Publicly Funded Agricultural Research and the Changing Structure of U.S. Agriculture (2002)

The influence of an innovation on the structure of agriculture depends not only on the nature of the innovation but also on who will use it. The characteristics of producers or farm operations influence the degree to which innovations are adopted. The first part of this chapter examines how the heterogeneity of producers and farm operations influences adoption of technology and innovation, and one approach for responding to this heterogeneity is offered. The second part of this chapter examines the way in which extension, the public-sector arm for transferring agricultural research results, can influence adoption—and hence structural change—through its public education and information programs. This section presents evidence demonstrating that what is transferred, to whom, and how it is transferred can have significant distributional effects. Finally, this chapter presents evidence that, at the state and the local level, extension is increasingly acknowledging the importance of and attempting to serve a greater diversity of farmers and other end-users. We highlight here structural and functional characteristics, innovative processes, and collaborative models of a more broadly “engaged” extension that should be investigated with regard to their structural implications.

The literature on the adoption of new technology in agriculture distinguishes various types of technologies and recognizes that heterogeneity

among producers and farm operations affects what is adopted, to what extent, and when. This section discusses the barriers to adoption that can affect producers differentially.

Differences in farm size may influence technology adoption. Figure 3–1, for example, shows the extent to which several technologies were adopted by dairy farms of different sizes. Some innovations, such as management-intensive rotational grazing—intensively grazing a portion of a pasture followed by a rest period to allow the forage to regrow—are used more by smaller farms than by larger farms. Larger farms tended to adopt others, such as total mixed-ration equipment and the use of milking parlors.

One characteristic that can affect the degree to which farms of different sizes adopt various new techniques or technologies is divisibility. The literature distinguishes between bulky and divisible innovations. Bulky innovations—such as tractors, combines, and other farm machinery—require a significant initial investment but reduce variable cost. It makes economic sense for a given farm to purchase a bulky technology only if its scale is above a critical level. There is a general assumption—and there is some supporting evidence (Feder et al., 1985; Marra and Carlson, 1990)—that larger farms tend to buy and adopt bulky innovations early. Small farms also might adopt the innovation if they collaborate and purchase equipment through a cooperative, for example, or if they rent equipment from dealers or obtain custom service from contractors. Homesteaders in the early days of U.S. agriculture demonstrated that smaller farms could benefit from machinery rental and custom services (Cochrane, 1979; Gross et al, 1996). Today, there is widespread use of custom services for harvesting and land preparation (for example, leveling fields using lasers). Some of the scale effects of technologies can be offset by institutional arrangements. Nevertheless, the introduction of bulky innovations affects the structure of agriculture significantly. The per-unit cost for the equipment owner is generally lower than for the renter or for the user of custom service. Those who purchase farm machinery also could have an extra incentive to augment the size of their operations to make full use of new equipment.

Many agricultural innovations are divisible in that they can be divided into small enough units to, in principle, be used on any size operation: chemical innovations (fertilizers, pesticides); biologic innovations (seeds, biologic pest controls); and managerial innovations (new techniques of pruning, modification of timing for some activities). Divisible innovations are ostensibly more scale neutral than are bulky innovations; indeed, in many cases, per-unit gain from the adoption of innovations, such as seed varieties, does not vary with size. However, adoption of divisible innovations can entail a large initial investment,

putting some farmers at a disadvantage (Feder and O’Mara, 1981). The fixed cost can come in several forms. Adoption of some new technologies requires training, so farmers with small farms and farmers who are less educated could be at a disadvantage. Evaluation of new technologies is time consuming and, again, less-educated farmers or individuals with smaller farms could be at a disadvantage. Training farmers can require additional expense. For example, adoption of modern irrigation technologies, such as drip irrigation, can require fixed cost to redesign and modify farm operations, preventing some farmers from adopting the technology.

Just and Zilberman (1988) assessed the distributional effects of introducing new divisible technologies within farms of varying sizes. They separated farmers into four groups with respect to adoption of more profitable but riskier divisible technologies: farmers with the smallest farms who are unable to adopt because they cannot cover the fixed cost of learning and adoption; farmers with small farms who are limited in their capacity to adopt because of credit constraints; owners of mid-sized farms who can fully or almost fully adopt; and owners of large farms who could be partial adopters because of risk considerations. Smaller, nonadopting farms can be worse off in absolute terms if research and development that introduces new technologies leads to reductions in output price. The credit-constrained farmers also can suffer relative to mid-size full adopters. Just and Zilberman’s analysis suggested that those who gain the most from the introduction of divisible technologies are farms large enough to fully adopt the technology but not so large that adoption will be hampered by risk considerations.

