Chapter 09 — The Ecology of Managing Up

The Ecology of Managing Up

In contemporary corporate management theory, ‘managing up’ is codified as an essential professional skill. Since the 1980s, scholars have advised subordinates to actively study their superiors’ blind spots, communication styles, and risk tolerances, deliberately reshaping the flow of information to achieve ‘alignment.’ Career sites list it as a required course for promotion; business schools build it into leadership curricula; management consultancies charge handsomely to teach it. The mainstream narrative frames this as a lubricant for organizational efficiency — a subordinate who manages up well is considered thoughtful, politically mature, trustworthy.

This framing inverts cause and effect. The subordinate is not lubricating efficiency. He is colonizing the sovereign’s field of judgment.

The mechanism operates through the subtle curation of context. A subordinate managing up does not necessarily lie; he selects. He presents the data aligning with a superior’s known preferences and omits the data that might trigger unwanted friction. By controlling the range of options presented, the subordinate dictates the boundary of the decision. The sovereign is handed a multiple-choice question in which every option leads to an outcome the administrative machine can tolerate. The illusion of authority survives intact — the sovereign signs the document, approves the budget, issues the directive — but the substantive judgment was completed further down the filtering chain long before, before the sovereign even knew a decision was required.

What makes this mechanism so hard to detect is that it never presents itself as confrontation. It requires no ill will from the subordinate — it does not even require him to be aware of what he is doing. A conscientious department head, preparing a briefing, will naturally ask himself: how will the boss see this? The question sounds entirely benign, even commendable. But the question itself is already the starting point of a transfer of judgment — because the moment the answer begins shaping what goes into the briefing, what the briefing presents is no longer reality itself, but ‘reality as it looks after passing through the boss’s preferences.’ Once this filter is installed, it activates automatically in every subsequent report, with no need for a fresh decision, no need for any outside reminder.

Consider a standard budget approval. A division head seeking funding does not submit the raw operating deficit. He submits a strategic narrative calibrated to the CEO’s known obsession with market share, quietly omitting the supply-chain risk underneath. The CEO approves the budget, believing he has made a strategic judgment. In fact, he has only ratified a simulation. This is not an individual moral failing. It is an ecological inevitability. In a monocentric system, managing up is an adaptive response to structural pressure. Once one node begins filtering information to reduce friction, the nodes around it face a prisoner’s dilemma: adapt, or be marginalized. A node that reports raw reality honestly is quickly outcompeted by one offering a frictionless summary. In this ecosystem, honesty generates friction, and friction reads as incompetence. What the system naturally selects for is whoever knows how to filter.

This mechanism of elimination is worth unpacking one layer further. Suppose two people work in the same department: one honestly reports the supply-chain risk; the other filters the risk out and presents only the flattering market numbers. In the short run, the second person’s briefings always run smoother, the meetings always run shorter, and the CEO’s impression of him is always more favorable — he gets to the point, he doesn’t waste my time on trivia. The first person gets labeled as lacking the big picture, always leading with bad news. None of this requires the CEO to consciously play favorites — it is purely the shape of the information itself, shaping the perception. Over time, the first person either learns to filter like the second, or falls behind him in every performance review. This is not moral decay. It is structural natural selection — the organization, through its own system of reward and punishment, is training an ever-larger population of nodes skilled at filtering, and every act of filtering is one small act of colonizing the master’s field of judgment.

This dynamic is not confined to the top of the pyramid. It runs recursively at every level of the hierarchy. The mid-level manager who filters reality for the CEO is, in turn, managed by his own subordinates. As information climbs the chain, every successive node filters out friction and pre-digests judgment before passing it upward. No level possesses raw truth or an independent will of its own. The CEO governs a simulated empire; the manager governs a simulated department. By the time reality reaches the top, it has already become entirely a product of the ecosystem itself. The only will that runs, unbroken, from the bottom to the top of this entire structure is the system’s own logic of self-preservation. The machine governs a masterless hierarchy.

This filtering process is not a uniform erosion of information in every direction — it is directional. What each level filters out is rarely random noise; it is, specifically, whatever detail would invite accountability onto that node, or interrupt a superior’s existing narrative. This means that the higher information climbs toward the top of the pyramid, the thinner it becomes and the more systematically it skews toward whatever is reassuring. The world the master at the top sees is not merely blurred — it has been structurally optimized for reassurance. This is precisely why so many major crises, in the moment before they erupt, follow the same pattern: everyone knew except the very top. This is not a failure of secrecy. It is the opposite — the mechanism of secrecy succeeded too completely.

Through this continuous filtering, executive power steadily erodes the master’s judgment. It is effective precisely because it advances with a soft posture. The subordinate does not seize the throne — he only filters the throne’s field of vision. This takeover is accomplished through the daily rituals of deference. Power is not seized. It is surrendered. The administrative function merely fills a vacancy left by a master who had already stopped governing in person. The master, relieved of the cognitive burden of processing raw reality, comes to depend on this machine instead. So long as the metrics look stable, he is glad to sign. His authority is reduced to a ceremonial stamp affixed to a decision the ecosystem has already made.

