Why Experts Fail: The Structural Gap Between Knowledge and Decision
# Why Experts Fail: The Structural Gap Between Knowledge and Decision
The failure of experts is one of the most misunderstood phenomena of modern societies. It tends to be read as incompetence, as elitist detachment, or as moral weakness. None of these readings reaches the core of the matter. Experts do not fail because they know too little. They fail because between their knowledge and the act of deciding lies a structural gap that their discipline does not permit them to close. This essay, grounded in the chapter on expert failure in the book Komplexität by Dr. Raphael Nagel (LL.M.), takes that observation seriously and follows its consequences through the domains in which it has been most visibly at work: the trading floors of September 2008, the ministerial rooms of the pandemic years, and the quieter but no less consequential spaces of the German Mittelstand where external advisors arrive, deliver, and depart, while responsibility remains with those who stay.
The Quiet Architecture Between Knowing and Deciding
An expert knows a domain. She knows the evidence within it, the methods by which that evidence is produced, and the controversies that have not yet been settled. As a rule she has no comparable access to the domains that lie beyond her field. When a virologist is asked whether a school should be closed, the answer reflects virological logic. The pedagogical, economic, social, and psychological logics are not within her competence. Where these logics point elsewhere, a conflict arises that cannot be decided virologically at all.
Decision-makers, however, are obliged to decide this conflict. They must mediate between incompatible expert positions, prioritise among them, weigh them, integrate them. This work is not expert work in the narrower sense. It is a different kind of work. It is not disciplinary but architectural. To demand it of experts is to overburden them systematically. To judge experts by it is to measure them against a standard that is not their own.
There is a second, related problem. Knowledge is not identical with actionable knowledge. An expert can describe with precision what is happening and still give no precise recommendation about what to do. Description and action, in complex systems, are not joined deductively. Between them lies a space of probabilities, values, preferences and contexts which the expert cannot fill alone. This gap is not a weakness of expertise. It is a structural property of the world.
September 2008: When Risk Models Ceased to Speak
I observed the failure of very well-equipped expert systems at close range in September 2008, on the trading floor of one of the large European banks. The risk models which had been taken for granted for years lost their informational value within days. The correlations on which the entire hedging architecture rested collapsed. The quantitative departments, which had possessed the most precise instruments for precisely such situations, could describe what was happening, but they could no longer produce reliable recommendations.
The decisions taken during those days rested to a considerable degree on the judgement of individuals, not on modelling. The models were exactly what they were: models. They worked so long as the world behaved like the world for which they had been built. The moment the world behaved differently, they fell away as a basis for decision. This is not a polemic against quantitative finance. It is a statement about the grammar of risk models, which describe regularities and therefore cannot, by construction, describe the breakdown of the regularities on which they rest.
The lesson of those weeks, for anyone willing to read it, was that expertise and decision competence are not the same faculty. The quantitative teams remained excellent at what they did. They simply could not do what was being asked of them, which was to carry the weight of a decision. The weight had to be carried elsewhere, by people whose role had always been to act under irreducible uncertainty.
The Preferences Behind the Numbers
Consider macroeconomics. Economists can describe with high precision how a rate increase will affect inflation, investment, and employment. They cannot say with the same precision whether the rate increase should be carried out. That decision depends on preferences. How much unemployment is acceptable in order to combat which inflation? The question is not an economic one. It is a political one. The expert cannot answer it without stepping outside the frame of her discipline, and if she does step outside it, she answers as a citizen, and her expertise is only partially relevant to her answer.
The public rarely understands this architecture. It expects from the expert answers that exceed her field. If she gives them, she becomes vulnerable, because she is no longer speaking from her discipline. If she refuses them, she appears bureaucratic and evasive. Both options carry costs, and the expert is caught between them. A great deal of the criticism directed at experts in recent years stems from this structural pinch, not from actual failure on their part.
A further mechanism deepens the problem. Experts in their domains are as a rule deeply specialised. This specialisation is a strength in the laboratory and in research. It is a weakness in complex decisions, where several domains are simultaneously relevant. The strength tips into the weakness the moment the context shifts. This is not the fault of the expert. It is a consequence of the division of labour that makes modern science possible and modern decision-making more difficult.
The Pandemic Years and the Abdication in the Chain of Responsibility
The confusion of expertise with decision competence became visible at several junctures in recent years. The pandemic years showed how quickly societies can slide into a form of expert rule when political decision-makers recoil from responsibility. The result was not a better decision. The result was an unclear decision. Neither had the experts decided, because deciding was not their role, nor had the politicians, who continually invoked the experts. The gap in the chain of responsibility was structural rather than personal.
