The Dunning-Kruger Effect and the Psychology of Choosing Politicians

I. The Double Curse of Incompetence

Rome, January 15, AD 69. Servius Sulpicius Galba has been emperor for seven months and dead for a few hours, struck down in the Forum. Tacitus writes him an obituary that first enumerates his virtues - military glory in Germania, moderate governance in Africa, justice in Hispania - and then turns on a single subordinate clause: “maior privato visus dum privatus fuit, et omnium consensu capax imperii, nisi imperasset.”¹ As long as he was a private citizen, he seemed greater than a private citizen; and by universal consent he would have been capable of ruling — had he never ruled.

Some nineteen hundred years before the study that gave the finding its name, a Roman senator had packed the heart of the problem into a single conditional clause, and in better style. Note how the sentence is built: first the merits, complete and without malice, then the reckoning. Anyone who writes about the conduct of public office can do it no other way — and this text will do the same.

In 1999, Justin Kruger and David Dunning published a study whose title already contains the entire problem: “Unskilled and Unaware of It.”² Their finding: people who are incompetent in a field systematically overestimate their performance in it — not out of vanity, but for a structural reason. The numbers are stark: participants who scored in the bottom quartile on tests of humor, grammar, and logic landed, on average, at the twelfth percentile - eighty-eight of every hundred participants did better - , yet placed themselves at the sixty-second, above average.² For the very ability needed to find the right solution would also be needed to recognize one’s own wrong solution as wrong. What those affected lack, Kruger and Dunning call metacognition — literally, knowing about knowing, that is, the ability to judge the quality of one’s own thinking. Those who have it notice when they do not understand something. Those who lack it do not notice, and they do not notice that either. Incompetence is therefore a double curse: it produces errors and at the same time prevents them from being noticed. The incompetent person does not cook the books; he is their most credulous reader.

In a conversation published in The New York Times on June 20, 2010, David Dunning said:³

David Dunning: “Well, my specialty is decision-making. How well do people make the decisions they have to make in life? […] Not just that people said these positive things about themselves, but they really, really believed them. Which led to my observation: if you’re incompetent, you can’t know you’re incompetent.”

Errol Morris: “Why not?”

David Dunning: “If you knew it, you’d say, ‘Wait a minute. The decision I just made does not make much sense. I had better go and get some independent advice.’ But when you’re incompetent, the skills you need to produce a right answer are exactly the skills you need to recognize what a right answer is. In logical reasoning, in parenting, in management, problem solving, the skills you use to produce the right answer are exactly the same skills you use to evaluate the answer. And so we went on to see if this could possibly be true in many other areas. And to our astonishment, it was very, very true.”

David Dunning in conversation with Errol Morris. The New York Times, 2010.³

One finding of the same study tends to be overlooked, and it is decisive for everything that follows. When Kruger and Dunning taught some of their participants the skill being tested - in this case, logical reasoning - , it was precisely the weakest among them who afterward judged their performance far more realistically. Those who learned saw, along with their new skill, its limits.² The double curse, then, is not fate. It can be broken, though not through introspection, since the instrument with which one would have to examine oneself is precisely the defective one. It is broken through learning, and learning, as a rule, begins with outside feedback that one takes in rather than fends off. Why some people fend off such feedback instead of using it has long been familiar to cognitive psychotherapy: anyone who ties his worth as a person to his reputation hears in every criticism of his work a verdict on himself — and one fends off such a verdict rather than examining it. That, too, is a thinking error, and a particularly insidious one: it protects itself by rejecting precisely the feedback that could correct it. Keep in mind the distinction between feedback one examines and feedback one fends off; this text will need it again — once in a success and once in its opposite.

Honesty requires noting that the finding is not undisputed among specialists. Some of the more recent research considers the size of the effect overstated and explains a good share of it statistically: those who score very poorly can hardly underestimate themselves, but they can overestimate themselves considerably, and that alone produces part of the pattern.⁴ For the argument of this text, however, the weaker claim suffices, one that even the critics do not dispute: self-assessment is no reliable substitute for a test. And that test, as Section II will show, is precisely what the electoral procedure does not provide.

Elsewhere, this journal has developed the finding as the epistemological axis of psychopathology. That sounds more learned than it is. Epistemology is the theory of knowledge; the point is that psychological suffering is, at its core, a problem of knowledge, more precisely: of false knowledge that takes itself to be true. The false certainty of the sick⁵ is not defiance but a defect of cognition — those who cannot see the thinking errors that make them ill take them for reality. The anxiety patient who takes his racing heart for the harbinger of a heart attack is neither lying nor exaggerating; he simply does not see that he is mistaken — and that is why no well-meant “Pull yourself together” will help him, only the correction of his error. What holds there for the individual also holds for institutions that select individuals. And here it becomes political: a democracy that fills its executive by election tacitly assumes that the electoral procedure tests fitness for office. That assumption is false, and not by accident: the flaw is built into the procedure itself.

This text makes its case in four steps. It first shows, from the procedure itself, why elections do not test fitness for office (Section II). It then shows, through a man who closed the procedure’s gap by hard work, that things can turn out well nonetheless (III). It next documents, through one tenure in office, what happens when no one closes the gap, and why such a state of affairs does not correct itself (IV and V). And it ends with the only government election at the federal or cantonal level that can still serve as a test: the election by the Federal Assembly, which on December 9, 2026, will choose the successor to Federal Councilor Guy Parmelin (VI).

II. The Test That Tests Nothing

Imagine an enterprise that chose its chief physician by a procedure measuring only how well the candidates speak on camera, how loyal they are to their professional association, and how reliably they can whip up an audience. One would wish a bankruptcy trustee on such an enterprise. Yet that is precisely the procedure of an election campaign — a talent show. It measures telegenic appeal, party loyalty, the ability to polarize, and name recognition — all of them abilities needed to win an office. It does not measure expertise, mastery of the issues, the leadership of people, or judgment under uncertainty — the abilities needed to do justice to an office. The two sets of competencies are not merely different; in part they work against each other, for whoever polarizes oversimplifies, and whoever oversimplifies has no business in a policy field like health care.

