Bumpy Road to Adoption of Precision Agriculture
November 18, 1997
Jess Lowenberg-DeBoer, Extension Agricultural Economist
Technological change is messy and disruptive. We often must change the way we think about things, as well as the way we do them. Frequently, techno-logical change has unexpected and unintended consequences that cause chain reactions throughout a production and marketing system. Technology does not suddenly appear from the laboratory or workshop fully formed and perfectly operational. It usually requires a period of adaption, with farm tinkerers, manufacturers and scientists all doing their part to make the technology both profitable and practical.
In the middle of this flux, farmers and agribusinesses must make technology choices. They can not wait until the dust is settled and the technology matures because that is too late. Most of the profits from any new technology go to the early adopter. The purpose of this paper is to identify some patterns of techno-logical change that will help farmers and agribusinesses make strategic decisions about one of the major technological changes in the 1990s, precision agriculture. The approach will be to look at what can be learned from previous technology, in particular the motorized mechanization of agriculture in the early twentieth century and the development of hybrid corn.
Hybrid corn came close to following the idealized S shaped adoption curve (Figure1). When hybrid corn became widely available in the Midwest in the early 1930s a few innovators tried it and it was widely publicized in yield trials and the farm press. As the benefits of planting hybrid seed became well known large number of farmers adopted it. In Indiana this rapid adoption phase occurred in the late 1930s and early 1940s. By about 1950 most farmers who would use hybrid seed had adopted the practice.
The key question is why did many US farmers waited until the late 1930s to adopt hybrid seed. The possibility of hybridizing corn was known in the 19th century. The first corn hybrid was tested at the Connecticut Experiment Station in 1908 (Figure 1) and yielded 202 bu./a., at a time when average corn yields were 40-60 bu./a. What held back the commercialization of hybrids? The barriers were partially institutional. Corn breeding had to be reorganized to identify adapted inbreeds and hybrids. A way had to be developed to get that seed to farmers. The science part of hybrid corn technology was available long before the organizations were available which would enable farmers to use it.
Before hybrids, corn breeding focused on mass selection of open pollinated varieties. Most of this selection was done by farmers. Selection by land grant university and seed industry researchers was a small part of the picture. One of the main activities of the USDA corn program was to coordinate a network of farmers selecting the best open pollinated ears from their corn fields. Trying to maintain this model, early corn hybrid researchers spent time trying to develop ways that farmers could develop and produce their own hybrids. Some suggested varietal crosses between different open pollinated varieties as a way that farmers could benefit from hybrid vigor. Others felt that farmers should buy inbreeds developed by universities and produce their own hybrids.
It was in the mid 1920s that Henry Wallace, founder of Pioneer Hi-bred Corn Company, Eugene Funk of Funk Brothers Seed and others came to believe that a new organizational structure was needed. Hybrid corn should be developed, produced and sold by specialized companies. It took a few years for these seed companies with the help of the USDA and the universities to identify adapted hybrids, as well as working out a marketing structure. By the early 1930s all the elements were in place and hybrid corn use expanded rapidly.
Seed corn became an early example of industrial specialization in agriculture. Hybrid seed is a science based product produced by specialists. The role of farmer tinkerers in developing hybrid corn technology is small. Even Lester Pfister, founder of Pfister Hybrid Corn Company, was as much an entrepreneur as a corn breeder. The adoption curve for hybrid corn is smooth in part because the technology came to the market in the 1930s in a relatively mature form. The “bleeding edge” of technology occurred earlier in the 1920s and at a very small scale.
Who are the early adopters in the hybrid corn story? Was it the New England farmers who grew the first hybrid corn in 1920 before there was a way to provide a regular supply of hybrid seed? Was it the innovators who experimented with early Corn-belt hybrids in the late 1920s and often found them to be poorly adapted, not much better than their best open pollinated varieties? Or was it those who purchased the first hybrids available on a large commercial scale in the 1930s? When it is said that “most of the profits from new technology go to the early adopters,” we mean the “early adopters who get it right.” In that sense, the “early adopter” who increased profits was the Midwestern farmer who used hybrids to increase production in the early 1930s before the impact of increased supply affected the market.
One of the precision farming parallels to the hybrid corn case involves data analysis. Most manufacturers and researchers involved in developing precision farming tools assume that farmers will analyze their own yield monitor, soil test and other site specific data. But is this like expecting farmers to produce their own hybrid seed, just because farmers have always produced their own seed? There are large economies of scale in data analysis that suggest that this function could be carried out more efficiently and at lower cost by specialized organizations. The question is what kind of organization is best suited to handle the data analysis: for profit business like fertilizer dealers and crop consultants, or nonprofit organizations on the model of farm business associations.
Motorized mechanization of North American agriculture presents a more complex adoption path than hybrid corn (Figure 2). The history is complete with false starts and stair step adoption patterns. Motorized mechanization of agriculture was built on a long history of animal powered mechanization. Initially, innovators wanted motorized machines to do what horse and ox power had done before on the size of farms that they were used to. One of the early problems for tractor manufacturers was to make tractors small enough. Eventually, with the development of new implements and the increase in farm size the comparative advantage of motorized mechanization could be exploited, but we are still living through the farm size adjustment.
Steam powered mechanization was essentially a false start that illustrates the problems with making decisions based on early prototypes. If someone had estimated the potential for mechanization of crop pro-duction based on the steam tractors available in the 1890s, the resulting analysis would have presented a relatively bleak picture. Steam power might be competitive with ani-mal traction on very large operations, with large stoneless fields, in areas with high labor costs. This was true in the 1890s in California and on the large bonanza farms in the Red River Valley of North Dakota. Given the information available at the time, mechanization of the Corn-belt was unlikely. The steam technology available would be awkward for primary tillage of the relatively small Cornbelt fields and it would be unusable for cultivation of growing crops.