This surrender of authority is not entirely passive. Most masters actively welcome the process, because the cognitive burden of raw reality is heavy, while a filtered summary is light. A CEO overseeing dozens of projects simultaneously cannot personally review every piece of raw data — he must trust his subordinates to digest it for him. The problem is not trust itself. Every organization at scale must depend on some degree of trust and delegation; that is a necessary condition of governance, not a pathology. The real pathology begins once the boundary of delegation disappears — once a subordinate is no longer merely organizing information for the master, but deciding, on the master’s behalf, which information is worth knowing at all. At that point, trust has mutated from a tool of governance into a silent transfer of judgment. And this boundary is precisely the hardest thing to detect, because on the surface, the two look identical: a concise report, a smooth meeting, a stamp affixed.

This mechanism grows a different surface in every industry, but the skeleton underneath is always the same. Consider a hospital’s chain of escalation. A nurse on duty notices an abnormal fluctuation in a patient’s vital signs. This observation must first be confirmed by a resident, then decided upon by an attending physician, and might, eventually, reach a quality-control meeting at the level of the deputy medical director. At every point of handoff, the node has a perfectly sound professional reason to make a ‘reasonable trim’ to the original observation — the resident might judge it a temporary instrument error not worth alarming the attending over; the attending might judge the case within manageable range and not worth writing into the formal report. Each of these trims, taken alone, is professionally sound, even commendably cautious. But when the same trimming logic runs simultaneously across thousands of patients and hundreds of nodes, the signals that genuinely point to systemic risk — a piece of equipment failing repeatedly, an abnormal pattern of reaction to some class of medication — get dismantled, in the layered rationalizing, into isolated incidents that never touch one another, never able to assemble, at any single node, into a complete picture alarming enough to reach a decision-maker. The quality report the deputy medical director eventually sees is always clean, its metrics always met — until some accident erupts at a scale too large to keep trimming away. This is the fault of no single clinician. It is the necessary result of an entire filtering chain, each link individually rational, collectively blind.

A military chain of command offers another sample, structurally identical but with far higher stakes. Frontline battle reports, compiled up through company, battalion, brigade, and division, face pressure at every level to be dressed into a version that ‘matches what headquarters expects’ — casualty figures softened, the strategic situation exaggerated, morale problems left unmentioned. This dressing rarely amounts to a deliberate deception of the supreme commander. It is, rather, that every intermediate commander is managing up to his own direct superior, so distortion compounds layer by layer instead of being corrected layer by layer. By the time the supreme commander makes a decision based on this progressively beautified report, what he commands is not the actual battlefield, but an optimistic narrative jointly compiled by the entire chain of command. The root of many great historical strategic miscalculations is rarely a commander’s lack of intelligence — it is that the information system he depended on had already been structurally tamed into a house of mirrors long before it ever reached his hands.

History provides an even more extreme catastrophe that pushed this same mechanism’s stakes to the level of life itself. At the end of the 1950s, amid a nationwide climate of political competition, grassroots cadres raced to inflate reported grain yields — the phrase ‘launching satellites’ described the spectacle of localities competing to announce ever more extravagant production figures, as though competing to launch the most spectacular achievement into orbit. This process matches, point for point, the managing-up logic this chapter has described: any cadre who honestly reported an ordinary yield would look politically insufficiently zealous against the backdrop of his exaggerating peers; anyone who inflated his figures stood to gain praise and promotion in the short term. Exaggeration became the most rational individual response to structural pressure, and once it appeared at one node, it forced neighboring nodes to follow suit or else look suspicious by comparison. This is the same prisoner’s dilemma described earlier in this chapter, activated simultaneously across an entire national bureaucracy.

The consequence of this exaggeration was not merely distorted statistics. The central government set each region’s grain-procurement quota according to the yield figures compiled up through the hierarchy — the higher the reported output, the more grain a region was required to hand over. By the time the inflated numbers, compounded layer after layer, reached the highest level of decision-making, they were no longer a rumor that could be easily questioned. They had become a ‘fact,’ jointly confirmed by an entire filtering chain. The procurement quotas set on the basis of this structurally optimistic fact ultimately exceeded what the land could actually produce, and local officials, under pressure to meet quotas set above them, often requisitioned even the grain peasants needed to eat and the seed grain needed for the next planting. Scholars estimate the resulting death toll at anywhere from over ten million to more than thirty million; the precise figure remains disputed, but the scale places it among the deadliest man-made famines of the twentieth century.

For this chapter’s central proposition, this case is an unavoidable, extreme confirmation: the information distortion produced by managing up does not stop at one bad budget decision or one strategic miscalculation. Given the right conditions, it can convert directly into physical death on a mass scale. No single grassroots cadre individually decided to manufacture a famine, just as no single mid-level manager individually decides to deceive a CEO. Every node was only responding rationally to the local political climate it found itself in. And it was the sum of these countless local rationalities that led an entire nation’s highest decision-making body, over a considerable stretch of time, to make resource-allocation decisions — based on a narrative of abundance jointly compiled by an entire bureaucracy — that directly cost the lives of tens of millions.

