Rare is the business executive who doubts the importance of responsiveness: to be acutely alert to business opportunities and threats, and to be capable of grabbing the opportunity or fending off the threat fast and effectively. Hence, when (re-)designing the organization structure, they tend to decentralize decision-making, so that decision rights are as close as possible to the people who deal with customers, competitors, front-line employees, and other stakeholders. By doing so they avoid the delays associated with information and approvals traveling up and down the management hierarchy.

In politics, this is known as the principle of “subsidiarity” or “the principle that a central authority should … [perform] only those tasks which cannot be performed at a more local level.” But which exactly are “only those tasks”?

It is an age-old question. As Henry Mintzberg noted in The Structuring of Organizations in 1979, “The words centralization and decentralization have been bandied about for as long as anyone has cared to write about organizations.” And that is a pretty long time, at least since 400 B.C., when Jethro advised Moses to distribute responsibility to various levels in the hierarchy.

In this article, we will present a simple logic to follow when addressing the centralization versus decentralization issue. We are not presenting a new path-breaking framework – but sometimes, it pays to get back to oft-forgotten first principles.

We find it useful to start with four qualities most executives want their organizations to have: responsiveness, reliability, efficiency, and perennity (e.g., the quality of being perennial, or continuing reliably in perpetuity). When deciding on the best level at which a given task should be done, assess the impact of the decision on these four qualities. For example, the efficiency gain from centralizing payroll processing may be much more important than the possible loss of responsiveness to changes in local labor laws that would be achieved by keeping this task decentralized; the centralization of payroll processing indeed enables the harmonization of systems and procedures, hence the realization of scale effects, possibly including the outsourcing to a single external service provider.

Let’s have a systematic look at the four qualities.

1. Responsiveness through immediacy. Responsiveness is all about taking the right action quickly in response to opportunities and threats. If the sources of these opportunities and threats (e.g., customers, competitors, suppliers, employees, regulators, partners, and so on) occur at the level of the operating unit, and if these interfaces are genuinely different between operating units, it makes sense to locate the corresponding tasks (e.g., sales, procurement, recruiting, regulatory affairs) and the accountability for proper execution at that level. For example, there is quite a difference between dealing with a regulator, say, in India and one in Indonesia. Decentralization allows immediacy in time and place, hence responsiveness.

2. Reliability through compliance. For some tasks, it is desirable or necessary to have common rules across the operating units: policies, standards, methods, procedures, or systems. Think of compensation and benefits policies, product design standards, quality assurance methods, fraud reporting procedures, financial reporting systems, and the like. These rules are meant to align the operating units with the company’s overall objectives, and make the business more predictable. Some organization unit then should assume the role of guardian: defining these rules, instructing employees properly, and monitoring conformity. It usually makes sense to assign that role to a centralized unit. For some tasks there is not even a choice: they must be done by law or statute by a central independent unit. Think, for example, of the internal audit or occupational safety functions in general, or the risk management function in banks, as the Basel III regulatory framework stipulates that this function should be “under the direction of a chief risk officer (CRO), with sufficient stature, independence, resources and access to the board.”

3. Efficiency through syndication. For some tasks the case for centralization is rather straightforward: a centralized unit can serve as the home for a task that is carried out more economically when aggregated in one unit than when all operating units take care of that task separately. There are various drivers of such efficiency gains:

• Traditional economies of scale. Having the same unit do more of the same task leads to continuity, standardization, specialization, leverage, and productivity. Treasury and IT procurement may be good examples.
• Minimum efficient scale. Some tasks require expertise or infrastructure that is scarce and for which demand from any individual operating unit may fluctuate over time. In that case, it is more efficient to have a pool of specialists than to replicate these in each unit. Think, for example, of legal, tax and technology experts, or costly test equipment.
• Avoiding duplication. Different operating units often have a common need for which the solutions can be (nearly) identical (e.g., an on-boarding manual for new employees, an instrument to monitor plant performance, a CRM tool). As a consequence, it is rather wasteful for each of the units to develop these solutions in parallel.

4. Perennity through detachment. There are certain tasks which, left to the discretion of the operating units, might not get done at all – this can be particularly true for tasks that are essential to the company’s long-term wellbeing, but do not serve a short-term function for the business units. Hence, a central unit with sufficient detachment from front-line operations may be required. Here are some examples of such tasks:

• Investments in initiatives with distant and uncertain benefits (e.g., radical innovation);
• Investments in initiatives whose benefits are contingent on everybody’s participation (e.g., knowledge management and talent management);
• Activities that involve cross-unit arbitration, i.e. weighing alternatives and setting priorities (e.g., product portfolio planning);
• Decisions to admit defeat and pull the plug (e.g., product withdrawal);
• Initiatives related to assets that are the sole property of the company (e.g., brands and capital).

There is no one-size-fits-all organization structure. Obviously, changes in a company’s environment or strategic priorities may have an impact on how it organizes itself. And getting the acclaimed qualities of responsiveness, reliability, efficiency, and perennity does not always require hard structural changes to the organization chart. In many cases, less disruptive changes can also do the job, such as physically dispersing central staff to the operating units, appointing permanent points of contact at HQ for the operating units, setting up a temporary task force of people from the operating units, and so on.

In an age where the concept of “self-managed organization” attracts much attention, the question of centralization versus decentralization does not go away. Nicolai Foss and Peter Klein argue in the article “Why Managers Still Matter” that “In today’s knowledge-based economy, managerial authority is supposedly in decline. But there is still a strong need for someone to define and implement the organizational rules of the game.” We hope that the simple logic presented in this article helps managers find solutions for achieving a balance between rules and responsiveness.