That this is more than hairsplitting can be measured. In 1969, the Canadian educator Laurence J. Peter, together with Raymond Hull, formulated the principle that bears his name: “In a hierarchy, every employee tends to rise to his level of incompetence.”⁶ The mechanism is simple. People are promoted for having proved themselves at their current level, not for being suited to the next; so they keep rising until they reach a level at which they no longer prove themselves — and there they stay. What began as satire was put to the test fifty years later by the economists Alan Benson, Danielle Li, and Kelly Shue, using performance data on salespeople at 131 firms. The firms preferentially promoted their best salespeople, and did so at the expense of other traits that would have better predicted success as a manager. The more salespeople had sold before their promotion, the less their teams sold afterward. The authors put a figure on the cost of this practice: “This behavior results in firms promoting workers who decrease subordinate performance by 30%, relative to a promotion policy that optimizes match quality.”⁷ The best salesperson, then, was not merely not the best boss; on average, the best salesperson made a worse one. An election campaign is promotion by sales figures.

The Swiss militia system once knew this. Under the militia system, public offices are held not by career politicians with specialized training but by citizens who exercise them alongside their profession or come to them from it. The system deliberately dispensed with formal qualification requirements — not out of indifference to competence, but because it presupposed competence elsewhere: in proving oneself in one’s profession, in local government, and in the military before entering politics. The diploma was dispensable because a proven record was its substitute. That substitute erodes to the extent that the pre-political career itself already consists of public relations: then the election tests only what it rewards anyway, and the loop closes. It works like this: whoever is well known gets elected, and whoever is elected becomes better known. The loop is self-confirming, and nowhere within it does the question arise whether someone can supervise a hospital, design a planning process with methodological rigor, or keep an experienced employee. The talent show is therefore not an abuse of the procedure - not a mishap that better candidates would remedy - , but is the procedure, once the proving grounds outside the show have disappeared.

Kruger and Dunning now supply the second, more uncomfortable half of the argument. If this procedure brought into office a miscast officeholder who could recognize the miscasting, little would be lost: such a person would work, delegate, learn. The double curse, however, holds that this very recognition is least likely precisely where it would be most needed. A miscasting that cannot recognize itself as such does not correct itself — it perpetuates itself. And it perpetuates itself, predictably, with the means it commands: communication instead of substantive work, control instead of delegation, decisions without a standard fixed in advance, fending off criticism instead of examining it. Whether this derivation holds is an empirical question. Switzerland obligingly supplies both arms of the experiment.

It would be some comfort if one could say that an election is, after all, a democratic procedure and therefore beyond all criticism. The inventors of democracy saw it differently, and in fact the other way around. Aristotle states in the Politics: “It is regarded as democratic to fill offices by lot, but as oligarchic to elect their holders.”⁸ In Athens, most offices were in fact filled by lot among the citizens; only the few that required proven expertise, such as the generalships, were filled by election. In 1748, Montesquieu repeated the observation almost word for word: selection by lot accords with the nature of democracy, selection by choice with that of aristocracy.⁹ The reason is simple: whoever votes wants not just anyone, but the best — and the claim that some are better is an aristocratic idea, not a democratic one.

None of this is an argument against elections, but it says everything about the condition on which they depend. The procedure can be justified only if it actually tests who is best; for election was conceived from the outset as a means of selection — the political scientist Bernard Manin speaks of the principle of distinction.¹⁰ The modern election has kept the aristocratic form and lost the test. It still promises the best and has come to measure visibility. The Athenians would presumably not have changed their verdict had someone explained to them that, two and a half millennia later, candidates would court the favor of viewers in television studios — they would merely have asked where in this procedure the test takes place. To this day, there is no answer.

Let us take stock of what has been shown so far, for everything that follows builds on it. First: an election campaign tests the ability to win an office, not the ability to do it justice. Second: where proving oneself before politics falls away, no one checks any longer whether the candidate is professionally and personally fit for the office at all. Third: according to Kruger and Dunning, those who reach office without that fitness cannot notice the deficiency themselves — unless they allow themselves to be taught from outside; and that becomes all the harder the more tightly they tie their worth as a person to their reputation. It does not follow that every elected official is unfit; it follows only that the procedure does not prevent it. Whether things go well in a given case then depends above all on what the person elected makes of the gap. The next two sections show the two possible responses in two cases whose starting points are strikingly similar: both without a university education, both coming from marketing, both from the same party.

III. Proof That There Is Another Way: Adolf Ogi, the Commercial-School Graduate from Kandersteg

When Adolf Ogi was elected to the Federal Council on December 9, 1987, the Neue Zürcher Zeitung (NZZ) objected that he had never attended university; his commercial-school education, it held, was narrow-gauge for the task.¹¹ The objection was factually correct: primary school in Kandersteg, three years of commercial school in La Neuveville, a period of language study in England — that was the entire formal education of the man who was to take over the Federal Department of Transport, Communications and Energy, a technically demanding infrastructure department for which not a single certificate qualified him.¹² His career before that: the Meiringen-Hasliberg tourist association, the Swiss Ski Association - from 1975 as its director - , then director general of Intersport.^12,13 An association official and sports marketer, elected to the National Council in 1979 as a darling of the tabloid press.¹² By the standards of the educated bourgeoisie, a provocation. He was elected only on the second ballot, with 132 votes; on the first, according to the Historical Dictionary of Switzerland, individual members of the center-right parties had withheld their votes because in 1983 he had supported the Social Democrat Lilian Uchtenhagen.^12,13 Ogi never forgot the NZZ’s reproach, and three decades later he reduced it to the tersest formula imaginable: “If you come into the Federal Council narrow-gauge, you can only surprise.”¹¹

What happened next is the most precise refutation of the diploma argument that Swiss history has to offer. Ogi treated his competence gap not as a communications problem but as a work assignment. He mastered the issues, surrounded himself with experts whom he listened to rather than replaced, and led the department to results one can visit — literally: on September 27, 1992, the electorate clearly approved the bill for the New Rail Link through the Alps.¹² A man without the Matura, the Swiss university-entrance diploma, won the popular vote on the century’s largest engineering project because he had understood it before he sold it. The Gotthard Base Tunnel, opened in 2016, is his monument — and, incidentally, the reply to the NZZ of 1987.

