The Charisma Trap and the Two-Sentence CEO

This is the capstone, and it begins with an error in the data you have been using your whole career. The CEOs you know about — the ones with biographies, keynote slots, magazine covers and quotable strategies — are not a sample of CEOs. They are a sample of visible CEOs, and visibility is produced by a machine that rewards drama, charisma, size, acquisitions and grand predictions. The quiet CEO who compounded a mid-sized manufacturer at 12% a year for eighteen years does not get a cover. Your intuitive model of "what great CEOs are like" was trained on a biased sample, and the bias runs toward extraversion, narcissism, boldness and story.

Two corrections follow. Visibility ≠ prevalence: most CEOs, including most excellent ones, are not like the ones you have heard of. Visibility ≠ effectiveness: the evidence on celebrity, awards and charisma does not show that visible CEOs perform better, and in some designs shows the reverse.

Then the module names the capability the whole course has been circling. A mature CEO can say, and mean, two sentences most people find psychologically incompatible: "We're going to do this" and "I was wrong. Change the plan." Charisma supplies the first. Only maturity supplies both.

In the Effectiveness Equation — Traits × Behaviors × Organizational Context × Current Moment — this module is about Behaviors, specifically the behavior of updating in public. It closes by asking you to write the plan that turns the course into changed behavior.

Visibility ≠ prevalence; visibility ≠ effectiveness. The rarest CEO capability is being able to say both "We're going to do this" and "I was wrong. Change the plan."

The first sentence requires conviction, agency, confidence and the willingness to lead people into uncertainty. The second requires intellectual humility, psychological security, evidence orientation and adaptability. The public rewards CEOs who say the first sentence loudly. Companies are built by CEOs who can say both — and the second sentence is the one that keeps the first from becoming fantasy.

FRAMEWORK Hayward, Rindova & Pollock (2004) developed the theory of CEO celebrity. Journalists attribute a firm's distinctive and consistent strategic actions to the CEO's disposition, which creates celebrity; CEOs who internalize that attribution become overconfident about their own efficacy and persist with the actions that generated the celebrity even when those actions stop paying off. The paper is theory-building — propositions, not tests — and should be cited for the mechanism, not for effect sizes. Its title is the mechanism in three words: believing one's own press.

RESEARCH FINDING Malmendier & Tate (2009) tested the mechanism using prestigious business-press CEO awards as a shock to status, comparing winners with predicted winners who did not win, in U.S. large firms from 1975 to 2002. Award-winning CEOs subsequently underperformed relative to their own prior record and to matched non-winners; they received higher pay, spent more time on outside activities such as board seats and book-writing, and their firms showed more earnings management. Effects were strongest in firms with weak governance. The matching design addresses mean reversion partly but not fully. What it supports: the moment a CEO becomes a "superstar" is, on average, the beginning of a worse period — larger where nothing constrains the star.

RESEARCH FINDING Chatterjee & Hambrick (2011), using the unobtrusive narcissism index (photo prominence, press-release prominence, first-person pronouns, pay gap — proxies, not clinical measures) on U.S. public-company CEOs, found that CEO risk taking generally rises after positive "capability cues" and falls after negative ones — but highly narcissistic CEOs are much less responsive to objective performance feedback and considerably more responsive to social praise such as media coverage and awards. Archival panel, associations. INTERPRETATION This is the charisma trap's engine: the CEOs most likely to attract praise are the CEOs whose behavior is most driven by praise and least by results.

RESEARCH FINDING Kaplan & Sorensen (2021), from 2,603 structured assessments of candidates for top roles (2000–2013), found four factors explaining over half the variance in thirty rated characteristics: general ability; execution versus interpersonal; charisma versus analytical; strategic versus managerial. Candidates with stronger interpersonal skills were more likely to be hired, yet the factors — including execution — predicted later advancement to CEO, suggesting boards may over-weight interpersonal skills at hiring. Kaplan, Klebanov & Sørensen (2012), on 316 PE/VC-backed CEO candidates, found execution abilities predicted subsequent success more strongly than interpersonal abilities. Single assessor, selected samples, observational.