The literature emphasizes the effect of differences in land quality and weather on technology adoption (Caswell and Zilberman, 1986; Green, 1995). Drip irrigation expanded California grape and avocado production to the foothills of the central and southern coasts and to sandy soils in Kern County. Center-pivot irrigation significantly expanded corn acreage to the sandy soils of western Nebraska and hillsides in Washington (Lichtenberg, 1989).

The development of such technologies that can benefit new farm owners in low-quality lands might not benefit owners of higher quality lands who do not need to adopt the technologies. In many cases, large farms on marginal land benefit from land-quality-augmenting technology (e.g., regions in the Central Valley in California), whereas small farms on high quality land, such as avocado growers in San Diego County in California, would not profit from the technology. The farms on higher quality lands are often well-established, traditional family farms that have been able to earn higher yields using less-advanced irrigation technologies. Thus, the research resulting in land-quality-

augmenting technology has a significant distributional effect among different farms.

Human capital is another source of heterogeneity that has a significant influence on adoption in the context of rapid economic and technical change. Schultz (1975) describes two dimensions of human capital—working ability and allocative ability. Allocative ability is education level, intellectual skills, and aptitude for learning and assessing new technologies; working ability is physical capacity for labor. Many technologies are management incentives that draw on a farmer’s allocative abilities. Huffman (1974; 2000) has related farmer schooling to decision making and adoption of technology. Wozniak (1993) has demonstrated that managers with more education are more likely to adopt new inputs and contact the extension service for adoption information than are operators with less education. Integrated pest management (IPM), for example, involves designing context-specific pest treatment as opposed to following a prescribed regimen of chemical pesticide application. Weibers (1992) shows that highly skilled farmers are more likely to adopt IPM, and, even after they seek the advice of consultants, educated farmers are likely to spray less pesticide and use the system more effectively.

Sunding and Zilberman (2001) reported that the tendency to adopt modern technology declines with age. McWilliams and Zilberman (1996) found that older growers owned fewer computers. Analysis of data from Tulare County, California showed that operator age level, along with education level and size of the farming operation, significantly influenced the probability of computer ownership (Putler and Zilberman, 1988). Green (1995) also found generational differences in modernization of farming practices, such as drip irrigation and automated irrigation, in California. Older farmers operate with shorter time horizons, so investing time and effort in adopting new innovations might not be practical. Younger farmers who operate with longer planning horizons often make a greater effort to acquire the skills or knowledge they need to adopt new technology. Older farmers might have less education or more limited familiarity with computers than do their younger counterparts. Older farmers, who own and operate a large percentage of farms in the U.S., are unable to take advantage of new technologies that are adopted by younger and more active farmers.

The literature emphasizes that contractual relationships, particularly tenure agreements (e.g., land ownership, rental), have a profound impact on adoption. Much of the research relating tenure arrangements to adoption has been performed in developing countries (Feder et al., 1985). Tenants with short-term contracts are less likely to adopt technologies requiring investment in land and assets that have a longer payoff. On the other hand, tenants might adopt equipment that is not tied to the land. Although farm management companies might not own land, they might own specialized, labor-saving equipment that will help them manage their operations more efficiently. Feder et al. (1985) surveyed a significant body of domestic and international literature and demonstrated that, under traditional tenure arrangements, when landlords are disengaged from farming activities and the contracts are of short duration, the adoption of modern technology is below average.

The previous section described factors that affect the adoption of innovation. These factors and others that characterize underserved groups, including race or ethnicity, are intangible policy parameters for research decision making. For example, there is such a diversity of small and medium-sized farms that it is difficult to generalize the essence of what they share in terms of research needs. It is much easier for a public research system to respond to the needs of underserved populations if it can target concrete production systems that have promise and can be funded readily. For example, in Wisconsin, management-intensive rotational grazing was used by about 22 percent of dairy farms in 1999, most of them small and medium-sized producers with herds of fewer than 100 cows (Buttel et al., 2000; Ostrom and Jackson-Smith, 2000). Hoop structures also have been used as a low-cost, labor-intensive alternative to confinement housing for smaller hog farms (Brumm et al., 1999). Thus, conducting research on rotational grazing or hoop structures might be useful for operators of small to medium-sized farms. If coupled with rigorous needs-assessment techniques to identify the production systems used by underserved populations, targeting nonmainstream or niche types of production systems can sometimes serve as a proxy for reaching underserved populations.

