The Green Revolution taught us that technological innovation - higher yielding seeds and the inputs required to make them grow - can bring enormous benefits to poor people through enhanced efficiency, higher incomes and lower food prices. Modern varieties of wheat, rice and maize were made available to millions of poor farmers in the developing world during the Green Revolution, first in Asia and Latin America but also in Africa, although later and to a lesser degree. By raising agricultural productivity, the Green Revolution lifted farm incomes and reduced food prices, making food more affordable for the poor. This virtuous cycle of rising productivity, improving living standards and sustainable economic growth has lifted millions of people out of poverty (Evenson and Gollin, 2003).

Critics of the Green Revolution claim that larger farmers in favourable agro-ecological zones benefited most from the new technologies, and that the resulting intensification of agriculture caused irreparable environmental harm. Several recent reviews have surveyed hundreds of Green Revolution economic impact studies conducted over the last 30 years in all parts of the world (Alston et al., 2000; Evenson and Gollin, 2003). Although these studies were carried out using a variety of different methods, they showed considerable consistency. The average social rate of return to public investments in agricultural research is in the region of 40-50 percent, supporting the assertion that agricultural research is a very good investment. Studies of the economy-wide impacts of Green Revolution technologies have shown that crop productivity gains quickly translate into higher demand and prices for land, labour and non-agricultural goods and services (Hayami et al., 1978; Hazell and Haggblade, 1993; Delgado et al., 1998; and Fan et al., 1998). Although larger farmers in the more favourable areas benefited first, smaller farmers in less favourable environments and landless labourers benefited through these second-round effects. Finally, perhaps the most important benefit of the Green Revolution for the poor was through shifting the food supply function, increasing output and reducing real food prices, because as argued by Alston et al. (1995), "only the poor go hungry".

The environmental record of the Green Revolution is more controversial. Critics argue that the Green Revolution high-yielding varieties required the use of irrigation and fertilizers to perform, and that they were often sold "bundled" with chemical insecticides that gave rise to pest resistance and pollution. Most scientists concede that insecticides were over-used, especially in the early years of the Green Revolution, but many dispute the environmental evidence regarding irrigation and fertiliser use, and argue that claims of environmental harm ignore the counterfactual problem: What would have happened to the environment without the Green Revolution? The central argument is that the Green Revolution technologies have allowed global cereal production to double since the 1960s on essentially the same amount of land. By reducing the need to expand cultivated area, the Green Revolution saved large tracts of virgin land from the plow (Borlaug, in FAO, 2004). The global population has doubled since the beginning of the Green Revolution but the proportion of the world's population that lives with chronic hunger has fallen by half, to 17 percent. Although the number of undernourished people remains stubbornly high, without the yield gains made possible by the Green Revolution, many more people would be hungry and much more land would be under cultivation.

The Gene Revolution offers the technical potential to further enhance the gains made in the Green Revolution. But there are some important differences between the Green and Gene Revolution paradigms that must be considered. The technologies that produced the Green Revolution were produced by public sector researchers at the national and international levels. These technologies were developed and disseminated freely as public goods, based on an explicit strategy for international technology transfer. Most biotechnology research, in contrast, is done by the private multinational sector. The resulting technologies are held under exclusive patents and are distributed on a commercial basis. This paradigm shift has important implications for the kind of research that is being done, the products that are being developed and their accessibility by poor farmers (Pingali and Traxler, 2003).

Intellectual property rights protection - through patents or other means - provide necessary incentives for technology developers and have greatly stimulated the growth of private agricultural research. As a result, however, developing countries are facing increasing transactions costs in access to and use of technologies. Existing public sector international networks for sharing technologies across countries and thereby maximizing spillover benefits are becoming increasingly threatened. The urgent need today is for a system of technology flows which preserves the incentives for private sector innovation while at the same time meeting the needs of poor farmers in the developing world.

The dominance of the private sector in biotechnology research is clear from a comparison of public and private research expenditures. The world's top ten transnational bioscience corporations spend about US$3 billion per year on agricultural biotechnology research and development. The CGIAR system has a total crop improvement budget of one-tenth that amount - about $300 million - and only about onetenth of that is devoted to biotechnology. Among developing countries, the three largest national agricultural research programmes (Brazil, China and India) have total budgets of less than $500 million each, of which about 5 to10 percent goes to biotechnology research (Byerlee and Fischer, 2001). China is the only developing country that has developed transgenic crop technologies independently of the international private sector. India and Brazil might develop this capacity, but very few other developing countries will be able to. Some of the CGIAR centres are working with national research systems and the private sector to develop transgenic crops for developing countries, but these programmes are small and poorly funded.

