The most consequential public problems — homelessness, pandemic preparedness, climate adaptation — share a structural feature that makes them nearly impossible to solve through conventional management. No single agency owns them. Authority is fragmented across organizations with different missions, different incentive structures, different cultures, and different accountability chains. The policy designer's challenge isn't just crafting the right intervention. It's architecting coordination among actors who have no obligation to cooperate.

Traditional public management assumes a hierarchy. A minister or director sets priorities, allocates resources, and holds subordinates accountable. But cross-boundary challenges dissolve that architecture. When a housing authority, a mental health agency, a municipal police department, and a nonprofit service provider all touch the same client population, who leads? Who decides? Who is responsible when things fall apart? The default answer — everyone and no one — is precisely the accountability vacuum that produces fragmented, duplicative, and ultimately ineffective service delivery.

Eugene Bardach's work on interagency collaboration established that craftsmanship in implementation requires deliberate design of cooperative structures. Cross-boundary leadership isn't a personality trait or a soft skill. It is a strategic function that must be engineered into the governance architecture itself. This article examines three dimensions of that engineering: finding where coordination actually creates value, building roles that span organizational divides, and constructing accountability frameworks that work when no one is fully in charge.

The instinct when facing fragmented authority is to call for comprehensive coordination — align everyone around a shared vision, build a common data platform, establish a joint governance board. This instinct is mostly wrong. Comprehensive coordination is expensive, politically fraught, and usually unnecessary. The strategic question isn't how to get all agencies to work together on everything. It's how to find the specific points where coordination creates enough value to justify the costs of collaboration.

Integration points are moments in a service delivery chain or policy process where the actions of one organization materially affect the outcomes of another. Consider a reentry program for formerly incarcerated individuals. The corrections department controls release planning. The housing authority controls shelter placement. The workforce development agency controls job training access. Each operates independently, but the sequencing of their interventions matters enormously. A person released without housing lined up is unlikely to show up for job training. The integration point isn't a grand unified reentry strategy. It's the forty-eight hours around release, where a handoff protocol between three agencies determines whether the entire system works.

Finding these leverage points requires what Bardach called a bottom-up mapping of operational interdependencies. You don't start with org charts. You start with the client experience or the problem trajectory, and you trace where one agency's output becomes another's input. This is analytical work, not political negotiation — though the findings inevitably have political implications, because they reveal whose cooperation is essential and whose is merely nice to have.

Once identified, integration points become the basis for targeted agreements rather than sprawling memoranda of understanding that no one reads. A well-specified integration protocol — who provides what information, by when, in what format, with what fallback if they don't — is far more valuable than a joint mission statement. It makes cooperation concrete, monitorable, and enforceable at the operational level where results are actually produced.

The discipline here is strategic selectivity. Not all fragmentation is a problem. Some organizational boundaries exist for good reasons — legal mandates, specialized expertise, political accountability. The goal is not to eliminate silos but to build bridges at the precise points where silos create value destruction. Every integration point you choose to manage is a point you must sustain with political capital and management attention. Choose poorly, and you exhaust your collaborative capacity on low-value connections while the critical handoffs remain broken.

Takeaway

Don't try to coordinate everything. Map where one agency's failure directly causes another agency's failure, and focus your collaborative capital on those specific handoff points.

Identifying integration points solves the where problem. The who problem is harder. Cross-boundary coordination doesn't happen by itself, even when everyone agrees it should. Organizations revert to their own priorities under operational pressure. Meetings get cancelled. Data sharing stalls. Someone has to make coordination their actual job — not their extra responsibility layered on top of a full portfolio, but the core of their professional role and identity.

Boundary-spanning roles come in several structural forms, each with different strengths. Formal liaison positions — embedded in one agency but with defined relationships to others — work well for stable, ongoing coordination needs. Joint program offices with shared staffing and blended funding work for high-priority initiatives where political leaders have signaled real commitment. Less formal arrangements — communities of practice, cross-agency working groups, secondment rotations — serve as lower-cost options for building relational infrastructure before formalizing structures.

The critical design variable is dual accountability. A boundary spanner who reports only to their home agency will inevitably be pulled back into that agency's priorities when resources tighten. But a boundary spanner who reports to a coordinating body with no operational authority becomes a meeting facilitator with no leverage. The most effective models create hybrid accountability: the boundary spanner has a home agency that provides career security and operational legitimacy, but their performance evaluation explicitly includes metrics defined by partner organizations. This is uncomfortable by design. The tension is productive.