RESEARCH FINDING Judge, Bono, Ilies & Gerhardt (2002) meta-analyzed 222 correlations from 73 samples — mostly students, military and mid-level managers, not CEOs — and found extraversion the most consistent Big Five correlate of leadership (.31, emergence and effectiveness combined), with conscientiousness .28, neuroticism −.24, openness .24, agreeableness .08 and a multiple R of .48. The criteria matter: "leadership" included peer-rated emergence (who gets seen as a leader) and supervisor-rated effectiveness, and the two diverge for some traits — conscientiousness related more to emergence (.33) than effectiveness (.16); agreeableness to effectiveness (.21) but not emergence (.05). INTERPRETATION Being chosen as a leader and being effective as one are different criteria with different correlates.

RESEARCH FINDING Harrison, Thurgood, Boivie & Pfarrer (2020), applying a validated language-based personality tool to nearly 3,000 S&P 1500 CEOs, found observed extraversion associated with higher market-perceived firm risk, and observed conscientiousness with lower perceived risk and better returns. Gow, Kaplan, Larcker & Zakolyukina (2016), in a working paper using a noisier language model on ~4,600 CEOs, found extraversion negatively related to contemporaneous and future ROA. Both are proxy-based and correlational. The fair counterpoint: Malhotra et al. (2018) found extraverted CEOs' acquisitions earned stronger abnormal announcement returns. Extraversion is not a defect. It is a trait the selection machine over-rewards.

RESEARCH FINDING— CONTESTED. Quigley & Hambrick (2015) found the share of performance variance attributable to CEOs rising over recent decades (7.8% to 15.7% of ROA variance by sequential ANOVA across three periods); Fitza (2014, 2017) argued that with tenures averaging about four years, a large part of any measured CEO effect is indistinguishable from chance — a press summary of the 2017 paper puts it at over 70%; Quigley & Graffin (2017) reaffirmed a significant effect "much larger than chance." Unresolved. INTERPRETATION Whatever the true number, it is far smaller than the attribution the press makes to any individual celebrity CEO.

RESEARCH FINDING On the second sentence: Ou, Waldman & Peterson (2018), in 105 U.S. tech SMEs, found CEOs rated as more humble by their top teams had more integrated teams, in turn linked to more ambidextrous strategy and better performance — correlational, small, single-industry. Owens & Hekman (2012), from 55 interviews, identified three humble-leader behaviors — admitting mistakes and limits, spotlighting others' strengths, modeling teachability — and noted that humility appears less effective under extreme threat. Milliken, Morrison & Hewlin (2003) found 85% of 40 interviewed employees could recall feeling unable to raise an important issue upward; Detert & Edmondson (2011) found implicit beliefs about the danger of speaking up suppress candor even in objectively safe settings. Neither is CEO-specific.

Where the evidence is weak

Celebrity is measured by awards and press coverage, narcissism by photo size and pronouns, extraversion by speech; every trait here is a proxy. The award study is the only quasi-experimental design, and it cannot fully rule out mean reversion. The Judge meta-analysis is not about CEOs. Humility's link to firm performance is less definitive than its link to team process. The Two-Sentence Test is a FRAMEWORK — the course's synthesis, not a construct any study has measured. Treat it as a lens to test on yourself, not a finding.

The machine that decides who you have heard of

INTERPRETATION Business media has a product to sell, and the product is narrative. Narrative needs a protagonist, a conflict, a bold move and a quotable line. The CEO who makes a giant acquisition, predicts the future on stage, founds a company at twenty-four, or fights a public war with a competitor supplies all four. The CEO who spent eighteen years improving working capital, promoting from within, saying no to nine acquisitions and yes to one, and compounding earnings at 12% supplies none. Hayward, Rindova & Pollock (2004) describe the first half of the loop: journalists attribute distinctive action to disposition, creating celebrity. Chatterjee & Hambrick (2011) describe the second half: the CEOs most responsive to that celebrity are the ones whose risk taking rises with praise. The loop selects for a type, then amplifies it.