A public policy approach would engage public research organizations and administrators in developing packages of research and development activities that target specific clienteles (including groups defined by scale characteristics and scale-related characteristics, such as race or ethnicity). Within that framework, research administrators would be justified in allocating funds and

resources to serve larger operators, provided the overall package of activities is transparent to the public and is balanced in terms of serving both medium-scale (or typical or representative) farmers and smaller, limited-resource or minority farmers.

Public-sector outreach activities, including extension, should serve a variety of producers, including limited-resource producers, organic producers, direct marketing producers, transition farmers, full- and part-time farmers, and cooperatives. The public sector should continue special efforts to reach out to underserved or minority communities.

Public-sector outreach, including extension, should take a proactive role in assessing the research and development and technology transfer needs of a variety of producers, including underserved and minority groups; designing appropriate strategies, such as applied, on-farm research, for serving those constituencies; and providing production assistance and other appropriate services, such as market development education for differentiated product markets, entrepreneurship education, financial strategies, value-added processing, and identification of opportunities for those working part time in agriculture.

The committee acknowledges that there are publicly funded research, extension, and education programs and projects for specific underserved populations, among them the U.S. Department of Agriculture (USDA) Small Farm Program, the Hispanic-Serving Institutions Educational Grants Program, the Tribal Colleges Research Program, and the 1890 Institutional Capacity Building Program. Minority-serving institutions, including the historically African American 1890s land grant universities¹, the 1994 Native American land-grant universities², and the Hispanic-serving institutions³ have provided important access to underserved groups. Other institutions also are responding in innovative ways (see Box 3–1). However, the committee also submits that the public-sector response to these populations has been less than proactive, initiated

only in response to considerable public pressure and to such litigation as the 1997 class-action lawsuit filed against USDA by African American farmers (Pigford v. Glickman, 1997). The committee encourages an even greater public-sector effort to engage these populations.

The public sector, at both the federal and the state level, should expand its programming focus with minority-serving institutions, which have unique access to underserved groups.

More effective communication with these groups would help research institutions move toward conducting research and extension that are relevant to their circumstances.

The previous section addressed heterogeneous producer characteristics that affect adoption of technology, and it proposed a public policy approach for responding to this heterogeneity. We now turn to a discussion of private- and public-sector technology transfer systems and the influence of public-sector technology transfer on the structure of agriculture. Technology transfer, defined in the broad sense, involves conveying information to the user and has traditionally been in the purview of Cooperative Extension and other areas within the public-sector system. Embedded within that definition, as a result of growth of private sector and patent protection legislation (U.S. Congress, 1980), is the protection of intellectual property. Strengthening of intellectual property rights over the past 20 years has had important effects on the technology transfer

area and, consequently, important distributional consequences. This area is discussed further in Chapter 5.

The private sector is increasingly important relative to the public sector in the delivery of agricultural technology. According to a survey conducted in 2001, more than 25 percent of agribusiness companies have total marketing budgets (including direct marketing, sales literature, farm shows, public relations, advertising, and internet marketing) exceeding $1 million dollars, and expenditures in agribusiness marketing budgets are increasing over 1998 levels (Agri-Marketing Association, 2001). In contrast, real aggregate federal expenditures for public extension have declined over time, from $332 million in 1991 to $280 million in 2000 (CSREES, Office of Extramural Programs).

Other surveys indicate that the private sector (e.g., input suppliers) is a significant source of technologic information. For example, in 1998, 38 percent of farmers using precision agriculture relied on input suppliers for advice compared to 17 percent of farmers who used the extension service as a source of information about the technology (Daberkow and McBride, 2001). In the 1998 Farm and Ranch Survey, 35 percent of farms relied on irrigation equipment dealers for information on water conservation and irrigation cost reduction compared to 41 percent who relied on extension agents or university specialists (USDA, 1999c). The 1991–1993 Area Studies survey data indicate that fertilizer company recommendations exceeded the extension service as a source of information on nitrogen fertilizer application decisions. The survey also shows that a greater percentage of pest management advice was provided by chemical dealers than by local extension for corn, cotton, potatoes, soybeans, and wheat (Caswell et al., 2001). Survey data from the 1993–1995 Chemical User Survey indicated that chemical dealers were also the most-used source of pest management information for a variety of fruit crops (USDA, 2000a).

The private sector contributes to the process of transforming an invention from discovery to application by investing additional resources for validation, manufacture, and distribution. In some cases, private sector groups provide additional technology support outreach in addition to or in concert with the public sector. New institutional and financing arrangements between public and private sectors for technology transfer are discussed later in this chapter. Private-sector vendors logically seek to work with farm operators who can contribute the most to their profits, and so they are more likely to seek out larger farm operations and develop products that make it easier to manage more land or animals with less labor.

