The Coordination Crisis: Why Multi-Agency Strategies Collapse Under Pressure

In May 2026, the Gulf Research Program awarded over $1.43 million through its ACT Initiative across fifteen community resilience projects on the Gulf Coast. Among them was a grant to the United Way of Greater Houston, explicitly titled "From Silos to Community-Centered Systems: Transforming Disaster Data and Communications Regionwide" — a project designed to break the data and communication barriers between government agencies, non-governmental organizations, and private sector actors that had repeatedly produced coordination failures during disaster response.

The grant's premise was precise: the participating organizations broadly agreed on the goal of coordinated disaster preparedness. They had never built the architecture to pursue it together.

This is the coordination problem that most multi-agency strategies never name directly. The planning process that convenes government, NGO, and private sector partners — or the pharmaceutical strategy session that aligns commercial, medical, market access, and patient support — produces a shared document and a shared sense of direction. What it rarely produces is the operational specification of how those parties will actually make joint decisions, share real-time information, and escalate conflicts under the conditions of genuine urgency. Those questions are treated as implementation details. They are the strategy.

Why shared intent is not enough

Laurence O'Toole and Robert Montjoy's foundational research on interorganizational policy implementation, published in the mid-1980s and subsequently validated across dozens of multi-agency settings, established the core finding that the field of strategic coordination has consistently failed to operationalize: the probability of effective joint action decreases predictably with each organizational boundary a decision must cross, and the rate of decrease is nonlinear.

Adding a third actor to a two-party coordination system does not increase coordination complexity by fifty percent. It can double or triple it, depending on how the decision authority among the three parties is structured — or, more commonly, left unstructured.

The mechanism O'Toole and Montjoy identified is not primarily one of misaligned incentives, though those exist. It is one of structural ambiguity: in the absence of pre-specified decision rules, each organization defaults to its own authority structure, its own reporting cadence, and its own definition of what constitutes a decision that requires escalation versus one that can be made locally.

When urgency compresses the timeline — when the hurricane is approaching or the product launch is two weeks away — the absence of those pre-specifications means that the parties who genuinely want to work together spend the available time negotiating the governance of their coordination rather than executing it. The coordination failure is not a failure of will. It is a failure of architecture.

The cognition problem under pressure

Louise Comfort's research on crisis management — developed across two decades of studying emergency response systems from the Northridge earthquake to post-Katrina Gulf Coast coordination — adds a critical dimension that pure governance analysis tends to underweight. Under crisis conditions, the cognitive demands on individual decision-makers escalate precisely when the organizational infrastructure for distributed cognition is most needed and most likely to be absent.

Comfort's framework identifies four interdependent requirements for effective multi-organizational crisis response: cognition (shared situational awareness), communication (reliable information flow across organizational boundaries), coordination (aligned decision-making), and control (clear authority structure). These four elements are mutually reinforcing when they are present and mutually degrading when they are absent. An organization that lacks shared situational awareness will make communication decisions that further distort it. Distorted communication degrades coordination. Degraded coordination produces authority conflicts that undermine control.

What Comfort's research makes visible is that the failure of multi-agency coordination under pressure is rarely attributable to a single missing element. It is a cascade — and it begins, almost invariably, with the absence of shared situational awareness. Organizations that have never agreed on a common operating picture of the challenge they are jointly addressing cannot build one in real time.

The data systems are incompatible. The definitions of key metrics differ. The reporting cadences do not align. The actors who need to share information do not have the established relationships required to share it quickly and trust its accuracy. These are not problems that emerge in crises. They are problems that were present before the crisis began and that the pressure of the crisis simply makes undeniable.

What the planning session needs to produce

The structural implication is one that most strategy processes resist because it adds apparent friction to a process that is already difficult: the multi-agency strategy session cannot conclude without producing an explicit coordination architecture, agreed upon by the actors who will need to operate within it. This is not the same as producing a strategy document. A strategy document describes what the parties intend to accomplish together.

A coordination architecture specifies how they will make decisions together when the situation is not going according to the strategy — who has authority to call a halt, what data each party commits to providing and at what cadence, how conflicts between organizational priorities will be escalated and by whom, and what the accountability mechanism looks like when a party cannot meet its commitments.

The organizations that escape the coordination crisis are not those with the most detailed strategy documents. They are the ones that have invested the facilitated time — before the crisis, before the launch, before the implementation pressure begins — in answering the questions that are too uncomfortable to answer in calm conditions and too urgent to answer under pressure.

The ACT Initiative's framing of this problem is correct and rare: the gap is not between organizations that want to coordinate and organizations that do not. It is between organizations that built the architecture before they needed it and organizations that assumed the intent would be sufficient. It never is.