Research Library

Deal-level data on large Western European buyouts by mature PE houses show abnormal (leverage- and sector-adjusted) performance that is positive on average, linked to greater improvements in sales and operating margin than at quoted peers. Partner background matters: ex-consultants and ex-industry managers add more value in deals built on organic/internal improvement, while ex-bankers and ex-accountants do better in M&A-driven deals.

Proprietary sample of large PE deals (selection toward mature, large funds), with performance decomposition that relies on assumptions about leverage and sector benchmarks; partner effects are correlational.

Founder-CEO status is endogenous: good past performance makes it less likely that a founder keeps the CEO title (founders step down after both very bad and very good results). Instrumenting founder status with founder deaths and number of founders, the estimated causal effect of founder-CEOs on performance is positive and larger than the OLS estimate.

Instrumental-variables regressions on large U.S. firms (Fortune 500-type sample); the instruments (founder mortality, number of founders) require exclusion restrictions that may be debated, and results concern large established corporations rather than young firms.

Module 1, Module 2, Module 9, Module 13

Founding families are present in about one-third of S&P 500 firms and hold about 18% of equity. Family firms perform better than non-family firms on accounting and market measures, the relationship is non-linear (rising then falling with family ownership), and firms whose CEO is a family member (founder or descendant) outperform those with outside CEOs. The authors conclude family ownership is an effective organisational structure that does not appear to harm minority shareholders.

Panel regressions on 403 S&P 500 firms, 1992–1999; endogeneity of family control and survivorship (only families that stayed in the S&P 500) limit causal interpretation, and later work (16a, 16b) finds descendant CEOs specifically underperform.

Using time-use diaries for 1,114 CEOs of manufacturing firms in Brazil, France, Germany, India, the UK, and the US, the authors apply an unsupervised machine-learning method (Latent Dirichlet Allocation) to reduce hundreds of activity types into two "pure" behavioral types: "managers" (more one-on-one meetings with operational/production staff, plant visits) and "leaders" (more multi-function, multi-participant meetings with C-suite executives; more communication and coordination). Each CEO receives an index (0–1) reflecting the mix. A one-standard-deviation increase in the index toward "leader" behavior is associated with about 7% higher sales, controlling for labor, capital, and other factors. Performance differences emerge gradually and become statistically significant only about three years after the CEO's appointment, which the authors argue is more consistent with behavior mattering than with pure reverse causality. Using a simple assignment model, the authors estimate that 17% of firms end up with the "wrong" type of CEO, with mismatch more common in lower-income countries (36%) than in high-income ones (5%); they estimate that this misallocation accounts for about 13% of the labor-productivity gap between high- and low-income countries.

One week of time-use per CEO, collected via daily phone calls (activities > 15 minutes), reported by the CEO (43%) or a personal assistant. Manufacturing firms only; one week may not represent a full year; the index is a data-driven classification, not a psychological trait measure; the 7% figure is an association, and the authors state it is consistent with either "leaders" being uniformly better or with horizontal matching frictions—they do not claim a causal effect and explicitly caution against interpreting "leader" as simply "better."

Module 3, Module 6, Module 7, Module 8, Module 12, Module 14

For 158 S&P 1500 manufacturing CEOs, the usual negative relationship between CEO option wealth at risk and strategic risk taking (behavioral agency) reverses to positive for CEOs high in extraversion, high in openness, or low in conscientiousness — i.e., personality changes how incentives translate into risk taking.

Modest sample; personality inferred from observer/linguistic profiling rather than self-report; cross-sectional moderation analysis.

Module 2, Module 6

Among 4,013 executives at 2,402 U.S. firms (1980–2006), the share of CEOs with military service fell from 59% in 1980 to about 6% in 2006. CEOs with military experience pursue more conservative policies (capital investment about 8% lower and R&D about 11% lower relative to the mean, lower leverage), are roughly 70% less likely to be involved in corporate fraud, and their firms perform better during industry downturns.

OLS with industry/year fixed effects plus IV using birth-cohort variation in draft exposure (e.g., Korean War cohorts); military service remains a self-selected experience, and the IV compares narrow birth cohorts, so effects may not generalise to volunteer-era veterans.

Using Danish firm data and the gender of the departing CEO's first-born child as an instrument for family succession (a first-born son raises the probability of family succession), the authors find family successions cause large negative effects: operating return on assets falls by at least 4 percentage points around the transition relative to unrelated-CEO successions. The gap is largest in fast-growing industries, industries needing skilled labour, and larger firms.

Instrumental-variables design on a large Danish administrative sample (mostly small and mid-sized private firms); external validity to large listed firms in other countries is uncertain.

Using changes in German bank executive boards driven by mandatory retirements, boards with younger executives take on more portfolio risk; a higher share of female executives is associated with higher risk (a weaker, less robust effect); a higher share of executives with PhDs is associated with lower portfolio risk.

Difference-in-differences exploiting mandatory retirement in German banks (many of them savings and cooperative banks), with portfolio risk measured from regulatory data; findings on gender in particular have been contested and may not generalise beyond the German banking system.

Using a manager–firm matched panel that tracks top executives who move across firms, manager fixed effects explain a significant share of variation in investment, financial and organisational policies (e.g., acquisitions, leverage, dividends, cost-cutting) after controlling for firm fixed effects. Managers exhibit consistent "styles"; older-cohort managers are more conservative, MBA holders more aggressive; styles are related to performance and better-performing styles earn higher pay in better-governed firms.

Identification relies on executives observed at two or more firms (a selected group), and fixed effects capture any persistent manager-specific difference (including endogenous matching), so "style" is not purely personality; large US-firm sample; no direct trait measures.

Module 1, Module 2, Module 6, Module 7, Module 11, Module 12, Module 13

Using a double-blind interview survey of 732 medium-sized manufacturing firms in the U.S., UK, France and Germany, the authors show management-practice scores are strongly associated with productivity, profitability, Tobin's Q and survival. U.S. firms score highest on average; every country has a long tail of very badly managed firms. Weak product-market competition and family succession by primogeniture (eldest son as CEO) are the two main correlates of poor management.

Survey-based scoring of 18 practices (monitoring, targets, incentives) is subjective and cross-sectional; the family-succession result is correlational and specific to manufacturing.

