Atlas of Innovation

Definition

A coordinated research program is an institutional funding framework designed to accelerate high-risk, high-reward research and overcome complex innovation challenges through active program management. The most well-known coordinated research programs are based on the example of the Advanced Research Project Agency (ARPA) model. Funders give program managers the autonomy and resources to coordinate a portfolio of research projects aimed at achieving transformative breakthroughs in pursuit of a clearly defined objective. At their best, programs typically focus on a single priority innovation area with a specific end user and often engage teams across government, industry, and academia.

Unlike most innovation funding approaches, coordinated research programs are not a single funding instrument. Instead, they can combine funding approaches such as grants, R&D contracts, and prizes in pursuit of a shared goal. Coordinated research programs stand out in some situations over individual tools alone due to coordination and flexibility.

Why might a coordinated research program be the right funding approach?

A coordinated research program works best for complex, high-risk, high-reward challenges that cross disciplines or require coordinated advances across multiple technologies, especially where other funding approaches struggle due to fragmentation. Program managers can balance centralization and decentralization of decision-making by picking different funding approaches and coordinating advances across disciplines and technologies. For disciplines and technologies where the program manager is less informed relative to performers, they can rely on decentralized approaches for downstream decisions to allow multiple teams to pursue diverse approaches, adapt quickly to new information, and apply specialized expertise to more narrow technical challenges.

A coordinated research program offers the following advantages:

Advancing “stacked” technologies: Coordinated research programs are particularly advantageous when they fund a portfolio of projects that together advance a group of complementary technologies, often described as a “technology stack.”.These technologies do not automatically fit together. Coordinated research programs can help these technologies fit together, while tailoring funding mechanisms for each. For example, a single program might fund university grants for early biotechnology research, support firms through R&D contracts to develop enabling software, provide targeted grants to support the development of shared standards, and use procurement contracts to build manufacturing capacity for resulting products.
Bridging the gap between basic research and private investment: Key technical advances often require later-stage research and tailoring for end-users (e.g., new crop varieties often need to be tailored to specific taste and aesthetic preferences to gain adoption). Coordinated research programs can help plug the remaining holes needed to make an innovation ready for commercial markets.
Avoiding consensus bias: Coordinated research programs empower program managers to pursue high-risk, high-reward ideas that consensus-based funding processes often avoid. This structure allows program managers to strategically choose a portfolio of technological approaches. They also have the authority to shut down projects that miss key milestones or have been outpaced by other research, which lets them take a more dynamic and responsive approach than conventional funding models.

What can go wrong?

Several issues can undermine the effectiveness of coordinated research programs if not carefully managed:

Coordination overhead: Coordinated research programs are most effective when focused on problems that require sustained coordination across actors, technical domains, and stages of development. When a problem can be addressed through a single grant or a small number of independent projects, coordinated programs add little value and introduce unnecessary overhead. Incremental projects underutilize key advantages: highly capable staff and flexible contracting authority.
Poorly chosen program managers: Effective program managers require an unusual combination of technical expertise, project management skills, a clear sense of the innovation area’s scope, and sound judgement regarding what is ambitious yet feasible within a typical 2-6 year program. Managers must also be willing to cut underperforming projects, as reluctance to do so undermines the value of coordination. Without these capabitlies, program managers may fund projects that are trivial or unrealistic.
Weak alignment with end use: If the prototype or technology developed under a program’s oversight fails to align with the needs of future commercial partners or end users, it risks being shelved despite technical success.
Challenges with time horizons: The fixed medium-term timeline of most coordinated research programs (2-6 years) creates a fundamental mismatch with certain types of research. The typical program duration is too short for breakthrough scientific research that often requires greater than 6 years of development.

To mitigate these challenges, programs can develop clear criteria for identifying problems that require coordination, recruit program managers with proven technical judgment and vision, coordinate with potential end-users early in the program design, and carefully assess whether research timelines align with the medium-term structure prior to launch.

Examples

Several agencies in the US and abroad have adopted the coordinated research program model. The defining example of an ARPA-style program the original at the Department of Defense:

Defense Advanced Research Projects Agency: Founded as ARPA in 1958, DARPA is the oldest and best-funded of the ARPA agencies. It invests in emerging technologies across many fields with potential applications for national security for the Department of Defense. DARPA has helped drive innovations such as the internet, GPS, and aircraft stealth technology. Programs are typically short-lived, and project managers serve in two-year terms, limited to six years.

