Introduction
Advances in agriculture have played a major role in reducing poverty, improving food security, and raising living standards around the world.
These improvements are part of a wider trend where global agriculture is becoming more productive. One way of measuring this productivity is total factor productivity (TPF). TFP compares how much output farmers produce with the amount of land, labor, machinery, and other inputs they use.
Since 2000s, global agricultural TFP grew by nearly 2 percent per year.
In addition, farmers face constant challenges in maintaining and increasing crop yields due to pests, diseases, and climate change. Thus, consistent adoption of improved crop varieties is essential. Figure 3.2 shows how farmers in developing economies across Asia, Africa, and Latin America and the Caribbean have adopted new crop varieties over the period 1961–2020, regardless of whether the improvements were developed by public research institutions, universities, or private companies.
More importantly, the figure tells a similar story on the uneven adoption of new technologies. In particular, countries in Sub-Saharan Africa are lagging significantly behind other regions in adopting improved crop varieties.
These patterns raise two important questions. Firstly, why do some countries adopt new agricultural technologies more quickly than others? And secondly, what conditions help the technologies spread?
This chapter examines how agricultural technologies spreads across countries and what drives their adoption. It focuses on two major modern technologies, namely crops with genetically modified (GM) technologies and precision agriculture technologies (PATs).
The section that follows examines what factors determine how agricultural technologies are diffused. It shows that successful diffusion requires interaction between three factors: namely, decisions made by farmers (demand-side), the innovation ecosystem producing the technology (supply-side), and the enabling environment. The third section of this chapter expands on the role of IP.
Three main lessons emerge. First, technology does not spread automatically. Farmers’ characteristics, national innovation capabilities, infrastructure, and policy choices all play a role. Second, many different factors affect adoption. Farm size is one example. Larger farms often adopt new technologies faster than smaller farms. Third, IP plays a complex role. It encourages private companies to invest in developing new technologies, but it can also affect affordability and availability in different ways depending on the country and technology.
Determinants of technology diffusion in agriculture
Many factors shape how agricultural technologies spread. These factors can be loosely categorized into three connected elements: demand-side factors, supply-side factors, and the enabling environment. Demographic shifts, pest and disease outbreaks, and consumer preferences also influence adoption and dis-adoption.
Demand-side factors: farmer decisions
Factors that influence a farmer's willingness to pay for new technologies are demand-side factors. Farmers face uncertainty regarding the performance of new crops, therefore have to weigh the upfront costs, learning requirements, financing constraints and market signals against the expected returns.
Farmers are willing to pay for new technologies, even if more costly, when they are shown to outperform what they currently have.
Supply-side factors: innovation capabilities
Agricultural technologies are agro-ecological and context specific. This is where the supply-side factors influencing technology diffusion are important. Strong innovation ecosystems are able to adopt, adapt and generate new technologies for local conditions.
Figures 3.3a and 3.3b show that advanced economies benefit the most from the diffusion of new to the world type of innovation and scientific knowledge. Northern America and Europe, followed closely by Asia, are the leading beneficiaries of technology diffusion in agriculture. These three regions produce 75 percent of new agricultural technologies based on novel patented technologies and nearly 95 percent based on novel scientific knowledge.
Enabling environment: institutions, markets and infrastructure
Supportive policies reduce risks and transaction costs in the diffusion of appropriate technologies. Infrastructure, functioning markets and appropriate IP frameworks enable diffusion. In addition, evidence shows that support for trusted intermediaries, such as extension services, farmer networks, suppliers, and cooperatives, play an important role in facilitating the use of new technologies.
Diffusion varies according to the technology
The local adaptation needs of crops with GM traits and the modularity of PATs drive their respective diffusion patterns.
Genetically modified crops
GM crops are plant varieties that scientists have genetically altered using genes taken from other species, or by synthetic construct, in order to make crops more robust and tolerant of specific farming challenges. Scientists can modify crops to resist pests, tolerate herbicide, extend shelf life, and/or produce higher yields, to name a few of the commonly found GM traits.
GM technologies must first be adapted to the local agro-ecological conditions through local R&D before they can be adopted by farmers.
Developing new crops can be lenghty. One study found that it can take about seven to ten years to develop stable lines using conventional breeding technique for fruits.