Ogi himself described how he went about it, and his words read like an antidote to the double curse of incompetence from Section I. He knew that being one hundred percent ready is not enough, because the others are too: “You have to be 110% ready!” Of negotiating he said that one must know how “to advance one’s strategy while at the same time building trust and listening to one’s counterparts, even when they talk nonsense!” About his education he was blunt: “Seen that way, my military career was my university.” And of himself: “I realized that I was ready to take on responsibility, even if that could sometimes mean taking blows for it.”¹⁴ Note what appears in these statements: preparation, listening, responsibility, and the price it exacts. And note what does not: the show. The military career as university is the militia system of Section II in a single sentence — a proven record as a substitute for the diploma. And listening, even when the other side talks nonsense, is the precondition without which the way out described in Section I cannot be had: those who listen to everyone will not miss the one piece of feedback that matters.

For what Ogi was doing, there is an older and sharper description than anything in the modern leadership literature. In the twenty-second chapter of The Prince, Machiavelli distinguishes three kinds of minds: the first understands on its own, the second understands what others show it, the third understands neither on its own nor through others. The first, he says, is outstanding, the second excellent, the third useless.¹⁵

The point of this threefold division lies in the middle. The second mind is no genius; it is something rarer, namely one that knows what it does not know and can therefore listen. A commercial-school graduate who grasps that he needs the engineer is, on Machiavelli’s scale, not the third but the second — and Machiavelli’s text there says, in so many words, “eccellente.” The third mind, by contrast, understands neither on its own nor through others — and consequently not which kind it belongs to. Four hundred and eighty years before Kruger and Dunning, and in shorter sentences.

The second part of Ogi’s tenure tends to be forgotten, and for the present day it is the more instructive. In the fall of 1995, after a reshuffle in the Federal Council, Ogi had to take over the Military Department because the center-right parties did not want a Social Democrat as defense minister; people spoke, as swissinfo noted, of a “relegation to the B league.”¹⁶ Under him, Switzerland in 1996 joined the Partnership for Peace, a NATO cooperation program for non-member states; in 1998 the department was expanded into the Federal Department of Defence, Civil Protection and Sport, the DDPS, as it is known today; and in 1999 Ogi played a decisive part in the Federal Council’s decision to reinforce the international Kosovo Force (KFOR) with a Swiss contingent, Swisscoy.¹² The Historical Dictionary of Switzerland concludes that, with a defense policy aimed at security through international cooperation, he “also contributed to overcoming the army’s crisis of purpose after the end of the Cold War.”¹² One need not share these policies to recognize the achievement: a man takes over a department regarded as a sidetrack and leaves behind one with a new mission. It is the same department for which the Swiss People’s Party (SVP) today expressly demands leadership capability; Section VI will return to this.

Honesty demands the other side of the ledger, and it supports the thesis all the more. Ogi also lost: in 1990 the vote on the nuclear moratorium, which passed against the Federal Council’s recommendation; in 1992 the vote on joining the European Economic Area - for him, as swissinfo wrote, a “black day” - ; in 1994 the vote on the Alpine Initiative, which likewise passed against the Federal Council’s recommendation.^12,16 The educated bourgeoisie smiled condescendingly at his fervor, his “Freude herrscht” (“Joy reigns”) became a catchphrase, his televised egg-boiling lesson an anecdote. But note the structure of his defeats: they were substantive defeats of a man who did the work, suffered in open battle and openly conceded — not a marketer’s staging mishaps that had to be spun away. Even at seventy-seven he called his failure by its name: the expansion of the national highway network. “We’re still working on that today.”¹¹ He has also revealed where this composure comes from: in sports, one learns “to win or to lose without the world coming to an end.”¹⁴ Those who can name their defeats possess what Kruger and Dunning deny the incompetent: metacognition, the ability to see one’s own shortcomings. Ogi’s case thus shows two things. First, and this it proves: the missing diploma is not the problem. Second, and this it makes plausible: the decisive variable is how one deals with one’s own gap — one can close it through work or paper over it through presentation. For the second variant, the present offers an object lesson whose initial conditions match Ogi’s almost point for point: commercial training instead of university, a career in marketing, a rise through the same party, taking charge of a specialized department without a background in its field. The starting point is the same. Only the way of dealing with it is not.

IV. The Counterexample: Natalie Rickli at the Health Directorate since 2019

Natalie Rickli, born in 1976, completed a commercial apprenticeship and from 2005 to 2017 worked in various roles at the Goldbach Group in Küsnacht, a company that markets advertising time in electronic media, eight of those years as Partner Relation Manager; she then struck out on her own as a communications consultant.¹⁷ From 2007 to 2019 she sat in the National Council for the SVP, where she made her mark in media policy and was, as the weekly WOZ wrote, one of her party’s most effective figures in the media.¹⁸ In 2019 she took over the Health Directorate of the Canton of Zurich — hospital planning, security of supply, oversight of one of the country’s largest university hospitals, the most complex policy field in the most populous canton. In 2023 she was reelected for the legislative term ending in 2027.

Since Federal Councilor Guy Parmelin announced his resignation in early October 2026, hers has been among the names mentioned as possible successors. On October 2, Swiss Radio and Television (SRF) counted her among the possible candidates; she had, it said, “for years been considered Federal Council material” and had “executive experience.”¹⁹ As of press time, it remains open whether she will run. The reader will recall how this text began: “Capax imperii,” capable of ruling, was what everyone judged Galba to be as well. In Tacitus, a subordinate clause follows; in the present case, that clause remains unwritten.

