The Case for AgARDA

Introduction

Once the leader in farm innovation, the U.S. government increasingly neglects agricultural R&D. American farm productivity growth is slowing, and the U.S. is losing ground to competitors such as China, India, and Brazil. Agricultural output needs to speed up to meet growing global demand, and disasters such as conflict-induced supply chain disruptions, pandemics, agricultural pests and diseases, and climate change threaten conventional production. A large suite of potential innovations could address these threats, but many risk failing to reach the market without a concerted effort to take longshot bets on early stage technologies. 

In defense, emergency medicine, and energy, the federal government has successfully implemented a distinctive funding model to do just this. Advanced Research Project/Development Agencies (ARPAs or ARDAs) are designed to tackle important technological challenges that require focus and agility that other programs lack. The 2018 Farm Bill authorized $50 million per year for a new Agricultural Advanced Research Development Agency (AgARDA). But in the following three years, Congress did not appropriate funds to the agency, and in FY2022 only gave it $1 million, so the USDA has not yet established AgARDA. This failure to fully fund AgARDA leaves promising solutions to the challenges facing U.S. and global agriculture unexplored and undeveloped.

Congress gave AgARDA four main objectives: 

• Prevent and prepare for natural and intentional threats to U.S. agricultural production
• Increase American agricultural sustainability, export competitiveness, and resilience to weather shocks
• Increase U.S. leadership in agricultural R&D
• Address areas that are currently neglected by existing funding sources

The scope of these tasks is vast, especially considering that the authorized $50 million annual budget is less than 1.7% of what the federal government currently spends on agricultural R&D. This is also far less than other ARPA programs receive. In FY2022, the budget for the Defense Advanced Research Project Agency (DARPA) was $3.8 billion, the Biomedical Advanced Research Development Agency (BARDA) $823 million, and ARPA-Energy (ARPA-E) $427 million. All three programs have proved to be well worth the investment. DARPA laid the groundwork for the internet, GPS, speech translation software, and many other advances in defense and civilian technology. BARDA’s work has led to 62 new and approved medical countermeasures, including the COVID-19 mRNA vaccines. Though founded only 13 years ago, ARPA-E has already turned $3 billion in total R&D funding into 25 marketed projects worth a collective $21.6 billion. If AgARDA focuses on a handful of key technology areas and follows the lessons from these ARPAs, it could have a large impact. However, early successes and partnerships with private and public stakeholders will be important for securing the political will to increase AgARDA’s funding to the level it will likely require to truly transform the food and agriculture sector.

ARPA programs succeed because of intense focus on a few important objectives, pursued nimbly with whatever technology is best suited to meet the goal. AgARDA should focus on three primary technological areas: 

• Passive monitoring of crop fields and livestock for known and novel diseases
• Genetically engineering more productive, stress-tolerant plants that are less reliant on farming inputs
• New ways of producing calories — such as precision fermentation or microbes that process waste — that are less susceptible to climate and disease threats than field crops and livestock

AgARDA must view its mission as helping new technologies cross the “valley of death” from early-stage research to commercialization. As such, the agency should focus not just on pioneering new advances, but also on forming an ecosystem of other public and private funders to scale up innovative technologies.

Problems in Agricultural R&D and Productivity Growth

Productivity growth is especially important in agriculture. While other sectors can grow by increasing inputs such as new factories, agriculture is constrained by nearly fixed inputs like land and water, particularly in developed countries. More than 82% of the growth in global food production since 1960 has been from crop yield increases. In the U.S., while the type of inputs have shifted from labor to intermediate inputs like pesticides and fertilizers, total inputs have remained essentially flat since 1948, increasing only 5% (data from USDA). Over the same period, outputs have nearly tripled. 

However, U.S. farm output and productivity growth is slowing. From the 1960s through the early 1980s, U.S. agricultural output grew at an average annual rate of 2.45%. This fell to 1.35% for the 1980s and 1990s, and has averaged just 0.69% since 2000 (data from USDA, shown below). This decline may pose serious supply constraints in the future, considering global agricultural output will have to grow by 1.34% per year through 2050 to meet growing demand.

