Agricultural Research

Abstract

This article reviews contributions made by agricultural research programmes in historical context. Before 1850, when the agricultural experiment station (AES) model was developed, most crop and livestock improvement was due to farmer selection of seeds and livestock breeding. By 1875, a number of plant breeding programmes were in place. Developed countries achieved a green revolution in the first half of the 20th century, developing countries in the second half. A number of countries are now benefiting from the gene revolution. An assessment of social returns to public spending on agricultural research shows these returns to be high.

Before 1850

The earliest form of agricultural research was agricultural invention. The patent systems in Europe date back to the Statute of Monopolies in 1623 in England. During the 18th century, England and France further developed their patent systems. Article 1, Section B of the US Constitution, drawn up in 1787, states that ‘Congress shall have the power to promote the progress of science and useful arts, by securing for limited times for authors and inventors the exclusive right to their respective writings and discoveries.’ The first Patent Act in the United States was enacted in 1790. Many of the earliest inventions, including Eli Whitney’s cotton gin, were agricultural inventions.

Prior to the development of the modern agricultural experiment station in 1843, the ‘botanic garden’ served as the chief research vehicle for plants. Botanic gardens were established in many countries, preserving and further classifying plants and trees in the tradition of Linnaeus. (Today there are 1,500 botanical gardens worldwide. Of these, 698 have germplasm collections for the conservation of ornamental species, indigenous crop relatives and medicinal and forest species, and 119 conserve germplasm of cultivated species, including landraces – that is, distinct types – and wild food plants.)

Both plant and animal improvement prior to the modern experiment station was achieved by farmers themselves. Prior to the 18th century, farmers selected seed from each crop to improve the productivity of crop species. (There are approximately 300,000 species of higher plants, that is, flowering and cone-bearing plants. Of these, 270,000 have been identified and described. About 30,000 species are edible and about 7,000 have been cultivated or collected by humans for food; 120 species are important cultivated crops, but 90 per cent of the world’s caloric intake is provided by only 30 species.)

As populations moved to new locations and production conditions, they created new landraces in each cultivated species. As new landraces were created, three distinct classes were identified. Landraces created in the centre of origin of cultivation were the first class. For rice, as many as three or four centres of origin (that is, locations of first cultivation) for the two cultivated species Oryza sativa and Oryza glaberrima have been identified. The second class includes landraces created in centers of diffusion (that is, locations where populations diffused the crop). The third class comprised landraces created in the New World countries in the Americas and Oceania.

These landraces were later collected and, along with mutants and uncultivated species in the genus, they constitute the genetic resources used in modern plant breeding programmes based on conventional methods of crossing parental plants. Table 1 summarizes contemporary ex situ genebank collections.

Agricultural Research, Table 1 Genebank collections (ex situ)

Animal improvement actually pre-dates crop improvement. It, too, was achieved by farmers and herdsmen. Most of the breeds of cattle, pigs, poultry, horses, sheep, and so forth were developed in the 16th through 18th centuries. Most were developed in Europe. Work animals, including oxen, horses and water buffalo, were particularly important in agriculture prior to the 20th century, when tractors became the dominant source of power in many countries. Work animals, including the powerful workhorses, important to cultivation, are sensitive to climatic conditions. Animal breeds used in Asia range from the powerful bullocks in North India, weighing more than a ton, to much smaller cattle in the Himalayan mountains.

1850–1900

Agricultural research programmes were changed dramatically with the development of the agricultural experiment station. It is generally accepted that the first truly scientific experiment stations were located in the UK, in the Rothamsted Experiment Station, established in 1843, and in Saxony, where several experiment stations were established in the 1850s.