Institution/regulation: Regulation shapes the diffusion of GM crop technologies. Most countries regulate both the farming and the importation of crops with GM traits. Crops with GM technology must first be approved by the regulators before importation or cultivation. Approval assures food safety and environmental protection of the locally adapted GM crop for consumption and cultivation.
However, obtaining regulatory approval can be expensive. Developing a crop with GM traits from discovery to commercialization costs USD 64.2 million on average. Regulatory approval per country costs an additional USD 43.2 million, which accounts for 37.6 percent of the total cost (i.e., from discovery to commercialization to regulatory approval).
Figure 3.4 shows more approvals for GM crop importation than for cultivation. This suggests that countries may be more willing to import GM crops than to allow farmers to grow them domestically. However, several countries have simplified the approval process whereby approval extends to both importation and cultivation, including Argentina and South Africa.
Regulatory approval does not guarantee adoption by farmers in a country. Pricing, input availability, farmer awareness, and infrastructure also matter.
Availability: Countries with strong innovation ecosystems are more likely to generate, adapt and adopt GM technologies for their local agro-ecological conditions. Argentina, Brazil, Canada, India and the United States are the largest adopters of GM technologies.
Affordability: Crops, trees and plants with GM technologies tend to be more expensive in comparison to improved varieties developed using conventional seed technology. Nevertheless, farmers might be convinced to purchase the more expensive GM technology if the expected profit outweighs the cost (see Table 3.1 for comparison of GM trait seed's technology costs and the average gross farm income from using the seed for cotton, soybean and maize crops).
Another cost farmers take into consideration is the price and availability of plant protection products. In Kenya, a survey of farmers on the adoption of hybrid maize and fertilizers found that changes in the price and availability of these inputs played an important role in determining whether they adopted the technologies or not.
Awareness and training: Farmer extension services and stewardship programs play an important role in demonstrating the efficacy GM crops to farmers and helping maintain the performance of the crops over its life cycle.
Familiarity with technology increases the speed of adoption. In India, farmers familiar with hybrid cotton cultivation adopted GM cotton faster. This same pattern was also seen in the case of China, Mexico and the US.
Infrastructure: Access to seeds with GM traits is important for farmers in rural areas. Kenya’s poor infrastructure, namely access to and availability of seed and fertilizer distributors, has contributed to a low adoption rate for hybrid maize and fertilizer among farmers, even when the estimated gross returns to adopting the technology were high.
Figure 3.5 shows how long it took for specific crops with GM technology to achieve widespread diffusion within selected countries, defined as 80 percent cultivation of the cropland. While the United States developed the GM technology, some developing economies reached widespread adoption faster, such as GM cotton in South Africa, GM soybean in Argentina and GM maize in Brazil.
Several factors can explain this. In Argentina, local pricing, input availability and institutional support help explain why its farmers were able to adopt the GM technologies quickly.
Meanwhile, consumer hesitation and regulatory caution slowed the adoption of GM technologies in the United States, particularly in the case where the GM crops were meant for food.
Precision agriculture technologies
PATs use sensors, satellite navigation, and data analytics to optimize farming operations. In general, there are three broad categories for PATs: (i) the data collection (sensors, satellite navigation), (ii) the data processing and/or analysis (yield monitoring, soil mapping), and (iii) the decision-making guidance (auto-steering tractors, variable-rate applications of fertilizers and pesticides).
Farmers in Australia, Canada, Europe and the United States lead in the adoption of PATs.
The US pioneered PATs in the 1980s, with adoption accelerating once global positioning systems (GPS) became widely available after 1983.
However, the adoption of PATs remains gradual. Studies show that farmers typically adopt individual PAT components rather than a complete system.
Less than one-third of US farmers use any PAT tools whatsoever and adoption occurs in modules rather than complete systems.
In addition, the PATs predominantly adopted vary according to agricultural need. Water scarcity led to the adoption of micro-irrigation in India, for example, whereas farmers in the US and Australia focus more on adopting guidance systems for large-scale cropping.
Figure 3.6 illustrates how farmers in the US are slower at adapting PATs in comparison to GM crops. It also shows how farmers can adopt the PAT in separate components, such as auto-guidance and VRTs.
Availability: Due to the modularity of PATs, compatibility with existing as well as new machinery matters. Farmers risk being locked into a single vendor ecosystem without interoperability standards.