Add the selection findings and the picture completes. Boards over-weight interpersonal skill and charisma at hiring (Kaplan & Sorensen, 2021). In general samples, the trait most associated with being seen as a leader is extraversion (Judge et al., 2002). The machine that produces CEOs and the machine that produces famous CEOs share a filter, and the filter is not "effective." It is "compelling."

Hence visibility ≠ prevalence. Bandiera et al.'s (2020) 1,114 manufacturing CEOs are a closer approximation of the real population than any list of famous names, and what distinguishes them in the data is how they spend their weeks, not how they perform on stage. Most CEOs are not founders, celebrities, acquirers or prophets. Most excellent CEOs are not either.

And visibility ≠ effectiveness. Malmendier & Tate (2009) is the cleanest test: win the award, then underperform, get paid more, write the book, manage the earnings — most where governance is weak. Harrison et al. (2020) and Gow et al. (2016) find observed extraversion associated with higher perceived risk and, in the working paper, lower ROA. Chatterjee & Hambrick (2007) find narcissism associated with variance, not mean. None of this shows that charisma hurts. It shows that charisma is not evidence of effectiveness, and that the celebrity it attracts is, on average, associated with a worse subsequent period.

What this means for your own model of a great CEO

INTERPRETATION You carry a picture of the CEO you would like to be, assembled from visible CEOs. Run it through the correction. The bold acquirer: Malmendier & Tate (2008) found overconfident CEOs roughly two-thirds more likely to acquire, with the market reacting more negatively. The grand predictor: Hayward et al.'s persistence proposition. The charismatic communicator: the trait boards over-weight and the trait most associated with emergence. What survives is less photogenic — execution (Kaplan, Klebanov & Sørensen, 2012), conscientiousness (Harrison et al., 2020), an integrated senior team (Ou et al., 2018) — and it compounds. This is not an argument for dullness; Module 4 established that overconfidence has adaptive uses in innovative settings, and Module 12 that ambition is half the bank CEO's job. It is an argument for base rates. Before you emulate a CEO, ask how you came to know about them.

The Two-Sentence CEO Test

FRAMEWORK A mature CEO can say, and mean, both of these:

"We're going to do this." This sentence requires conviction — a belief strong enough to commit resources and people against uncertainty; agency — the sense that outcomes depend on what you do, which Module 1 identified as core to CEO temperament; confidence — enough to absorb the doubt of others without absorbing their paralysis; and leadership — the willingness to be the person who is wrong if it fails. It is the sentence of the Player, the founder and the turnaround CEO. Without it, nothing starts. Owens & Hekman's (2012) observation that humility is less effective under extreme threat is the reminder that there are moments when only this sentence will do.

"I was wrong. Change the plan." This sentence requires intellectual humility — holding a belief and its refutation at the same time; psychological security — an identity that does not depend on the belief being true, so that abandoning it is not abandoning yourself; evidence orientation — asking what would change your mind before you have to; and adaptability — redirecting an organization you have just pointed somewhere else. It is the sentence of the Architect, the bank CEO and the second-tour turnaround CEO. Without it, the first sentence eventually becomes fantasy.

INTERPRETATION Why is doing both so rare? Because each sentence is easiest for a temperament that finds the other hardest. The person who says "we're going to do this" with the most conviction is often the person for whom being wrong is most threatening — the narcissist who discounts results and amplifies praise (Chatterjee & Hambrick, 2011), the overconfident acquirer, the celebrity who persists (Hayward et al., 2004). The person who says "I was wrong" most easily is often the person who never committed hard enough to be wrong at scale. Charisma supplies the first sentence in abundance and often an inability to say the second. That is why the ability to do both is far rarer than charisma — and why it, not charisma, is what the course means by a calibrated CEO.

Module 4's rubric applies. Productive conviction: "We can probably accomplish this. Here is the evidence, the assumptions, and what would make us change course." Epistemic arrogance: "We will accomplish this. People who disagree simply don't understand the vision." The first version contains the second sentence in advance. A CEO who states, at the moment of commitment, what would make them reverse has made "I was wrong" a planned event rather than a humiliation.