With the experiment station and its formal structure of experiments with ‘treatments’ and ‘controls’, agricultural research became scientific, and by 1900 agricultural science was established as a mature applied science. The application of statistical methods to experiments furthered this development. R. A. Fisher, the statistician at the Rothamsted Experiment Station in the UK from 1919 to 1933, is credited with numerous methodological developments, many of them relevant to modern-day econometrics. Early experiments focused on agricultural chemistry, including the application of chemical fertilizers and related soil amendments. By 1875 or so, formal plant breeding programmes were beginning to be established. It is often thought that formal plant breeding did not take place until after the ‘rediscovery’ of Gregor Mendel’s work, first published in 1856, in 1900. But that is not the case: breeding programmes in sugar cane, wheat and many other crops were established before 1900. Sugar cane breeders in Java and Barbados simultaneously discovered techniques to induce flowering in sugar cane plants in 1878, and by 1900 the ‘noble’ canes from their breeding programmes were beginning to transform sugar cane production in several countries.

In the United States, the Hatch Act of 1887 provided funds for experiment stations in every state. Most state experiment stations recognized the synergistic relationship between research and graduate teaching, and formally linked experiment stations with land grant college programmes. It is widely thought that legislation such as the Hatch Act reflected exceptional wisdom on the part of legislators. This was not the case. Prior to the Hatch Act, many states had considerable experience with experiment stations. This was also true for the Land Grant College Act – the Morill Act – in 1862. Some 20 states had established colleges of agriculture prior to 1862. As these programmes matured, veterinary medicine colleges were established in land grant colleges. By 1900, sufficient experimental data were available from state agricultural experiment stations to answer many questions of importance to farmers in the US.

1900–1940

The period 1900 to 1940 was a one of extraordinary achievements by agricultural experiment stations. Plant breeding gains were achieved in most crops planted in temperate zone countries (in effect, temperate-zone developed countries realized a green revolution in this period). Plant breeding gains in sugar cane, coffee, tea and spices (the Mother Country crops) were also achieved in tropical regions. Brazil and Argentina in Latin America realized major gains (Brazil became the world’s major producer of coffee and sugar; Argentina the major exporter of beef).

Two major scientific developments in plant breeding were achieved during this period. The first was the development of techniques to produce hybrid crop varieties to take advantage of the ‘heterosis’ effect in crops. The early development of hybrid techniques took place at Harvard and Yale Universities, but the major achievement was made by Donald Jones at the Connecticut Agricultural Experiment Station in New Haven. Jones developed the ‘double cross’ method for seed production. Hybrid seed production requires ‘selfing’ or ‘inbreeding’ for several generations. Prior to Jones, a single cross was made between two inbred lines to produce hybrid seed; the seed cannot be saved by farmers because the heterosis effect is present only in the hybrid generation. Jones used four inbred lines in a double-cross to produce seed more efficiently. Since Connecticut is not a major corn production state, it was several years before hybrid corn was available to farmers in Iowa. Henry A. Wallace, later a vice-president of the US, was an early leader in developing private industry production of hybrid corn. He established the Pioneer Hybrid Seed Company in 1926.

Zvi Griliches (1957) analysed the adoption of hybrid corn by farmers in different US states. Farmers in Alabama had access to hybrid corn varieties 20 years after farmers in Iowa. This was not because hybrids suited to Iowa farmers were not exhaustively evaluated in Alabama. Alabama farmers did not have hybrid varieties until seed companies established breeding programmes in Alabama to develop varieties suited to Alabama production conditions. Corn has a high degree of photo-period sensitivity. Varieties suited to Alabama were also varieties with longer growing seasons. This same principle applies to the green revolution (see below). No country without a functioning plant breeding programme has realized a green revolution.

The second scientific development was another form of hybridization, interspecific hybridization or ‘wide crossing’. Until the gene revolution, based on ‘recombinant DNA’ techniques, all plant breeding entailed a ‘sexual’ cross between two ‘parent’ cultivars (this continues to be the case for achieving continuous plant improvement). Inter-specific hybridization entails a sexual cross between different species, usually members of the same genus. This was first achieved in sugar cane in 1919 when breeders achieved crosses between Saccharum officianaram, the cultivated species, and Saccharum spontaneum, an ornamental species of sugar cane. Later a third species, Saccharum barberie, was added.

By the 1980s, inter-specific hybridization techniques (chiefly embryo rescue techniques) had been developed for most crop species. With these techniques, sexual crosses have been achieved between cultivated species and most or all uncultivated species in the same genus for all important crop species.