Affordability: Cost and complexity are key barriers to PAT adoption. PATs require upfront investments in hardware, software and connectivity. Table 3.2 compares the cost of using three different mapping technologies to analyze crop health and manage plant protection inputs accordingly for vineyard crops. It shows the different PAT components needed: a platform, sensing device and a satellite navigation system corrector (RTK-GNSS) across the technologies. Regardless of their farm size, farmers face costs starting at over EUR 16,500 to implement any one of these mapping technologies. Yet, as shown in Table 3.2, the per-unit cost of adopting any of these technologies declines as farm size increases.
In some countries, farmers have abandoned using PATs because of a lack of profitability.
Awareness and training: Skills and training influence the use of PATs. A survey of farmers in Europe found PATs not to be user-friendly and this and a lack of support in using the technologies are two of the main barriers to adoption.
Nevertheless, those farmers who have adopted certain entry-level PATs are more likely to adopt other PAT components. For example, the widespread adoption of GPS for auto-steering tractors can also be used to guide agricultural input applications, such as variable rate fertilizers. This probably explains why variable rate fertilizer technology has only been modestly adopted.
Infrastructure: Infrastructure limits the adoption of PATs. This agricultural technology is highly dependent on the physical and digital infrastructure present within a given country, such as having reliable electricity and cellular or satellite connectivity for data flows.
PATs that rely on the latest broadband and wireless technology 5G will not work in areas where the electrical and digital connectivity is weak.
Figure 3.7 illustrates the infrastructure challenge globally. Most large farms have access to 4G connectivity, whereas the majority of small farms are connected to the older technology, 2G.
The spread of mobile technologies can facilitate access to PAT, especially when tailored to digital connectivity constraints. India, for example, is investing in R&D to adapt PATs for local farmers in regions with weak digital connectivity.
Diffusion varies according to farm size
Large farms adopt GM crops and PAT faster than their small-sized counterparts because they are abe to spread fixed costs over more output and have better access to finance and information.
By contrast, small farms face credit constraints, higher per unit costs, and lack awareness of new technologies and their use. These factors affect profitability and increase the risks and uncertainty that arise from adopting new technologies.
This divide is present within countries, especially in relation to PATs. Large farms in Australia and the US report a higher usage of guidance auto steering due to their farms requiring large-scale cropping. Ten percent of large farms in Argentina, occupying 78 percent of total agricultural land, use advanced PATs.
A similar pattern appears in other regions.
Globally, small farms – less than two hectares – are the majority by number, but occupy only a small share of total farmland.
Technological disparity across farm sizes affects lower income farmers disproportionately. Table 3.3 summarizes the different determinants of GM and PATs technology diffusion. For small farms these factors can be more acute and pose a bigger barrier to adopting these new technologies.
A slower adoption of new technologies implies a slower gain in productivity, and a lower increase in income levels.
Without targeted support, technological progress can widen income gaps and reduce inclusiveness.
Diffusion happens faster within a supportive environment
Government support through policies, regulations and funding schemes can narrow the technology diffusion adoption divide. Even industry-led initiatives help technology diffuse faster.
Regulation sets the boundary conditions for diffusion
GM crops require risk assessment for human health and the environment. Efficient, transparent processes reduce uncertainty and speed safe adoption. Where processes are slow or unpredictable, firms delay investment and farmers wait.
Policies actively promote adoption
Many governments help fund new technology purchases or demonstrations, support local adaptation research, and invest in extension. In GM technologies, the public sector has been an active enabler of the technology diffusion of yield-improving crops through farmer extension services and training.
They may also set sustainability targets to encourage PATs use. For example, farmers in Brazil and Canada have benefitted from public-support programs and policies designed to encourage the adoption of PATs. In the US, various funding schemes and public–private partnerships help increase PAT uptake.
The private sector, as well as farmer-led initiatives, also plays an important role in the speed and spread of GM crops and PATs among farmers. In some countries, the agricultural technology producing firms engage with the public sector to provide extension farmer services, as well as training to help maintain GM crop integrity over its life cycle. For example, industry-led initiatives have facilitated PATs use by farmers in Australia.