Module 8, Module 12, Module 13

Using World Management Survey data on over 11,000 firms in 34 countries, the authors argue management practices behave like a technology: better-managed firms are more productive, and differences in management account for roughly 30% of TFP differences both across and within countries. A structural model rationalises why management quality rises with competition and firm age.

Survey-based management scores plus a structural model; cross-country comparisons rest on survey comparability and the model's assumptions.

Using the U.S. Census Bureau's Management and Organizational Practices Survey of about 35,000 manufacturing plants (2010 and 2015), the authors find large dispersion in structured management practices, with about 40% of the variation occurring across plants within the same firm. Management practices account for more than 20% of the variation in productivity, comparable to R&D, ICT or human capital. Business environment (proxied by right-to-work laws) and learning spillovers from the arrival of large plants are identified as drivers.

Large administrative survey with plant-level productivity data; the drivers are identified through natural-experiment-style variation but remain suggestive rather than definitive.

Module 8, Module 12

In a sample of 92 U.S. commercial-bank CEOs (2006–2014), more narcissistic CEOs adopted riskier bank policies before the 2008 shock, especially when their pay was option-heavy; strong board monitoring dampened this. Banks led by more narcissistic CEOs before September 2008 recovered more slowly to pre-shock performance afterwards; TARP capital injections moderated that slower recovery.

Narcissism measured with unobtrusive archival indicators (the Chatterjee & Hambrick style index: prominence of CEO photo, first-person pronoun use, pay gap, etc.); small sample, single industry and single crisis, observational design.

Module 2, Module 4, Module 11, Module 12

Using an unobtrusive narcissism index (prominence of the CEO's photo in the annual report, prominence in press releases, first-person-singular pronoun use in interviews, and cash/non-cash pay relative to the second-highest-paid executive) for 111 CEOs in computer hardware/software (1992–2004), narcissistic CEOs showed greater strategic dynamism and grandiosity, made more and larger acquisitions, and produced more extreme and fluctuating firm performance. On average, however, their firms performed neither better nor worse than those led by less narcissistic CEOs.

Archival, single-industry sample; narcissism is measured by indirect proxies (not a clinical or self-report instrument), which later reviewers criticise (e.g., pay ratios may reflect firm size or governance); causal direction (narcissists selecting into dynamic firms) cannot be ruled out.

Module 2, Module 4, Module 7, Module 11, Module 14

Introducing "capability cues" (contextual signals about one's current ability), the authors show that CEO risk taking generally rises after positive cues and falls after negative ones. Highly narcissistic CEOs are much less responsive to objective performance feedback (recent firm results) than other CEOs, but are considerably more responsive to social praise (media coverage, awards). Tested on risky capital outlays of US public-company CEOs (1992–2006) and on acquisition premiums paid (2001–2008).

Archival panel with the same unobtrusive narcissism index as 2007; two US samples; risk taking proxied by capital-spending and acquisition-premium measures; associations under fixed-effect/panel controls, not experimental evidence.

Module 1, Module 3, Module 4, Module 5, Module 6, Module 7, Module 11, Module 14

Combined narrative review and meta-analysis of 37 studies; the authors identify five measurement approaches and conclude that findings are "mixed and potentially dependent upon methods." Reported meta-analytic correlations (ρ) include: firm financial performance ρ ≈ .06 (k = 19, N ≈ 8,307; significant but small, with ROA and Tobin's Q individually non-significant); innovation/growth composite ρ ≈ .09 (k = 19); R&D spending ρ ≈ .11 (k = 7); new-product innovation ρ ≈ .14 (k = 6); risk taking ρ ≈ .03 (k = 7, not significant); financial leverage ρ ≈ .07 (not significant); M&A frequency/size ρ ≈ .01 (k = 5, not significant); CEO duality ρ ≈ .09 (significant); firm size ρ ≈ .08. The direction for innovation/R&D is positive and small; risk-taking and acquisition effects are not reliably different from zero in the pooled data.

Small k for most outcomes, heavy reliance on the Chatterjee–Hambrick unobtrusive index, and heterogeneity across measures; the authors explicitly warn that results depend on measurement method. A later meta-analysis (Kraft, 2022, JPIM, 39(6), 749–772, 68 studies) finds narcissism raises performance via innovation but that this is offset by other harmful behaviours (net negative).

Module 2, Module 4, Module 11

Building a General Ability Index (GAI) from S&P 1500 CEO résumés (1993–2007) that captures number of positions, firms, industries, conglomerate experience and prior CEO roles, the authors find generalist CEOs earn a pay premium of about 19% (roughly $1 million per year) over specialists. The premium is largest when firms hire externally and switch from a specialist to a generalist, and when the CEO is tasked with complex assignments such as restructurings or acquisitions.

Panel regressions of pay on the GAI with firm and CEO controls; pay premia are a market price of skills, not a direct measure of performance, and the index counts breadth of experience rather than quality.

Module 7, Module 10, Module 12, Module 13

Milliken et al. interviewed 40 full-time employees across industries: 85% reported at least one occasion on which they felt unable to raise an important issue with a superior; the most common withheld issues concerned a supervisor's or colleague's competence, organizational process problems, pay, and ethical concerns; the dominant reasons were fear of being labeled negatively (e.g., as a troublemaker) and of damaging relationships, followed by perceived futility and fear of retaliation; 74% of those who stayed silent said colleagues who knew of the same issue also stayed silent. Detert & Edmondson, across four studies (190 interviews in a high-tech firm; 185 executive-education participants; 265 online/MBA respondents; a three-wave study of 116 executive MBAs), identify five "implicit voice theories"—taken-for-granted beliefs that speaking up is risky: presumed target identification (the boss will take it personally), need solid data or solutions before speaking, don't bypass the boss upward, don't embarrass the boss in public, and negative career consequences of voice. These beliefs predict silence over and above personality and current context, and persist even where the environment is objectively safe. Tourish & Robson argue (conceptually) that both managers' sensemaking (overcommitment to chosen courses of action, dismissing dissenters as an out-group) and employees' self-censorship systematically filter critical information out of upward communication, producing "iatrogenic" problems caused by leaders' own decisions.