Why the second sentence is an information problem, not a virtue

INTERPRETATION Module 5 established that the CEO's information environment is distorted: most employees can recall withholding an important concern (Milliken et al., 2003), and implicit rules against speaking up persist even where it is safe (Detert & Edmondson, 2011). Every time a CEO defends a failing plan, the organization learns that evidence against the plan is unwelcome, and the next piece of evidence does not arrive. The charisma trap and the information bubble are the same trap, viewed from inside and outside.

The practical response is mechanisms, not intentions: pre-stated reversal criteria; a person whose job is to bring the bad news; a regular review whose only agenda is "what has stopped being true"; and the discipline of saying the second sentence early and visibly on small things, so that saying it on a large thing is not unprecedented. The Personal Calibration Plan asks you to name yours.

Fictional composite.

Two CEOs ran industrial filtration companies of similar size in the same decade. You have heard of one of them.

Kip Marchand became CEO of Vireo Systems, a $600M-revenue listed filtration maker, at 44. Within three years he had announced a "platform" strategy on stage at an industry conference, acquired four companies for a combined $900M, and been profiled by two national magazines as the executive who would "reinvent an unglamorous industry." He won a trade-press CEO-of-the-year award in year four. Revenue reached $1.6B. Vireo's stock tripled. He gave a well-received talk about conviction.

Ingrid Sallow became CEO of Alder Filtration, a $520M-revenue listed competitor, at 51, promoted from operations. She did not give conference talks. In seventeen years she made one acquisition, for $140M, after declining nine others; she wrote in an annual letter that the nine had each failed a written test the board had agreed in advance. She replaced the sales compensation plan twice — the second time reversing the first, with a paragraph explaining why the first design had been her mistake. Alder's revenue grew to $2.4B, almost entirely organically; operating margin rose from 9% to 16%; the stock compounded at roughly 13% a year. No magazine profiled her. A search partner once described her, not unkindly, as "the least interesting excellent CEO I know."

In year six of Marchand's tenure, two of the four acquisitions were written down by a combined $380M. The platform strategy had been persisted with for two years after the integration data turned; three senior executives who had raised the data had left. The board, which had combined the chair and CEO roles in year four, appointed an independent chair in year seven, and Marchand departed in year eight with revenue at $1.3B and the stock below where it had been when he arrived. The magazines ran a piece on hubris.

When Marchand was on the cover, which company was performing better? Vireo — on the metrics the cover measured. When did the divergence become visible to the public? Around year six, two years after it was visible inside Vireo. What did the public learn? Marchand's name twice — once as a hero, once as a warning — and never Sallow's at all.

Sallow said both sentences. "We're going to do this" — one acquisition, two compensation redesigns, a $200M plant. "I was wrong. Change the plan" — in writing, about her own design, in a year when nothing forced her to. Marchand said the first sentence brilliantly and never managed the second until a board said it for him. Visibility ≠ prevalence: for every Marchand there are several Sallows you will never read about. Visibility ≠ effectiveness: the cover was a lagging indicator of the first half of the story and a leading indicator of the second.

Put four archetypes in Marchand's position in year four — the award just won, the integration data just turned, the platform strategy publicly identified with the CEO.

The Visionary CEO reads the integration data as noise and the award as signal — the pattern Chatterjee & Hambrick (2011) describe. Gain: if the platform thesis is right and merely early, this CEO's persistence carries it through the trough. Cost: this is the archetype for whom the second sentence is hardest, because the vision is the identity; two years of persistence after the data turned is its characteristic failure, and the combined chair role removed the only external source of the second sentence.

The Operator CEO would have read the integration data as the only thing that mattered. Gain: the two failing acquisitions are restructured or divested in year four, at perhaps $120M rather than $380M, and the executives who raised the data are promoted rather than lost. Cost: the operator is likely to over-correct — abandoning the platform idea rather than narrowing it — and to say the second sentence so readily that the organization stops believing the first. Never being wrong because never being committed.

The Founder CEO — imagine Vireo founder-led — has the highest discretion and the deepest identification with the strategy (Modules 9 and 11). Gain: exploratory conviction and a long horizon. Cost: with no board that can say the second sentence for them, the founder's ability to say it is the whole game. Lee, Hwang & Chen (2017) found founder CEOs measurably more optimistic on language- and option-based proxies; that optimism is why the second sentence is hardest here.

