural bureaus in the region and from interviews of scientists from Biotechnology Research Institute of CAAS. We estimated the impacts separately by region because bollworm and other insect diseases differ among the three cotton production regions. The national level figures are the aggregation of the regional data based on the area shares observed in 2001. Table 3. Hypothesized yield and input difference (%) between GM and non-GM crops and GM adoption in 2001–2010 Source: author's estimates. 3.1. Projecting adoption rates Chinese farmers have adopted Bt cotton at an impressive speed. The question is whether and how the adoption behavior develops in the future and how the associated productivity differentials can be expected to behave. While we have the benefit of historic observations on Bt cotton, the likely technology diffusion of GM rice must necessarily be based on some assumptions. Existing theory on technology diffusion provides some guidance. New technologies with superior characteristics compared to their predecessors are typically not adopted at once by all potential users (see e.g. Karshenas and Stoneman, 1995; Geroski, 2000 and Sunding and Zilberman, 2001 for overviews). One approach that describes innovation adoption as a process of information spread is the epidemic diffusion model. , 6 An alternative approach is to take different characteristics of potential adopters into account in a decision theoretic framework (see e.g. Griliches, 1957; Hategekimana and Trant, 2002; Diederen et al., 2003a and Diederen et al., 2003b). Potential adopters vary over characteristics like farm size, market share, market structure, input prices, labor relations, farm ownership, and current technology. These factors affect the profitability of adoption, and hence the adoption behavior. Given the uncertainty about adoption patterns we follow a rather stylized approach to the projection of adoption rates. Our basic projection assumes that technical change in GM technologies is higher than in non-GM technologies. The new technologies are assumed to be so attractive to farmers that the maximum technically feasible adoption will be realized. As this assumption may be too optimistic we subsequently subject the adoption rates to a sensitivity analysis. For the impacts after 2001, we assume that the technical progress of Bt cotton will be continued as there is a range of forthcoming improved technologies (Table 2). Based on the above empirical study on Bt cotton adoption and its impacts on various inputs and yield, we hypothesize the future patterns of Bt cotton adoption by region and its impacts on inputs and yield as those presented in Table 3. All figures in this table represent the difference (in percentage) of input and yield between Bt cotton and non-Bt cotton. For Bt cotton adoption, we estimated them by region as bollworm and other insect diseases differ among three cotton production regions. The national level figures are the aggregation of the regional data based on the area shares observed in 2001. Because the commercialization of GM rice has not been approved yet, examination of its impacts on rice production inputs and yield are impossible from the farm level survey. However, the government has approved a number of insect, disease and herbicide resistant GM rice varieties for field trial and environmental release since the late 1990s. Interviews were conducted in the trial and environmental release areas by the authors. The results from these interviews are used to hypothesize the impacts of GM rice commercialization on rice yield and input uses (Table 3). It should be noted that Table 3 assumes the seed price difference between GM and non-GM varieties to be constant over time. This is a conservative assumption, which will tend to an underestimation of GM gains if seed prices will in fact converge to a lower level in the future. On the other hand, the hypothesized adoption rates for rice are perhaps overestimating the speed of GM rice adoption. 4. Methodology and scenarios 4.1. Baseline The impact assessment of Chinese biotechnology developments has been done with the help of the well-known GTAP modeling framework. This is a multi-region, multi-sector computable general equilibrium model, with perfect competition and constant returns to scale., 7 The model is fully described in Hertel (1997). This model enables us to incorporate the detailed factor specific GM cost savings as estimated in Section 3. In addition, the multi-sector framework captures backward and forward linkages between the GM crops and the using and supplying sectors. In the GTAP model, firms combine intermediate inputs and primary factors land, labor (skilled and unskilled) and capital. Intermediate inputs are composites of domestic and foreign components, and the foreign component is differentiated by region of origin (Armington assumption). On factor markets, we assume full employment, with labor and capital being fully mobile within regions, but immobile internationally. Labor and capital remuneration rates are endogenously determined at equilibrium. In the case of crop production, farmers make decisions on land allocation. Land is assumed to be imperfectly mobile between alternative crops, and hence allow for endogenous land rent differentials. Each region is equipped with one regional household that distributes income across savings and consumption expenditures. Furthermore, there is an explicit treatment of international trade and transport margins, and a global banking sector, which intermediates between global savings and consumption. The model determines the trade balance in each region endogenously, and hence foreign capital inflows may supplement domestic savings. The GTAP database contains detailed bilateral trade, transport and protection data characterizing economic linkages among regions, linked together with individual country input–output databases which account for intersectoral linkages among the 57 sectors in each of the 65 regions. All monetary values of the data are in USD million and the base year for the version used in this study (version 5, public release) is 1997 (Dimaranan and McDougall, 2002). For the purposes of this paper, the GTAP database has been aggregated into 12 regions and 17 sectors. The aggregation scheme is found in Appendix Table A. The comparative static model has first been used to generate a so-called baseline projection for 2001–2010. In the second step, the impact of alternative biotechnology scenarios is assessed relative to the baseline projection for 2010. The baseline is constructed through recursive updating of the database such that exogenous GDP targets are met, and given exogenous estimates on factor endowments—skilled labor, unskilled labor, capital and natural resources—and population. For this procedure see Hertel et al. (1999), the exogenous macro assumptions are from Walmsley et al. (2000). The macro assumptions for Asia have been updated with recent information from the ADB economic outlook 2002. The baseline projection also includes a continuation of existing policies and the effectuation of important policy events, as they are known to date. The important policy changes are: implementation of the remaining commitments from the GATT Uruguay round agreements, China's WTO accession between 2002 and 2005; global phase out of the Multifibre Agreement under the WTO Agreement on Textiles and Clothing (ATC) by January 2005; and EU enlargement with Central and Eastern European countries (CEECs). Next to those macro- and policy assumptions, the baseline incorporates new data for the Chinese economy. We have incorporated an updated Input–Output table for China, which better reflects the size and input structure of agriculture. An important feature of the new table is an improved estimate of primary factor cost shares in agriculture and improved estimates of crop yields. The new estimates use micro data from farm surveys conducted by a number of ministries led by the State Price Bureau. Another feature of the adjusted database is a drastic adjustment to agricultural trade data for China, which incorporates trade information for 2001. Between 1997 (the base year for GTAP version 5) and 2001 the structure and size of Chinese trade has changed dramatically, and we have adjusted the GTAP data to reflect these changes. We also incorporated econometric estimates for income elasticities for livestock products, rice and wheat (Huang and Rozelle, 1998). The updated estimates for income elasticities are lower than the original GTAP estimates, and are provided in the Annex. This matters especially for the medium-term projections for livestock consumption. Given all this base information for 2001, we project the model in two steps: 2001–2005 and 2005–2010. Summary information on the baseline projection is provided in the Annex. 