Milliken et al. is a small exploratory interview study (n = 40; not CEO-specific). Detert & Edmondson combine inductive interviews with survey scale development and multi-wave prediction—stronger, but mostly self-reported voice intentions/behavior in non-CEO samples. Tourish & Robson is a conceptual/theoretical paper with no new data. None directly measures CEO information environments; the inference to "CEO isolation" is an extrapolation.

Module 3, Module 5, Module 8, Module 14

Introduces "team psychological safety"—a shared belief that the team is safe for interpersonal risk taking—and shows it is associated with team learning behavior (seeking feedback, discussing errors, experimenting), which in turn predicts team performance; learning behavior mediates the safety–performance link. Team-leader coaching and context support are associated with psychological safety. Team efficacy did not predict learning behavior once psychological safety was controlled.

Multi-method study of 51 work teams (functional, self-managed, product-development, and project teams) in one U.S. office-furniture manufacturer; surveys of 427 team members and 135 external observers, plus interviews and observation. Single company, cross-sectional (author explicitly notes causality cannot be established), modest number of teams, and some low-reliability scales; some construct overlap with trust is acknowledged.

Module 3, Module 5

Among 95 U.S. bank holding companies and investment banks with 2006 compensation data, banks whose CEOs had incentives better aligned with shareholders (larger dollar ownership) performed worse in the crisis (roughly 10 percentage points lower stock return and ROE). Option-based pay and cash bonuses did not predict worse performance. CEOs did not reduce holdings before the crisis and lost on average about $31.5 million (median $5.1 million), consistent with their having believed the risks were value-creating ex ante.

Cross-sectional regressions of crisis-period (mid-2007 to end-2008) returns on pre-crisis incentive measures; small sample (95 banks), one crisis, correlational.

Module 4, Module 12

About 11% of the largest U.S. public firms in the sample are run by a founder-CEO. Founder-CEO firms invest more in R&D and capital expenditure and make more focused (fewer diversifying) acquisitions than successor-CEO firms. An equal-weighted portfolio of founder-CEO firms over 1993–2002 earned a benchmark-adjusted return of roughly 8.3% per year; after controlling for firm and CEO characteristics the abnormal return is about 4.4% per year.

Panel of large U.S. public firms, 1993–2002, comparing founder-CEO and non-founder-CEO firms via calendar-time portfolio returns and regressions. Founder status is not randomly assigned; founders who stay are the survivors of earlier selection, and the return anomaly is specific to one decade and one market, so the result is an association, not a causal estimate.

Quigley & Hambrick partition variance in performance for 1,015 U.S. firms in 30 industries (18,467 firm-years, 2,732 CEOs, 1950–2009) and find the share of performance variance attributable to individual CEOs rose substantially across three periods—e.g., by sequential ANOVA, the CEO effect on ROA rose from 7.8% (1950–69) to 10.8% (1970–89) to 15.7% (1990–2009); by multilevel modeling, from 13.8% to 17.8% to 22.9% (increases significant at p < .001). Fitza (2014) argues that variance-decomposition "CEO effects" are inflated by chance: simulating data in which performance is random, he shows that a sizable apparent CEO effect emerges from randomness alone given short CEO tenures (averaging about four years, too short for luck to average out); a press summary of his 2017 follow-up states that over 70% of the measured CEO effect could be attributable to chance. Quigley & Graffin (2017) reply that Fitza's simulation assumptions overstate the random component and, using multilevel modeling, reaffirm a CEO effect "significant and much larger than chance"; Fitza (2017) rejoins that under more realistic assumptions the CEO effect remains close to chance regardless of estimator.

All are variance-decomposition studies of archival accounting/market performance of large U.S. firms. The "CEO effect" is a statistical residual associated with CEO identity, not a measure of any specific CEO action, and its magnitude is sensitive to model choice (sequential ANOVA vs. multilevel), industry/year controls, tenure lengths, and assumptions about random shocks. The debate is unresolved.

Module 1, Module 11, Module 14

Quigley & Hambrick partition variance in performance for 1,015 U.S. firms in 30 industries (18,467 firm-years, 2,732 CEOs, 1950–2009) and find the share of performance variance attributable to individual CEOs rose substantially across three periods—e.g., by sequential ANOVA, the CEO effect on ROA rose from 7.8% (1950–69) to 10.8% (1970–89) to 15.7% (1990–2009); by multilevel modeling, from 13.8% to 17.8% to 22.9% (increases significant at p < .001). Fitza (2014) argues that variance-decomposition "CEO effects" are inflated by chance: simulating data in which performance is random, he shows that a sizable apparent CEO effect emerges from randomness alone given short CEO tenures (averaging about four years, too short for luck to average out); a press summary of his 2017 follow-up states that over 70% of the measured CEO effect could be attributable to chance. Quigley & Graffin (2017) reply that Fitza's simulation assumptions overstate the random component and, using multilevel modeling, reaffirm a CEO effect "significant and much larger than chance"; Fitza (2017) rejoins that under more realistic assumptions the CEO effect remains close to chance regardless of estimator.

All are variance-decomposition studies of archival accounting/market performance of large U.S. firms. The "CEO effect" is a statistical residual associated with CEO identity, not a measure of any specific CEO action, and its magnitude is sensitive to model choice (sequential ANOVA vs. multilevel), industry/year controls, tenure lengths, and assumptions about random shocks. The debate is unresolved.

Module 11, Module 14

Georgakakis & Ruigrok, using 109 CEO succession events at large international firms, find that outsider succession is associated with better post-succession performance under specific conditions: when the outsider is demographically similar to the incumbent top team (easing integration), when the outsider has a diverse career background, and when the firm is already performing well in a favorable industry environment—concluding that origin effects should be assessed in interaction with multilevel context, not in isolation. Shen & Cannella, drawing on power-circulation theory and 228 U.S. successions, distinguish "followers" (insiders succeeding a CEO who left voluntarily), "contenders" (insiders succeeding a dismissed CEO), and outsiders; successor type interacts with post-succession senior-executive turnover to affect ROA, and departing-CEO tenure has an inverted-U relationship with post-succession ROA.

Both are archival, observational studies with modest samples (109 and 228 successions) and accounting-based performance outcomes; categorizations of successor type and "outsider" are proxies; causal inference is limited by selection (boards choose outsiders in particular situations).