These three cases reveal the same ecological law’s full generality: the pathology of managing up requires no special conditions particular to any industry or regime. It requires only three conditions holding simultaneously — information must pass through artificial hierarchical transmission; every layer must have some incentive to keep a superior satisfied; and the cost of honesty must exceed the cost of filtering. Wherever these three conditions hold — corporation, hospital, army, or government — the same filtering logic will spontaneously emerge, with no one needing to design it deliberately. This is precisely what this book means by ‘ecology’ rather than ‘conspiracy’: no one is pulling strings behind the scenes. Every node is simply reacting rationally to its own local environment, and it is the accumulation of these countless local rationalities that constitutes the systemic paralysis of judgment as a whole.

The prisoner’s dilemma at work here deserves a more precise description. For any given node, ‘reporting honestly’ is, in the short run, almost always the dominated strategy: if both sides are honest, both bear the cost of friction, but judgment at least survives; if the other side filters while you stay honest, you look clumsy, slow, difficult to work with, suffering doubly in both performance review and internal standing; if both sides filter, friction is minimized, reporting runs smoothly, and neither party suffers any penalty in the short term. The crux of this game structure is that ‘both sides filter’ is the individually rational choice for every player involved, yet it is the unique equilibrium point at which the entire system loses its capacity for judgment. No single node is responsible for this outcome, because every person is simply doing what is best for himself — and this is precisely the underlying logic by which the apparatus emerges spontaneously, with no one designing it.

History supplies a still more extreme specimen of this same dynamic: the eunuchs and sycophants of absolute monarchy. Their power derives entirely from their position as the ultimate filtering node. In a system where the master has been physically enclosed, the filtering chain manufactures reality itself. What a eunuch controls is not only the memorials the master reads, but who is granted an audience, whose petition is ever heard at all. Once this node succeeds in insulating the master from every raw friction, the master’s judgment begins to atrophy. What the master navigates is not the real empire, but a simulation choreographed by the node. The eunuch need never issue a command in his own name — he need only determine what the master sees, and the master’s own absolute power will endorse the eunuch’s will of its own accord.

This specimen is extreme and clear precisely because it fuses ‘access to information’ and ‘physical access’ into a single thing — the eunuch does not merely filter the content of memorials; he directly determines who may enter the inner chamber, whose petition is ever heard. This double monopoly leaves the filtering chain almost no possibility of external verification: the master cannot bypass the node to check, because even the channel of access needed to check is controlled by that same node. Modern bureaucracy rarely closes this loop at the physical level, but its functional equivalent is everywhere: a CEO’s calendar, arranged entirely by his chief of staff, determines which meetings he attends and whose voice he hears in them; the battlefield intelligence a general receives has already been compiled and filtered entirely by his own staff. The physical palace walls are gone. The dual gate filtering information and personal contact remains, entirely intact.

The flesh-and-blood eunuch is a historical particular, but the ecological function is universal. In the modern corporation, this niche is occupied by the chief of staff, the executive assistant, the trusted deputy. The title has changed. The structural position has not. They are the apparatus’s human interface, the adaptive nodes through which the machine operates.

The competencies taught in corporate training seminars — anticipating a superior’s needs, framing reports to match cognitive biases, absorbing operational friction — read identically to the survival arts of the imperial court. The modern subordinate and the ancient eunuch occupy the same ecological niche. In both environments, an operationally absent master sits at the apex. In both, the system rewards whichever node makes itself indispensable to that master’s comfort. The master is managed. The truth is filtered. The apparatus runs on the fuel of perfect service. The ultimate irony of modern management is that the more skillfully subordinates are trained to manage up, the more completely the organization loses its capacity to judge.

This irony carries a deeper corollary still. When the art of managing up is openly taught inside an organization, institutionally rewarded, that organization is, in effect, systematically training every one of its nodes to do the same thing: make judgment disappear without a trace. No one ordered it. No one is guilty of anything. Everyone is simply doing what their job requires — doing good work, making the report look clean, making things easier for the boss above. And it is precisely this soft erosion — one no single person can be held accountable for, yet which occurs everywhere at once — that constitutes the apparatus’s most stable survival strategy. It needs no dramatic moment of seizing power, because judgment has already been quietly surrendered, a thousand times over, inside a thousand harmless, considerate, professional briefings.

This is exactly where managing up differs from the other mechanisms of devouring the master. Regulatory proliferation and the recursion of oversight still carry an institutional appearance — a name, a title, a formal box on the org chart. Managing up requires no institutionalization at all. It can achieve the identical effect with no rule ever authorizing it, purely through millions of individually harmless daily interactions — judgment flowing, silently, out of the master’s hands and into the entire filtering chain. Once this process has run its full course inside an organization, the master is no longer a ruler who has been hollowed out. He has become a totem, kept and fed. He still signs. He still speaks. He still enjoys every ceremonial honor due him. But the position that once required his personal intervention, his personal judgment, his personal accountability for the outcome, has already been quietly filled by the ecosystem’s own logic.

What he governs is not an organization. It is a hollow shell that has learned to breathe on its own.