What is remarkable is the yardstick by which her own party intends to measure candidates for the succession. On October 2, 2026, the SVP set down its standard in writing: “The departments with the greatest future problems and the greatest leadership deficit are the Federal Department of Justice and Police (FDJP) and the Federal Department of Defence, Civil Protection and Sport (DDPS). Since one of these two departments is the most likely to become vacant and could fall to the new SVP nominee, the SVP will pay particular attention, in the succession, to the ability to assume leadership responsibility for the FDJP and the DDPS.”²⁰ By way of explanation for readers outside Switzerland: which department a new member of the Federal Council takes over is decided by the Federal Council itself after the election; the SVP expects that it could be one of the two named. Read that statement twice, for it contains more than it says at first glance. The party diagnoses a “leadership deficit” in two departments and, as the remedy, demands the ability to assume leadership responsibility. That is precisely the ability that, as Section II showed, an election campaign does not test, and that, as Section III showed, Ogi demonstrated. Crucial for everything that follows: the yardstick does not come from this journal. It comes from the party to which the health director belongs. This text merely applies it.

What had to be said about Ogi must be said here too: nothing about this background is a reproach. The militia system requires no diploma, and the commercial-school graduate from Kandersteg proved that none is needed. The only question is how the officeholder has dealt with the gap between what the office requires and what her training provided — and, measured by her party’s yardstick, whether her conduct of office shows the ability to assume leadership responsibility. That question can be answered on the basis of documented events, without peering into anyone’s inner life. Four groups of findings, each with what the sources record against her and what they record in her favor.

First group of findings: the drain of expertise.

In April 2025, after conversations with former and current employees as well as outside parties, the WOZ documented the state of the directorate.¹⁸ At the time, the secretary general was Labinot Demaj, the fourth since Rickli took office — a man from the IT industry who had been with the canton for only a few months when he was unexpectedly placed in the most important post in the directorate. In a Swiss cantonal directorate, the secretary general is the top administrator, heading the director’s staff and serving as the link between the director and the various offices. Demaj’s predecessor, Daniel Roth, moved to the Federal Office for the Environment after less than a year. The chief financial officer left. Peter Indra, head of the Health Office and one of the most seasoned experts on the Swiss health care landscape, left the directorate shortly before reaching retirement age; the WOZ report does not give the reasons. According to the report, the list of departures from key positions is long; the WOZ’s sources describe a leadership style marked by “micromanagement, compulsive control, habitual overruling, lack of respect”; a frequently heard verdict, they say, is that the many personnel changes are due above all to Rickli’s leadership style.¹⁸

For the sake of completeness, since a robust finding includes what speaks in the other direction: according to the directorate’s own figures, its overall staff turnover was around fourteen percent in 2022 and 2023 but most recently fell below nine percent, and thus below the long-term average. The directorate attributes the many personnel changes to the aftermath of the pandemic, after which there was “a lot of catching up to do,” and to reorganizations that “not everyone liked”; it rejects the criticism of its conduct of office as the view of unrepresentative interlocutors pursuing interests of their own.¹⁸ All of this belongs in the picture. But reorganizations are themselves leadership decisions, and a normalized overall turnover rate does not explain why the top positions in particular — where the institutional memory resides, that is, the knowledge of why things are arranged the way they are — had to be refilled so often that the directorate counted four secretaries general in just under six years. Personnel statistics are an average; expertise is not. Measured by the SVP’s yardstick, this finding is the first and perhaps the most important. The party does not demand leadership responsibility - everyone receives that along with the office - , but “the ability to assume leadership responsibility”: the capability to lead a department as its subject matter requires. That capability shows itself not at press conferences but in whether experienced professionals want to remain under a given leadership. They are voters of a special kind: they examine the conduct of office every day at close range — and they vote with their feet.

Second group of findings: the provocation cycle.

Only a few days after the directorate had been obliged to announce the departure of Roth, its short-lived secretary general - a departure that had caused a stir internally - , Rickli suggested in the SonntagsZeitung that even the abolition of mandatory health insurance should be considered.^18,21 The proposal concerned not some detail but the basic architecture of the Swiss health care system since 1994 — the core of her own policy field. All the major media reported on it, the Social Democrats (SP) collected signatures for an open letter, the outrage machine ran at full throttle. Then: nothing. Nothing more was heard of the proposal, as the WOZ noted in April 2025.¹⁸ The directorate denies any connection with Roth’s departure; the request for the interview, it says, came from the newspaper.¹⁸ That may be so. But the pattern repeated itself. In July 2025, Rickli called on the federal government to examine a ban on gender-reassignment surgery for minors, “so that we can continue the discourse”; in the canton, eighteen such procedures had been performed on sixteen- to eighteen-year-olds in the preceding two years. An inquiry commissioned by her directorate into allegations against the child and adolescent psychiatry service had found no systemic deficiencies a year earlier; the directorate had nonetheless set up a center of expertise afterward.²² Whether the demand was justified on the merits need not be decided here; the headline, in any case, was perfect, the Zurich SVP cheered, and the problems of the Health Directorate receded into the background for a news cycle.²² Once is an interview. Twice is a method: the maximalist proposal followed - as far as is publicly known - by no further step is the instrument of a politics that manages attention rather than issues. It is — the reader may savor the irony — precisely the core competence of advertising sales and communications consulting, the fields from which the health director comes: placing messages so that they achieve maximum effect. The question was never whether Rickli possesses this competence. The question was whether the second competence, the one a department demands, kept pace with it.

At this point it is worth glancing at the oldest manual of image management that European literature possesses. It was written by a Jesuit from Aragon, appeared in 1647, and is called Oráculo manual. Baltasar Gracián devotes aphorism 130 to the subject, “Hacer y hacer parecer” — doing, and letting it be seen: things pass not for what they are but for what they seem. What is not seen is as if it were not.²³

Note the order of the two verbs, for Gracián’s entire moral lies in it: first do, then show. He writes for the prudent, who should also make their abilities visible because the world, after all, judges by appearances. The modern application has spared itself the first half of the instruction and works only with the second — and the astonishing thing is how well it works on its own. A proposal that no one pursues generates the same headlines as one that is implemented, and causes considerably less work.

