The empirical development of
the Emergent Coordination Model
and the Agreement Accelerator

The Agreement Accelerator was not designed in the abstract. It emerged through iterative action research across four very different environments.

Since 2012, Sovereignty First has conducted four field tests—in Indonesia, Syria, Afghanistan, and Liberia. Each tested different elements of the theory and infrastructure.

These tests provide preliminary evidence. They do not constitute a controlled scientific demonstration. That is the next step.

INDONESIA · 2012

Question
Could our cooperative-capacity methodology improve the effectiveness of a troubled three-country, $60 million forest-preservation demonstration project?

Situation
The project had three competing lines of authority, little shared understanding of its purpose or definition of success, no adequate implementation plan, strained institutional relationships, growing political controversy, and a serious threat of cancellation.

Intervention
We helped the leadership establish a single line of authority and shared vision, develop an implementation plan, repair important institutional relationships, and build support within Indonesian and Australian ministries.

Result
The internal intervention largely achieved what it was designed to achieve. The project began producing and publishing results from the environmental, financial, and political intervention it had been established to test.

Then the external environment changed. A new Australian government sharply reduced foreign aid, and the project was cancelled.

What We Learned

The operation was a success. The patient died.

Our methodology had successfully translated from American businesses and partnerships to a large international collaboraiton. But we had treated the project’s organizational capacity as the principal probem.

It wasn't enough.

The project's survival depended upon a larger network of influential actors whose support the project could not command. We had not systematically identified and cultivated that entire network.

What Remains Unknown
Would systematically identifying and building relationships with every consequential actor have changed the outcome? Could a sufficiently broad network of support have made the project more resilient to a major political change?

We don't know.

SYRIA · 2015–2016

Question

Could we recruit a diverse group of influential actors involved in the Syrian conflict into a recurring information-sharing process, publicly report their aggregated perspectives in Arabic and English, and thereby increase shared understanding, trust, and the formation of peacebuilding agreements?

Situation

Syria was years into a civil war involving numerous domestic armed groups, regional powers, international governments, NGOs and other institutions. The information environment was extraordinarily opaque.

Intervention

We began with a desk review and deliberately prioritized recruiting non-state armed actors, whom we expected would be among the hardest influential actors to reach.

We then conducted three rounds of interviews over eighteen months, aggregating and publishing the results in Arabic and English.

What Happened

Approximately two dozen influential actors participated, including non-state armed actors, government foreign ministries and NGOs.

The interviews revealed something we had underestimated: participants weren't simply interpreting the same situation differently. In many cases, they appeared to be describing different realities.

The resulting reports made these differences visible.

However, we do not know how extensively the reports circulated beyond the participants themselves.

What We Learned

Syria provided preliminary evidence that even highly divided and difficult-to-reach actors could be recruited into a recurring shared-information process.

But recruitment and information production weren't enough.

We made the information publicly available through a web portal. Influential actors could pull the information if they chose to seek it out. We had not yet developed a mechanism to push relevant information directly to the people positioned to use it.

The experiment therefore exposed a deeper problem of common understanding: How do you get influential actors to attend to and understand the perspectives of other influential actors—especially when those perspectives differ substantially from their own?

What Remains Unknown

Could the process recruit an even broader range of consequential actors—including financial institutions, intelligence organizations and other difficult-to-access institutions?

Would repeated participation and exposure to aggregated perspectives actually increase shared understanding?

Would increased shared understanding contribute to greater trust?

And, ultimately, would those changes increase the formation and durability of peacebuilding agreements?

We don't know.

AFGHANISTAN · 2021

Question

Could we establish Agreement Accelerators in ten villages across four Afghan provinces? If so, could they strengthen inclusion, common understanding, and trust—and increase the emergence and persistence of self-governing agreements?

Situation

Afghanistan was twenty years into a foreign-led reconstruction effort. The national government remained weak and heavily dependent on foreign support. Decades of war, civil war, internal displacement, and returning refugee populations had disrupted communities and their institutions.

Intervention

We trained our Afghan partners remotely by videoconference. They then recruited and trained local interviewers and facilitators as staff.

Our staff conducted recurring rounds of interviews with influential actors and published the aggregated information on a password-protected web portal.

We then invited influential actors assessed by their peers as highly committed to their villages to participate in structured dialogues. Two facilitators guided participants through a sequence of questions informed by Elinor Ostrom's Core Design Principles, helping them identify shared problems, develop solutions, and craft self-governing agreements.

What Happened

The Agreement Accelerators completed two rounds in some villages and three in others. Interview participation was strong. Participants engaged in the structured dialogues and began identifying problems, developing solutions, and crafting agreements.

Several of these problems had remained unresolved for a decade or more.

Within a few months, a district governor asked us to expand the process from individual villages to the district level.

Then the political environment changed dramatically.

As the United States withdrew and the Taliban advanced, our Afghan partners reached out to religious leaders whom the Taliban respected. We explained the Agreement Accelerator and sought permission to continue operating.

After the Afghan government fell and the U.S. evacuation was completed, we received permission to continue in one province and were negotiating permission in two others.

Our funder, the U.S. Institute of Peace, supported continuing the work. But the Biden administration subsequently ordered U.S. government-funded work in Afghanistan to stop, ending our program.

What We Learned

First, the Agreement Accelerator showed preliminary evidence of doing what it was designed to do.

Influential local actors participated voluntarily in the process and began working together on problems that had remained unresolved for years. They moved beyond information-sharing into problem selection, solution development, and the formation of self-governing agreements.