A survey of 79 PE investors managing over $750 billion finds they evaluate deals mainly on IRR and multiples rather than DCF; capital structure choices blend trade-off and market-timing logic; they expect to create value more through growth than cost reduction; and they place heavy weight on management, frequently recruiting their own senior management teams (the paper reports that PE investors commonly replace or add top managers post-deal). Practices cluster into distinct strategies related to the founders' backgrounds.

Self-reported survey responses from a non-random set of PE firms; describes stated practice rather than measured outcomes.

The authors train a statistical-learning model linking linguistic features of CEOs' unscripted earnings-call Q&A to Big Five scores from ghSMART assessments, then apply it to ~70,000 calls by ~4,600 CEOs. Openness is positively related to R&D intensity and negatively to leverage; conscientiousness is negatively related to growth; extraversion is negatively related to contemporaneous and future ROA and cash flow. Out-of-sample correlations between predicted and assessed traits range from about .23 (agreeableness) to .49 (neuroticism).

Language-based inference of personality validated on a small set of assessed CEOs (a 28-CEO validation sample is cited); moderate validity correlations mean substantial measurement error; not peer reviewed; associations only.

Module 2, Module 3, Module 14

Using psychometric instruments administered in a large survey of CEOs and CFOs of public and private firms, the authors find that CEOs are markedly more risk-tolerant and more optimistic than population norms (US executives more optimistic than non-US ones; ~80% of US CEOs classified "very optimistic" vs ~65% of CFOs). Firms with risk-tolerant CEOs initiate more M&A and are more often growth firms; optimistic CEOs use more short-term debt; risk-averse CEOs receive pay with a higher fixed-salary share and lower pay-performance sensitivity (consistent with agency theory).

Cross-sectional survey with validated psychometric scales (e.g., Scheier–Carver–Bridges optimism; lottery-style risk-aversion items) matched to firm data. Self-selection into responding, self-report measures, and cross-sectional design mean the trait–policy links are associations, not causal effects; comparisons to "lay population" rely on external benchmark samples.

Module 1, Module 2, Module 4, Module 7, Module 12

Using psychometric instruments administered in a large survey of CEOs and CFOs of public and private firms, the authors find that CEOs are markedly more risk-tolerant and more optimistic than population norms (US executives more optimistic than non-US ones; ~80% of US CEOs classified "very optimistic" vs ~65% of CFOs). Firms with risk-tolerant CEOs initiate more M&A and are more often growth firms; optimistic CEOs use more short-term debt; risk-averse CEOs receive pay with a higher fixed-salary share and lower pay-performance sensitivity (consistent with agency theory).

Cross-sectional survey with validated psychometric scales (e.g., Scheier–Carver–Bridges optimism; lottery-style risk-aversion items) matched to firm data. Self-selection into responding, self-report measures, and cross-sectional design mean the trait–policy links are associations, not causal effects; comparisons to "lay population" rely on external benchmark samples.

Module 8, Module 9

Hambrick & Finkelstein introduce "managerial discretion"—the latitude of action available to executives—as a bridge between views that executives determine outcomes and views that they are largely constrained; discretion is argued to derive from the task environment, internal organization, and the executive's own characteristics, and to determine how much executives matter. Wangrow et al. review the empirical literature and conclude that the theory remains incompletely validated: most studies operationalize discretion via industry/task-environment factors, with far less attention to organizational and individual (managerial) sources; measurement is inconsistent; and constraints on discretion may act as complements or substitutes—an agenda for future work.

1987 is conceptual (no data). 2015 is a narrative empirical review; conclusions depend on the studies available and highlight that discretion is frequently proxied rather than measured directly.

Module 11, Module 12

Hambrick & Fukutomi propose that CEO tenures pass through five "seasons"—response to mandate, experimentation, selection of an enduring theme, convergence, and dysfunction—driven by shifts in commitment to a paradigm, task knowledge, information diversity, task interest, and power; they argue the model implies performance tends to rise and then decline over a long tenure. Miller finds empirically that CEO tenure is inversely related to the prescribed match between organization (strategy/structure) and environment—long-tenured CEOs were less likely to have organizations aligned with environmental demands, especially in uncertain environments and under concentrated ownership—and that better match was associated with better financial performance.

Hambrick & Fukutomi is a conceptual model with no data. Miller is a cross-sectional study of a modest sample of firms using questionnaire and archival measures of environment, strategy, and structure (sample size not confirmed from an accessible source); "match" is inferred from contingency-theory prescriptions; correlational.

Module 6, Module 10, Module 13

Theoretical paper proposing that organisational outcomes (strategic choices and performance) partly reflect the values and cognitive bases of top executives; because these are hard to observe, observable demographics (age, tenure, functional background, education) can serve as proxies; the top management team, not only the CEO, is the appropriate unit.

Conceptual; no data. Demographic proxies were later criticised as crude ("black box"); the theory's predictions are conditional on managers actually having discretion.

Module 1, Module 11

Restates the premise that executives' experiences, values and personalities shape how they interpret situations and therefore their choices, and adds moderators: managerial discretion (effects appear only where executives have latitude), executive job demands (heavy demands increase reliance on heuristics and personal dispositions), and TMT behavioural integration (team-level effects strongest when the team actually works jointly). Calls for psychological, not only demographic, measures of executives.

Conceptual review; empirical support for moderators is drawn from subsequent literature rather than tested in the paper.

Module 1, Module 2, Module 4, Module 6, Module 7, Module 8, Module 11, Module 12

Develops and validates (against videometric/observer ratings on a CEO sample) a machine-learning linguistic tool that scores CEOs' Big Five traits from earnings-call speech, then applies it to 3,000+ S&P 1500 CEOs. CEO traits predict strategic change, and the effects depend on recent firm performance — person and situation interact.

Observer-rated (not self-report) personality used as ground truth; spoken language in scripted-adjacent settings; large archival sample but correlational; trait-specific effect sizes on strategic change are modest.

Module 2, Module 6

Develops and validates (against videometric/observer ratings on a CEO sample) a machine-learning linguistic tool that scores CEOs' Big Five traits from earnings-call speech, then applies it to 3,000+ S&P 1500 CEOs. CEO traits predict strategic change, and the effects depend on recent firm performance — person and situation interact.