America’s slowing agricultural output growth appears tied to declining public investment in agricultural R&D. Over the past two decades, total federal and state funding for agricultural R&D has fallen by more than 32%, reaching its lowest levels in fifty years. In addition to declining funding levels, public agricultural R&D also appears to be less effective on a per-researcher basis. While agricultural R&D has long been a valuable investment — a recent review put the social rate of return between 17% and 67% — the returns to R&D, measured as agricultural output per researcher, have fallen by 3.7% per year since 1970. 

Declining returns might be due to the shift from public to private R&D in agriculture. While the private sector accounted for 35% of total U.S. agricultural R&D in 1970, by 2014 its share had risen to 58%. The federal government devotes 42% of its funding for agricultural research to basic discoveries. By contrast, only 7% of private R&D in the U.S. goes to basic research; the rest is applied research or experimental development. Thus the move from public to private funding also represents a shift from basic to applied research, potentially reducing the return on investment in agricultural R&D.

While falling government support for agricultural R&D is a global problem, the U.S. is still losing ground to economic competitors. In OECD countries, the share of public R&D devoted to agriculture fell by 43% from 1981 to 2013, less than the 55% decline in the American share. In 2008, China overtook the U.S. as the largest funder of agricultural R&D. Since the CRISPR-driven gene editing revolution, China has filed twice as many agricultural biotechnology patents as the U.S. Global agricultural productivity has increased by 33% in the last twenty years, but only by 13.5% in the U.S. 

In the last fifty years, America’s share of global grain and oilseed production has fallen by 22%. America’s declining share of production is due both to slower productivity growth and fundamental land constraints, highlighting the need for increased U.S. agricultural innovation. While a general increase in public R&D investment is important to reverse falling agricultural productivity trends, the new funding model promised by AgARDA could address the declining returns to current R&D approaches and help tackle important but neglected challenges in agriculture.

Advantages of the ARPA Model to Agriculture

The ARPA model was pioneered by the DOD during the height of the Cold War to establish U.S. technological superiority and create disruptive innovations. It has since been successfully applied to other sectors, like emergency medical countermeasures and energy, and other governments have sought to copy the model. One key feature of the ARPA model is its horizontal, minimally bureaucratic structure. ARPA projects are headed by outside experts — borrowed from industry or academia — who report to the head of their project office (such as DARPA’s Microsystem Technology Office or Tactical Technology Office) and to the director of their ARPA program. ARPA directors are responsible only to the head of their respective agency (i.e., DOD, DOE). Each project manager is given a limited term, typically three to five years, and broad latitude to pursue a technological objective by funding industry or academic groups. 

ARPAs can hire personnel without having to go through the standard federal hiring process, allowing them to quickly bring on new researchers as the mission requires. These agencies put out funding opportunities by announcing an ambitious goal and then allowing applicants to propose solutions. This means that projects aren’t bound to one particular technology, but can flexibly use whatever approach best achieves the objective. Additionally, USDA funding decisions are guided by advisory committees, and the charters for specific subcommittees set aside seats for industry representatives. While stakeholder feedback is important to ensure that public research addresses the sector’s needs, ushering in disruptive technologies may be best done by a research agency free from industry input. 

The food production industry is highly concentrated and dominated by entrenched interests, meaning that the sector is slow to bring new technologies to market. Of twelve major industries, food and agriculture had the lowest market share occupied by startups, at just 2%. Agriculture relies on many rural producers, making technological diffusion much slower than in more centralized industries. This low level of industry dynamism and slow technological change are poorly suited to the challenges facing food production. The ARPA model, with its ability to innovate disruptive technologies focused on challenging and neglected challenges, can have an especially potent impact on agriculture.