Infrastructure matters
Governments also play an important role in providing the right infrastructure for the adoption of GM crops and PATs, and across farming in general. For example, irrigation infrastructure provides an important and stable growing condition that has supported the adoption of GM cotton in China, Mexico and the US. In the case of PATs, the lack of or weak physical and digital connectivity limits the usability of PATs across and within countries.
Standards and competition policy matter
For PATs, open standards enable interoperability across devices and brands. Standard essential patents (SEPs) licensed on fair, reasonable and non-discriminatory (FRAND) terms can unlock compatibility and reduce switching costs (see Chapter 5 for a discussion of SEPs).
Competition oversight can deter anti-competitive bundling which locks farmers into proprietary ecosystems.
The role of intellectual property
IP protection has a nuanced impact on agricultural technology diffusion. It provides incentives for private R&D investment, but can create barriers to access proprietary technology and to develop follow-on inventions. The effects vary according to technology, country context and market structure.
IP protection drives innovation
IP protection provides a temporary exclusive right to restrict the use and sale of protected technologies.
In plant innovation, much of the progress historically has come from public-sector breeding programs. This remains largely true across many developing economies. But in some economies, such as the United States, private investments in plant innovation have surpassed public spending.
Technological advances in plant breeding can be protected using the following intellectual property (IP) instruments in the US:
Plant Patent Act (1930): provides exclusive rights over asexually reproduced plants, including ornamental plants and fruits.
Plant Variety Protection (PVP) Act (1970): provides protection for plant varieties of sexually reproduced crops, which include oilseed crops, grasses and grains.
Patents (utility patents): The US Supreme Court decision on Diamond v. Chakrabarty (1980), and the United States Patent and Trademark Office (USPTO) Board of Patent Appeals and Interferences (1985) decision on Ex parte Hibberd extended that jurisdiction’s patent protection system to include seeds, parts of plants, plant varieties, genetically engineered organisms, and gene products.
These gradual changes in how innovation in plant breeding can be protected made innovation in the area attractive to the private sector.
For private firms, IP rights make it possible to capture returns on R&D investments, thereby creating stronger incentives to innovate.
In addition, the field is risky due to high attrition rates and no guarantee of success. One study found that only five genetic traits were eventually commercialized for farmers out of the 560 genetic traits identified in an R&D process.
The public sector uses IP protection. For example, public research institutions in the United States file for patent protection on agricultural innovation to enable technology transfer to private firms and to facilitate its commercialization.
However, the public sector does not rely on IP protection to the same extent as the private sector. In many instances, the scope of agricultural innovations is different, with the private sector focusing on those inventions that are scalable, while the public sector focuses on region-specific types of inventions.
IP protection shapes collaboration. Licensing agreements and joint ventures give local firms and NARSs access to foreign technologies, enabling them to adapt innovations to local germplasm and strengthen domestic capabilities. It also helps to build the innovation capabilities of the local innovation ecosystems of countries. In many cases, private companies have partnered with public research institutions to tailor GM technology to specific agro-ecological conditions, thereby building stronger local innovation ecosystems.
Yet IP can restrict access
Technologies protected by IP rights can be costly to access, affecting both the development of new innovations and their eventual use.
Farmers often seek crop varieties that combine multiple desirable traits, such as pest resistance, drought tolerance, and high yields. Developing such multi-trait seeds can be expensive for seed producers, particularly when the traits are owned by different IP holders. Negotiating licenses with multiple parties can lead to “stacked” royalty fees, with each rights holder seeking a significant share. These cumulative costs can squeeze margins for seed producers, and discourage the commercialization of seeds with optimal trait combinations.
In the case of crops with GM technology, one study indicates that royalty rates account for more than half of the total markup on seeds.
Concerns about IP protection in plant innovation have instead focused on broader societal and ethical issues. Questions have been raised about whether biological materials should be subject to private ownership, and whether IP restrictions limit access to agricultural technologies, particularly in ways that may conflict with humanitarian objectives. A notable area of tension lies between patent holders on one side, and farmers’ privileges on the other, reflecting ongoing debates over how to balance innovation incentives with equitable access (see Box 3.2).
In 1996, Monsanto (now part of Bayer) developed RoundUp ReadyGM soybeans—genetically modified to tolerate the company's Roundup herbicide. Farmers adopted this GM technology quickly due to its proven efficacy and performance.
However, between 2007 and 2023, the company faced several lawsuits regarding how it enforced its patent rights.