Observer-rated (not self-report) personality used as ground truth; spoken language in scripted-adjacent settings; large archival sample but correlational; trait-specific effect sizes on strategic change are modest.

Module 2, Module 14

The paper develops the construct of CEO celebrity: journalists attribute a firm's distinctive and consistent strategic actions to the CEO's disposition, creating celebrity. CEOs who internalise this attribution become overconfident about their efficacy (hubris) and tend to persist with the actions that generated the celebrity, even when those actions stop paying off.

Theory-building paper with propositions; it does not itself present empirical tests, so it should be cited for the framework rather than for effect sizes.

Module 4, Module 11, Module 14

In 120 Ecuadorian SMEs, extraversion and openness relate to initiating strategic change only; emotional stability and agreeableness relate to both initiation and the performance of implementation; conscientiousness has opposing effects — it hinders initiation but helps effective implementation.

Small-firm, single-country survey sample; cross-sectional; Big Five measured by survey instruments.

Module 2, Module 3, Module 6

Firms led by overconfident CEOs (options- and press-based proxies, 1993–2003) invest more in R&D, obtain more patents and patent citations, and achieve more innovative output per R&D dollar; they also have higher stock-return volatility. There is no evidence of lower sales, ROA or Tobin's Q. The innovation advantage appears only in innovative industries.

Large-panel archival study relying on Malmendier–Tate proxies; innovation measured by patents/citations (industry-specific relevance); results are conditional correlations and the authors note volatility/riskiness rises alongside innovation.

Module 2, Module 4, Module 9, Module 10

In a panel of listed U.S. banks 1994–2009, banks run by overconfident CEOs loosened lending standards and raised leverage more than peers in the run-up to the 1998 Russian crisis and the 2007–09 crisis (e.g., faster loan growth, especially real-estate lending; market leverage roughly 5 percentage points higher). During the crises these banks suffered larger loan-default increases, bigger drops in operating and stock performance, higher CEO turnover and higher failure rates (about 10% vs 4% failing).

Overconfidence measured with the Campbell et al. (2011) option-holding proxy (CEOs who hold deep-in-the-money options). This is an indirect behavioural proxy; the design is observational, and overconfident CEOs may sort into banks with riskier strategies.

Module 2, Module 4, Module 12

Using historiometric Q-sort ratings of archival material (interviews, biographies, writings) for 17 CEOs of 9 large US firms (e.g., IBM, Coca-Cola, Disney, GM, Xerox, Kodak), CEO traits correlated with TMT dynamics: conscientiousness with legalism, control and centralisation; neuroticism with leader dominance/weakness, factionalism and rigidity; agreeableness with cohesion and decentralised power; extraversion with leader dominance; openness with risk taking and intellectual flexibility. TMT intellectual flexibility, optimism, cohesion and risk taking in turn correlated with income growth.

Very small sample (17 CEOs) with many tests; Hollenbeck et al. (2006) showed the findings have low statistical power and unstable parameter estimates (results change materially when a single case is dropped). Ratings are third-party historiometric judgments, not assessments of the CEOs. Treat as illustrative, not as reliable effect sizes.

Meta-analysis of 222 correlations from 73 samples. Corrected correlations with leadership (emergence and effectiveness combined): extraversion .31, conscientiousness .28, neuroticism −.24, openness .24, agreeableness .08; multiple correlation of the Big Five with leadership R = .48. Extraversion is the most consistent correlate; conscientiousness relates more to emergence (.33) than effectiveness (.16); agreeableness relates to effectiveness (.21) but not emergence (.05).

Samples are mostly not CEOs (students, military, mid-level managers); "leadership" criteria include peer-rated emergence and supervisor-rated effectiveness rather than firm outcomes; correlations are corrected for unreliability; moderate heterogeneity.

Module 1, Module 2, Module 3, Module 14

From 2,603 ghSMART assessments of candidates for CEO, CFO, COO and other top roles (2000–2013), four factors explain over half the variance in 30 rated characteristics: general ability; execution vs. interpersonal; charisma vs. analytical; strategic vs. managerial. CEO candidates score higher on all four (CFO candidates lower). Candidates with stronger interpersonal skills are more likely to be hired, yet the factors — including execution — predict later career advancement to CEO, suggesting boards may overweight interpersonal skills. Patterns hold across public, PE-backed and VC-backed firms; gender differences on the factors are small, yet women are less likely to become CEO.

Single-assessor (ghSMART) interview-based ratings; the sample is candidates who were assessed, not all executives; outcome is subsequent hiring/career progression rather than firm performance; observational design.

Module 1, Module 2, Module 3, Module 7, Module 14

Using detailed ghSMART assessments of CEO candidates for private-equity (buyout) and venture-capital-backed companies rated on 30+ abilities, factor analysis yields two main dimensions: general ability, and an execution/"hard" versus interpersonal/"soft" (team-oriented) contrast. Both buyout and VC firms tend to hire CEOs high on general and execution ability; success (subjective and objective) is more strongly related to execution skills than to interpersonal skills; incumbency is only marginally related to performance once measured ability is controlled.

Assessment data from one firm (ghSMART) on 316 CEO candidates at 224 companies involved in PE/VC transactions, 2000–2006 (confirmed from the authors' conference version); performance outcomes are partly subjective investor ratings; sample is selected (candidates PE/VC firms chose to assess) and not representative of public-company CEOs; associations rather than causal estimates.

Module 1, Module 2, Module 3, Module 7, Module 9, Module 14

Reconceptualizing insider/outsider status as a continuum ("outsiderness," reflecting distance from the firm and industry), the study finds no direct main effect of CEO outsiderness on post-succession performance. Outsiderness helps when pre-succession performance is poor and when the environment is turbulent, and its effect depends on post-succession context, including concurrent strategic change and top-team turnover.

Longitudinal archival data from U.S. airline and chemical industries, 1972–2002. Two industries only; "outsiderness" and strategic change are archival proxies; observational.

Module 6, Module 7, Module 10, Module 13

Across S&P 1500 firms, founder CEOs use more optimistic language on Twitter and in earnings conference calls, are more likely to issue earnings forecasts that turn out too high, and are more likely to hold deep-in-the-money options (interpreted as believing their own firm is undervalued) than professional CEOs. The authors argue that investors do not appear to fully price in this founder overconfidence.