Most years, pests and pathogens reduce global yields for major crops by 20 to 30%. In 1971, a single fungal disease, the Southern Corn Leaf Blight, reduced U.S. corn yields by more than 15% beyond typical losses. In 2016, wheat yields fell by 30% in France, the world’s fifth largest producer, due to unprecedented disease spread. In 2019, the ongoing African Swine Fever epidemic in China cost the country almost one percent of its total GDP. Beyond natural dangers, American agriculture is also at increased risk of engineered attacks. Over the past century, every state biological weapons program developed agricultural bioweapons, a worrisome precedent in our era of increased geopolitical tensions. A RAND report prepared for the Secretary of Defense calls the food and agriculture industry the country’s “soft underbelly,” given how easily it could be targeted. But crop and livestock disease monitoring remains mostly manual, with several promising technologies that would allow early detection remaining unexplored. 

Climate change likewise poses a serious threat to food supplies. Some modeling shows that with 2 ^oC of warming, there will be a 7% annual chance of a 10% drop in the global production of corn — a crop that provides a fifth of humanity’s calories — and other crops are similarly vulnerable to synchronous production shocks across multiple continents. Given these and other threats to food production, we must invest in an ambitious portfolio of food and agriculture technologies. 

Strategic Goals for AgARDA

The ARPA funding model is not wedded to a particular technology, and operates by onboarding expert project managers with fixed terms during which they can pursue a targeted development goal. This time-bounded focus and high level of independence between projects has been crucial for the success of other ARPAs. But individual project managers aren’t just pursuing a hodgepodge of different technologies. Rather, DARPA was tasked at its founding  with finding innovations that furthered a few high level strategic goals known as “Presidentials,” and projects are chosen based on their potential to contribute to these goals. 

Congress has already given AgARDA overarching objectives, and defined its scope as encompassing any improvements to the productivity of all parts of the food supply chain and defense against agricultural diseases. However, a set of broad technological priorities has not yet been set for the agency. While AgARDA must continue the legacy of giving individual project managers discretion over how they choose and pursue specific technologies, it should set strategic goals in three areas: 1) automated, passive monitoring of known and emerging livestock and crop diseases, 2) paradigm shifting applications of genetic engineering and synthetic biology to enhance plant productivity and stress tolerance, and 3) novel, decentralized food production systems that do not rely on traditional agriculture. 

Plant pathogens can be carried by winds across continents or spread by highly mobile vectors. In the U.S., 98% of animal products come from facilities with more than 1,000 animals, making the industry extremely vulnerable to disease spread. The federal government sees livestock disease monitoring as a matter of national security, and Presidential Policy Directives 7 and 9 order DHS to assist in surveillance efforts. However, the main monitoring strategy is veterinary reporting, meaning that diseases are only detected when a farmer notices an animal displaying symptoms. Even then, a new disease may not be recognized. The USDA’s National Center for Foreign Animal and Zoonotic Disease Defense focuses on developing diagnostic kits for known diseases, rather than innovating ways to detect emerging diseases. Likewise, plant disease surveillance centers on certifying that U.S. fields are free from a select group of known diseases so that America can export produce. This process depends on manual inspection that is impossible to scale to the level needed to closely monitor the nearly 400 million acres — 16% of U.S. land — occupied by crops. However, advances in remote sensing, environmental DNA sequencing, machine learning, and other fields promise new ways to detect disease occurrence at scale. Many of these innovations, however, are likely to be neglected by private companies and producers since much of the benefit to early disease detection is to the sector as a whole. If a farmer detects a disease early, they may nonetheless have to cull their herd or plow over their field, so they still suffer the economic loss, but their neighbors benefit from this early action. AgARDA should fill in this funding gap and focus on technologies for passive agricultural disease monitoring. 

Genetically modified crops, first introduced to the U.S. in 1994, have been an economic windfall for farmers and provide large environmental and economic benefits to society as a whole. Adopting GM crops increases farmer profits an average of 68%, and they have added $150 billion to the global economy while reducing pesticide use and CO₂ emissions. Crop biotechnology has the highest research intensity of any agricultural input subsector,¹ and much of the rise in private R&D investment has been driven by publicly funded advances in biotechnology. However, modern gene editing and synthetic biology could allow for much more fundamental changes to plant physiology.