Third group of findings: hospital planning in court.

This finding requires a little background, because the Swiss health care system is not self-explanatory terrain for outsiders. Every canton maintains a hospital list. Only a hospital on that list can bill its treatments to mandatory basic health insurance and to the canton; a hospital that is struck from the list thus loses essentially all patients who have only basic insurance. For a clinic whose patients mostly lack supplementary insurance, delisting comes close, economically, to closure. Drawing up the hospital list is therefore the most consequential planning task of a cantonal health directorate — and one for which, according to the case law, federal law requires a proper procedure: the requirements against which hospitals are measured must be clearly defined and communicated in advance.

In August 2022, the cantonal government adopted the hospital list for the years from 2023 onward. The Adus Clinic in Dielsdorf, an orthopedic hospital in the Zurich Unterland, no longer received any service mandates on it; it was, as the Swiss news agency SDA wrote, struck from the list.^24,25 The Health Directorate deemed it “not relevant to the provision of care.”²⁶ The justification given was that it was a purely elective clinic - one that performs only scheduled procedures, no emergencies - with a selective range of services that played no role relevant to the provision of care.²⁵ Added to this was the accusation of “cherry-picking”: the clinic, it was said, concentrated too heavily on lucrative procedures.²⁷ The clinic describes itself as the lowest-cost hospital in the canton; ninety percent of its patients, it says, have basic insurance only, that is, no supplementary coverage for a private or semiprivate ward.²⁵ No court had to decide who is right on this point. Formally, the decision was made by the full cantonal government; the planning was prepared and publicly defended by the Health Directorate.

The clinic fought back, and in a judgment of July 23, 2025, the Federal Administrative Court ruled in its favor; the decision became public at the end of August.^24,26 Anyone who reads the reasoning finds no dispute about cherries but a finding about the procedure, and it can be summarized in four points. First, the canton had “not clearly defined and communicated in advance the specific requirements for the service mandates to be awarded here.” Second, the criteria it ultimately applied were “strikingly vague and open-ended.” Third, it had relied on case volumes that captured not the treatments of Zurich’s population but all discharges from Zurich hospitals, regardless of where the patients lived; it was therefore “not comprehensible in methodological terms” why it had relied on these figures. Fourth, it had not raised the argument that larger hospitals offer a broader medical spectrum until its submission to the court, “and thus too late.”²⁴ The trade press summed up: “methodological flaws with regard to the case volumes and a lack of transparency in the criteria,” a decision that did “not meet the federal-law requirements for the hospital planning procedure.”²⁶ The NZZ ran the headline “Schlappe für Natalie Rickli: Spitalschliessung im Kanton Zürich war unrechtmässig” (“Defeat for Natalie Rickli: Hospital Closure in the Canton of Zurich Was Unlawful”).²⁸ In September 2026, it became known that the clinic would be back on the hospital list as of January 2027, with a service mandate for orthopedics, “definitively and without time limit.”^25,27

One need not be a lawyer to grasp the core of this judgment; it is enough to have taken an exam once. An exam whose standard is not known to the candidates in advance, whose criteria are so loosely worded that they can support almost any result, and whose reasoning is in part supplied only in the appeal proceedings is not an exam one can rely on — regardless of whether its result happens to be right in a particular case. And anyone who gauges a village’s need for bread by the sales of its bakery counts the travelers passing through and takes the result for the needs of the residents. These are precisely the flaws the court records in legal language. The court did not examine whether the result was settled beforehand, and this text makes no such claim; the finding weighs heavily enough without that insinuation. Note the kinship with Section II. There the subject was an electoral procedure that does not measure what it purports to measure. Here the subject is a planning procedure about which a court has found the same. And those who fail to notice the flaws of a measuring procedure before applying it, noticing them only when a court points them out, provide a textbook case of the double curse of incompetence from Section I: the same care that would have produced sound planning would have been needed to recognize unsound planning as unsound.

Here, too, the points in the directorate’s favor belong in the picture, for there are some. In the same planning cycle, the canton won another case in November 2025: the Canton of Thurgau had appealed against the rehabilitation list; the Federal Administrative Court dismissed the appeal without reaching the merits, holding that the neighboring canton lacked standing because it had no interest worthy of protection in litigating over the occupancy of its own clinics.²⁹ This victory, however, says nothing about the method of Zurich’s planning, since the method was not examined. Fairness also requires noting that the canton has implemented the Adus judgment. Correction, then, is possible. In this case it came after a court had set the price for it — and that is not self-correction but correction from outside. Section V will show why such a system rarely knows any other kind.

Fourth group of findings: criticism as attack.

Cardiac surgery at the University Hospital is the case the public most strongly associates with the health director — and at the same time the one in which the record speaks most clearly in her favor. Between 2016 and 2020, an above-average number of patients died at the Department of Cardiac Surgery of the University Hospital Zurich (USZ); at the center of the reporting was a controversial implant that the then head of the department had helped develop.^18,30 In May 2026, an independent investigative commission appointed by the hospital board and chaired by former Federal Supreme Court judge Niklaus Oberholzer presented its report. On the basis of 4,500 operations, it established a statistically demonstrated excess mortality of 68 to 74 deaths and judged the use of innovative medical devices to have been inappropriate in 13 cases.³¹ Excess mortality means that this many more patients died than would have been expected given the risks of comparable procedures. The hospital summarized the report as follows: the hospital board at the time, the hospital’s highest governing body, “underestimated the gravity and inadequately pursued the misconduct identified.” The Health Directorate, by contrast, had “according to the investigation report repeatedly raised the question of patient safety and pointed out the various irregularities and leadership shortcomings at an early stage.”³¹