Cross-sectional/panel comparison of founder vs. professional CEOs in S&P 1500 firms using four indirect overconfidence proxies (language tone, forecast error, option-exercise behaviour, market reaction). Proxies capture expressed optimism rather than a validated psychological trait; founder status is endogenous (founders who remain CEO of large listed firms are a selected group).

Module 1, Module 2, Module 4, Module 9, Module 14

Using U.S. public firms whose CEO died suddenly (1979–2002), the authors report that an exogenous switch from a founder CEO to a professional CEO is associated with about a 43.8% decline in the firm's citation-weighted patent count. Founder-led firms pursue more exploratory innovation and produce more patents at both quality extremes (breakthroughs and duds). Inventive employees leave at higher rates after the transition. R&D spending does not differ significantly, suggesting the difference lies in how innovation is managed, not how much is spent.

Difference-in-differences around sudden CEO deaths (a plausibly exogenous shock), with patents as the innovation measure. Sudden-death samples are small and drawn from an older period, patents are an imperfect proxy for innovation, and effects may partly reflect disruption from any unplanned succession rather than founder-specific traits.

Module 4, Module 9

CEO hubris is positively related to firm risk taking, and this relationship is stronger when the CEO has more managerial discretion. Discretion-enhancing conditions that amplified the hubris–risk link included munificent but complex markets, low organizational inertia, greater intangible resources, CEO–board chair duality, and CEOs who were not politically appointed.

Survey data on 2,790 CEOs of Chinese manufacturing firms. Large sample, but single country in a transitional economy; hubris and risk taking are survey-based proxies; cross-sectional and correlational, with possible common-source concerns.

Module 4, Module 11

Using linguistic analysis of unscripted remarks by 2,381 S&P 1500 CEOs (10,166 CEO-firm-years, 2002–2012; 1,710 deals), more extraverted CEOs make more, and larger, acquisitions; board-network participation partly mediates the effect; the effect is stronger in less competitive industries and where managerial entrenchment is higher; extraverted CEOs' deals also earn stronger abnormal announcement returns.

Language-based extraversion measure; archival panel; correlational with fixed effects; performance evidence is short-window market reaction.

Module 2, Module 11, Module 14

CEOs classified as overconfident — those who persistently fail to reduce personal exposure to company-specific risk (e.g., holding deep-in-the-money options past the point a rational diversifier would exercise; "Longholder") — show investment that is significantly more sensitive to internal cash flow, especially in equity-dependent firms. This is consistent with overconfident CEOs overestimating project returns and viewing external finance as overly costly, producing over-investment when cash is available and under-investment when it is not.

Panel of Forbes 500 CEOs' personal portfolio decisions (1980–1994) linked to corporate investment; "overconfidence" is a revealed-preference proxy that could also reflect inside information, risk tolerance or signalling (the authors test and argue against these); large, established firms only.

Module 2, Module 4

Using both option-holding and press-portrayal proxies, overconfident CEOs are about 65% more likely to make an acquisition, with the effect strongest for diversifying deals and when internal financing is available. The market reacts more negatively to their deals (announcement return about −90 basis points vs −12 bp for other CEOs).

Same Forbes 500 panel and proxy approach; press-based measure relies on journalists' word choice; alternative explanations (private information, empire building, risk preference) are addressed but not eliminated; sample predates 1995.

Module 2, Module 4, Module 7, Module 10, Module 14

Using prestigious business-press CEO awards as a shock to status, the authors find award-winning CEOs subsequently underperform relative to their own prior record and to matched non-winning CEOs. After winning, they receive higher pay, spend more time on outside activities such as board seats and book-writing, and their firms show more earnings management. The effects are strongest in firms with weak governance.

Matched-sample comparison of award winners vs. predicted winners who did not win, U.S. large firms, 1975–2002; mean reversion is partly addressed by the matching design but cannot be fully ruled out.

Module 1, Module 4, Module 11, Module 14

Overconfident CEOs (those who hold deep-in-the-money options late into expiry, the "Longholder" measure) view external finance, especially equity, as overpriced: they are 37–49% less likely to issue equity when raising external capital and lean more on debt, leaving their firms with higher leverage. CEOs who grew up during the Great Depression are debt-averse and rely more on internal finance; CEOs who served in WWII, particularly in combat, choose more aggressive leverage.

Panel of large U.S. firms (Forbes 500 lists, 1980–1994) plus a later ExecuComp sample; overconfidence is inferred from option-exercise behaviour rather than measured directly, and early-life effects rely on birth-cohort variation.

Module 4, Module 12

Hambrick & Fukutomi propose that CEO tenures pass through five "seasons"—response to mandate, experimentation, selection of an enduring theme, convergence, and dysfunction—driven by shifts in commitment to a paradigm, task knowledge, information diversity, task interest, and power; they argue the model implies performance tends to rise and then decline over a long tenure. Miller finds empirically that CEO tenure is inversely related to the prescribed match between organization (strategy/structure) and environment—long-tenured CEOs were less likely to have organizations aligned with environmental demands, especially in uncertain environments and under concentrated ownership—and that better match was associated with better financial performance.

Hambrick & Fukutomi is a conceptual model with no data. Miller is a cross-sectional study of a modest sample of firms using questionnaire and archival measures of environment, strategy, and structure (sample size not confirmed from an accessible source); "match" is inferred from contingency-theory prescriptions; correlational.

Module 6, Module 10, Module 11, Module 13

Milliken et al. interviewed 40 full-time employees across industries: 85% reported at least one occasion on which they felt unable to raise an important issue with a superior; the most common withheld issues concerned a supervisor's or colleague's competence, organizational process problems, pay, and ethical concerns; the dominant reasons were fear of being labeled negatively (e.g., as a troublemaker) and of damaging relationships, followed by perceived futility and fear of retaliation; 74% of those who stayed silent said colleagues who knew of the same issue also stayed silent. Detert & Edmondson, across four studies (190 interviews in a high-tech firm; 185 executive-education participants; 265 online/MBA respondents; a three-wave study of 116 executive MBAs), identify five "implicit voice theories"—taken-for-granted beliefs that speaking up is risky: presumed target identification (the boss will take it personally), need solid data or solutions before speaking, don't bypass the boss upward, don't embarrass the boss in public, and negative career consequences of voice. These beliefs predict silence over and above personality and current context, and persist even where the environment is objectively safe. Tourish & Robson argue (conceptually) that both managers' sensemaking (overcommitment to chosen courses of action, dismissing dissenters as an out-group) and employees' self-censorship systematically filter critical information out of upward communication, producing "iatrogenic" problems caused by leaders' own decisions.

