The Management Tax

Why the middle manager is not merely unnecessary, but corrosive

---

There is a particular kind of corporate meeting that anyone who has worked in a large organization will recognize immediately. A dozen people sit around a table or stare into a grid of video thumbnails. An agenda circulates. Slides are presented. Someone asks a clarifying question that is not really a clarifying question but a bid for relevance. A follow-up meeting is scheduled. Nothing is decided. The people who actually know things (the engineers who built the product, the salespeople who spoke to customers last Tuesday) were not in the room. They were, in the idiom of the age, “executing.”

This theatre has a name: middle management. And after decades of quiet tolerance, the corporate world is finally running out of patience with it.

Amazon, Alphabet, and Meta have all announced significant culls of managerial layers in recent years. Bayer, the German chemicals giant, went further: it eliminated its entire middle-management tier in a restructuring it called “Dynamic Shared Ownership,” a phrase that is either inspiring or alarming depending on your tolerance for corporate vocabulary. The results, by most measures, were impressive. Decision cycles shortened. Accountability sharpened. The company spent less time explaining itself to itself.

These are not isolated experiments. They reflect a hypothesis that is rapidly hardening into orthodoxy among the most competitive companies in the world: that middle management is not a neutral feature of organisational life, but an active drag on it. A tax, levied on every idea that passes through it, on every decision that requires its approval, on every employee whose career depends on keeping it happy.

---

The case against is damning

Begin with what middle managers actually do, as opposed to what their job descriptions claim. In theory, they translate strategy into action, develop talent, and co-ordinate across functions. In practice, research consistently finds that the majority of their time is consumed by activities that add no discernible value: attending meetings, producing status reports, reviewing work that did not require reviewing, and managing upward with an assiduousness that would impress a courtier.

A 2023 study by the consulting firm Gartner found that fewer than one in four managers felt confident their organisation’s structure supported the company’s strategy. This is a remarkable finding. Middle managers are, by definition, the mechanism by which strategy is meant to become reality. If they themselves do not believe they are serving that function, the obvious question is: what exactly are they for?

Part of the answer is historical accident. The middle-management layer metastasized during the twentieth century not because it was efficient but because it was legible. Hierarchies made accountability visible, or at least made it appear so. A manager could be held responsible for the output of their team in a way that an amorphous flat structure could not. When organizations were slower, less technically sophisticated, and less capable of generating real-time data on performance, this arrangement made a kind of rough sense.

That world no longer exists. Modern organizations swim in data. An engineering team’s velocity, a sales team’s pipeline, a product’s user-retention curve: all of these are observable, measurable, and trackable in ways that would have seemed like science fiction to a 1970s executive. The informational rationale for the middle-management layer, the idea that someone needs to gather signals from the front line and carry them upward, has been largely eliminated by the tools that the front line itself now uses.

What remains when you strip away the informational function? Largely, a political one. Middle managers, in the absence of genuine strategic contribution, tend to make themselves useful by managing relationships, filtering information, and (this is the part that rarely appears in the literature) protecting their own position. This is not a moral failing. It is a structural inevitability. Give anyone a role whose primary value is ambiguous and they will fill the ambiguity with activity that signals value rather than creates it.

---

The binary that clarifies

The most cogent alternative to the managerial layer is also the most radical: replace it with a simple question. In a modern company, does this person build something or sell something? If neither, the burden of proof lies with them.

This is a deliberately blunt instrument, and its bluntness is the point. Every organisational layer that cannot be classified as product creation or revenue generation introduces friction by default. It must justify itself not merely by existing but by demonstrably improving the output of those who do build or sell. The default assumption should be that it does not.

This framework does not eliminate leadership. It redefines it. In a flat organisation, the role of the senior executive is not to manage managers who manage managers who manage people. It is to do four things, and four things only: deliver clear judgment on difficult decisions, raise the quality of judgment across the organisation, choose the strategic direction, and allocate capital toward it. Everything else is either a distraction or a task that belongs to an individual contributor.

The leadership function, understood this way, is genuinely demanding, more demanding arguably than the traditional model. Without a buffer of middle management absorbing noise and smoothing over uncertainty, senior leaders must be directly exposed to the texture of the work. They must know enough about what engineers are building to give useful feedback on it. They must understand the sales process well enough to interrogate a quarterly shortfall. They must engage with individual contributors not as a performance of accessibility but as a genuine mechanism of information gathering and talent development.