Rickli herself said after the publication: “I was informed at the end of January 2020 about the latest developments and incidents at the heart clinic. We reacted immediately, put questions to the hospital management and the hospital board, and demanded documents.” And: “What I lacked at the time were the legal means to order an administrative investigation myself.”³² Both statements are consistent with the chronology as the NZZ has reconstructed it: when the first warnings reached her, Rickli had been in office for less than nine months; at the handover, no one had told her about the conspicuous mortality; she learned of problems at the end of January 2020 through an internal whistleblower report — just before, as the cantonal government later found, the pandemic overwhelmed her directorate.³⁰ Afterward she demanded information from the hospital board more frequently, saw to it that the hospital’s internal investigation was expanded to include the whistleblower’s questions, and commissioned a governance report that triggered reforms.³⁰ When, in early 2021, an external investigation commissioned by the hospital concluded that no patients had been endangered and that nothing was criminally relevant, the directorate, by its own account, saw no reason to doubt it.³⁰ The instrument of an administrative investigation of its own - an investigation that the supervisory authority orders itself instead of leaving it to the hospital - will become available to the directorate, under a decision of the cantonal government, only in 2027.³⁰ Whether more would have been possible under these circumstances is disputed; the NZZ has raised the question and compiled the arguments on both sides.³⁰

Therein lies the irony of this record. In the case for which the health director was most fiercely attacked, an investigative commission headed by a former Federal Supreme Court judge has certified that her directorate asked the right questions early. Anyone who wants to blame her for the deaths in cardiac surgery will find no support in this report, and this text does not do so. What is documented, by contrast, is something else: how the directorate dealt with the press coverage.

When the SonntagsBlick wrote about a whistleblower report that had allegedly been left lying unattended with Rickli, a counterstatement under the right of reply went out, according to the WOZ’s account, “immediately.”¹⁸ The right of reply is the statutory right to answer a statement of fact published in the media with one’s own version; invoking it is legitimate. When the financial journalist Lukas Hässig reported at the end of April 2024 on his portal Inside Paradeplatz on the deaths and on the question of when Rickli had been informed, Rickli and the Canton of Zurich - after conciliation had failed - filed suit with the district court on November 14, 2024, for violation of personality rights, coupled with the demand that the journalist pay CHF 10,000 in non-pecuniary damages to the organization Reporters Without Borders.^33-35 Here, too, the other side of the ledger: Hässig’s article contained an error, which he himself acknowledged — it gave 2021 as the date on which Rickli had received mortality figures for the clinic; the correct date is March 2022.^33,35 A member of a cantonal government may defend herself against false statements of fact; that is entirely her right. In the summer of 2025, the proceedings ended in a settlement: Inside Paradeplatz published a clarification; according to Hässig’s account, Rickli waived the damages and paid the court costs.³⁴ That the damages were originally to go, of all places, to an organization that champions press freedom worldwide remains a punch line no one would need to invent. And while the WOZ was working on its report on the directorate, there was, according to its account, an “intensive search” behind the scenes for people who had given the paper information.¹⁸ Should that be true, it would not be the most productive use of administrative resources.

Taken on its own, each of these reactions may be permissible, and in the Hässig case the objection was justified on one point. What is revealing is not the individual reaction but their common direction. The counterstatement and the lawsuit ask whether the picture a report paints of the head of the directorate is accurate; the search reported by the WOZ does not even ask that, only who talked. None of the three reactions asks what a report says about the institution. Those who treat criticism as information examine the substance first. Those who treat it as an attack examine their own image first and then the source — and an examination of sources is more likely to find a suspect than an error. This does not contradict the finding of the investigative commission; it complements it. Of the hospital, the directorate asked the right questions. Of the press, it asked different ones. This is the way out described in Section I, turned into its opposite: outside feedback, fended off instead of heeded.

The Romans with whom this text began would have asked at this point: Cui bono — who benefits? The answer sounds paradoxical and is nonetheless compelling: criticism benefits no one more than the person criticized — provided that person reads it. It supplies what one’s own organization does not, because every organization tends to please those at the top. Whoever fends it off forgoes the cheapest quality control there is; it does not even cost a consultant’s fee. Plutarch has recorded how costly such forgoing can become. When the first messenger reported to the Armenian king Tigranes that the Roman army under Lucullus was advancing, the king had his head cut off. After that, no one reported anything to him; Tigranes listened only to those who flattered him and sat in ignorance while the war was already blazing around him.³⁶ The methods have grown milder since then, but the mechanism has remained the same: whoever punishes the messenger abolishes not the bad news but the next messenger. Measured by the SVP’s yardstick, this means: leadership capability also shows itself in whether someone can read an unpleasant message before looking for its messenger — and in whether, under that person’s leadership, the next messenger still finds the courage to come forward.

V. The Logic of the System: Why Nothing Corrects Itself

Before the four groups of findings are brought together, a self-disclosure is due, one that Robert Musil placed at the beginning of his lecture “On Stupidity” (“Über die Dummheit”) in Vienna in 1937: to speak about stupidity, one must presuppose that one is not stupid oneself, and thereby display that one considers oneself clever — which is generally regarded as a sign of stupidity.³⁷ The objection is justified and is expressly not refuted here; it applies to this text and to its author as well.