Milliken et al. is a small exploratory interview study (n = 40; not CEO-specific). Detert & Edmondson combine inductive interviews with survey scale development and multi-wave prediction—stronger, but mostly self-reported voice intentions/behavior in non-CEO samples. Tourish & Robson is a conceptual/theoretical paper with no new data. None directly measures CEO information environments; the inference to "CEO isolation" is an extrapolation.

Module 3, Module 5, Module 8, Module 14

In 195 small and medium-sized Indian BPO firms, CEO Big Five traits predict strategic flexibility (the ability to adapt quickly to environmental change), and strategic flexibility mediates the relationship between CEO personality and firm performance. Each trait either enhances or inhibits flexibility.

Survey-based measures of personality and flexibility in one industry and country, small/medium firms with high CEO discretion; cross-sectional design; common-method concerns partially addressed by multiple respondents; generalisation to large public firms is uncertain.

CEO humility (defined as self-awareness, openness to feedback, appreciation of others, low self-focus, and self-transcendent pursuit) is positively associated with the CEO's empowering leadership behaviors, which relate to greater TMT integration. TMT integration in turn relates to an empowering organizational climate perceived by middle managers, which is associated with middle managers' work engagement, affective commitment, and job performance. The paper also conceptually elaborates the humility construct via interviews.

Data from 328 TMT members and 645 middle managers in 63 private Chinese companies, collected at two time points, supplemented by interviews with 51 CEOs; multilevel modeling and SEM. Chinese private-firm context (possible cultural specificity of humility norms); humility rated by others; correlational, so causal direction is inferred rather than demonstrated.

Module 2, Module 3, Module 5, Module 7

Drawing on upper-echelons, power, and paradox theories, the authors propose and largely support a mediation model in which CEO humility is associated with greater top-management-team (TMT) integration (collaboration, information sharing, joint decision-making, shared vision) and lower CEO–TMT pay disparity; these in turn are associated with a more ambidextrous strategic orientation (pursuing both exploration and exploitation), which is associated with better firm performance.

Multi-wave survey plus archival data from 105 small-to-medium U.S. computer software/hardware firms. Humility is rated by TMT members (perceptual measure); the sample is small, single-industry, and SMEs, so generalization to large public companies is not established; design is correlational with time separation, not causal.

Module 3, Module 5, Module 7, Module 14

2012: From 55 in-depth interviews with leaders across settings (high tech, banking, hospitals, financial services, religious organizations, manufacturing, military), the authors induce three core humble-leader behaviors: (a) acknowledging personal limits, faults, and mistakes; (b) spotlighting followers' strengths and contributions; (c) modeling teachability (openness, listening, seeking feedback). Humility is said to work by "modeling how to grow," legitimizing followers' development and uncertainty; contingencies include perceived leader competence and sincerity, and extreme threat/time pressure (humility is less effective in crisis). 2016: Across three studies (607 participants in 161 teams; 84 lab and 77 field teams), leader humility spreads via social contagion to produce "collective humility," which fosters a collective promotion focus and, through it, higher team performance.

2012 is qualitative/inductive (no effect sizes; theory-building only). 2016 combines an experiment manipulating leader humility, a longitudinal team simulation, and a multi-stage field study in health services—stronger causal grounding than most humility work, but samples are teams and lower-level leaders, not CEOs.

Module 3, Module 5, Module 6, Module 7, Module 8, Module 10, Module 14

2012: From 55 in-depth interviews with leaders across settings (high tech, banking, hospitals, financial services, religious organizations, manufacturing, military), the authors induce three core humble-leader behaviors: (a) acknowledging personal limits, faults, and mistakes; (b) spotlighting followers' strengths and contributions; (c) modeling teachability (openness, listening, seeking feedback). Humility is said to work by "modeling how to grow," legitimizing followers' development and uncertainty; contingencies include perceived leader competence and sincerity, and extreme threat/time pressure (humility is less effective in crisis). 2016: Across three studies (607 participants in 161 teams; 84 lab and 77 field teams), leader humility spreads via social contagion to produce "collective humility," which fosters a collective promotion focus and, through it, higher team performance.

2012 is qualitative/inductive (no effect sizes; theory-building only). 2016 combines an experiment manipulating leader humility, a longitudinal team simulation, and a multi-stage field study in health services—stronger causal grounding than most humility work, but samples are teams and lower-level leaders, not CEOs.

Module 3, Module 5

Using historiometric Q-sort ratings of archival material (interviews, biographies, writings) for 17 CEOs of 9 large US firms (e.g., IBM, Coca-Cola, Disney, GM, Xerox, Kodak), CEO traits correlated with TMT dynamics: conscientiousness with legalism, control and centralisation; neuroticism with leader dominance/weakness, factionalism and rigidity; agreeableness with cohesion and decentralised power; extraversion with leader dominance; openness with risk taking and intellectual flexibility. TMT intellectual flexibility, optimism, cohesion and risk taking in turn correlated with income growth.

Very small sample (17 CEOs) with many tests; Hollenbeck et al. (2006) showed the findings have low statistical power and unstable parameter estimates (results change materially when a single case is dropped). Ratings are third-party historiometric judgments, not assessments of the CEOs. Treat as illustrative, not as reliable effect sizes.

Tracking 27 CEOs of large (mostly public, multi-billion-dollar) companies 24/7 for 13 weeks each (about 60,000 hours), the authors report that CEOs worked on average 9.7 hours per weekday and did business on 79% of weekend days and 70% of vacation days; 72% of work time was spent in meetings; 61% of interactions were face-to-face and 24% electronic; 46% of time involved at least one direct report; about 30% of time went to external constituencies (customers only ~3%, investors ~3%). They frame the CEO job as managing dualities (direct vs. indirect influence, internal vs. external, formal authority vs. legitimacy) and recommend deliberate agenda-setting, since only ~43% of time was spent on agenda-advancing activities.