For example, in the late 1960s, DARPA set out to connect researchers across the country through a new kind of computer network. ARPA funded research institutions and a small private firm. A year later, ARPANET connected four research institution sites as the precursor to the modern internet.

In the early 1970s, DARPA program managers pursued radar signature reduction by partnering with aerospace companies outside the industry’s leading contractors. While leading contractors focused on established aircraft technologies, other companies were more willing to take risks on unconventional ideas. DARPA leveraged this dynamic to fund advances across the entire technology stack, from new materials and electromagnetic modeling to new manufacturing techniques. The program succeeded by maintaining a clear military focus while allowing flexibility regarding which technologies were funded.

The coordinated research program has proliferated across the US government and abroad. While each instantiation adapts to different missions, many core features remain:

Advanced Research Projects Agency – Energy: As an agency within the Department of Energy, ARPA-E invests in high-risk, high-reward energy innovations to promote “economic security, national security, and environmental well-being.” Program managers serve renewable three-year terms. Since its launch in 2009, ARPA-E has invested over $4 billion in projects ranging from grid-scale energy storage to carbon capture, with a focus on bringing new technologies to market. For example, ARPA-E’s focus on speculative clean energy solutions enabled breakthroughs in battery technologies that were later commercialized for wider use.
Advanced Research and Invention Agency: Founded in 2023, ARIA is the United Kingdom’s agency for high-risk, high-reward research across a broad range of fields, including artificial intelligence and biology. ARIA’s £59 million Safeguarded AI program brings together expertise from computer science and philosophy to develop tools to make frontier AI models safer. Unlike US programs, ARIA’s remit is not focused on one topic and has greater autonomy (e.g., through exemptions from the UK’s standard contracting regulations that require broader solicitation).
Additional US government and international examples: The US has implemented this model across various agencies, authorizing AgARDA for agricultural technologies, ARPA-H for health, ARPA-I for infrastructure, IARPA for the intelligence community, and HSARPA for homeland security. Many proposals for new ARPA-style programs across the US government hope to replicate the successes of DARPA, but many focus areas lack conditions that make for effective ARPA-style programs. In addition to ARIA in the UK, the European Union founded the Joint European Disruptive Initiative (JEDI), and Germany has established the Federal Agency for Disruptive Innovation (SPRIN-D).

Program managers at ARPA-style programs use a variety of funding approaches to disburse funding to researchers, including:

[[Research Grants]]: Program managers use research grants to fund universities, nonprofits, and private companies. For example, ARPA-E funds university researchers to simulate nuclear fusion dynamics.
[[Milestone Payments]]: Coordinated research programs can tie payments to the achievement of specific performance benchmarks. For example, ARPA-H’s Healthcare Rewards to Achieve Improved Outcomes (HEROES) program offers payments to providers who meet specific care improvement benchmarks across several priority health areas, including heart health and addiction.
[[Prizes]]: Prize-based mechanisms allow ARPA agencies to crowdsource innovation. For example, DARPA ran a “Cyber Grand Challenge” prize competition to help develop automatic defense systems that can detect and correct software flaws.
[[Government-research lab ventures]]: Like government-research lab joint ventures, coordinated research programs can involve coordination across government, academia, and the private sector. However, coordinated research programs involve one lead program manager for a specific program, whereas government-research lab joint ventures require broader oversight.
[[Intramural science]]: Intramural programs rely on in-house government researchers and infrastructure. They are most useful when research involves sensitive data, requires large fixed investments, and/or demands close alignment with policy goals. In contrast, coordinated research programs are more effective for high-risk, high-reward and time-limited challenges rather than longer-term research goals.
[[Government-research lab joint ventures]]: Joint ventures bring together government and private-sector capabilities, particularly in areas that require close collaboration with industry. For example, NOAA’s work with SpaceX on satellite collision avoidance leverages both federal infrastructure and commercial capabilities. While they facilitate resource sharing and co-development, they often require more complex coordination. Coordinated research programs offer a more centralized, agile alternative.

Coordinated Research Program (aka ARPA model)

Definition

Why might a coordinated research program be the right funding approach?

What can go wrong?

Examples

Further reading

Definition

Why might a coordinated research program be the right funding approach?

What can go wrong?

Examples

Related funding approaches

Further reading