The Cooperative Extension Service was established under the Smith-Lever Act in 1914 (U.S. Congress, 1914). Extension receives public support through CSREES and from state and local governments. A major function of the extension service has been to translate information from research for farmers and other citizens through adult education. State extension organizations provide administrative support and subject matter specialists. A local extension agent system with county offices throughout the nation was developed to produce and distribute information on applied problems. The Cooperative Extension Service is accountable to county governments, state governments, land grant universities, and the federal government through CSREES and by myriad grants awarded through other private, state, and federal entities.

Cooperative Extension is an important link between research and the structure of agriculture. Huffman and Miranowski (1981) demonstrated that extension can be a substitute for formal education and human capital in adopting technology. Thus, extension can ensure that the structural implications of new technologies are considered. Extension can affect agricultural structure through what is communicated (or what is not communicated), to whom, and how that information is communicated.

For several decades there has been a substantial literature on the relationship between farm size (and other indicators of farmers’ socioeconomic status) and contact with Cooperative Extension and other “change agents.” The early literature is summarized in Everett M.Rogers’ Diffusion of Innovations (1995). More recently, Huffman and Evenson (2001) showed that public R&D and education (including extension) have been at least as important as private R&D and market forces for changing livestock specialization, farm size, and farmers’ off-farm work participation from 1953–1982 (discussed in Chapter 2). A survey relating herd size and five dimensions of contact with extension among Wisconsin dairy farmers reported that farmer use of extension services was highly variable and appeared to be correlated with the size of the farm (Ostrom et al, 2000).

As shown in Table 3–1, the disparity in extension contact by herd size is smallest for reading extension publications or articles but is very large in the case of the extension agent visiting the operator’s farm⁴. The study also reported that, as herd sizes increased, farm operators were more likely to report that extension programs had been beneficial to their farm business. A disproportionately large share of the dairy farmers who were “unsure” whether they had benefited from extension came from farms with smaller herds. Using data from a 1989 survey of North Carolina farmers, Flora et al. (1993) found that several factors were associated with frequency of contact with extension agents and extension information: race (whites much more than African Americans or Native Americans), gender (men much more than women), size (large much more than small, measured either by acres or by gross sales), and type of agriculture (conventional more than alternative). Although some extension programs, such as Missouri’s Small Farm Family Program, have worked with limited-resource and small farmers, extension’s audience is composed mostly of larger-than-average farm operators.

Some groups traditionally have received little support from extension. African American and Native American farmers have traditionally not fully accessed public-sector support in the area of technology transfer (USDA, 1997a; 1998b). When minority farmers are approached, the method may not be culturally or socioeconomically appropriate.

The next section illustrates that the public-sector technology transfer system is changing through the development of partnerships with the private sector and through increasing engagement among public-sector institutions.

Public-sector technology transfer is increasingly resulting from novel partnership arrangements with the private sector. Public-sector involvement has the potential to ensure that the results of research are publicly accessible. Mechanisms of technology communication are improving because of public sector networking with private (e.g., National Pork Board) and other public institutions. Technology transfer arrangements may include the following:

• Universities invest in development and commercialization. Some universities invest selectively in product development through partnerships with industry, specialists, and farm advisors. These team efforts can be partially financed through private-sector contracts or contributions. University teams have been involved in seed development and commercialization for wheat, cotton, and other agricultural crops. University task forces also have developed IPM and waste disposal strategies.

• Individuals who made discoveries while working in the public sector engage in private development of a product. University or USDA researchers sometimes start private businesses based on work conducted in the public sector. In other cases, public-sector employees serve as consultants to private enterprises. Many entrepreneurs commercialize the knowledge and findings they obtain while working toward graduate degrees. Some consulting companies established by university researchers and graduates offer managerial or agronomic expertise. In other cases, for example in the biologic control business, enterprises provide expertise and a product (beneficial insects).

• Private-sector entrepreneurs and companies invest in development and commercialization of university findings. The classical processes of

technology transfer are still common. Commercial entities develop technology innovations reported by government-supported researchers in the scientific literature, frequently enlisting public-sector researchers as consultants.

• Private companies finance university research in exchange for the right to develop and commercialize the resulting innovation. For example, Iowa State University’s Pig Genome Mapping program is facilitated by private industry through provision of financial and genetic resources.

• Private entities buy the rights to commercialize university patents or varieties. In recent years, the use of formal processes of agricultural technology transfer has increased. For example, the University of California at Davis received more than a million dollars one year for the right to use its strawberry varieties. The chemical industry bought the patent rights for dibromochloropropane (DBCP) from the University of Hawaii to develop nematode control strategies.