Private sector biotechnology research is naturally focused on developing technologies suitable for the major commercial agricultural input markets in the temperate-zone production environments of North America and Europe. Some farmers in developing countries have been able to take advantage of "spillover" benefits from private sector research aimed at farmers in the developed world. These farmers are located primarily in temperate production zones in South America, South Africa and China. Transgenic crop research and development is being carried out on more than 40 crops world-wide and dozens of innovations are being studied by both the public sector and private sector. But there is clear evidence that the problems of the poor are being neglected. Barring a few initiatives here and there, there are no major public sector or private sector programmes to tackle the critical problems of the poor or targeting crops and animals that they rely on. This includes the crops that provide the bulk of their food supply and livelihoods rice and wheat - but also a variety of "orphan crops" like sorghum, pearl millet, pigeon pea, chickpea and groundnut that are largely neglected in conventional or biotechnology research programmes. Traits of particular interest to the poor include resistance to production stresses like drought, salinity, disease and pests, as well as nutritional enhancement. Concerted international efforts, including public/private partnerships, are required to ensure that the technology needs of the poor are addressed and that barriers to access are overcome.

A further difference between the Green Revolution and the Gene Revolution lies in the health and safety concerns and the regulatory requirements surrounding GMOs. Many developing countries lack the necessary capacity to formulate and implement their own regulatory procedures. The State of Food and Agriculture reviews the scientific concerns and evidence associated with transgenic crops and summarizes the international scientific consensus, where it exists.

The transgenic crops that are currently being grown commercially and the foods derived from them have been evaluated by the national food safety authorities of several countries using procedures consistent with internationally agreed principles (ICSU, 2003). These foods have been judged safe to eat and the methods used to evaluate their safety have been deemed appropriate (FAO/WHO 2000). To date, no verifiable untoward toxic or nutritionally deleterious effects resulting from the consumption of foods derived from GM crops have been discovered anywhere in the world (ICSU, 2003). Scientists acknowledge that little is known about the potential long-term safety effects of foods derived from transgenic crops (or any foods) and they recommend continued monitoring. Scientists agree that foods derived from emerging, more complex genetic transformations may require additional food safety procedures.

The safety of any whole food is very difficult to determine scientifically due to the natural variation in the foods themselves and the complexity and diversity of human diets and metabolism. Thus, the safety of conventional foods is usually established on the basis of a history of safe use. Risk analysis of foods derived from GMs is conducted using principles similar to those used for food additives: i.e. differences from the conventional food are identified and those differences are tested. If no harmful effects are found, the food is deemed to be as safe as its conventional counterpart. The concept of 'substantial equivalence' is the cornerstone of this approach which is endorsed by the FAO/WHO Codex Alimentarius Commission. Substantial equivalence is only one step in the risk analysis process, not the conclusion (i.e. a product is substantially equivalent to its conventional counterpart except for the identified differences which are then tested). Some critics argue that risk analysis based on the concept of substantial equivalence is not sufficient for GM foods because complex, unexpected differences arising from genetic modification may be missed.

The main food safety concerns regarding GM crops and foods derived from them are:

