The System Drives the Behavior

Leadership Capability and the System That Constrains It

Edith A. Hughes, D.Sc.  ·  Adaptive Value Design LLC  ·  2026

I. The Case for a Whole-System View

Federal transformation frameworks are calling for the right changes. Adopting product thinking, aligning teams to value streams, and empowering cross-functional delivery are the practices a government built for a digital era requires. Where the opportunity lies is in taking a whole-system view — one that examines not only the digital-era practices to adopt, but the conditions across the operating model that determine whether those changes can take hold, scale, and sustain. Those conditions include governance structures, performance management incentives aligned to mission outcomes, and the organizational design that determines how work actually flows.

One part of that whole-system view is leadership capability. Initiatives like Tech Force are actively addressing a real and urgent need: building the technical workforce of software engineers, data scientists, and hands-on delivery practitioners that modern government services require. That investment is necessary and valuable. But the workforce that executes delivery also needs leaders who know how to build the organizational conditions in which that talent can flourish — leaders who understand how to design for flow, align authority to mission domains, and remove the structural friction that technical talent alone cannot overcome. Developing that leadership capability, and designing the operating model that enables it, is the complementary investment that makes the technical workforce investment pay off.

The federal government has many sophisticated, experienced leaders who understand what good looks like in a digital-era organization. Attracting and developing more of that capability — and enabling those leaders to have sustainable impact — is precisely what the system needs to encourage. This article examines how to design the governance structures, incentive frameworks, and organizational design that reinforce the behaviors transformation requires.

When federal executives do not define value together, the explanation is rarely a lack of capability. The more consequential question is whether the current system creates the conditions that make that work possible — whether the norms expect it, the incentive structures reward it, the development pathways build toward it, and the organizational calendar makes room for it. Designing a system that does all four is what makes collective leadership behavior the norm rather than the exception.

Article 1 establishes what leaders need to do before a product operating model can work: define value collectively, map their customers, and develop a shared understanding of the system they jointly manage. This article examines the system conditions that make that foundational work possible, and what changes are needed to make it the norm.

II. Two Agencies, Two Outcomes

The contrast between two federal agencies illustrates the argument more clearly than any structural diagnosis can.

Several years ago, the Census Bureau faced a leadership challenge that many federal agencies encounter but rarely address directly: a senior executive team that had not developed a shared understanding of the problem space they collectively owned. Individual leaders were capable and committed. The enterprise picture — the full scope of what the agency was trying to deliver, who its customers were, and how value flowed across organizational boundaries — was not complete. It existed primarily as a linear sequence of steps referred to as the Survey Development Life Cycle.

The head of the Bureau made a deliberate decision: he directed his group of executive leaders to work as a team to address that gap, and he backed that directive with what it required: time, budget, and permission to bring in external facilitation. That combination matters. Even capable, motivated leaders needed to be explicitly asked. Without protected time and resources, day-to-day work fills every available hour. The system had not asked them, and it had not resourced them. Their director did both. He also encouraged them to think not in terms of how the bureau should work, but how best to meet customer needs and desires. He encouraged innovation, taking risks, and thinking differently.

What followed was a series of structured full-day working sessions, held bi-weekly over five months, enabling a sustained but not all-consuming investment of leadership attention. The cadence was deliberate: sessions close enough together to build momentum, spread out enough to allow reflection between them. The executive team was not briefed on a framework someone else had built. They built it together. They jointly defined the problem space, mapped their customers across all three categories, and developed a shared understanding of the system they collectively managed. The word “jointly” is doing important work here. The goal was shared understanding, not distributed slides.

The engagement worked. The executive team developed the cohesion and shared language needed to move forward. They performed the core work that makes operating model improvement possible: defining value, understanding customers, and mapping the system. It became the foundation for what came next.

As Steve Jost, co-chair of the Census Bureau’s Data Dissemination Task Force, later reflected, the engagement ventilated the siloed activities that had prevented enterprise thinking. Two legacies from the work are documented and still visible today. The first is organizational: the Center for Enterprise Dissemination, an enterprise-wide solution that modernized the Census Bureau’s approach to data delivery.¹ The second is cultural: a Bureau-wide Transformation Initiative that a decade later the agency’s own leadership described in terms that echo what the foundational work made possible — breaking down silos, thinking collaboratively, delivering value to their data users together.² The ventilation held.