This did not establish that the process strengthened inclusion, common understanding, or trust, or that the agreements would persist. The intervention was terminated too early to establish those claims.

But it provided preliminary evidence that the infrastructure could generate the behavior it had been designed to support: Influential actors voluntarily forming agreements to address shared problems.

Second, we encountered the boundary problem again.

Unlike Indonesia, we had deliberately cultivated support outside the immediate intervention. When the governing environment changed radically, our Afghan partners successfully engaged actors influential with the incoming Taliban government and obtained or pursued permission to continue.

But we had not adequately mapped the sovereignty structure of the funding system.

Our managers at USIP supported continuation. USIP itself, however, did not possess final authority over whether U.S. government-funded work could continue in Afghanistan.

Once again, the survival of the intervention depended upon a consequential actor outside the system we had adequately incorporated.

What Remains Unknown

Would the Agreement Accelerator continue producing agreements over multiple cycles?

Would those agreements persist?

Would repeated cycles measurably strengthen inclusion, common understanding, and trust?

Could the process operate effectively above the village level—at district, provincial, national, or larger scales?

Could Agreement Accelerators be funded and governed in ways that make the infrastructure less vulnerable to unilateral termination by a single external funder or political authority?

We don't know.

LIBERIA · 2022–2024

Question

Could we establish Agreement Accelerators across three counties and nine towns in northeastern Liberia? If so, could they strengthen inclusion, common understanding, and trust—and increase the emergence and persistence of self-governing agreements?

Situation

Liberia had been rebuilding since the end of its second civil war in 2003, following roughly a quarter-century of civil conflict.

National, county, and municipal governments had limited capacity to provide security, stability, and basic services. Economic opportunities were scarce, particularly outside agriculture, while a young and growing population increasingly sought other livelihoods.

At the same time, Liberians were continuing the long process of building a national society across sixteen ethnic groups and the country's historical divisions.

Intervention

The program had three phases.

First, we offered intensive psychosocial training designed to strengthen emotional self-regulation, interpersonal capacity, and the ability to plan for the future. Hundreds of young people participated in fifty-hour, five-week trainings.

Second, we trained approximately one hundred young people in Elinor Ostrom's Core Design Principles for crafting self-governing agreements.

Third, we began establishing Agreement Accelerators.

What Happened

We had expanded into six towns across two counties. The work was struggling in all six.

Local leaders offered passive support. We had also made a consequential design mistake: Participants received modest transportation stipends to attend the trainings.

Those payments gradually became income streams for some families. We could no longer tell whether participants were attending because they valued the training or because they were being paid to attend. When we eliminated the payments, most stopped coming.

Meanwhile, our Liberian partners had been telling us for roughly two years to work in what they called communities.

Liberian towns contain dozens of communities, typically comprising roughly 500–1,000 people and organized around a small number of extended families. Communities have established leadership structures, including leaders representing men, women, and youth.

We changed course.

We largely stopped organizing the intervention and infrastructure through towns and began sending teams directly into communities. Anyone who wanted the training could participate. We began establishing Agreement Accelerators there.

And we stopped paying participants.

The response changed quickly. Communities began requesting the program, and we developed a waiting list.

Seven months into this redesigned approach, the Trump administration dismantled USAID, which had funded the program, and our funding ended.

What We Learned

First, financial incentives can corrupt the signal we most need to observe.

If people are paid to participate, participation itself no longer tells us whether they value the infrastructure. Eliminating payments gave us a much cleaner signal: communities were voluntarily requesting the work.

But the larger discovery was more important.

We had been operating at the wrong social scale.

The formally recognized towns, cities and counties did not necessarily possess the local governing capacity required for self-governing agreements to emerge and persist.

The communities often did.

Their governance arrangements were less formally institutionalized but appeared more socially embedded: networks of families and established leaders through which people had likely been organizing collective life across generations.

This suggested a fourth necessary condition:

Functional sovereignty—the practical capacity of a group to govern the matters over which it is attempting to reach agreements.

Inclusion, shared understanding, and trust may not be enough. Self-governing agreements also require a social structure capable of authorizing, implementing, maintaining, and adapting them.

What Remains Unknown

When a social system has sufficient functional sovereignty, inclusion, shared understanding, and trust, will self-governing agreements emerge and persist at measurably higher rates?

(The accumulated field evidence is suggestive. It is not sufficient to establish the claim.)

Can the four conditions be deliberately and reliably strengthened?

How much strengthening is enough?

Does strengthening them cause more agreements to emerge—or are the observed associations explained by other factors?

Do agreements formed through the process persist?

Can the process scale from communities to larger systems containing multiple centers of functional sovereignty?

And again:

Can Agreement Accelerators be funded in ways that do not leave the infrastructure vulnerable to the unilateral decisions of a single external institution?

We don't know.

Thirty Years of Development.
Four Field Tests.
One Unanswered Question.

Functional Sovereignty: Can participants actually govern the domain in question?

Inclusion: Are the consequential actors sufficiently represented?

Shared Understanding: Do the actors understand enough of the system—and one another's perspectives—to coordinate?

Trust: Are relationships sufficient to support cooperation despite disagreement and uncertainty?:

Across four field environments, we accumulated evidence that these conditions matter, and developed infrastructure intended to strengthen them.

We have not demonstrated that strengthening these four conditions causes self-governing agreements to emerge and persist.

That is the next question.