The Public Company CEO who has internalized the calibrated model would have done something none of the others do by instinct: at the moment of announcing the platform in year one, stated what would make the company reverse it. Gain: when the data turned, the reversal was a pre-planned event, announced as the execution of a stated criterion rather than the confession of a failure — and the executives who brought the data were doing the CEO's stated bidding. Cost: the year-one announcement is less thrilling; the profile is less likely; the award may never come. That is the trade. It is a good one.

FRAMEWORK— the Overuse Ladder for the charismatic CEO. Confidence → arrogance; vision → fantasy; persistence → stubbornness; dominance → intimidation; optimism → delusion. Each rung is a place where the first sentence has been said and the second cannot be.

The characteristic sequence is not a collapse but a substitution: social feedback replaces performance feedback. Chatterjee & Hambrick (2011) describe it — responsiveness to results falls, responsiveness to praise rises. Malmendier & Tate (2009) describe it from outside — after the award, more pay, more outside activity, more earnings management. Hayward et al. (2004) describe why it is stable — the CEO experiences reversing the celebrated strategy as reversing themselves. Milliken et al. (2003) and Detert & Edmondson (2011) describe why no one stops it — the people who could are calculating, correctly, that the CEO does not want to hear it.

The opposite failure exists. The CEO who can say "I was wrong" and cannot say "we're going to do this" has climbed a different ladder: humility → hesitation; adaptability → strategy-of-the-month; inquiry → paralysis. Owens & Hekman (2012) found humility less effective under extreme threat. An organization led by a CEO who reverses easily and commits rarely learns that nothing is real, and stops executing.

Early warning signs

For the CEO:

• You have been described in print as a visionary, and you found it accurate.
• Your last three major decisions were each called "bold," and none was accompanied by a written statement of what would reverse it.
• You can name the last award; you cannot name the last time you reversed a decision in public.
• The executives who disagreed with the flagship strategy have left, and their departures were described as "not a fit."
• Your time on boards, media and speaking has risen, and your pay has risen faster than your company's results.

For the board:

• The CEO's public identification with a strategy is so strong that reversing it would read as a vote of no confidence in the CEO, and the board has started to weigh that.
• Board materials have become presentations; the numbers arrive after the narrative.
• Nobody on the board can state the criterion under which the flagship strategy would be abandoned.
• The board has been asked to combine the chair and CEO roles during a period of praise (Module 11).

1. List the five CEOs who most shaped your picture of the job. For each: how did you come to know about them? What does the list tell you about the machine that selected it?
2. When did you last say "We're going to do this" about something genuinely uncertain? When did you last say "I was wrong. Change the plan" in front of your senior team? Which was longer ago, and why?
3. Which sentence is harder for you — because of temperament, or because of what you believe it would cost you in others' eyes?
4. Name a decision you are currently defending. Write, now, the evidence that would make you reverse it. If you cannot, what does that tell you?
5. Who in your company is paid, in effect, to bring you the news you do not want? If no one, what would it take to create the role?
6. Has praise — press, awards, a board's admiration — ever changed a decision you made? Be specific.
7. Which of the course's frameworks described you most uncomfortably? That is where the calibration plan starts.

• Your model of "great CEOs" was trained on a biased sample. Media and boards both select for charisma, boldness and story; the machine that makes CEOs famous and the machine that makes them CEOs share a filter, and the filter is not effectiveness.
• Visibility ≠ prevalence: most CEOs, including most excellent ones, are not founders, celebrities, acquirers or prophets. Visibility ≠ effectiveness: awards precede underperformance on average, narcissists respond to praise over results, and observed extraversion is associated with higher perceived risk.
• The CEO effect is contested and, on any estimate, far smaller than the individual attributions the press makes. Hold your own story loosely.
• The Two-Sentence Test names the rarest capability: conviction and agency to say "We're going to do this," and humility and security to say "I was wrong. Change the plan." State your reversal criteria at the moment of commitment, and the second sentence becomes a plan rather than a confession.
• The course ends in a plan, not a diagnosis. Write it.