A decade later, a large federal agency was embarking on a far more ambitious transformation. A historic funding influx presented real opportunity. McKinsey had proposed a “product and platform” operating model — a consulting-firm framework distinct from the practitioner-grounded body of work that IT Revolution has built from documented enterprise transformations, starting with Mark Schwartz and Jason Cox’s work on Thinking Environments³ and evolving through the research behind Accelerate⁴, Team Topologies⁵, and related disciplines. The language was sophisticated. The ambition was real.

But no one fully understood what the model terminology meant or what foundational practices the model required — the Agile principles, the project-to-product shift, the team topology concepts that give a product and platform operating model its substance and its scale. So implementation defaulted to local interpretation rather than shared understanding. The CIO stepped in, and a fundamentally business transformation became an IT initiative.

The Deputy CIO understood what changes were needed. Working with the Senior Technical Advisor on the business side, he developed a clear-eyed diagnosis of what steps the transformation required. But that understanding never reached the agency head or the broader C-suite. The insights existed, but the organizational culture made it difficult to surface them to where they could produce a decision. That culture does not emerge in a vacuum — it is shaped over time by what the system rewards, what it tolerates, and what it makes safe to say out loud. A system that offers no incentive for risk-taking and no pathway beyond project management will produce leaders who stay within those boundaries. The culture reflects the system that formed it. This story is not about bad leaders. It is a story of a leadership team that was never given the conditions, the design, or the opportunity to do the foundational work together.

The gap between the two agencies reflects more than a difference in conditions. It reflects a difference in moment. In 2013, the engagement design that made the Census Bureau change possible drew on emerging business design practices that were at the frontier of what large organizations were beginning to apply. The consulting-firm model that the other federal agency was attempting to adopt in 2023 had only been formally introduced in 2016. By 2023, leading practice thinking had already continued to evolve well beyond it, building on lessons from hundreds of enterprise transformations. Federal government was a decade behind the leading edge. Without the foundational work needed to make a large-scale adoption of any model viable, the gap was difficult to close.

III. Vertical by Design

The Census Bureau story is the exception. Understanding why requires looking at how federal agencies are structured, not as they appear on an org chart, but as they function in practice.

Federal agencies are organized vertically. Senior leaders are appointed to run functions: a program office, a technology organization, a financial management division, a human capital function. Each has a defined lane, a budget, a set of statutory authorities, and a performance framework oriented around that lane. The system is designed for vertical accountability. It is not designed for horizontal alignment.

The flaw is not in the people who lead these organizations. It is a feature of how the organizations were built. Most federal organizational structures were built in an earlier economic era, when industrial-age management principles of specialization, hierarchy, and functional control were the dominant model for large institutions. Agencies evolved over decades in response to statutory mandates, oversight requirements, and management frameworks that rewarded functional excellence over enterprise coherence. A chief financial officer who manages the budget well is doing the job. A chief information officer who delivers technology on time and within cost is doing the job. The job, as defined, does not include sitting together to jointly define what value the enterprise exists to deliver.

The result is a leadership culture in which cross-functional collaboration at the executive level is the exception rather than the expectation. When it happens, it typically happens informally, through relationships built over time, or in response to a crisis that forces coordination. It is rarely designed into the operating model as a structural requirement. And without that design, it depends entirely on the individual initiative of leaders who are already fully occupied managing their own domains.

IV. A Performance Architecture Built for Functions

The structural reality described in Section III is not accidental. It is, in significant part, a product of the legal and policy frameworks that govern how federal executives are managed and evaluated.

The Senior Executive Service was established to create a cadre of senior leaders capable of moving across agencies and managing complex government programs with a government-wide perspective. That ambition was real and the SES has produced generations of capable federal leaders. But the performance management system built around it evaluates executives primarily on the results they produce within their own organizational domains. Individual executives receive individual performance plans, individual ratings, and individual pay adjustments. The system is designed to hold people accountable for what they control, not for what the enterprise collectively produces.

The Government Performance and Results Act, modernized in 2010 as GPRA-MA, strengthened strategic planning and performance reporting across the executive branch. Agencies now produce comprehensive strategic plans, annual performance plans, and performance reports that are more rigorous and more transparent than anything that existed before. That progress is genuine. But GPRA-MA does not require agencies to translate statutory purpose into a jointly owned, quantified definition of public value. It requires reporting on performance. It does not require the upstream leadership work that would make that reporting meaningful.