Typically, farmers pay technology fees when purchasing improved crop varieties. In many countries, this purchase exhausts the patent holder's rights, allowing farmers to use the technology as they wish, including saving unused seeds for the following season.
Monsanto took a different approach. When commercializing its GM technology, the company required seed producers, grain companies, and farmers to sign technology licensing agreements that prohibited farmers from saving and replanting purchased seeds beyond one season. This restriction contradicted farmers' traditional practice of seed saving.
Farmers who used seeds containing GM traits—knowingly or not—were required to pay indemnity fees based on their harvest. This led to lawsuits against Monsanto in countries such as Brazil, Canada, and the United States, among others.
In Brazil, farmers argued the indemnity fees "hurt the collective rights of millions of farmers" and filed a class action lawsuit on April 9, 2009. After years of legal proceedings, the Brazilian Appeals Court ruled in favor of Monsanto in 2019, determining that farmers' privilege to save seeds do not apply to patented technologies.
For PATs, the absence of interoperability standards can slow adoption. PATs’ modular design means farmers can purchase each component separately. However, farmers may encounter compatibility problems between hardware and software (digital) elements. In addition, they may face challenges integrating PAT systems with existing farm equipment.
Many digital components in PATs rely on overlapping and complementary proprietary technologies, which can be costly to license. This could be solved with agreed-upon interoperability standards. If PAT providers adopted standards enabling seamless integration between PAT systems and existing machinery, adoption could accelerate. This would mirror successful standardization in other digital technology sectors (see Chapter 5 on standard essential patents).
But, unclear IP ownership creates its own barriers
While IP protection may raise the issue for the adoption of new technologies in agriculture, a lack of certainty as toIP ownership can also create a barrier by complicating commercialization and limiting the potential benefit of the innovation.
In advanced plant breeding, unresolved ownership rights concerning foundational technologies can create tension for follow on innovators. A prominent example is the ongoing licensing dispute involving the CRISPR Cas9 gene editing tool. This technology allows scientists to make highly targeted changes to the genetic code of plants, animals, and other organisms, correcting genome errors or introducing beneficial traits. CRISPR Cas9 has gained rapid popularity due to its relative ease of use, efficiency, and flexibility, enabling development of improved traits roughly twice as fast as older breeding technologies.
The dispute centers on overlapping patent claims and unclear licensing rights for follow-on inventions. Two foundational patents underpin CRISPR Cas9: one filed by Jennifer Doudna and Emmanuelle Charpentier at the University of California, Berkeley, and the University of Vienna, and another filed by Feng Zhang at the Massachusetts Institute of Technology on behalf of the Broad Institute. Both patents cover CRISPR Cas9’s use, but apply to different classes of DNA- based organisms. The Doudna/Charpentier rights are licensed to Corteva; Zhang’s patent is held by the Broad Institute.
Because the claims appear to overlap, each rights holder could, in theory, prevent the other from commercializing follow on inventions that infringe on their patents. This creates serious licensing uncertainty — potential innovators may not know which party to approach for commercialization rights.
The complexity deepened when the U.S. Court of Appeals for the Federal Circuit sent the case back to the Patent Trial and Appeal Board to determine priority of invention, prolonging resolution.
Some researchers avoid the licensing gridlock by turning to alternatives such as CRISPR 3.0 technologies, including CRISPR Cms1 and CRISPR C2c2, owned by Benson Hill Systems. While these tools may not match CRISPR Cas9’s efficiency, they provide a clearer path to commercialization because patent ownership is well established.
Similar ownership uncertainty also affects PATs, particularly concerning farm level data. Sensors embedded in PAT systems collect millions of data points daily, covering soil conditions, input usage, weather patterns, and other farm specific metrics. Aggregated at regional or national levels, this data could be critical for improving farm management practices and guiding agricultural investment decisions.
However, ownership and access rights of this farm level data remain poorly defined. For instance, who controls this data — the farmer, the PAT manufacturer, the provider of specific system components, the cloud service operator storing the information, or the sensor developer that initially captured it?
Potential solutions include data cooperatives and data trusts, which establish governance frameworks for data use while ensuring farmers share in the benefits. By clarifying ownership and usage rights, such structures could unlock the full potential of farm level data for both individual and regional agricultural improvement.