Peter Drucker, Management: Tasks, Responsibilities, Practices — Chapter 46: Result-Focused Design — Federal and Simulated Decentralization
Introduction
I worked closely with a senior executive whom I deeply respected. He had one peculiar rule: he refused to let any report include the phrase “as instructed by the Chairman.”
In a Korean conglomerate, this was countercultural. Every department wants its work elevated to the Chairman’s agenda. A report that opens with “the Chairman’s directive” survives; one that doesn’t gets buried under more urgent priorities. That was the operating logic I had grown up with.
His position was different. “If you drag the Chairman into everything, the organization can’t function.”
I understood his point intellectually, but living by it was another matter. He may have been right in principle. Whether you could uphold that principle and still survive as a Korean conglomerate executive was a separate question.
In one of our recent sessions, another participant shared his own experience. He had served as CFO of a publicly listed company shortly after a change in controlling shareholders. The new owner began intervening deeply in operations — twice a week, every project across every department had to be reported in roughly a dozen Excel sheets. Six executives spent more than ten hours each week building those reports. For three months, he himself spent two to three days a week just on the reporting. “I couldn’t do my actual job as a CFO.”
These two leaders were looking at the same problem from different angles. Top management does not exist as a distinct task. Everyone — at every level — is doing operations.
This chapter pinpointed the discomfort I had carried for years. The fact that you can’t get anything done without involving the Chairman is only the surface. The deeper problem lies elsewhere.
Federal Decentralization Is Not About Delegating Authority
Consider Alfred P. Sloan’s General Motors in the 1920s. Chevrolet, Buick, Oldsmobile, and Cadillac were not just different car lines. Each was an autonomous business with its own market, its own profit and loss, and its own management.
This is the prototype of what Drucker calls federal decentralization.
The most common misunderstanding is that decentralization means pushing authority downward. Many read it as a weakening of the center. Drucker says the opposite.
Paradoxically, the more he decentralized, the stronger the center became. Once each division took ownership of its own market, top management was freed from operational supervision. Instead of chasing Chevrolet’s quarterly numbers, they could focus on the direction and strategy of GM as a whole.
Here lies Drucker’s central insight. He distinguishes three management tasks: operations, innovation, and top management. These are fundamentally different in nature. Yet in most organizations, top management spends its days directing operations. The work that only top management can do — what should our business be? — gets squeezed out.
The purpose of decentralization is not to push authority down. It is to design a structure in which top management can finally do top management’s work.
What Most Organizations Don’t See
This is where the problem starts. The idea that “top management” exists as a distinct task — separate from operations — is barely recognized in most large organizations.
A senior participant in our group, who has spent decades observing this pattern, put it bluntly:
“Top management isn’t understood as a task. People assume that running operations is what the chief executive does. The notion that there is a top-management task that exists apart from operations — that idea simply isn’t there. It’s not that people reject it. They don’t know it exists. And you can’t do well what you don’t know exists.”
This is the heart of the matter.
What I observed throughout my years as an executive matches this exactly. A Chairman who immerses himself in operational detail earns praise. Executives, in turn, treat dragging the Chairman into operational decisions as part of their own role.
Meanwhile, the work that should have existed alongside this — what should our business be ten years from now? Where do we allocate capital? How do we develop our key people? — gets handled in the cracks, or quietly omitted. When the senior executive said “don’t drag the Chairman in,” he had only half the prescription. The other half — what then should the Chairman do? — was left empty.
Another participant added his own observation. “I find it surprisingly hard to make a meaningful contribution from where I sit now. Almost every decision ends up requiring alignment with someone else. And those around me — they have no real interest in any work that isn’t already on the Chairman’s agenda.”
The absence of decentralization is not just a problem at the top. At the top, the Chairman drowns in operations. In the middle, executives spend their time engineering Chairman-level agenda items. At the bottom, business unit leaders feel they cannot make meaningful decisions from their own seats. Everyone behaves rationally within their own square, and the whole moves irrationally.
Even So, the Spirit Can Be Built
Federal decentralization is not easy to implement. GM’s structure was not designed from scratch — it emerged through decades of mergers and acquisitions, and Sloan chose to refine the resulting structure rather than dismantle it.
Concentrated ownership, complex group structures, founding-family involvement in day-to-day operations — any of these can collide with Drucker’s three reserved areas (entry and exit decisions, capital allocation, and key personnel) being held by top management while everything else is delegated to autonomous units.
So what should organizations do when full federal decentralization is structurally out of reach?
One participant, who has spent years inside this question, offered a useful framing. When a major Korean automaker acquired its smaller competitor, a full hundred-percent merger would have made federal decentralization possible. But capital market conditions ruled that out, and the acquired company had to remain a separately listed entity. Within that constraint, the two companies have evolved a distinctive structure — a group headquarters function that attempts to capture the spirit of federal decentralization without being one. It is not pure federalism. It is an attempt to preserve what federalism is for.
You don’t need to be a textbook federation. But the spirit must be preserved: a structure in which each business unit head can focus on their own market, and through that, top management is freed from operations to do the actual work of top management. The form is path-dependent. The spirit is a choice.
And to preserve that spirit, one recognition has to come first. Top management is a task that exists apart from operations. Without that recognition, no organizational redesign reaches the heart of the problem.
Closing
Drucker defined the purpose of organization as “liberating human energy.” A structure in which nothing moves without involving the Chairman liberates neither the Chairman’s energy nor the executives’. Everyone waits for approval, everyone watches for cues, everyone behaves rationally within their square — and the whole moves irrationally.
I still don’t know whether that senior executive was entirely right. Surviving in my world often required playing the very game he refused to play. But what he prescribed — don’t drag the Chairman in — was only half of the answer. The other half is this: the more the Chairman steps back from operations, the more clearly the work that only the Chairman can do comes into view.
If you are at the top, ask yourself: on the desk in front of you, are there decisions that only you can make — or are there mostly decisions that anyone could have made?
If you are an executive below the top, ask yourself: the time you spend turning your work into the Chairman’s agenda — do you really believe that is your job?
The more you decentralize, the stronger the Chairman becomes. But only if the Chairman decides to stop being an operator and start being a top executive.
From the Monday morning Drucker reading group, 7 a.m.