In most settings, the efficiency with which plants use photosynthesis to convert solar energy into chemical energy does not exceed 1%. The theoretical potential for photosynthesis efficiency, however, is about 12%, although several factors such as energy loss to respiration bring this down to between 4.6% and 6%, depending on a species’ photosynthetic pathway. Several synthetic biological approaches aim to close this gap between potential and actual productivity efficiency by changing plant photosynthetic machinery and metabolism. However, translating these proposals into marketable breakthroughs requires significant focused investment, as reengineering even a microbe’s central metabolism is estimated to cost $50 million over six to eight years as of 2016. The USDA’s Agricultural Research Service currently conducts research into photosynthetic efficiency improvements, but it focuses on gene mapping to find existing variation, rather than betting on new pathways to engineer better plants. Advanced biotechnology could also reduce the inputs required for agricultural production, such as by engineering soil microbes that fix atmospheric nitrogen to associate with non-leguminous plants, decreasing our dependence on imported fertilizers. AgARDA is the agency best suited to aggressively tackle the challenge of revolutionizing crop biotechnology, meeting the world’s growing demands with fewer resources and making our agriculture more resilient to production shocks.

Two core attributes of conventional agriculture make our food supplies difficult to fully secure. First, production is concentrated in a few breadbaskets that are often far from major population centers, leaving consumers vulnerable to transportation disruptions and dependent on regional or international trade. The top five states producing corn, soy, and wheat — crops that occupy nearly 90% of U.S. cropland — contribute 60%, 53%, and 59% of the nation’s supplies respectively. The food in a typical American meal travels 1,500 miles, meaning that disruptions such as truck driver shortages can quickly cause food supply shortages. 80% of people worldwide live in a net food importing country, leaving them vulnerable to trade restrictions and conflict. Second, agriculture occupies huge, outdoor spaces and so is at the mercy of the weather and natural or intentional disease spread. Since these two features are shared by all of our food sources, a single disruption could affect the whole food supply chain at once. Advances in cellular, microbial, and other indoor agricultural technologies could diversify food supplies. DARPA has two ongoing projects on alternative food production. However, while DARPA aims to pioneer dual-use technologies — those with both defense and civilian applications — its focus with these projects is providing food supplies to U.S. troops cutoff from supply lines. As a result, it’s not optimizing for highly scalable, cost effective innovations. AgARDA should fill this gap and focus on developing food technologies such as cellular agriculture or precision fermentation that have public appeal and are readily affordable.

Learning from other ARPAs

Beyond copying parts of the formal organizational structure from other ARPAs, a successful implementation of AgARDA hinges on learning important lessons from its predecessors: DARPA, ARPA-E, and BARDA. Unlike defense, the energy and agriculture sectors are overwhelmingly private and highly distributed, and unlike emergency medical countermeasures, their products are used daily. This required ARPA-E to develop strategies for bringing these innovations into a highly established market. The agency has had to quickly create alliances to maintain political support within the DOE, where it could be seen as competing for limited funds, and among private actors who exert influence over policymakers and could see ARPA-E’s exploits as disruptive to the industry. Under the FY2022 budget, funds for AGARDA, which only ended up being $1 million out of the authorized $50 million, came from the budget of USDA’s Office of the Chief Scientist. This type of funding structure can make other agencies and offices within USDA view AgARDA as competing for resources, potentially handicapping AgARDA’s political support within the agency. This possible conflict furthers the need for AgARDA to mimic ARPA-E’s strategy of forming close partnerships with other government agencies and industry to build a base of support. The agency should also have a diversified portfolio of investments and concrete pathways for finding a market for new technologies so it can quickly demonstrate its usefulness. 