4.2. Scenarios The central question of this paper is the assessment of economic benefits of research and commercialization of GM crops in the face of likely international policy developments. Towards this end four scenarios have been developed. The first scenario is designed to study the impact of Bt cotton adoption. This impact consists of the part that is already realized in 2001 (Table 1 and Table 3) and the subsequent productivity gains during the period 2001–2010, as summarized in Table 3. Since the potential cost savings affect only farmers who have adopted the GM crop varieties, we weigh the productivity and seed cost estimates by adoption rates to arrive at an average impact on the cotton sector. The second scenario adds the commercialization of GM rice during 2002–2010 to the adoption of Bt cotton. Again, we use the productivity estimates and adoption rates from Table 3. Given the uncertainty in the magnitude of the GMO impacts on input usage and yields and the uncertainty with regard to the adoption rates we conduct a sensitivity analysis on these parameters. The third scenario focuses on a possible import ban on GM products from China. Given that China has commercialized both Bt cotton and GM rice, an import ban on GM rice by the main trading partners is simulated. Finally, we investigate the effects of the recent regulation on labeling of imported soybeans that came into effect in March 2002. This scenario is unfolding in the situation where both the cotton and rice crops have been commercialized. In addition to labeling imported soybeans, the scenario includes labeling of domestic GM rice. The scenario design is ‘additive’, by adding new elements one at a time, and we disentangle the separate effects of each new element where appropriate. 5. Economic impact assessment 5.1. The impacts of commercializing Bt cotton The farmers' decision to adopt Bt cotton weighs the cost savings due to its increased yields, labor cost savings and reduced pesticides cost against increased seed costs. Table 4 shows the total impact of adopting Bt cotton and the contributions of these components to the supply price of cotton, relative to the situation without Bt cotton in 2010. Table 4. Main sectoral effects of adopting Bt cotton (percent change, relative to situation without Bt cotton in 2010) Source: model simulations. The supply price will be 10.9% lower in 2010. The yield increasing and labor saving impacts of Bt cotton contribute, respectively, 7%-point and 3.3%-point to this total effect. The pesticides saving impact lowers the price with 1.7% while the higher seed price increases the supply price with 1.1% (Table 4). The lower supply price increases demand. Domestic demand increases with 4.8% and exports with 58%. However, the share of exports in total demand is very low at 0.24%, and export growth does therefore contribute only mildly to the total cotton demand growth. The rise in domestic demand is almost completely caused by increased demand from the textiles sector. The lower domestic price also implies that cotton imports decrease with 16.6%, relative to the ‘no-Bt’ case. Higher exports and lower imports imply that the trade balance for cotton will improve with 389 million USD (Table 4). The textiles sector is the other main benefiting sector from adopting Bt cotton. The lower supply price of cotton implies that the supply price of textiles decreases with 0.3%. The cost share of cotton in textiles amounts to 2.5% of total cost. The 10.9% decrease in cotton price leads to 0.27% (−10.9%×2.5%) decrease in textiles costs. Output and exports increase with 0.7% and 0.9%, respectively, while imports decrease with 0.3%. This causes the textiles trade balance to improve with 1067 million USD. 5.2. The impact of commercializing both Bt cotton and GM rice 5.2.1. Impact on the rice sector This scenario assumes GM rice commercialization on top of the adoption of Bt cotton during 2002–2010. This mimics the current adoption process, where Bt cotton continues its rapid adoption path, but GM rice is yet to be released for commercial purposes. Consequently, the results incorporate both the Bt cotton effect and the GM rice effect, but the interaction effects between rice and cotton are negligible. This becomes evident by comparing the second and third column in Table 5. The adoption of GM rice generates cost savings due to its yield increasing, labor saving and pesticides saving impact. If the adoption will take place according to the assumed scenario the supply price of rice will be 12% lower in 2010. Almost 8%-points can be contributed to the yield increasing impact of GM rice, 4.4% to the labor saving impact, and 0.9% to pesticides saving (Table 5). The higher seed price increases the supply price with 1.1%. Despite the sharp decrease in price the output response is only 1.4%. This is due to the low income and price elasticities of domestic demand. People do not demand much more rice if the price decreases or their income increases. The increase in exports is very high (67%), but the impact on output is limited since only a small portion (1.2%) of production is exported. Table 5. Impacts on rice sector of adopting GM rice (percent change, relative to situation without GM products in 2010) Source: model simulations 5.2.2. Macro impact The commercialization of both GM crops has substantial welfare effects. Table 6 separates aggregate macro effects into the Bt cotton and GM rice components. The adoption of Bt cotton enhances welfare in China by 1097 million USD in 2010. (equivalent variation, EV). The adoption of GM rice enhances welfare in China by 4155 million USD (Table 6). The impact is therefore four times larger than in the case of Bt cotton, which is explained by the larger size of the rice sector in 2010 (EV in terms of sectoral value added is for both sectors about 15%). This implies that with the same productivity gains more resources are saved in the rice sector. Table 6. Macro impact of adopting Bt cotton and GM rice (a) Source: model simulations. (a) Numbers do not exactly add up to the ‘Total’ column because of small interaction effects. The impact on factor prices varies across factors. Land is a ‘sluggish’ production factor that is not easily reallocated between alternative uses. Hence we allow for land rent differentials across crops. Land prices decline because factor demand is lower due to the yield increasing effect of the GM technology. At the same time, the output expansion falls short of the yield increase, and consequently less land is demanded in the aggregate. Labor and capital are perfectly mobile across domestic sectors. Although the demand for labor decreases in both crops, the aggregate demand for labor increases. In the cotton case the additional labor demand originates mainly from the unskilled labor intensive textiles sector. Due to the positive technical change impact the real exchange rate, 8 improves in both experiments, and this leads to a deterioration of the trade balance. 5.2.3. Impact on other sectors The two major price effects of adopting GM rice are the lower price of rice itself and the lower land price. Sectors that use rice or land intensively will therefore achieve the biggest cost gains and can lower their prices and expand output. Land intensive sectors such as wheat, coarse grains, cotton and other crops can use the extra land that is not necessary anymore to produce the demanded quantity of rice. Animal products (mainly pork and poultry) output will grow because they use land and can use the cheaper coarse grains. Especially the other food sector (mainly food processing) can lower its price because the rice they use as inputs has become much cheaper. This generates an output growth in the other foods sector, which in turn leads to more intermediate demand for its inputs such as wheat and other crops. Although not apparent from Table 7, it should be noted that the effects of GM adoption differ in one important aspect between the two crops. Not only is rice a much larger sector than cotton in terms of its contribution to agricultural output and employment, we also observe completely different demand side effects. Consumers demand not much more rice if price is lower or income higher. This means that consumer can spend their increased income and money they save on buying rice on other products. These income effects increase the demand for many other sectors. Such indirect demand effects are not much observed for Bt cotton. Table 7. Impacts of adoption of Bt cotton and GM rice on other sectors in 2010 (percentage change relative to situation without GM products) Source: model simulations. 