During 1900–40, developed country agriculture (and some developing country agriculture) was also being affected by the development of farm machinery and tractor power. Stationary tractors and steam engines were developed before 1900. After 1900 the row crop tractor was developed along with improved harvesting and planting machinery. By the 1930s these developments were changing the structure (farm size, off-farm work) of US agriculture. These developments were produced largely by private sector firms in the farm machinery and farm chemical industries. Patent incentives existed for mechanical, electrical and chemical inventions in this period. They were not developed for genetic inventions until after 1980.

1940–1965

At the end of the Second World War, agricultural research experienced a renaissance in developed countries. This was at least in part because of synergism between public sector agricultural research and private sector R&D in the farm machinery and farm chemical industries. By 1965 supermarkets had crowded out the ‘mom and pop’ grocery stores in most US cities. Poultry production was effectively industrialized by 1965 as confined housing units became the norm. Dairy production was subject to scale economies, and herd size was increasing. Feed management had improved greatly. The widespread use of United States Department of Agriculture grades and standards for livestock was transforming the meat packing industry. By 1965, in all OECD countries total factor productivity growth was faster in the agricultural sector than in the rest of the economy, and this continues to be the case today.

In developing economies, a sense of alarm had been created by the growing recognition that developing countries were in for a population explosion. With improvements in public health measures, death rates, particularly among children, began to decline and life expectancy began to increase. With even modest delays until the birth rate declined, this meant rapid increases in population. The alarm in question centred on food security. Many alarmists of the 1950s, notably Paul Ehrlich (1968), concluded that food production growth could not keep pace with population growth.

The international community (including the World Bank, regional banks, foundations and bilateral aid organizations) responded by developing a system of international agricultural research centers (IARCs). The first two IARCs were the International Rice Research Institute (IRRI) in the Philippines and the International Wheat and Maize Improvement Center (CIMMYT) in Mexico. These two centres were credited with creating a ‘green revolution’ based on high-yielding varieties of rice and wheat introduced to farmers in 1965. Other IARCs, however, contributed to green revolutions in all major food crops.

The Green Revolution: 1965–2004

The period 1965–2004 was truly extraordinary for agriculture. In 1991 the Soviet Union collapsed, leaving the former Soviet republics in severe recession. This included the agricultural sector. Most, but not all, developing countries experienced a green revolution during this period.

Table 2 summarizes the production of green revolution modern varieties (GRMVs) by five-year period. These data show that the production of GRMVs is increasing over time. Thirty-six per cent of all GRMVs were crossed in an IARC programme. Twenty-two per cent of GRMVs crossed in national agricultural research system (NARS) programmes utilized an IARC-crossed parent or other ancestors. Non-government organizations (NGOs) did not produce GRMVs. None were crossed in developed country programmes and transferred to developing countries. Private sector firms did produce hybrid maize, sorghum and millet varieties (five per cent of GRMVs) but only after improved open-pollinated varieties (OPVs) had been produced by IARC programmes. GRMVs were produced in public sector IARC programs and in NARS programmes in developing countries.

Agricultural Research, Table 2 Average annual varietal releases by crop and region, 1965–2000

Table 3 summarizes the economic consequences of the green revolution. Production increases are separated into increases from higher crop area planted and increases from higher yields. Yield increases are further separated into GRMV contributions and other input (fertilizer, labour) contributions. In the early green revolution period, production increased by 3.2 per cent a year. Yield increases account for 2.5 per cent a year. In the late green revolution period, production increased by 2.2 per cent per year. Yield increases accounted for 1.8 per cent per year. The sub-Saharan Africa region was an outlier in both periods, with low modern varieties (MV) contributions. The green revolution for sub-Saharan Africa was not accompanied by increased inputs, as it was in Asia and Latin America. (At least 12 countries – Afghanistan, Angola, Burundi, Central African Republic, Congo (Brazzaville), Gambia, Guinea Bissau, Mauritania, Mongolia, Niger, Somalia and Yemen – did not have a Green Revolution. Most are in sub-Saharan Africa.)