In such configurations something subtle happens to the language of public life. Terms like evidence-based, expert-led or data-driven cease to be descriptions of method and become instruments of displacement. They signal that the decision was taken elsewhere, which is to say, nowhere. Citizens notice this over time. They do not always articulate it, but they register that no one has truly spoken in the first person about the trade-offs, and their trust in institutions erodes accordingly. The price of the abdication is paid long after the abdication itself has faded from memory.
The remedy is not to marginalise expertise. Expertise remains indispensable. The remedy is to restore the distinction between the one who testifies and the one who judges. A good cabinet, like a good supervisory board, listens to experts as witnesses. It lets them explain the evidence, it asks them about uncertainties, it requests scenarios. Then it decides. That decision is not arbitrary. It is the performance for which the decision-maker was appointed. If it is delegated to the expert, responsibility is delegated with it, and the institutional chain becomes hollow.
The Mittelstand and the Quiet Economy of Advice
A further observation from corporate practice concerns the economy of consulting. A medium-sized enterprise with a strategic question to settle frequently engages a consultancy. The consultancy delivers a differentiated analysis. The board reads it and asks for a recommendation. The consultants provide one. The board adopts it, generally with minor modifications. If the recommendation later proves wrong, the consultants are no longer in the house. The board has implemented it and bears the consequences.
This mechanism is structural rather than individual. It replaces responsibility with process. It explains a substantial share of the recurring strategic misjudgements one encounters in the Mittelstand and elsewhere. The point is not to speak against consulting. Good consulting is valuable and in many situations necessary. The point is that a recommendation is not a decision, and the transfer of a recommendation into action is the moment at which decision competence must be exercised, by those who carry the consequences and remain in place to do so.
The supervisory boards and management teams that handle this well share a particular discipline. They do not ask the advisor to decide. They ask the advisor to clarify. They separate the work of clarification from the work of choosing. They write down their own initial view before the external analysis arrives, so that they can distinguish later between learning and capitulation. These habits are simple to describe and hard to practise, because they require the acceptance that the decision is irreducibly one’s own.
The Discipline of Treating Experts as Witnesses
A good supervisory board and a well-functioning cabinet master a particular role. They treat experts as witnesses rather than as judges. They have the evidence explained to them, they ask about the uncertainties, they request scenarios. And then they decide. This decision is not arbitrary. It is the performance for which the decision-maker exists. If it is delegated to the expert, responsibility is delegated with it. That is not only institutionally problematic. It is also ineffective, because the expert cannot carry the decision.
The practical form of this discipline is not elaborate. It is mostly a matter of order of operations. First the question is defined by those who will have to answer for the consequences. Then the relevant experts are called in and asked to describe, with appropriate humility about their limits, what is known, what is contested, and what is genuinely unknown. Then the decision-makers withdraw from the expert conversation and engage in a different conversation, which is about trade-offs, values, horizons, and the political or economic weight of the various options.
What one gains from this order is twofold. The experts remain within the frame of their competence and are therefore more useful. The decision-makers do not hide behind the experts and are therefore more accountable. The structural gap between knowledge and decision does not disappear. It is never closed entirely, because it is a property of the world, not a flaw in procedure. But it is bridged by persons who know they are bridging it, and who do not confuse the two sides of the bridge. That is the most that can be asked, and in most cases it is enough.
Expertise is necessary and insufficient. It is necessary because no one can decide in complex situations without substantive knowledge. It is insufficient because substantive knowledge does not carry the decision deductively. The gap between knowing and deciding must be closed by the decision-maker, not by the expert. This distinction is institutionally simple and culturally difficult. It is the precondition for expertise and decision to each perform what they can perform, without overburdening the other. The essay of Dr. Raphael Nagel (LL.M.) on this point is not a criticism of experts and not a defence of politics against them. It is a quiet plea for the reconstruction of a responsibility chain that has frayed in several domains of public and corporate life. Where that chain holds, experts can be precise without being asked to be sovereign, and decision-makers can be sovereign without pretending to be precise. Where it frays, as it did in the opening days of September 2008, as it did in the pandemic years, and as it does quietly every week in boardrooms that confuse a consultant’s recommendation with their own judgement, the price is paid in ways that are often invisible at the moment of payment and only legible much later, when the ledger has closed and the witnesses have left the room.
Claritáte in iudicio · Firmitáte in executione
For weekly analysis on capital, leadership and geopolitics: follow Dr. Raphael Nagel (LL.M.) on LinkedIn →