5.2.4. Impact in different periods Table 8 shows the impact of adopting Bt cotton and GM rice over time. The incremental contribution of adoption within three periods is given. The first two columns show the impact of past adoption that is already achieved in 2001. In 2001 the welfare gain due to the adoption of Bt cotton is more than one third of the total welfare gain of Bt cotton realized by 2010. The additional gains from adopting Bt cotton in the other two periods slow down, as most farmers that potentially adopt have already switched to the new varieties. For GM rice all the benefits have still to come. Between 2001 and 2005, as adoption of GM rice starts to pick up, about one third of the welfare gains in 2010 are realized. In the period 2005–2010 the adoption rate increase from 40% to 95% and China is expected to arrive at the steep part of the adoption curve and a large part of the potential gains will be realized. Fig. 1 shows the cumulative land productivity gains obtained endogenously from the simulations. Land productivity is defined here as the ratio of output to land use. Fig. 1 displays the change of this ratio, cumulated over the simulation period. Again, the S-shaped curvature for Bt cotton and GM rice indicates that the productivity gains will level off in the future. This pattern is well known from the ‘green revolution’ that dramatically improved rice yields in the 1970s. The productivity growth is not perpetual. Table 8. Impact in different periods: adoption of Bt cotton and GM rice (incremental contribution of adoption within a period in percent changes) Source: model simulations. (5K) Fig. 1. Simulated land productivity growth rates over time. The graph is obtained from a Spline interpolation of simulated ratios of output growth over land use in 1997, 2000, 2005 and 2010. 5.2.5. Trade impact on other regions Although China witnesses rising exports and/or reduced imports as a consequence of rapid GM adoption, the patterns of global trade in both the textiles and garments and the rice sectors are not affected very much. Table 9 presents the changes in the regional trade balance relative to the ‘no-GM’ case in 2010. The impact is negligible on major rice importers such as Africa and some rice deficit developing countries in Asia. Major rice exporters in South-East Asia (i.e., Thailand, Vietnam and Burma) may witness a drop in net export revenues. The Chinese biotechnology research strategy has in the first place concentrated on crops that are of great importance to rural livelihoods, and not on those that are important in terms of export earnings. Rice exports from China represent only a small share in international rice trade. Table 9. Impact of adoption of Bt cotton and GM rice in China on the commodity trade balance in various regions (year 2010, comparison against situation without GM crops) Details of country groups are provided in Appendix Table A. There is an immediate negative impact on other major cotton exporters, most notably India and Pakistan, which are part of our OthAsia region. The cost savings and yield increases from Bt cotton translate into lower production cost for the Chinese textiles and garments industry, but these cost reductions are not of such orders of magnitude that other garments producers (e.g., India and Bangladesh) are affected very much. The phasing out of the multifibre agreement by 2005 is of greater importance for global textiles and garments trade than Bt cotton commercialization in China., 9 5.2.6. Robustness of results: sensitivity analysis on productivity shocks In this section we conduct a systematic sensitivity analysis (SSA) on the productivity parameters, given the uncertainty in the magnitude of the GMO impacts on input usage and yields and given the uncertainty with regard to the adoption rates. In Section 3.1 we have argued that the maximum technically feasible adoption of the GM technology may not be realized. As a starting point for the SSA, we have taken a more conservative projection of the adoption rates of Bt cotton, which are obtained from estimating a logistic equation for each region. , 10 The estimates that are based on historical adoption data do not take fully into account that the benefits of GM technologies over non-GM technologies increase over time and are therefore lower than those reported in Table 3. According to the logistic model 56% of the area would be Bt cotton by 2010, rather than the 92%, which is believed to be technically feasible given increased benefits of GM technologies. For rice, the same procedure has been followed, albeit that we do not have historical observations available. Here, we have also reduced the mean adoption rate in 2010 to 56%. The simulation results of the previous estimates, as described in Section 5.2.2, and a scenario where the adoption rates are 56% for both cotton and rice are given in Table 10. The average effects vary almost linearly with the adoption rates. The lowering of adoption rates by about 40% compared to the optimistic scenario of Table 3, results also in a reduction of values of key variables by about 40% (compare the first two columns called "Total impact"). Table 10. Results of sensitivity analysis: adoption of Bt cotton and GM rice Source: model simulations. Next, we performed an SSA to test the robustness of our results with regard to the productivity shocks due to uncertainty in GM impacts and adoption rates. The SSA procedure follows Arndt (1996), and uses a Gaussian quadrature. A main advantage of the SSA is that it produces estimates of means and standard deviations of model, while requiring only a limited number of model runs. This approach views the adoption rates as random variables with associated distributions. We assume that the productivity shocks fall within a band of plus and minus 60% of the mean and the distribution is assumed to be triangular around the mean. Table 10 shows that the standard deviation around the mean values is generally low, and the SSA results are very encouraging as regards the robustness of the simulation estimates. For example, if we subtract two times the standard deviation from the mean EV estimate, we still observe a positive macro economic welfare gain, of 1.4 billion USD in 2010. 5.3. Assessment: benefits of GM adoption In the discussion above we have referred to the equivalent variation (EV) concept to provide a summary indicator of the potential economy-wide benefits of GM adoption. Of course, the conventional EV measure of welfare changes does not take into account other important aspects of human well-being. The welfare measurement is based on a comparison of utility derived from consumption with and without the simulated changes. The utility function, 11 does not account for intrinsic, positive or negative, utility that might be attached to the introduction of new crop varieties. Another clear benefit of Bt cotton adoption is the reduced application of insecticides. According to Huang et al. (2000) pesticide poisoning affected between 30,000 to more than 70,000 persons in farming each year in China in the past decade. On average China had about 500 deaths due to pesticide poisoning every year, and the number has increased significantly since the late 1980s and reached 741 in 1995. Bt cotton has an enormous potential to reduce the health risks of insecticide use. According to Pray et al. (2002) Bt cotton has significantly reduced the number of farmers who are poisoned each year. Based on surveys these authors show that 22%–29% of the non-Bt cotton farmers reported poisonings in 1999 and 2000, while only 5–7% of the Bt cotton farmers reported poisonings. , 12 Self-reported ailments are only the tip of the iceberg. Both visible acute health impairments and invisible chronic health diseases of rice farmers are closely linked with the extent of their exposure to pesticides (Hunbag et al., 2000). The estimated macro-economic welfare gains of adoption far outweigh the biotech research expenditure in China. The optimistic scenario, with high adoption rates, results in an annual income gain of roughly 5 billion USD in 2010, while the lower range estimate with lower adoption rates still delivers 3 billion USD. These gains are recurring annually and may be compared to R&D expenditures reported in Huang and Wang (2003). They estimate biotech research expenditures in 2000 at about 40 million USD. The accumulated expenditure between 1986 and 2000 amount to about 450 million USD (in real 2000 prices). The implied social rates of return to research are certainly very high. The question addressed in this section is whether it is still worthwhile for China to commercialize GM rice if consumer concerns in the enlarged EU, Japan, Korea and South East Asia lead to a ban on GM food products. Technically, this is modeled as a non-tariff-barrier against Chinese rice imports that reduces these countries' imports of Chinese rice to zero. Under this scenario