This analysis of mechanization would be accurate given the technology available at the time, but it would have missed the mark completely because it did not reflect the potential for technology change in the form of smaller internal combustion engines, power take off (PTO), tricycle type tractors which could be used in row crop cultivation and rubber tires. These innovations were not on the market in the 1890s, but their precursors were being discussed.
At the turn of the century most parts of the U.S. economy were mechanizing because science and innovation was supplying ever more convenient motorized power and because higher labor costs were driving a demand for automation. With hindsight we can see that mechanization of agriculture was inevitable, it was just a question of how. In a market economy it is unlikely that agriculture will be permanently different from other sectors of the economy.
It can be argued that precision farming is currently at the steam tractor stage. Currently available precision technology fits on some farms in some areas, just as steam mechanization was feasible in California and in the Red River Valley in the 1890s. But most economic studies indicate that current technology often fails to cover costs. Using current technology requires consider-able time and dedication. Packaged precision farming systems are not yet available. At the same time, precision farming technology currently in the development stage could greatly increase its profitability and ease of use (e.g. soil sensors, decision support systems). Global Positioning System (GPS) based technology is being applied throughout the economy wherever activities are scattered over a large geographical area. GPS systems are being applied in trucking, forestry and security serv-ices. It is likely that agriculture will be able to find a profitable use for GPS. We just do not know yet which use will be most practical and profitable. Current precision farming technology may be a false start and we can ask if precision farming innovators are the early adopters who get it right and reap the profits of innovation, or if they are the tinkerers who prepare the way for later widespread adoption?
Precision farming is a new technology with a long history (Figure3). Farmers have long tried to maximize crop yields and profits by spatially varying input applications. Scientists have studied spatially variability since at least 1915. Mechanization made it profitable to treat large areas with uniform inputs. GPS and other precision farming technology promises to reverse the trend to standardized crop recipes and make it economically feasible to manage crops on a more site specific basis.
Precision agriculture is an intuitively appealing concept and some expect it to have a smooth, rapid adoption path similar to hybrid corn in the 1930s and 1940s (Figure 3.). Several characteristics of the technology and the economic context suggest that the dynamic of change for precision farming will not fit the classic S curve model:
1) the technology is immature —farmers can not buy a complete precision farming system. Partial systems may be profitable for some farmers, but not for everyone.
2) the technology lends itself to tinkering — precision farming is not a yes or no choice. Farmers can and will modify the technology substantially.
3) institutions are not ready to deal with precision farming data —getting the most from precision farming will probably require data pooling, but we lack the organizational structure to do this pooling effectively.
4) Midwestern agriculture is increasingly risky — with reduced government price stabilization precision farming adoption may be subject to the kind of eco-nomic ups and downs that delayed mechanization of many farms in the 1920s and 1930s. An alternative scenario for precision farming adoption suggests (Figure 4).
- a) the current burst of precision farming enthusiasm may not be sustainable — many farmers and agribusinesses bought into precision farming on a promise of profitability. While crop prices are high and the technology is a novelty, they may be willing to wait. When corn drops below $2/bu., Cornbelt farmers and agribusinesses will ask harder questions about precision farming returns.
- b) lack of decision support systems may constrain wide spread use of precision farming technology. Nice colored maps are not enough to make it pay.
- c) full adoption may only occur after complete on-the-shelf systems are available and supporting institutions are developed to manage and analyze data. The current effort to develop systems in which farmers can analyze their own data might turn out to be like the efforts of corn breeders to develop systems which would allow farmers to grow their own hybrid corn seed.
Conclusions and Implications
Experience with technology adoption in agriculture suggests that the relatively rapid, smooth adoption process exemplified by hybrid corn in the Midwest is the exception, rather than the rule. The dynamics of adoption are often bumpy, with false starts and periods of stagnation because of technological or institutional barriers.
Science based technologies which are presented to the farmer as a package for an either/or choice often have the smoothest adoption path. If they are profitable, not too risky and within the resources of farmers, they are rapidly adopted. If profitability is questionable, risk is too high and/or resource requirements too great, they disappear. But even in the case of these package technologies, institutional factors may affect adoption. Widespread use of hybrid corn was delayed by the lack of organizations for corn breeding and commercialization. Use of genetically engineered crop varieties is facilitated by the organizations created for commercialization of hybrid corn.
Technologies which come to the market in an incomplete and immature form may have long adoption periods with many ups and downs. Farmer and agribusiness innovators often play a major role in the development of these technologies. The adoption of the tractor in the U.S. is an example. Precision farming technology has many of the same adoption characteristics as motorized mechanization. It came on the market in an incomplete form and many questions remain about its profitability. It lends itself to farmer modification.
In a strategic planning perspective, the dynamics of technology adoption may derail the best laid plans. The cases of hybrid corn and motorized mechanization suggest the following lessons:
1) be ahead technologically, but not too far ahead — evidence is that profits go to the early adopter, not necessarily to the tinkerer.
2) be prepared for bumps — technology change is rarely smooth. Precision agriculture may have more ups and downs than most.
3)development of management expertise is often the most durable product of trying new technology — in the case of precision farming, we know that the hard-ware and software will change rapidly. Building the capacity for spatial management will be the longest lasting investment.