Musil then distinguishes honest stupidity, which is simply slow, from a second kind, which he calls the higher or pretentious kind and describes as the true disease of education — the most dangerous disease of the mind. What is meant is not primarily a lack of intelligence but an education that has gone awry: a mismatch between what someone knows and what that person passes judgment on. In the language of 1937, this is the double curse of incompetence from Section I. Of this higher stupidity Musil writes the sentence that explains the four groups of findings better than any organizational analysis: there is, he says, simply no significant idea that stupidity would not know how to apply; it is mobile in every direction and can put on all the garments of truth — truth, by contrast, has only one garment and one path at a time and is always at a disadvantage.³⁷

This is precisely the asymmetry that brings a directorate into such a state. Correct work on the issues has a single form: the right one. Communication has unlimited forms and can dress any of them up as substantive policy — as the courage to reform, as concern for young people, as protection of the institution against false reporting. The Italian economic historian Carlo M. Cipolla formulated a related and even more uncomfortable rule of behavior: the stupid person is more dangerous than the bandit, for the bandit follows an understandable incentive and is therefore predictable, whereas the stupid person causes harm without gaining anything from it.³⁸ Administrations know this from experience: against self-interest there are controls; against sincere conviction, none. “There is nothing more frightful than ignorance in action,” Goethe wrote.³⁹

Four groups of findings, one pattern. The drain of expertise, the provocation cycle, hospital planning without criteria fixed in advance, and the handling of criticism are not four problems but four manifestations of the same problem — they correspond precisely to the compensations that Section II derived from the double curse. Those who do not steer an issue at the level where its substance is dealt with steer what remains steerable: perception. Where a leadership style of the kind the WOZ’s sources describe allows little room for expert dissent, experienced people leave — and with them precisely those who would be the corrective. Those who do not set the standards of a planning process in advance measure with a yardstick that comes into being only as they measure — and learn from a court what their own people should have told them. Those who do not use criticism as information process it as an attack: with a counterstatement, a lawsuit, and, according to the WOZ’s account, a hunt for informants. None of these observations requires a look into anyone’s inner life; they describe documented official conduct whose logic follows from the structure. That is the difference between an analysis of conduct and a diagnosis from a distance — and at the same time the reason why the finding is so inescapable: a weakness of character could be remedied. A structure that prevents its own correction cannot be remedied from within.

The historian Barbara Tuchman described this pattern as wooden-headedness: assessing a situation in terms of fixed preconceptions while ignoring all contrary signs. For a finding of folly, she requires, among other criteria, proof of two points — that the conduct was recognized as harmful in its own time and that another course was available.⁴⁰ Both apply here: the criticism of the four sets of events is dated and published, in one case even as a court judgment, and the other course consisted in each case of doing the obvious — retaining the experienced senior staff, following up on the proposal, fixing the criterion before the evaluation, examining the criticism instead of asking where it came from. This text does not claim that her third criterion is met, namely a policy pursued by an entire group and outlasting a single tenure in office; it describes a tenure, not an epoch.

This brings the argument full circle, back to Kruger and Dunning. Such a system can hardly correct itself from within — by the logic of the double curse, the deficiency hides itself from view, and with the drain of senior staff, those who could have pointed it out are gone. Corrections therefore come, as a rule, from outside, and, as the hospital planning shows, they then come with a docket number. The only outside authority, however, that decides who holds the office, and that need not first be called upon to do so, has demanded no correction: in 2023 the incumbent was reelected. Some of the events documented here still lay in the future at the time. That is true, but it exonerates only the voters, not the procedure: even an election held in full knowledge of all four groups of findings would have measured only what elections measure. The procedure that produced the casting has confirmed it — and with that, the talent show delivers its final punch line: it judges the show, and the show was flawless.

Many things may therefore be said of Zurich’s voters, but not that they were deceived. They chose their health director themselves, twice, in free and secret elections, the second time after four years in which they could watch her at work. The field the elected official comes from calls this customer loyalty. Pity would be out of place here, for it would imply that the voters did not know what they were doing — and that would be the real insult. They pay nonetheless, as taxpayers, as premium payers, and, if it comes to that, as patients, and they pay three times over: with the turnover costs of an administration whose top positions do not settle down; with the opportunity costs of substantive policy that gets short shrift amid the headlines, that is, with the value of what could have been accomplished in the same time; and with the costs of a planning decision that only a court put right. The audience chose the program. In doing so, it also signed up for the costs. In 2027 it may decide anew — unless the Federal Assembly takes the decision off its hands in December.

For the next season may already be in the works, this time on the federal stage. The difference is a procedural one, but it decides everything: this time the hall holds not an audience but a jury.

VI. The Test That Remains

What follows from this? Not a call for formal qualification requirements for elected offices — that would be as illiberal as it would be ineffective, and Adolf Ogi is living proof that diplomas are the wrong criterion. The commercial-school graduate and the commercially trained advertising-sales professional started from the same line: without a university education, without a background in the field, without any proof of fitness for the office they took on. The former turned unproven fitness into proven fitness by working, listening, and conceding his defeats. The latter’s conduct of office shows, in the four documented groups of findings, the alternative. The difference was never the diploma. The difference lies in how one deals with one’s own gap — and behind that, if Kruger and Dunning are right, stands the metacognition that can be neither elected nor delegated nor marketed.

And yet there is one place where a government election can serve as a test, and it will itself be put to the test in the coming weeks. Of Switzerland’s twenty-seven governments, twenty-six cantonal and one federal, exactly one is not elected by the people: the Federal Council. It is elected by the United Federal Assembly, that is, the 246 members of the National Council and the Council of States sitting together, many of whom have known the candidates for years from committees, parliamentary group meetings, and floor debates. The parliamentary groups interview the applicants before nominating; the record in office of cantonal government members is documented and publicly accessible; and those who vote know the files, or can. That is not a talent show but an examination board — at least when it behaves like one.

The Swiss people, incidentally, expressly wanted this division of labor. On June 9, 2013, with just over three quarters of the vote and in every single canton, they rejected a popular initiative from within the ranks of the SVP that would have had the people elect the Federal Council.⁴¹ Without offending the voters, one may recognize in this a form of collective metacognition: a sovereign that knows it cannot examine candidates for the national government leaves the examination to those who can. It is the insight Machiavelli attributed to the second mind — it understands what others show it. And it ranks among the more gratifying ironies of this story that the same party which in 2013 wanted to make the people the electoral body is, in 2026, handing the Federal Assembly a standard for its examination.

The timetable is set. By noon on October 23, the SVP’s cantonal branches submit their candidates; the search committee then interviews them and presents its recommendation to the parliamentary group’s executive committee by November 13; the parliamentary group is expected to make its nomination on November 20, and on December 9 the Federal Assembly votes.²⁰ That leaves almost seven weeks in which to test what no election campaign tests.