exports of GM rice from China decline substantially. Whereas an increase of rice exports volume of 67% was projected when both GM rice and Bt cotton are adopted, the trade ban results in a drop to just 5% above the baseline result for 2010 (Table 11). This follows immediately from the export shares in the baseline situation in 2010 (without all the biotech shocks), which amount to 21%, 8% and 9% for South East Asia, Japan–Korea and the EU27 (enlarged EU with 27 counties), respectively. Rice output is also declining, by 0.5% points (1.4–0.9%=0.5%, Table 11). The drop is limited, because the share of exports in production is only 1.2%. China is developing the largest public plant biotechnology capacity outside of North America. The international debate on GM technologies has its influence on Chinese policy making and on agricultural industry. Adoption of Bt cotton has been proceeding at a rapid pace in recent years. The largest part of the potential productivity gains from Bt cotton will be realized already by 2005, thereafter the productivity growth is slowing down. In contrast, GM rice is not yet available to farmers on a commercial basis, and our estimates indicate that large productivity gains are yet to be realized between 2005 and 2010. The economic gains from GMO adoption are substantial. In the most optimistic scenario, where China commercializes both Bt cotton and GM rice, the welfare gains amount to an additional annual income of about 5 billion US$ in 2010. This amounts to about 3.5 USD per person. This is not a small amount in a country, where according to the World Bank 18% of the population had to survive with less than 1$ per day in 1998., 18 If actual adoption rates are lower, we still observe an income gain of 3 billion USD in 2010. Given the importance of rice for agricultural production, employment and food budget shares, the gains from GM rice adoption are orders of magnitude larger than the Bt cotton gains. The estimated macro economic welfare gains far outweigh the public biotechnology research expenditures. Trends affecting the next generation of U.S. agricultural biotechnology: Politics, policy, and plant-made pharmaceuticals Patrick A. Stewart , a, , and Andrew J. Knightb a Department of Political Science, Masters of Public Administration Program, P.O. Box 1750, Arkansas State University, State University, AR 72467, USA b Department of Criminology, Sociology and Geography, Arkansas State University, State University, AR 72467, USA Received 7 December 2003; Revised 3 March 2004; accepted 4 March 2004. Available online 11 September 2004. Abstract This paper analyzes the structure and history of regulatory policies in the United States, focusing on recent regulatory changes due to the promise and threat posed by plant-made pharmaceuticals (PMPs). PMPs are the latest advance in the genetic engineering of plants and promise to produce medicines inexpensively and abundantly by using a range of different plants as factories to express active medicinal ingredients; however, PMPs may pose a risk to the public's health if they enter the food supply. How the benefits and risks of PMPs are addressed by the respective government's regulation and how this will affect what, if any, products make it to the marketplace and their ultimate success are of great concern to many different parties, ranging from consumers and farmers to health and food production industries. As a result, this paper addresses the history of agricultural biotechnology regulatory policy since 1972, arguing that three distinct periods may be identified: (1) from 1972 to 1986 when the new biotechnology was focused on scientific self-regulation in the laboratory; (2) from 1987 to 2002, as the technology was being developed and widespread release of certain technologies became more common and was not perceived as an environmental threat, regulations became increasingly laxer; and finally, (3) we argue that we are entering a third phase with a series of controversies over regulatory infractions involving genetically engineered (GE) plants and the potential threats posed by PMPs. 1. The next generation in U.S. agricultural biotechnology While genetically engineered (GE) crops, such as Round-Up Ready soybean and Bacillus thuringiensis (Bt) corn and cotton, have become a pervasive part of agricultural production in the United States over the past 7 years, their place in the market is by no means assured. With many nations following the lead of Europe by not accepting goods derived from GE plants into their markets, or demanding their labeling, consumers will not have an opportunity to purchase these products, as these markets will remain closed. In countries that are willing to embrace genetic engineering plants, like the United States, if consumers are unwilling to buy these products and the public is unwilling to accept the risk of GE plants being grown, it is unlikely that GE crops will survive as part of the agricultural system. Public opinion is often the key driver in regulatory change. As a reaction to public perception of potential threats and not experienced events, the biotechnology regulatory arena has experienced a good deal of change since 1986. Because of the lack of substantive experience with health and/or environmental threats from the release of biotech products, federal government agencies established an amalgam of existing regulations to respond to potential, but not established, threats. These regulations use genetic engineering as the trigger and have undergone a series of alterations, as more knowledge of the risks associated with the release of GE plants has been accumulated, as well as in response to public reactions, or lack thereof, to perceived risks. Likewise, change in the field testing of plant biotechnology has occurred since the regulatory regime was put in place in 1986 and field releases began in 1987. Three different generations of alterations to plants have been identified as likely taking place. First-generation biotechnologies alter the characteristics of plants so that they require less agricultural inputs such as herbicides, pesticides, and fertilizer as well as other chemicals. Second-generation biotechnologies focus on improving product quality so the plants are more nutritious, tastier, or stay fresh longer. Third-generation GE plants are ones in which cash crops act as "factories", producing industrial goods, pharmaceuticals, and other products more efficiently and cheaper than traditional approaches [1]. Author Keywords: Genetic engineering; Agricultural biotechnology ; Regulation; Field releases; Plant made pharmaceuticals, PMPs; Plant-made industrial products, PMIPs First-generation products, such as Round-Up Ready herbicide tolerant plants and Bt insecticidal crops, are used extensively by farmers. While crops exhibiting product quality characteristics have been given regulatory approval, the second-generation crops have yet to catch on in the marketplace. For instance, Calgene's Flavr Savr tomato, which was designed to have a longer shelf life and a better taste than traditional tomatoes, appeared briefly in grocery stores but was eventually pulled from the shelves due to marketing and transportation problems. Finally, the third generation of GE crops includes plant-made pharmaceuticals (PMPs) and plant-made industrial products (PMIPs). PMPs are designed to produce vaccines and antibodies for a wide range of diseases like rabies, traveler's diarrhea, cholera, hepatitis B, antibodies to fight cancer, and tooth decay, and therapeutic proteins for cystic fibrosis, liver disease, and hemorrhages. PMIPs can be used for a variety of industrial purposes, such as to accumulate heavy metals in the plant to clean up soil, perform as biosensors for hazardous materials such as explosives found in landmines, produce enzymes and epoxies for industrial uses and plastics to replace petroleum-based products, and to produce cosmetics [2]. GE plants, however, have not been embraced by all segments of society, as criticism and controversy have attended their production, particularly as issues surrounding environmental and health risks have become publicized. 