number of centers increased to twelve, and the funding per center increased. This led to a tenfold increase in nominal spending to US$141 million in 1980. During the 1980s, spending continued to grow, more than doubling in nominal terms to reach US$305 million in 1990. The rate of growth had slowed but was still substantial. In the 1990s, however, although the number of centers still grew, funding did not grow enough to maintain the level of spending per center and growth rates declined. Since 2000, funding has grown in total but with a continuing trend toward earmarked sup­port for specific projects and programs of research involving multiple centers and other research providers outside the CGiAR. In 1980, the share of the four founding centers of total CG spending was 54%, but by 2004 it had slipped to only 36% (Pardey et al., 2006a).

8.1.2 Determinants of public and private R&D investments
A conceptual model of the factors that influence these in­vestments is needed to make a critical assessment of research investments trends.

8.1.2.1 Determinants of public research
In the absence of public intervention, private firms will un-der-invest in research when the output of that research has the characteristics of a public good—that is, the outputs of research are often non-rival and non-excludable. Because of the public goods nature of research, the social benefits are much higher than the private benefits, and hence the justification for public intervention. While many public in­vestments have high social benefits, public investment will only be justified if the return is higher than other forms of public investment. A review of the RORs to research (see 8.2) shows that public investments do have high payoffs, of­ten 40 to 50% or more. Considering that private companies and governments usually can obtain credit at interest rates below 10% and the public RORs on other types of govern­ment investments are considerably lower than 40%, these returns are very high.
      Whereas studies that show high social RORs to research investments may convince economists that agricultural re­search is a good investment, most policy makers who actu­ally do not appear to have been sufficiently convinced that these high social RORs warrant large investments. Rather, public investments in agricultural research respond to many of the same forces that influence the amount and direction of private research (Hayami and Ruttan, 1985).
      Public agricultural R&D increases when there are ad­vances in basic knowledge and technology in fields such as biology, chemistry, engineering, and information technol­ogy that increase the possibility of an innovation or reduce the cost of developing an innovation. This is referred to as an increase in technological opportunity. The discovery of dwarfing genes in rice and wheat created the opportunity for plant breeders around the world to produce many new types of varieties that would respond to higher doses of nu­trients and water. These opportunities increased the poten­tial return to research and led to major increases in public sector plant breeding research around the world. Likewise, the tools of biotechnology have created a major shift in the innovation possibility curve for plant and animal breeding,

pest control, and abiotic stress tolerance, and many govern­ments have again responded by increasing their investments in research (Box 8-1).
         Changes in the demand for agricultural products by farmers and consumers induced public R&D investments (Hayami and Ruttan, 1985). Historically in Asia, popula­tion and per capita income increases increased the demand for basic food grains such as rice; farmers were unable to increase production rapidly due to limited land and agricul­tural prices increased. Private firms did not attempt research to fulfill this demand because profits were projected to be insufficient. When farmers and consumers were sufficiently well organized and demanded a solution, Asian governments invested in agricultural research. For example, following World War I Japan had very high rice prices and consumers demanded cheaper food, but Japanese farmers had no land for expansion. The government responded by investing in research that eventually led to nutrient-responsive rice va­rieties; this resulted in biological technologies and inexpen­sive fertilizers increased yields per unit of land. In the late 1950s, national governments, nonprofit foundations, and aid donors responded in a similar manner to the food crisis and high food prices caused by rapid population growth in other Asian countries and invested in the international agricultural research centers and the national agricultural research systems.
         Demand for solutions to specific problems, such as a new disease or pest or the shortage of a key input (such as the aforementioned land shortages in Asia and resulting de­velopment of land-saving technologies), also lead to public sector research investments and can direct the allocation of investments. The worldwide public sector response to Avian Influenza is a current example. There are also demands that receive insufficient investments, for example, research on diseases such as malaria or investment in appropriate ag­ricultural technologies for poor people. Whether these fac­tors will actually lead to more or less R&D investments by governments depends on the structure of the government, its ability to raise money, and the power of various interest groups to influence government spending decisions.
          Some governments are more committed to R&D as a major tool for economically sustainable development. They will put a larger share of their budget into research of all types including agriculture (Anderson, 1994). The structure of the research system will also influence the size and direc­tion of agricultural research (Morris and Ekasingh, 2002). Some governments have structured their R&D system to be more responsive to the demands of the agricultural sector while others are more responsive to demands of the food consumers, agricultural scientists, or foreign aid donors. The size and power of different interest groups can also have a major impact on the size and direction of agricultural R&D. Commercial farmers can push their governments for large investments in research that is likely to concentrate on reducing costs crop production or increasing demand. If the textile industry is strong, research will be focused on bring­ing down the cost of cotton. Strong consumer lobbies are likely to lead to research that lower food prices. In countries where private research on topics such as maize breeding or poultry breeding become strong, the companies that are do­ing this research will lobby for government to stop compet-