Over the past four years, Congress has appropriated just half a percent of the funds authorized for AgARDA in the 2018 Farm Bill. In that bill, AgARDA was only chartered as a pilot project. This shows how tenuous the agency’s existence is. If it is fully funded, it will have to quickly establish its usefulness and build a base of support within and outside of government. Making these connections will also advance AgARDA’s goal of bringing new technologies to market. ARPA-E hosts annual summits where they feature awardees, attracting VC investors and pioneering researchers. This helps create an ecosystem across business and academia that benefits from ARPA-E’s leadership and investment, forming a supportive coalition. AgARDA should host similar events, and tap into the existing R&D infrastructure within the USDA to form intra-agency partnerships. Potential partners already exist: the USDA’s Science Council has an Emerging Technology Team that looks for opportunities for technology adoption among USDA stakeholders. The team focuses on sensor information for agricultural production assessments, so it could benefit from disease monitoring technologies pioneered by AgARDA and in turn would help find early market adopters for these innovations. The USDA’s National Institute for Food and Agriculture has a small business grants program that could provide seed funding to food and agriculture startups that AgARDA helps launch. This partnership could mimic ARPA-E’s SCALEUP (Seeding Critical Advances for Leading Energy technologies with Untapped Potential) program, which provides support to small companies using ARPA-E generated technologies. To direct these and other efforts, AgARDA should adopt ARPA-E’s model of a Tech-to-Market Team, a separate group of staffers working full time to find marketing opportunities for novel technologies.

AgARDA can also learn valuable lessons from DARPA’s model of using government investment and technology adoption to create early markets to spur further development. DARPA has been able to rely on DOD procurement to provide an early market for its technologies. This makes it easier for DARPA to launch innovations into use than for ARPA-E, and a fully funded AgARDA, since the energy and agriculture sectors are largely private. However, as noted above, DHS and USDA² share responsibility for monitoring America’s crops and livestock for disease. These agencies could trial new technologies, helping scale them up and serving as testbeds to demonstrate efficacy to private investors. 

The USDA could also provide various incentives for farmers to adopt early stage technologies. Crop producers already have to comply with conservation standards to receive federal insurance subsidies, and farmers growing GM crops have to set aside part of their field to other varieties to reduce disease spread. The USDA could offer further subsidies to farmers willing to set aside small portions of their land to growing new crops under development by AgARDA or to those who implement disease monitoring technologies before these prove their commercial viability.

Knowing that it had to prove its utility quickly, just as AgARDA will have to do, ARPA-E made sure to invest in a diversified set of technologies. Some fell into the traditional domain of ARPA: longshot gambles on high-risk, high-reward technologies. Others are more surefire bets with shorter timelines that are likely to produce valuable developments in the near future. AgARDA could benefit from a similar portfolio approach to quickly demonstrate its usefulness. DARPA, BARDA, and ARPA-E all also have the authority to issue prize rewards for outside research teams that achieve some technical objective. Prizes are often extremely successful in attracting outside investment, such as a $10 million prize for innovations in private space fairing technology attracting more than $100 million in investment from competitors. Prizes also reduce agency overhead, since evaluators only have to assess final outcomes without first preselecting teams based on likely success. As currently proposed, AgARDA does not have prize authority, but this should be changed so that it can better replicate the success of other ARPAs. AgARDA could also learn from ARPA-E’s “second shot” review process, where reward applicants can respond to reviewers comments and submit revised proposals. This could be especially important for agricultural R&D because reviewers may not be familiar with all of the technical background of a proposal; more than half of citations in patents for new agricultural technologies in the last forty years are of other patents from outside of agriculture. This means that AgARDA will likely draw on insights and experts from a large range of fields, limiting the utility of the traditional, “one-shot” peer review process.

Early on, AgARDA should consider hiring alumni of other ARPA programs to incorporate these and other lessons about organizational structure, partnership formations, and funding approaches.

Representative Technologies

One of the ARPA model’s central strengths is that it doesn’t prespecify specific technologies, so its research focuses can’t be known in advance. However, some general technological directions that may well be part of a future AgARDA portfolio illustrate the potential for advances in food and agriculture technology to revolutionize the industry. Potential technologies include:

• Agricultural disease surveillance
• Nucleic acid monitoring
• Reengineered rubisco for CO₂ efficiency gains
• Engineering synthetic chloroplast genomes
• Artificial photosynthesis
• Microbial-based foods