2.1. In the lab: Asilomar (1972) to Coordinated Framework (1986) The impetus for regulation of the new biotechnology came not from an experienced catastrophe or crisis, but from public concerns about potential environmental disaster. Well-meaning, but politically inexperienced, scientists called for self-regulation to address public concerns they inadvertently kindled. Specifically, in 1974, a meeting called for by a group of eminent scientists in one of the most visible and important journals in the scientific world (Science) was attended by 150 carefully selected participants at the Asilomar Center in Pacific Grove, CA [5]. This meeting, which was held to calm public concerns over the use of recombinant DNA technology, instead highlighted the uncertainty of elite scientists and their desire to restrict debate to within the scientific community by limiting public involvement and press coverage [6 and 7]. While the result, scientific self-regulation with restraints only enforced on Federally funded projects [specifically, by the National Institutes of Health (NIH)], was as intended, the Asilomar conference and the events attending it served to set in motion a risk-averse perspective in which the threat of the new biotechnology was assumed before it was proved. This in turn led to it being the first technology to be regulated before risk was shown to exist ( [8 and 9] p.223). Over the next decade, most research tended to be laboratory-based. However, as the new biotechnology started moving from the lab to the field, concerns over field releases of GE organisms were raised, especially by such interest groups as the Foundation on Economic Trends (FET). One GE organism in particular raised concern—a soil bacterium genetically altered to reduce the likelihood of frost damage by lowering the point at which ice forms on a plant, in turn preventing an estimated US$1 billion in losses annually. Unfortunately dubbed "ice-minus", the perceived threat of the bacterium escaping, proliferating, and altering the environment was used as a focusing event to draw attention to the potential dangers raised by the new biotechnology, especially as the FET brought suit against the Environmental Protection Agency (EPA) for not protecting the environment against this threat. This, combined with the need to clarify administrative turf who had regulatory primacy over the nascent industry and broader environmental concerns, led to the Reagan Administration's OSTP proposing the Coordinated Framework for the Regulation of Biotechnology (hereafter the Coordinated Framework) in 1985, and its being promulgated in 1986 ([6] p. 192–197). The resultant Coordinated Framework coordinated the regulatory jurisdictions of the Food and Drug Administration (FDA), NIH, EPA, USDA, and the National Science Foundation (NSF). In all cases, a "pragmatic" approach was used in which preexisting regulations were utilized on a product-by-product basis, but with the use of genetic engineering processes to set off the regulatory trigger ( [10] p. 79). The Coordinated Framework put in place dealt with jurisdictional overlap between the USDA, EPA, and FDA 1 with GE plant products as they move from the field to consumers. The first line of regulatory oversight with the field release of GE organisms was and remains the USDA's Animal and Plant Health Inspection Service (APHIS), primarily through the Plant Quarantine Act (PQA) and the Federal Plant Pest Act (FPPA), although USDA also claims oversight through the Federal Meat Inspection Act (FMIA), the Poultry Products Inspection Act (PPIA), the Virus, Serum, Toxin, and Analogous Products Act (VSTA), and the Federal Seed Act (FSA). The EPA's regulatory oversight comes into play when products reach the commercial stage of development through the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) and the Toxic Substances Control Act (TSCA). Finally, FDA regulates the new biotechnology under the Federal Food, Drug and Cosmetic Act (FFDCA) and the Food Quality Protection Act (FQPA), which also affects EPA to a lesser extent. 2.2. In the fields: Coordinated Framework (1987) to widespread field release (2002) As previously stated, the initial point where GE organisms are regulated is by the USDA as field releases, or the movement of GE organisms into or through the United States, under 7 CFR part 340 of the FPPA and the PQA. Under these acts, APHIS asserts broad regulatory authority over organisms, products, and articles that are plant pests or could harbor plant pests, whether they are genetically engineered or not. Although the Coordinated Framework explicitly states federal agencies should focus on characteristics of risk posed by the product, APHIS uses genetic engineering to trigger regulatory oversight. With the permitting process, which was established in 1987 to allow for field testing of GE plants, the process applies to organisms using genetic materials from organisms defined as plant pests, unknown or unclassified organisms, or organisms that the APHIS deputy administrator determines to be or has reason to believe is a plant pest [11]. These regulations, however, are not encompassing of all GE plants. While the use of recombinant DNA inserted through Agrobacterium is a trigger for regulation, recombinant DNA inserted through a gene gun, genes inserted that do not come from a listed plant pest, or a plant whose pest status is undetermined do not trigger the same regulations. Although creators of such plants have, to date, sent courtesy notifications or permit applications ([12 and 13] p. 107), there is no certainty that this practice will continue. In March 1993, the permitting process was changed by APHIS to include a notification track in order to simplify the process. Six plant species, corn, cotton, potato, soybean, tobacco, and tomato, which were considered genetically well characterized, and in which the transmission of GE characteristics were seen as limited due to the lack of wild relatives in North America, were given notification status. The reduction of paperwork through the use of the notification track, instead of the permitting procedure, led to a decrease in the average waiting period from 120 to 30 days, and costs from US$5000 to US$250 dollars. As can be expected, there was a sudden upturn in field release activity, especially regarding these crops (see Fig. 1). (10K) Fig. 1. USDA-APHIS field releases. In May 1997, further changes to the field release regulations were put in place by APHIS to allow the introduction of the great majority of GE plants under the notification procedure. With this approach, a plant is eligible for the notification process if it meets the following requirements: the plant species is not listed as a noxious weed in the area where it is to be released; the inserted DNA is stably integrated into the host genome; the inserted DNA's function is known and does not cause production of an infectious entity; encoded substances are not toxic to nontarget species likely to feed on the plant or encode products for pharmaceutical use; virus derived sequences must be unlikely to facilitate virulence and spread in plants; and, finally, the new genes must not be derived from human- or animal-disease-causing agents ([13] (p. 109–109); [11]). In other words, unless there is seen to be an environmental threat from the new plants, the less rigorous data collection standards of the notification approach is applied. As can be seen in Fig. 1, this led to an increase in use of the notification track as well as a decrease in utilization of the permitting track. Statistical analysis supports the contention that policy changes put in place since 1987 have led to greater field release activity. Regression analysis of the effect of policy change on total field release activity over 15 years, measured as permits plus notifications, suggests this, as the model is highly significant and explains a good proportion of the variance (adjusted R2=0.955), while not showing autocorrelation (Durbin–Watson=2.070). Analysis of the parameter coefficients suggests that all variables are significant at the 0.10 level and have a positive effect. Specifically, the year variable is highly significant and shows an increase in, on average, 52 field releases a year since the APHIS program was put in place. The two regulatory changes also had a significant, though lesser, effect on the amount of field releases with the 1993 policy change accounting for an additional 229 field releases per year and the 1997 policy change leading to an added 190 field releases a year (Table 1). Table 1. Field releases of genetically engineered plants Further analysis of trends in field release through consideration of the utilization of the permitting track bolsters the contention of the notification track replacing the permitting approach. While the model, which incorporates both regulatory changes, does not meet model statistics, removing the effect of the first regulatory policy change in 1993 leads to the model reaching statistical significance at the 0.10 level, although it only explains a fraction of what can be expected of a time series regression analysis and there is suggestion of autocorrelation, meaning the model does not have the correct variables for specification. However, what can be gleaned from the model is an increase of about eight permits a year. The change to regulatory policy in 1997 led to a decrease in permitting activity by