• Cooperative Extension forms partnerships with the private sector, particularly for the production of public goods. For example, the Farm Bureau, Trees Forever, chemical companies, and Extension are collaborating to install riparian buffers in Iowa. Nonpublic sources of extension funding, including grants and contracts, often are used in technology transfer.

The structural implications of public-private sector partnerships are an important issue needing further analysis but are beyond the charge of this committee.

A recent report by the Kellogg Commission, Returning to Our Roots: The Engaged Institution (1999), endorses the concept of institutional engagement. The report’s recommendations encourage institutions to go beyond conventional one-way outreach and service and to become “more sympathetically and productively involved with their communities” with a “commitment to sharing and reciprocity”. As extension considers this model, its technology transfer activities should be more effective in reaching a more diverse audience, and that could have implications for the structure of agriculture.

Consistent with the model of engagement, the structure, function, and processes of extension are changing and may provide access to a wider set of expertise within the university community, engage farmers and others in the research process, and facilitate improved accessibility of options.

Extension is increasingly positioned at a higher administrative level than the college of agriculture within many universities. In 1995, extension played a university-wide role outside colleges of agriculture in 44 percent of 186 land grant universities surveyed (Warner et al., 1996). A more recent survey found the same distribution for land grant universities (Luft, 2000). Examples of university-wide positioning can be found at the University of Wisconsin, Oregon State University, and Iowa State University. From this administrative vantage point, extension can access a wider range of university resources and expertise. An institution-wide arrangement has fostered new interdisciplinary research and more holistic outreach that broaden the range of users, expand the utility of research, and change the research agenda.

Extension is increasingly responding to problems that go beyond its traditional focus on agricultural production and farm programs. The private sector has focused on transferring technologies from which value can be captured in the production or sale of seeds, machinery, agrochemicals, and plant and animal nutrients. Extension consistently engages in higher risk endeavors, particularly those that involve alternatives or farm groups considered less profitable for the for-profit sector, such as the production of public goods, from which revenue cannot be captured.

Surveys in Missouri and focus groups in Minnesota have demonstrated the broad set of expectations for extension and needs identified by local stakeholders (Warner et al., 1996). CSREES has operated to connect other federal agencies with its state extension partners (Box 3–2).

A more broadly defined extension service in terms of constituencies and stakeholders can ensure that structural dimensions of research results and technologies are considered.

Extension should continue to reach out to other programs within universities, to draw wider networks of human resources, and to work with broader arrays of partners in the federal, private, nonprofit, and client sectors. CSREES should continue to facilitate more interdisciplinary and interagency activities that involve its state extension partners. CSREES should evaluate the potential and effectiveness of these extension approaches to serve diverse constituents.

Stakeholder participation is growing in extension, and extension workers are changing to be more receptive to farmers’ perspectives. The definition of “stakeholder” is broadening, too, although land grant universities are still somewhat hesitant to broaden the range of stakeholders because of the perceived increased transaction costs and the fear of alienating some client groups (Extension Committee on Organization and Policy, 1996).

Whereas the “demonstration plot” model of Extension generally engaged farmers whose larger operations were considered exemplary, a participatory model is likely to involve a broader variety of farmer circumstances along the research-extension continuum. In California, for example, extension workers are facilitating highly applied, on-farm research. The SARE program (see Box 3–3) demonstrates that extension processes are increasingly engaging stakeholder participation.

Extension is increasingly supporting farmer-to-farmer networking, although that has long been the basis of technology transfer, for example, through field days and demonstration plots. Examples include extension’s support of the group Practical Farmers of Iowa and work with farm stewardship groups in Minnesota and farmer marketing groups in Illinois. Alternative forms of outreach and engagement, including use of the Internet, also are resulting in greater stakeholder participation (See Box 3–4).

The literature on adoption suggests that various producers and farm operations adopt innovations differently. Different degrees of adoption can be signaled by characteristics of producers or farm operations, such as farm size, regional differences in land quality, availability of human capital, producer age, and tenure arrangements. Some research innovations are more likely to be adopted by specific groups of producers, with structural implications. An approach to setting priorities for research, based on needs assessment of a variety of users, is proposed as an avenue for targeting heterogeneous producers and farm operations.

This chapter discussed the structural impacts of extension through the disproportionate support for specific farmer groups. The chapter also contrasted the structural dimensions of the conventional “technology transfer” model of extension with new models characteristic of more engaged institutions. These new models are characterized by increasing collaboration with the private sector, changes in extension’s position within universities, a broadening of the extension mandate through linkages with other federal agencies, and greater stakeholder participation in setting priorities for research and extension activities. Research is needed to analyze the structural effects of these collaborative approaches.