This is, in the language of Silicon Valley, “high-context” leadership. It requires people at the top of organizations who are comfortable being wrong in public, who can absorb contradictory information from the front line without retreating into the comforting abstractions of the dashboard, and who understand that developing judgment in others is as important a contribution as exercising it themselves.

---

The counterarguments deserve acknowledgment

Defenders of middle management are not without ammunition. Scale is the most credible objection. A company of twenty people can plausibly operate without managerial layers; a company of twenty thousand almost certainly cannot. At some threshold of complexity, the co-ordination costs of pure flatness exceed the costs of the hierarchy it replaces. Nobody has precisely identified where that threshold lies, which is itself revealing: if the benefits of middle management were as clear as its proponents claim, you might expect the evidence to be more legible.

The second objection concerns human development. Middle managers, the argument runs, serve as training grounds for future senior leaders. Strip them out and you deprive the organisation of the developmental infrastructure that creates the next generation of executives. There is something to this, but it proves less than it appears to. Much of what passes for “development” in traditional hierarchies is, in practice, the cultivation of political skill: the ability to navigate upward, manage stakeholders, and survive reorganisations. These are not the same as the judgment-based, technically grounded capabilities that flat organizations actually need. If anything, the hierarchical model may be selecting for the wrong traits.

The third objection is cultural. Removing middle management is easy to announce and hard to execute. Self-managing teams require trust, clarity of purpose, and a tolerance for ambiguity that most organisations have spent decades systematically discouraging. The transition can be ugly. Bayer’s restructuring, for all its philosophical elegance, was accompanied by legal challenges, employee protests, and the kind of operational turbulence that does not appear in a slide deck about dynamic shared ownership.

These are genuine difficulties. They are not, however, arguments for the status quo. They are arguments for managing the transition carefully, which is different from not making it at all.

---

What the data suggests

The empirical record on flat organizations is more favorable than the sceptics like to admit. A study of more than 300 companies published in the American Economic Review found that every additional layer of management was associated with a measurable reduction in the speed of decision-making and a modest but consistent decline in employee engagement. Research on self-managing teams, most notably W.L. Gore, the manufacturer perhaps best known for Gore-Tex, which has operated without traditional hierarchy since its founding in 1958, suggests that flat structures can sustain not just productivity but genuine innovation over long periods.

The productivity gains most commonly cited (studies have found improvements of 25-35% in comparable organizations) come with an important caveat: they materialise only when the cultural infrastructure supports them. Autonomy without clarity is not liberation but chaos. The organizations that have made flat structures work have invested heavily in making expectations explicit, in building shared decision-making frameworks, and in creating what might be called institutional judgment: a widely distributed understanding of what the organization values and how it makes trade-offs.

This is, incidentally, the deepest indictment of the middle-management model. Organizations that rely on hierarchical layers to enforce coherence are, in effect, admitting that they have not done the harder work of creating genuine alignment. The manager becomes a substitute for culture, and a poor one: expensive, slow, and prone to departure.

---

The real argument

The debate about middle management is, at its core, a debate about trust. Traditional hierarchies are built on a latent assumption that most people, left without supervision, will do less than they could. The flat organization inverts this: it assumes that most people, given clarity about what matters and the autonomy to pursue it, will do more.

The evidence increasingly favours the optimists. What has changed is not human nature (people have always been capable of self-direction when properly motivated) but the technical infrastructure that makes self-direction visible and co-ordinated. The tools that allow engineering teams to work asynchronously across time zones, that give salespeople real-time visibility into pipeline, that let leadership see the texture of the work without being physically present for it: these are not merely conveniences. They are the preconditions for a new form of organisation.

Middle managers filled an informational void. The void no longer exists. What remains of the role is, at best, a scaffolding that has outlasted the building it was meant to support, and, at worst, an active impediment to the speed and clarity that modern competition demands.

Companies that recognize this early will move faster, spend less, and attract the kind of talent that finds meaning in the work itself rather than in the management of it. Companies that do not will continue scheduling meetings to discuss the findings of the meeting held to prepare for the meeting. The outcome, eventually, should be obvious.