an order of about 96 a year, suggesting that other factors are at work. Finally, the 1993 APHIS policy change put in place a petition process that allowed for the determination that certain plants are no longer regulated articles. Furthermore, an extension process, whereby closely related plants are ascribed a nonregulated status, was put in place ([13] p. 104). Once it has been decided by APHIS that a transgenic plant has nonregulated status, APHIS cannot exercise additional oversight over the plant and its descendants, even if separate deregulated lines are crossbred conventionally. This might lead to wild species with potential weediness problems ( [13] p. 111–112). Change during the period from 1986 to 2002 can be effectively seen as occurring within the agricultural biotechnology subsystem with minimal public input. Specifically, changes in 1992 and 1997 to USDA-APHIS field release regulations, while spurred by OSTP directives, did not incorporate public input to any great extent. The lack of negative events, along with increased knowledge and experience with rapidly advancing and diversifying GE plant field testing, precipitated the easing of regulations. Furthermore, the lack of public input and likely recalcitrance to allow deregulated field experimentation of an uncertain technology certainly accelerated this trend towards relaxed regulations. 2.3. In the public eye: on the precipice of changes to the regulatory framework 2002—?? The most recent changes to the regulatory structure concerning agricultural biotechnology are coming about, due in great part to concerns "that the expansion in agricultural biotechnology increasingly will put pressure on seed production and commodity handling systems" ([14] p. 50578) to segregate and control its products. Further, the concomitant diversification of GE plants with agronomic properties, consumption traits, and industrial production qualities that may enter into the environment have stirred doubts as to their safety. Specifically, concerns over the current regulatory scheme, with its relatively insulated policy-making approach, have been raised by three separate events at the turn of the century that have called into question the scientific basis for regulation, the effectiveness of regulatory enforcement, and the integrity of the food system. , 2 The first of these focusing events occurred in 1999 when a laboratory study published by Losey et al. (1999) [15] in the eminent peer-reviewed scientific journal Nature called into question the environmental safety of Bt, which was engineered to express a protein that kills targeted insects that attack economically important crops such as corn, cotton, and potato by eating through their guts, leading to sepsis and the inability to digest food. This article suggested that the monarch butterfly, a highly visible symbol of the environment, as well as other beneficial insects, would be harmed by Bt corn pollen while in their larva stage. While technically correct and seen by those in the industry as acceptable collateral damage due to its having a negligible effect on these butterflies, this study led to a debate and follow-up studies that lasted for over 2 years and drew a good deal of media coverage ([16] p. 189–192). Additionally, it pointed out potential flaws in the Coordinated Framework, as the effect of pollen that expressed Bt was not considered until after Bt corn was in the field. Specifically, the Bt corn in question moved through the APHIS field release regulatory process, which only considers the likelihood that a plant will become a plant pest and only indirectly considered potential harms to nontarget species, without consideration of potential harms to such species as monarch butterflies. EPA regulations, inasmuch as they deal with plant-incorporated protectants (PIPs), 3 such as Bt crops under FIFRA, does have regulatory authority if the pesticidal substance (the crops with PIPs) harms nontarget species. While regulatory action was not taken, the result was that by the 2001 field season, Ciba Seeds (Novartis), the company producing the type of Bt corn most toxic to the monarch butterfly, removed that particular Bt corn from the market in spite of it being "a significant market force during 1996–1999" ([13] p. 72–75). The second controversy garnering national attention and concern likewise dealt with Bt corn. A variant of Bt, which is expressed in Starlink corn and not deemed fit for human consumption due to potential human allergic reactions but seen as safe for use as animal feed, found its way into the human food supply. The public interest group "Genetically Engineered Foods Alert" performed tests on taco shells and other corn-based products being sold in grocery stores, like Safeway, and in fast food restaurants, such as Taco Bell, and found that these products contained Starlink Bt corn [16]. Indeed, within a single year, of 110,000 grain tests by Federal inspectors, Starlink corn showed up in one tenth [1]. The resulting uproar led to actions by EPA to cancel the registration of this corn in spite of Starlink's parent company Aventis attempting to win approval based on its safety as Generally Recognizable As Safe (GRAS) from FDA. However, when this was discarded as an option, Aventis and USDA bought back existing grain supplies and recalled food with Starlink corn in it. Further, EPA no longer allows "split" registrations in which PIPs may be registered for animal feed but not for human consumption. As a result of this, public attention was drawn to flaws in the regulatory system, especially the ease in which food security may be breached, and Congressional hearings were held to discuss this and other concerns with agricultural biotechnology [16]. The final focusing event, that of Prodigene's PMP corn, has likewise led to public concern over the safety of the food supply. In September and October of 2002, in Iowa and Nebraska, respectively, APHIS found "volunteer" corn plants genetically engineered to produce a pharmaceutical to prevent "traveler's diarrhea" growing in soybean fields in violation of permit conditions. Specifically, Prodigene did not abide by the conditions of their field release of PMPs from the previous year as small quantities of this corn ended up in soybean that was to be processed and sold for human consumption. As a result of this, Prodigene had to pay a civil penalty of US$250,000, destroy 500,000 bushels or $2.7 million dollars worth of soybean in Nebraska, and incinerate 155 acres of corn in Iowa due to concern that cross-pollination occurred, as well as post a US$1-million-dollar bond and accede to higher compliance standards for future field tests [17]. Further, and perhaps more important in terms of long-term political implications, the Grocery Manufacturers of America (GMA) and other food processing interest groups expressed concern over plant made pharmaceutical field test regulations, with John R. Cady, CEO of the National Food Processors Association commenting, "nothing short of alarming to know that at the earliest stages of development of crops for PMPs, the most basic preventative measures were not faithfully observed. This apparent violation of rules…very nearly placed the integrity of the food supply in jeopardy." [18]. As a result of these focusing events, especially the Prodigene fiasco, a certain degree of uncertainty over the shape of the federal regulatory system was experienced,, 4 with a concomitant drop in permit activity (see Fig. 1) and experimentation with PMPs (see Fig. 2). To address the decreasing trust in the regulatory structure, OSTP published "Proposed Federal Actions To Update Field Test Requirements for Biotechnology Derived Plants and To Establish Early Food Safety Assessments for New Proteins Produced by Such Plants" in August 2002. Specifically, the notice was published to provide guidance to USDA, EPA, and FDA to update field-testing requirements for food and feed crop plants and establish early food safety assessments for new plant proteins, most specifically PMPs and PMIPs, in line with the 1986 Coordinated Framework. (13K) Fig. 2. Industrial use GE Plants and PMPs. According to the document, three principles are relied upon in updating the Coordinated Framework. First, the level of field test confinement should be consistent with the level of environmental, human, and animal health risk associated with the introduced proteins and trait(s). Second, if a trait or protein presents an unacceptable or undetermined risk, field test confinement requirements would be rigorous to restrict outcrossing or commingling of seed. Further, the occurrence of these genes or gene products from these field tests would be prohibited in commercial seed, commodities, and processed food and feed. Finally, even if these traits or proteins