Agricultural Research, Table 3 Economic consequences of the green revolution (growth rates of food production, area, yield, and yield components, by region and period)

The Recombinant DNA (rDNA) Gene Revolution

In 1953 Watson and Crick published work (Watson 1968) that identified the ‘double helix’ structure of DNA and established DNA as the carrier of genetic information. In 1974 Cohen at Stanford and Boyer at the University of California at San Francisco achieved recombinant DNA ‘transformation’ or insertion of ‘alien’ DNA into organisms, and the field of genetic engineering was born (Cohen 1997).

Within a few years many ‘crop biotech’ companies were established. Large agricultural chemical companies were early entries into the field. Today seven life science firms (Monsanto, DuPont, and Dow in the US, Syngenta, BASF, and Bayer in Europe, and Savia in Mexico) dominate the genetically modified (GM) crop products industry. The first GM products introduced in the late 1980s were commercial failures. But bovine somatotrophin hormone (BsT), a product to stimulate milk production, was successfully introduced in 1993.

In 1995 several companies introduced GM crop products for canola (rapeseed), soybeans, maize and cotton. These products fall into two classes: herbicide tolerance and insect resistance (Bacillus thuriengensis, B_T). Herbicide tolerance (soybeans, canola and maize) enables weed control with traditional herbicides. This trait has been highly valued by farmers and rapidly adopted. Most of the world's canola and soybeans now have this trait, as does considerable acreage of maize. Insect resistance is achieved by engineering maize and cotton plants to produce B_T toxins that limit insect damage to the plant. This has a particularly important effect on cotton, where insects cannot readily be controlled by insecticides.

GM crop products enable farmers to reduce production costs. Cost reductions depend on mechanization status and insect pest status. Estimates of cost reduction vary by country, with Western European countries having negligible cost reduction potential (less than one per cent, because they produce little cotton, canola or soybeans). The US has significant cost reduction potential, as do many developing countries. It should be noted, however, that cost reduction gains are ‘static’ in nature (that is, they do not cumulate over time). Dynamic gains can be produced only by the development of generations of modern varieties, as reflected in Table 2 for GRMVs. The gene revolution is not a substitute for the green revolution.

The gene revolution has become strongly politicized in recent years. A clear division has emerged between the original European Union countries and North American countries. The European Union position is that the ‘precautionary principle’ should apply, while the North American position is that, in the absence of scientific evidence to the contrary, farmers should be allowed to adopt GM crops (see FAO 2004).

Returns to Agricultural Research

Griliches (1958) was the first economist to measure ‘returns to research’ by computing returns to hybrid corn research. To do this, he created a cost stream and a benefit stream, and applied present value methods to them. (At a five per cent discount rate the present value of benefits was roughly seven times the present value of costs. Some interpreted this as a 700 per cent rate of return. Of course, it was in fact a benefit–cost ratio.) Griliches computed an internal rate of return to hybrid corn research of 43 per cent.

Evenson (2001) reviewed more than 300 studies of returns to research in the decades after the Griliches studies. Table 4 reports a summary of internal rates of return reported in these studies. The project evaluation studies utilized methods similar to those used by Griliches. The statistical studies generally regressed measures of total factor productivity on research stock variables. Some studies were focused on specific commodities, others on aggregate research programmes. Several studies made a distinction between pre-invention science and applied science, and several studies were undertaken of the private sector contribution to agriculture.

Agricultural Research, Table 4 Returns to agricultural research studies

The studies are characterized by great diversity in internal rates of return (IRRs), ranging from IRRs of zero to very high levels. Median IRRs are high for all categories. This diversity is consistent with the fact that research is a highly uncertain activity.

Finally, Table 5 utilizes data from the green revolution where GRMV adoption rates were available. The method applied was similar to that which Griliches originally used. These data confirm the estimates in Table 4. Very high returns to IARC research are shown. Returns to NARS programmes are lower, especially in sub-Saharan Africa where many countries did not achieve a green revolution.

Agricultural Research, Table 5 Green revolution returns to research

See Also

• Agriculture and Economic Development

• Population and Agricultural Growth

• Technology