Many new technologies promise to revolutionize agricultural disease surveillance. Automating and scaling up farmers’ ability to notice diseased crops could be a huge boon to the industry and secure food supplies. Plants produce volatile organic compounds (VOCs) as end metabolites, releasing these into the air at low levels. The composition of VOCs that plants release changes when they are attacked by pests and pathogens. While the reasons for these changes are not fully understood, in some instances they induce insects to preferentially eat infected leaves, and other times VOCs trigger defense mechanisms in neighboring crops. No single VOC is diagnostic of plants suffering infection or herbivory, but many studies have shown that emissions changes are detectable by assessing the relative proportion of different compounds. However, further work is needed to characterize these changes and to create cheap, portable air sampling machinery capable of detecting multiple compounds at once and automatically carrying out multifactor chemical analysis to provide a warning of a plant pest or pathogen. The USDA has supported research into using VOCs for disease detection, but has focused on narrow challenges like detecting a single disease. AgARDA could aim for more ambitious goals such as developing sensors that can detect any type of VOC composition change, potentially indicating the spread of any disease rather than just one. The agency would also work to develop ready-for-market sensors and help commercialize them.

Another set of emerging approaches for disease detection involve monitoring the environment for nucleic acids (RNA and DNA). Remote collection devices can sample air, soil, and water for genetic traces of organisms in the environment and search for the sequences of pests and pathogens. While several techniques have been successfully deployed in ecological monitoring, such as for invasive species, there remain several hurdles in both the physical design of effective sample collection and the processing of genetic information.³ It is increasingly recognized that human disease surveillance systems must be able to detect not just known pathogens but also new and unknown threats. The same principle should be applied to agricultural disease surveillance. This daunting technical challenge requires technology that can search for and identify new nucleic acid sequences without knowing what to look for. However, several approaches have been proposed, and this is the type of longshot, high-reward target that AgARDA would be especially well suited to pursue. The ultimate goal should be to have cheap, highly scalable sampling systems that can collect RNA and DNA from the agricultural environment — air, runoff water, soil, and livestock waste — and automatically search for known and unknown pathogens.

Multiple ambitious engineering approaches to fundamentally improving plant efficiency have been proposed, but these targets remain far off. Plants are not optimized for biomass production, but rather for survival and competition, and this means that many parts of plant physiology could be improved. For example, two protein complexes involved in photosynthesis use light to catalyze reactions, but in terrestrial plants they compete for the same photons while only using half of the available light spectrum. One of these could potentially be replaced with a reaction center found in photosynthetic bacteria that uses a different portion of the light spectrum, doubling the efficiency. Another strategy could be to reengineer rubisco, the most abundant protein on the planet and the enzyme responsible for converting CO₂ into organic compounds. Rubisco does not differentiate well between carbon monoxide, the molecule it is supposed to bind to, and oxygen gas. Plants have to expend some of the energy they get from photosynthesis getting rid of oxygen-bounded rubisco, causing a roughly 25% efficiency loss. A rubisco engineered to reliably differentiate oxygen gas and carbon monoxide (as other enzymes already do) would provide massive agricultural benefits. 

Synthetic biology approaches to plant improvements could have many other advantages. Complex cells acquired the ability to photosynthesize by absorbing a less complex cell that is now the chloroplast. The chloroplast relies on the plant nucleus to code for many of the proteins that it needs to function. Proteins produced elsewhere in the cell have to travel to the chloroplast, an inefficiency exploited by many plant diseases that disrupt chloroplast-nucleus communication. Engineering synthetic chloroplast genomes (plastomes) could therefore both increase crop efficiency and reduce disease susceptibility. These and other similarly fundamental changes to crop physiology hold immense promise, but are far from commercial application and require large concerted research investment to bring from being theoretical to practical. Several initiatives in this arena, such as changing the carbon fixation pathway of rice, have been underway for more than a decade with continual but marginal successes. AgARDA would be able to quickly trial multiple approaches, solicit the creativity of a diverse array of research teams, and quickly pivot to whatever technological option appears most promising, allowing the agency to potentially revolutionize crop production.