do not present a health or environmental risk, field test requirements should still minimize the occurrence of outcrossing and commingling of seed, although low levels of genes and gene products could be found acceptable based upon meeting applicable regulatory standards ([14] p. 50 579). In light of concerns raised by increased experimentation with PMPs and plants expressing industrial compounds and addressed by OSTP in their notice [14], USDA-APHIS changed rules concerning their field testing of PMPs in March 2003 [19]. The amount of comments in response to this Federal Register notice reflects the changing salience concerning the field release of GE plants. While the changes to the APHIS regulations in 1993 garnered 84 comments and the even more wide-ranging changes in 1997 attracted only 50 comments ( [13] p. 104–105), the Federal Register notice concerning PMP field-testing requirements attracted at least 847 comments (of which 77 were late), many of them from concerned citizens. A high percentage of comments were sent by individuals not commonly associated with the agricultural biotechnology debate, when compared with comments to the previous two Federal Register notices. While critiques were raised in many comments by those who appeared to have ties with the organic movement or with environmental groups such as Greenpeace, as evidenced by the large number of comments received via email, concerns were raised by other politically powerful groups. GMA and affiliated groups expressed concern over uncontained field releases of PMPs and PMIPs, especially in food and feed plants, which account for 75% of all field releases under APHIS notification and permit regulations. Interestingly enough, while support for a total ban on PMPs was expressed by a small number of individuals, concern by consumer groups and traditional biotechnology opponents was tempered, likely mitigated by the potential for medical benefits from this new technology. While the resulting regulations are expected to be modified further over the coming years, they currently incorporate significant changes in how PMPs and PMIPs are regulated [20]. Specifically, for all plants genetically engineered to produce pharmaceutical and/or industrial compounds and field-tested under permit, APHIS established seven conditions that can be grouped into three categories. The first considers field test siting, the second considers the dedication of equipment and facilities to their production, and the third considers procedural matters. Field test siting regulations proposed by APHIS provide two conditions to be met, with special consideration for pharmaceutical corn. First, the perimeter fallow zone will be increased from 25 to 50 ft to prevent inadvertent commingling with plants to be used for food or feed. Second, production of food and feed plants at the field test site and perimeter fallow zone will be restricted for the following season to prevent inadvertent harvesting. Furthermore, specific permit conditions for pharmaceutical corn have been instituted, likely due to corn being the organism of choice, accounting for three quarters of PMP field releases [1]. The large percentage of experiments with corn derives from a variety of reasons, including farmer experience and expertise with raising it, the ideal storage nature of its seeds, the large amount of scientific knowledge concerning its genetics, and the ease in which its genetics are transferred [1]. The first permit condition requires no corn grown within 1 mile of the test site during any field tests involving open pollinated corn—an eightfold increase from standards for foundation seed. When pollen flow is controlled by bagging, the spatial buffer is reduced to 1/2 mile, and a temporal buffer is established with pharmaceutical corn not to be planted less than 28 days before or 28 days after corn grown in the zone from the 1/2- to 1-mile boundary. With the establishment of these buffers, whether they are 1/2 or 1 mile out, border rows will not be allowed to reduce the isolation distance. A second theme concerns the dedication of farm equipment and facilities to the production of such crops. First, APHIS requires planters and harvesters to be dedicated to the test site for the duration of the tests, and although tractors and tillage attachments do not have to be dedicated, they have to be cleaned in accordance with APHIS protocols. The equipment and regulated articles must be stored in dedicated facilities for the field tests duration. The final three requirements from the proposed rules concern procedural aspects of dealing with field tests of PMPs and plants producing industrial compounds. First, APHIS requires the submission of cleaning procedures to minimize risk of seed movement. Second, procedures for seed cleaning and drying are required to be submitted and approved to confine plant material and minimize risk of seed loss or spillage. Finally, permittees will be required to implement an APHIS-approved training program to successfully comply with the stated permit conditions [19]. A key factor in any regulatory arrangement is the ability to ensure that those regulated are complying with the requirements set forth. As a result of the potentially contentious nature of PMPs and PMIPs, APHIS plans to increase the number of field site inspections "to correspond with critical times relevant to the confinement measures." ([19] p. 11338) Therefore, in addition to maintaining records of activities related to meeting permitting conditions and increasing the likelihood of auditing them to verify that required permit conditions were met, APHIS might inspect permitted field tests up to five times during the growing season—once at preplanting to evaluate the site location, once at the planting stage to verify site coordinates and adequate cleaning of planting equipment, at midseason to verify reproduction isolation protocols and distances, at harvest to verify cleaning of equipment and their appropriate storage, and again at postharvest to verify cleanup of the field site. In addition, two postharvest inspections may occur to verify that the regulated articles do not persist in the environment. Finally, APHIS may inspect more frequently if deemed necessary. ( [19] p. 11338–11339). Possibly due to the number of responses received as a result of the Federal Register request for comments concerning APHIS changing their PMP field release regulations and/or the vehemence of concern voiced by those participating in the process, the potential for both PMPs and PMIPs entering into the food supply were cited as points of concern. As a result, and using the PMP regulatory changes as a starting point, APHIS took immediate action to remove the notification track option, requiring complete permit track review in their recent (August 6, 2003) interim rule and request for comments. As stated in the Federal Register notice, "…we believe it is prudent and necessary to remove the notification option for all industrials pending the completion of our ongoing review of part 340." ([21] p. 46435). The rationale given in the interim rule and request for comments was that while 14 field releases (nine notifications, five permits) have been carried out to date, the type of genetic engineering being carried out was to enhance such nutritional components as oil content. However, recent genetic modifications have been for "nonfood traits with which APHIS has little regulatory experience or scientific familiarity." ([21] p. 46434) As such, the definition of PMIPs has three criteria: (1) the plants produce compounds new to the plant; (2) this compound has not normally been used in food or feed; and (3) the compound is being expressed for nonfood/feed purposes ( [20] p. 46435). An administrative reorganization of how USDA-APHIS regulates biotechnology recently created the Biotechnology Regulatory Services (BRS). This reorganization can be seen as another move to address concerns raised by PMPs and PMIPs specifically and GE organisms generally. According to USDA-APHIS, "Given the growing scope and complexity of biotechnology, now more than ever, APHIS recognizes the need for more safeguards and greater transparency of the regulatory process to ensure that all those involved in the field testing of GE crops understand and adhere to the regulations set forth by BRS." Changes instituted by BRS include new training for APHIS inspectors in auditing and inspections of field trials, the use of new technologies such as global positioning systems, and analysis of historical trends to inform monitoring and inspection. According to APHIS, there are six overarching goals that the changes will serve with nine key components being (1) enhanced and increased inspections in which risk-based criteria, along with other factors, will