Equally important is the informal dimension. Research on collaborative governance consistently finds that personal relationships and trust networks are the actual infrastructure of cross-boundary work. Formal roles create the permission and the space for these relationships to develop, but they don't guarantee it. Smart governance designers invest in relationship-building opportunities that feel mundane — shared training programs, co-located office space, joint problem-solving exercises on low-stakes issues — because these create the social capital that makes high-stakes coordination possible later.

One underappreciated design choice is where boundary spanners sit in their organizations' hierarchies. Too senior, and they lack operational knowledge of where coordination breaks down. Too junior, and they lack the authority to compel their own agency to change behavior. The sweet spot is typically at the program management level — senior enough to make commitments, operational enough to know what commitments are realistic. Multiple boundary spanners across different levels create a richer coordination architecture, but they must be connected to each other to avoid simply replicating the fragmentation problem within the collaborative structure itself.

Takeaway

Coordination requires dedicated roles, not good intentions. Design boundary-spanning positions with hybrid accountability — grounded in one agency for legitimacy, but evaluated by partners for collaborative results.

Here is the governance paradox that haunts every cross-boundary initiative: the problems that most require collaboration are the problems where accountability is hardest to establish. When a homeless individual cycles through emergency rooms, shelters, and jail without receiving sustained intervention, who failed? The health system? The housing authority? The criminal justice system? Each can point to its own mandate and argue — often correctly — that it performed its defined function. The system failure belongs to no one, which means no one is held responsible, which means no one has an incentive to fix it.

Traditional accountability runs vertically — agencies report up to elected officials through performance metrics and budget processes. Cross-boundary work requires horizontal accountability mechanisms that don't naturally exist in public governance architectures. Building them is one of the most sophisticated design challenges in public management, because you're creating obligations between organizations that have no formal authority over each other.

The most effective approach is shared outcome metrics with disaggregated contribution tracking. Partner organizations jointly own a population-level outcome — say, a reduction in chronic homelessness in a metropolitan area. But the joint commitment is supplemented by a clear specification of what each organization contributes to that outcome: the housing authority commits to a certain number of permanent supportive housing units, the health system commits to mobile mental health teams, the workforce agency commits to supported employment slots. This structure preserves individual organizational accountability while making the interdependence explicit and visible.

Public reporting is the enforcement mechanism. When outcome data is published regularly and attributed to the collaborative — not to any single agency — political leaders and the public can see whether the system is working. Individual agencies that underperform their commitments face reputational pressure from partners, not just from their own hierarchical chains. This is softer than statutory enforcement, but in practice it's surprisingly powerful, because senior public managers care deeply about their standing among peer agencies whose cooperation they need for other purposes.

The deeper insight is that perfect accountability in cross-boundary work is impossible, and pursuing it is counterproductive. Overly rigid accountability frameworks discourage risk-taking and adaptation. They incentivize agencies to protect themselves rather than solve problems. The goal is not to eliminate accountability gaps but to make them small enough and visible enough that they motivate cooperation rather than blame-shifting. This requires a tolerance for ambiguity that is foreign to most public audit cultures — and a leadership willingness to defend collaborative structures against demands for tidy attribution of responsibility.

Takeaway

When no one owns the whole problem, design shared outcome metrics that make interdependence visible — but resist the urge to assign blame with false precision. Productive accountability in collaborative governance is about transparency, not control.

Cross-boundary leadership is not a charisma problem. It is an architecture problem. The leaders who achieve coordination across fragmented authority do so because they design structures that make cooperation rational — targeted integration points, dedicated boundary-spanning roles, and accountability mechanisms calibrated to the reality of shared responsibility.

The recurring temptation is to treat collaboration as a cultural aspiration rather than an engineering challenge. But goodwill dissipates under operational pressure. Only designed-in incentives, roles, and information flows sustain coordination when agencies face competing demands on their time and resources.

The strategic discipline is knowing that you cannot integrate everything. You choose your integration points, invest in the connective roles that sustain them, and build accountability frameworks honest enough to tolerate ambiguity while transparent enough to motivate action. That is the craft of cross-boundary governance.