Several technologies that could produce food without relying on traditional crops and livestock have attracted significant attention. However, all approaches remain in early stages and face considerable challenges to reaching economic competitiveness with conventional food production. AgARDA could deliver the needed focus and resources to break through the most important bottlenecks. Indoor agriculture, such as vertical farming, is a frequently proposed solution to sustainability challenges, and has attracted significant venture capital investment. However, the energy demands given current technological constraints are massive.⁴ This is largely because photosynthesis is inefficient at turning light, in this case artificial light, into food energy. Recent research, however, has shown that photovoltaic cells can be used to more efficiently create chemical energy by fixing CO₂ into an organic compound that can then be “fed” to plants. This artificial photosynthesis system is at least four times more efficient as biological photosynthesis, and further work could provide efficiency gains that might make indoor agriculture economically viable at scale.

Other non-traditional food production technologies avoid reliance on crops and livestock altogether. One approach, cellular agriculture, grows animal cells in nutrient solutions, thereby providing a biologically identical food to conventional meat. While the unit price of cultivated meat has fallen precipitously, faster than Moore’s law, the technology faces a host of obstacles to cost effective production. These include expensive cell culture media and bioreactors, and the potential need to engineer cells to have a higher metabolic efficiency. Microbial food production, such as precision fermentation where yeast are engineered to produce specific proteins or other desired nutrients, avoids some of these constraints.⁵ Several companies are already bringing microbial-based foods to market. This technology is therefore not as far off as other areas that AgARDA may work on, but the agency could focus on innovating new uses for microbial foods or on production and supply chain logistics. These and other distributed food production technologies should be aggressively pursued to reduce our vulnerability to a wide range of food supply shocks.

One of the important features of the ARPA model is the ability to trial multiple technologies to tackle a given challenge and to pivot to whatever approach demonstrates the greatest promise. This means that the technologies that AgARDA will develop can’t be prescribed ahead of time, and those described above are a non-exhaustive list of potential technological directions intended to illustrate the potential for R&D to provide breakthrough innovations in our food system.

Lessons for AgARDA

The U.S. food and agriculture sector faces four major challenges:

1. America is losing ground to other agricultural producers in its share of the global market and its contribution to innovation.
2. American agriculture is not growing at the rate needed to meet rising global demand.
3. The U.S. R&D ecosystem is both receiving less public funding and is becoming less effective at producing productivity gains.
4. U.S. and global food supplies face multiple threats.

By providing a new funding structure and incorporating expertise from across academia, government, and industry, a fully funded AgARDA could address all of these challenges. If the agency is focused on a few key strategic goals — monitoring for new agricultural disease, drastically improving crop productivity, and pioneering new food production technologies — it can help secure the U.S. against threats to food supplies and restore America’s lead in agricultural production and innovation. To accomplish this, AgARDA should implement lessons learned from other ARPAs in how to successfully innovate and commercialize new technologies. 

Among those lessons: AgARDA should host regular summits that bring together researchers and industry funders (and showcase AgARDA grant winners). It could partner with existing USDA teams and departments that have existing ties with industry stakeholders to promote new technologies. For instance, the USDA’s Emerging Technologies Team (ETT) could work with farmers interested in new remote sensing technologies for agricultural monitoring.

Other government agencies and departments could act as early adopters of new technologies, helping to test and scale innovation. DHS and APHIS partner to monitor U.S. agricultural for pests and pathogens, so they could potentially be an early market for new disease surveillance technology.

AgARDA could also model a program after ARPA-E’s SCALEUP, supporting projects transitioning to successful companies. AgARDA should create a team dedicated to matching private funders to launch new companies or improve the technological portfolio of existing ones, similar to ARPA-E’s Tech-to-Market Team. These activities could be paired with existing USDA efforts to support small businesses, such as the small businesses grants program.

More generally, AgARDA should consciously attempt to diversify the risk level of technologies in its portfolio, with some being ambitious longshots and others being more surefire bets. It should offer prizes for achieving important technical milestones, attracting diverse talent and private funding. Implementing a “second shot” review process would allow applicants for AgARDA funding to respond to reviewer feedback and revise their proposals.

All of these proposals take their cue from existing, successful ARPA-style programs in the federal government, and all of them are eminently feasible if AgARDA is appropriated in the upcoming Farm Bill. If Congress chooses to fund AgARDA, the program will have to move quickly to build a track record of success. Focusing on a few tractable areas for research and mimicking the best features of its predecessors will make that success far more likely.