be used to assess field test sites, with higher-risk sites being inspected at least once a year and other sites being randomly selected for yearly inspections; (2) auditing and verification of records of businesses and organizations to verify accuracy and implementation; (3) remedial measures to protect "agriculture, the food supply, and the environment in the event of compliance infraction" with the establishment of a "first-responder" group to deal with serious infractions; (4) standardized infraction resolution in which criteria will be established to determine the extent of an infraction and the response, whether this be further investigation, the issuance of a guidance letter, the issuance of a written warning, or referral to APHIS' Investigative and Enforcement Services (IES) unit for further action; (5) documentation, in which a database will be set up to track field test inspections and resulting compliance infractions; transparency to keep stakeholders and the public informed on the regulatory decision-making process; (6) continuous process improvements, where as the science of biotechnology advances, regulations and permit conditions to allow safe field testing will also do so; (7) an emergency response protocol, being developed with input from EPA and FDA, in which a quick response plan will be put in place "to counteract potential impacts on agriculture, the food supply, and the environment"; (8) training for field test inspectors in their dealings with PMP and PMIP field test sites, as well as the latest in auditing; and (9) certification concerning compliance with the highest level of auditing standards [22]. Although the reorganization can be seen as streamlining and focusing enforcement efforts, the potential for unduly high levels of workload stresses placed on this 26-member unit can be foreseen. First, BRS draws on APHIS inspectors to inspect field tests; however, more than 2600 of these agriculture quarantine inspectors have been transferred to the Department of Homeland Security (DHS) [20]. The current agreement between USDA-APHIS and DHS allows for continued access by APHIS and BRS, although it can be expected that problems might occur as a result of split responsibilities and duties. 3. Conclusions The awareness of the potential for agricultural biotechnology to transform the landscape of American farming through the development of economically important new products, including PMPs and PMIPs, has long been recognized. Just less than 10 years ago, this journal devoted a special issue to "Biotechnology and the Future of Agriculture and Natural Resources" [23]. Then, uncertainty over the future of agricultural biotechnology was based upon the lack of financial support for research and development as well as vague and unfocused regulations [24]. These same concerns exist now in spite of better characterized biotechnology-based science and technology and a better understanding of economic and ecological risks and benefits. The concerns over the new agricultural biotechnology are often termed as one in which the issue is less about the science of GE crops and more about the social issues in which this technology is nested. This "surrogate for safety" is a reflection on the idea that "in many areas of life there is less and less control. For some segments food offers some control." [25]. The threat of drugs and medicines, as well as a variety of industrial compounds, entering the food supply through normal production channels can be seen as particularly dreaded by the American public, which, while largely unaware of the extent of genetically modified products in their food supply, have been attenuated to threats to their security since 9-11. In spite of the lack of evidence of human disability through consumption of GE foods, concern has increasingly been raised in the European Union, which is establishing labeling standards, and Africa, where GE corn destined for famine relief was turned down due to health and ecological concerns. While the history of field release of genetically modified plants had been one of technical domination by insiders, with regulatory change largely ignored by the general public, recent events involving threats to monarch butterflies by Bt corn, potentially allergenic Starlink Bt corn meant solely for animal feed entering the U.S. food supply, and PMPs produced by Prodigene nearly entering the American food system have alerted the American public to potential threats, rupturing the previously insular policy subsystem. While these events provide evidence that the regulatory system is being successfully implemented, their occurrence has drawn attention to gaps in the Coordinated Framework. At least two recent events have the potential to further expand the scope of concern and thus conflict. A report by the Center for Science in the Public Interest (CSPI) called into question the enforcement of guidelines set by EPA requiring growers using Bt corn to set aside land as refuge for pest management purposes [26]. Here, corporations have been called upon to regulate farmers directly due to the use of preexisting pesticide regulations under the Coordinated Framework—a task for which they are not well suited [27]. And most recently, on November 12, 2003, a coalition of environmental groups and consumer advocates sued USDA in federal court to stop the field testing of PMPs due to lack of risk assessment concerning other crops, wildlife, and humans [28]. In light of these concerns and reflected in the rapidly changing field release regulations of PMPs and PMIPs put forward for comment in the Federal Register in March and August of 2003, there is a high likelihood that the Coordinated Framework for the Regulation of Biotechnology will continue to change. Whether this change will occur in the form of marginal alterations in the regulatory approach by EPA, FDA, and USDA, especially in the case of the latter with the newly constituted APHIS-BRS, while retaining the Coordinated Framework, or a major change in the regulations through the creation of a new agency or approach, remains to be seen. As more becomes known about this still young technology and its potential for health, ecotoxicological and ecological effects, as well as the complex and nonlinear environment it operates in, the more likely negative side effects will be discovered and dealt with. Already, both USDA-APHIS and EPA are strengthening their ties with each other with monthly coordinated phone calls and are enhancing transparency and ties with stakeholders through public workshops and meetings. Additionally, greater attention is being given to different means of approaching ecological control of these products, in light of a newly released National Academy of Sciences report on the biological confinement of GE organisms [29]. Regardless, new agricultural–environmental biotechnologies stand on a precipice of change. Over the next 15 years, they may continue to change how food, drugs, and industrial products are produced, or they may be yet another failed technology along the lines of nuclear power with its plants withdrawn from farmers' fields, depending on how issues dealing with public trust in regulations are addressed. In either case, it is social support for the technology and trust in regulatory institutions that matter most. Approaches to the regulation and safety assessment of genetically modified (GM) crops have been developed in a very proactive manner. The first international and national provisions for the safety assessment and regulation of genetically modified organisms (GMOs), including GM crops and derived foods were drawn up by scientific experts in the mid-1980s (OECD, 1986 and US OSTP, 1986). This was nearly a decade before the first regulatory approval of a genetically modified crop in 1995. Since then, the global area of commercial cultivation of such crops has risen to 58.7 million hectares in 2002 ( James, 2002). Commercially cultivated GM crops include soybean, maize, cotton, canola, potatoes, and tomatoes. At present, the most widely grown GM crops contain new genes that confer herbicide tolerance or insect resistance. Other crops are being developed that have improved nutritional characteristics for their food or feed use; GM soybeans and oil seed rape with altered fatty acid profiles, for example, have already undergone regulatory review. Future advances in genomic sciences promise the discovery of new genes conferring desirable characteristics to crops that may fundamentally alter a crop's metabolic functions, promising further nutritional enhancement and resistance to abiotic stresses. It is important that we should continue to proactively assess whether current approaches to safety assessment are appropriate also for future GM crop products with more complex traits