• Allergens and toxins: Gene technology - like traditional breeding - may increase or decrease levels of naturally occurring proteins, toxins or other harmful compounds in foods. Traditionally developed foods are not generally tested for these substances although they often occur naturally and can be affected by traditional breeding. The use of genes from known allergenic sources in transformation experiments is discouraged and if a transformed product is found to pose an increased risk of allergenicity it should be discontinued. The GM foods currently on the market have been tested for increased levels of known allergens and toxins and none have been found (ICSU). Scientists agree that these standard tests should be continuously evaluated and improved and that caution should be exercised when assessing all new foods, including those derived from transgenic crops (ICSU, GM Science Review Panel).
• Antibiotic resistance: antibiotic resistance is a food safety concern because many first generation GM crops were created using antibiotic resistant marker genes. If these genes could be transferred from a food product into the cells of the body or to bacteria in the gastrointestinal tract this could lead to the development of antibiotic resistant strains of bacteria, with adverse health consequences. Although scientists believe the probability of transfer is extremely low (GM Science Review Panel). The use of antibiotic resistance genes has been discouraged by an FAO and WHO expert panel (2000) and other bodies. Researchers have developed methods to eliminate antibiotic resistance markers from genetically engineered plants.
• Other unintentional changes: unintentional changes in food composition can occur during genetic improvement by traditional breeding and/or gene technology. Chemical analysis is used to test GM products for changes in known nutrients and toxicants in a targeted way. Scientists acknowledge that more extensive genetic modifications involving multiple transgenes may increase the likelihood of other unintended effects and may require additional testing (ICSU, GM Science Review Panel).
• Potential health benefits of transgenic foods: Scientists generally agree that genetic engineering can offer direct and indirect health benefits to consumers (ICSU). Direct benefits can come from improving the nutritional quality of foods (e.g. golden rice), reducing the presence of toxic compounds (e.g. cassava with less cyanide) and by reducing allergens in certain foods (e.g. peanuts and wheat). However, there is a need to demonstrate that nutritionally significant levels of vitamins and other nutrients are genetically expressed and nutritionally available in new foods and that there are no unintended effects (ICSU). Indirect health benefits can come from reduced pesticide use, lower occurrence of mycotoxins (caused by insect or disease damage), increased availability of affordable food and the removal of toxic compounds from soil. These direct and indirect benefits need to be better documented (ICSU, GM Science Review Panel).

Transgenic crops were grown on 67.7 million hectares of land in 2003, in a total of 18 countries (James, 2003). Ninety-nine percent of the global area planted in transgenic crops in 2003 was accounted for by just six countries, four crops and two traits. These same crops and traits are the subject of most of the transgenic crop research underway in both developed and developing countries and public and private sectors.

Some transgenic crops, especially insect resistant cotton, are yielding significant economic gains to small farmers as well as important social and environmental benefits through the changing use of agricultural chemicals. Estimated farm-level benefits vary widely from country to country and year to year, depending primarily on pest pressures, seed prices and the availability of effective alternative pest-control measures. Prices for Bt cotton seed tend to be higher than for conventional seed, but are more than compensated by higher effective yields and lower pesticide costs. Reduced chemical pesticide use helps farmers and their families avoid the health and environmental dangers associated with pesticides.

With the exception of China, all transgenic crops commercialized to date have been developed and distributed by private companies. The evidence suggests that, despite fears of corporate control of the sector, farmers and consumers so far are reaping a larger share of the economic benefits of transgenic crops than the companies that develop and market them. The evidence also suggests that small farmers are just as likely as large farmers to benefit from the adoption of transgenic cotton. Farmers in other countries, however, where transgenic cotton is not grown have experienced small economic losses as a result of lower cotton prices.

It must be considered that this economic evidence is based on only two or three years of data for a relatively small number of farmers in just a few countries. The short-term farm-level gains may not be sustained over time as larger numbers of farmers adopt the technologies. More evidence is required to determine what the level and distribution of benefits from transgenic crops will be in the longer run.

Biotechnology - including genetic engineering - can benefit the poor when appropriate innovations are developed and when poor farmers in poor countries have access to them on profitable terms. So far these conditions are only being met in a few developing countries.

Biotechnology should be part of an integrated and comprehensive agricultural research and development programme that gives priority to the problems of the poor. Biotechnology can complement but not substitute for research in other areas like plant breeding, integrated pest and nutrient management and livestock breeding, feeding and disease management systems.

The public sector - developing and developed countries, donors, and the international research centres - should direct more resources to agricultural research, including biotechnology. Public sector research is necessary to address the public goods that the private sector would naturally overlook and to provide competition in technology markets.

Governments should provide incentives, institutions and an enabling environment for public and private sector agricultural biotechnology research, development and deployment. Public-private partnerships and other innovative strategies to mobilize research and technology delivery for the poor should be encouraged.

Regulatory procedures for GMOs should be strengthened and rationalised to ensure that the environment and public health are protected and that the process is transparent, predictable and sciencebased. Appropriate regulation is essential to command the trust of both consumers and producers, but duplicative or obstructionist regulation is costly and should be avoided. GMOs should be evaluated on a case-by-case basis, commensurate with the level of risk.

Capacity building for agricultural research and regulatory issues related to biotechnology should be a priority for the international community. FAO has proposed a major new programme to ensure that developing countries have the knowledge and skills necessary to make their own decisions about the use of biotechnology.

(Table 1, 2, 3, 4, 5, 6)