The Office of Management and Budget, which oversees the implementation of federal performance management, issues Circular A-11 each year as the primary guidance document to govern budget formulation, strategic planning, and performance management across the executive branch. Circular A-11 has incorporated requirements for evidence-building, program evaluation, and alignment between budget decisions and performance outcomes. That guidance matters. But the 2026 version took notable steps backward: rolling back specifications on the use of Technology Business Management principles in Section 55 and significantly reducing Customer Experience requirements in Section 280. Both retreats move agencies away from the discipline of aligning investments to mission domains and using evidence to understand how the customers they serve define value, which is precisely the foundational work this series argues must come first.

V. They Rarely Have the Opportunity

The structural and statutory constraints described in Sections III and IV would be difficult enough to overcome on their own. A third constraint compounds both: federal executives are rarely given the directive, the time, or the structured opportunity to do this work together even when they are individually motivated to do so.

Leadership development in the federal government is extensive by some measures. The Federal Executive Institute, which was founded in 1968 as the government’s dedicated residential development center for senior leaders, represented the most deliberate institutional investment in federal executive capability the government had made. The FEI curriculum was oriented around individual leadership competencies instead of the collective enterprise capability this series describes. But it existed. In February 2025, the current administration abolished it by executive order. The 14-acre campus in Charlottesville has since been transferred to the University of Virginia. The government’s primary institutional home for federal executive development is gone.

What the FEI never fully addressed, and what its absence makes more urgent, is the collective capability this series describes: working as an enterprise team to jointly define value, map customers, and develop a shared understanding of the system the team collectively manages. That capability is not in the SES qualification standards. It is not assessed in SES candidate development programs. It is not modeled at the top of most federal agencies.

The Census Bureau story is instructive in a way that goes beyond the specific outcome. The director’s directive was necessary but not the only driver of success. He also gave his team what the directive required to be meaningful: protected time on the calendar, budget for external facilitation, and the organizational cover to step away from day-to-day operations long enough to do the work. What followed was a deliberate design: structured sessions with a clear purpose, a cadence that allowed for reflection, and a facilitation approach that required executives to build shared understanding rather than receive it. That design did not emerge from the federal leadership development system. It was brought in from outside, drawing on emerging business design practices that were only beginning to gain traction in leading organizations at the time, and applied deliberately. Without the resources to support it, the directive would have remained an aspiration.

VI. The System Can Be Redesigned

The point of this article is that even the most capable federal executives are constrained by a surrounding system that does not foster the behaviors that effective operating model transformation requires. That diagnosis is more hopeful than it might appear. Systems can be redesigned. Conditions can be changed. And the conditions that made the Census Bureau outcome possible are not mysterious; they can be identified, replicated, and eventually institutionalized.

The starting point is the simplest and most overlooked: executives need to be asked. The Census Bureau director’s directive was not complicated. He told his team to work as a team, gave them the time and resources to do so, and brought in the design that made it possible. That act — explicit, authorized, and backed by the agency’s senior leadership — is what created the opening for everything that followed. Most federal executives never receive that directive. Building it into agency operating norms, and eventually into the expectations placed on agency heads by OMB and the Office of Personnel Management, would change the default condition for every agency it touched.

The second condition is deliberate engagement design. The Census Bureau sessions worked because they were designed to produce a specific outcome: shared understanding, built together, through structured work. That type of engagement is a learnable, teachable discipline. But currently it is not part of the federal leadership development curriculum. The abolition of the Federal Executive Institute significantly weakened the institutional infrastructure for delivering that curriculum. Rebuilding that infrastructure, with this collective capability at its center, is one of the most important investments the federal government could make in its own operating performance.

The third condition is structural: the SES performance management system needs to include shared accountability for enterprise outcomes. Individual performance plans that measure only what individual executives control will continue to drive individual behavior. Until senior leaders are collectively evaluated on whether their agency is delivering measurable public value, not just whether their function is operating efficiently, the incentive to stay in lane will outweigh the incentive to look across it.

A fourth condition compounds all three: executive tenure. Federal agencies are disproportionately led by political appointees whose average tenure is measured in months rather than years. An executive team that turns over before it has completed the foundational work cannot build on what it started. A team that inherits an incomplete transformation from its predecessors rarely has the shared context to continue it. The structural work of aligning domains, governance, and architecture requires sustained leadership attention over many years, not a single administration’s initiative. Reforms that extend the tenure of transformation leadership — whether through longer appointment terms, stronger career executive continuity, or explicit transition requirements — are among the most consequential changes the system could make.

These changes are complex. Each one touches legal frameworks, administrative structures, and cultural norms that are deeply embedded. But none of them requires replacing the people currently leading federal agencies. They require redesigning the conditions those people operate within: the norms, the incentives, the development pathways, and the structured opportunities that either ask executives to work as an enterprise team or don’t.