For this test, the benchmark is ready at hand, and it is the same one from which this text drew its thesis. The department for which the SVP today expressly demands leadership capability was once before taken over by a man without a university education, amid talk of a “relegation to the B league”; in the judgment of the Historical Dictionary, he helped overcome the army’s crisis of purpose after the end of the Cold War.^12,16 The yardstick, then, is not abstract. It has a name and a record, and against it the four groups of findings in this text can be translated into four questions that any parliamentary group can put to any candidate: Do experienced people stay under your leadership? Do you follow up on your proposals once the cameras are gone? Do you set the criteria before you evaluate? And do you read the criticism before you look for the critic?

The Federal Assembly should know two temptations before it succumbs to them. The first is the Peter Principle from Section II: the tendency to promote those who have held their own at their current level instead of examining whether they are fit for the next. In this sense, a reelection is no proof of capability; whoever reads it that way confuses confirmation with examination and sales figures with leadership capability. The Roman Curia captured the second temptation in an elegant formula: Promoveatur ut amoveatur — let him be promoted so that he may be removed. For this reason, many a Zurich resident may look upon a candidacy with quiet sympathy. For the Federal Assembly, it would be the worst of all maxims, because it would turn a Zurich question into a federal one. The people of Zurich elected their health director themselves, twice; that was their right, and it remains their business. The Federal Assembly has the advantage of not having to follow that choice — neither in the sense of understanding it nor in the sense of following suit.

And finally, the hospital planning has supplied a lesson that finds its last application here: one fixes the standard first and then applies it, not the other way around. The SVP has fixed it, in writing and dated: the ability to assume leadership responsibility. All that is missing is its application. In everyone’s judgment, Galba was fit to rule as long as he did not rule; what had been examined was only how he served, not how he would rule. Whether such a sentence will one day be written about Guy Parmelin’s successor, or its opposite, will not be decided in office alone. It will be decided in the examination beforehand — and this time there is one.

Where the electoral procedure does not test this ability and the person elected does not possess it, what remains in the end - except in elections to the Federal Council - is only one examining authority that decides on the person: the thinking public. Journalists who document rather than insinuate — the WOZ’s investigative reporting and the coverage of the cardiac surgery affair and of hospital planning show that they exist. And citizens who read, compare, and remember instead of voting for whatever glitters brightest. That is strenuous, and therein lies the difficulty: the talent show exists because it relieves the audience of this effort. Those who want to abolish it must take that effort upon themselves. Musil ended his lecture by pointing to the last and most important remedy for stupidity: modesty.³⁷ That is no exercise in humility but the practical side of metacognition — knowing where one’s own judgment ends and someone else’s is needed. It cannot be elected, delegated, or marketed; it can be acquired only along the path that Kruger and Dunning found in their laboratory: by learning and allowing oneself to be corrected along the way. Thinking for oneself remains an imposition. It requires no one’s permission, only one’s own willingness to take on the effort of thinking.

Disclosure: The founder of this journal is affected by a pending administrative proceeding conducted by the Health Directorate of the Canton of Zurich; this text relies exclusively on the listed, publicly accessible sources.

Frequently Asked Questions

What Is the Dunning-Kruger Effect?

The observation, described by Justin Kruger and David Dunning in 1999, that the least competent overestimate themselves the most. Kruger and Dunning attribute this to a lack of metacognition: the same ability that produces a correct solution is needed to recognize one’s own wrong solution as wrong. Incompetence therefore produces errors and at the same time prevents them from being noticed.

Why Don’t Elections Test Competence?

An election campaign measures telegenic appeal, party loyalty, the ability to polarize, and name recognition — the abilities needed to win an office. It does not measure expertise, leadership, or judgment, the abilities needed to do justice to an office. The Swiss militia system once relied on a record established before entry into politics to supply this test; where the career itself consists of public relations, the test disappears.

How Is a Member of the Swiss Federal Council Elected — and Can Competence Be Tested in the Process?

The members of the Federal Council are elected by the United Federal Assembly, that is, the 246 members of the National Council and the Council of States sitting together. The parliamentary groups interview the candidates before nominating them, and many members of parliament know them from years of working together. This makes the Federal Council election the only government election at the federal or cantonal level that is conducted by a body able to test competence. The Federal Assembly will elect Guy Parmelin’s successor on December 9, 2026.

Does a Government Office Require a University Education?

No. Without a university education, Adolf Ogi led a technically demanding department to results one can visit, because he treated his competence gap as a work assignment: mastering the issues, listening to experts, conceding defeats. What matters is not the diploma but how one deals with one’s own gap.

What Criticism of the Zurich Health Directorate Is Documented?

For the period since 2019, independent media investigations document numerous departures from key positions, a leadership style described by sources as micromanagement, spectacular proposals that, as far as is known, were followed by no further steps, and a lawsuit, settled in 2025, against a journalist in connection with the cardiac surgery scandal at the University Hospital. In 2025, the Federal Administrative Court found that the removal of the Adus Clinic from the hospital list rested on case volumes whose use was not comprehensible in methodological terms and on vague criteria not set in advance. In the directorate’s favor, the investigation report on the cardiac surgery affair records that the Health Directorate repeatedly asked about patient safety. The directorate rejects the criticism of its leadership style; all sources are listed in the article.

What Can Citizens Do?

Take on the examination that the electoral procedure does not provide: read, compare, remember — pay attention to journalism that documents, instead of voting for whatever glitters brightest. Thinking for oneself remains an imposition, but it is the only authority that can correct the talent show.

Your Comment

Are you interested in contacting the author? Do you have remarks, suggestions, or additions regarding this article? Do you have personal therapy experiences? We welcome substantial feedback.

Website

Name * Email *

Your email address will not be published. It is used solely for potential inquiries by the editors.

Comment *

I agree to the processing of my data in accordance with the Privacy Policy. *

Submit Comment