Descriptive time-use study of a small, self-selected sample of 27 CEOs; assistants logged time in 15-minute increments. Published in HBR without peer review; no performance outcomes are linked to the time-use patterns, so it describes what CEOs do, not what works.

In U.S. public firms where the departing CEO was a member of the controlling family, firms that appoint a family-related successor experience large declines in return on assets and market-to-book relative to firms that promote unrelated CEOs. The underperformance is concentrated among family heirs who did not attend a selective undergraduate institution, consistent with nepotism limiting the talent pool.

Event-study comparison of roughly 300+ CEO successions in U.S. family-controlled public firms (1980s–1990s/early 2000s); succession choice is not random, and the selective-college proxy for ability is coarse.

Quigley & Hambrick partition variance in performance for 1,015 U.S. firms in 30 industries (18,467 firm-years, 2,732 CEOs, 1950–2009) and find the share of performance variance attributable to individual CEOs rose substantially across three periods—e.g., by sequential ANOVA, the CEO effect on ROA rose from 7.8% (1950–69) to 10.8% (1970–89) to 15.7% (1990–2009); by multilevel modeling, from 13.8% to 17.8% to 22.9% (increases significant at p < .001). Fitza (2014) argues that variance-decomposition "CEO effects" are inflated by chance: simulating data in which performance is random, he shows that a sizable apparent CEO effect emerges from randomness alone given short CEO tenures (averaging about four years, too short for luck to average out); a press summary of his 2017 follow-up states that over 70% of the measured CEO effect could be attributable to chance. Quigley & Graffin (2017) reply that Fitza's simulation assumptions overstate the random component and, using multilevel modeling, reaffirm a CEO effect "significant and much larger than chance"; Fitza (2017) rejoins that under more realistic assumptions the CEO effect remains close to chance regardless of estimator.

All are variance-decomposition studies of archival accounting/market performance of large U.S. firms. The "CEO effect" is a statistical residual associated with CEO identity, not a measure of any specific CEO action, and its magnitude is sensitive to model choice (sequential ANOVA vs. multilevel), industry/year controls, tenure lengths, and assumptions about random shocks. The debate is unresolved.

Module 11, Module 14

Quigley & Hambrick partition variance in performance for 1,015 U.S. firms in 30 industries (18,467 firm-years, 2,732 CEOs, 1950–2009) and find the share of performance variance attributable to individual CEOs rose substantially across three periods—e.g., by sequential ANOVA, the CEO effect on ROA rose from 7.8% (1950–69) to 10.8% (1970–89) to 15.7% (1990–2009); by multilevel modeling, from 13.8% to 17.8% to 22.9% (increases significant at p < .001). Fitza (2014) argues that variance-decomposition "CEO effects" are inflated by chance: simulating data in which performance is random, he shows that a sizable apparent CEO effect emerges from randomness alone given short CEO tenures (averaging about four years, too short for luck to average out); a press summary of his 2017 follow-up states that over 70% of the measured CEO effect could be attributable to chance. Quigley & Graffin (2017) reply that Fitza's simulation assumptions overstate the random component and, using multilevel modeling, reaffirm a CEO effect "significant and much larger than chance"; Fitza (2017) rejoins that under more realistic assumptions the CEO effect remains close to chance regardless of estimator.

All are variance-decomposition studies of archival accounting/market performance of large U.S. firms. The "CEO effect" is a statistical residual associated with CEO identity, not a measure of any specific CEO action, and its magnitude is sensitive to model choice (sequential ANOVA vs. multilevel), industry/year controls, tenure lengths, and assumptions about random shocks. The debate is unresolved.

Fitza (2014), Fitza (2017)

Module 1, Module 11, Module 14

Georgakakis & Ruigrok, using 109 CEO succession events at large international firms, find that outsider succession is associated with better post-succession performance under specific conditions: when the outsider is demographically similar to the incumbent top team (easing integration), when the outsider has a diverse career background, and when the firm is already performing well in a favorable industry environment—concluding that origin effects should be assessed in interaction with multilevel context, not in isolation. Shen & Cannella, drawing on power-circulation theory and 228 U.S. successions, distinguish "followers" (insiders succeeding a CEO who left voluntarily), "contenders" (insiders succeeding a dismissed CEO), and outsiders; successor type interacts with post-succession senior-executive turnover to affect ROA, and departing-CEO tenure has an inverted-U relationship with post-succession ROA.

Both are archival, observational studies with modest samples (109 and 228 successions) and accounting-based performance outcomes; categorizations of successor type and "outsider" are proxies; causal inference is limited by selection (boards choose outsiders in particular situations).

Module 10, Module 13

Milliken et al. interviewed 40 full-time employees across industries: 85% reported at least one occasion on which they felt unable to raise an important issue with a superior; the most common withheld issues concerned a supervisor's or colleague's competence, organizational process problems, pay, and ethical concerns; the dominant reasons were fear of being labeled negatively (e.g., as a troublemaker) and of damaging relationships, followed by perceived futility and fear of retaliation; 74% of those who stayed silent said colleagues who knew of the same issue also stayed silent. Detert & Edmondson, across four studies (190 interviews in a high-tech firm; 185 executive-education participants; 265 online/MBA respondents; a three-wave study of 116 executive MBAs), identify five "implicit voice theories"—taken-for-granted beliefs that speaking up is risky: presumed target identification (the boss will take it personally), need solid data or solutions before speaking, don't bypass the boss upward, don't embarrass the boss in public, and negative career consequences of voice. These beliefs predict silence over and above personality and current context, and persist even where the environment is objectively safe. Tourish & Robson argue (conceptually) that both managers' sensemaking (overcommitment to chosen courses of action, dismissing dissenters as an out-group) and employees' self-censorship systematically filter critical information out of upward communication, producing "iatrogenic" problems caused by leaders' own decisions.

Milliken et al. is a small exploratory interview study (n = 40; not CEO-specific). Detert & Edmondson combine inductive interviews with survey scale development and multi-wave prediction—stronger, but mostly self-reported voice intentions/behavior in non-CEO samples. Tourish & Robson is a conceptual/theoretical paper with no new data. None directly measures CEO information environments; the inference to "CEO isolation" is an extrapolation.

Hambrick & Finkelstein introduce "managerial discretion"—the latitude of action available to executives—as a bridge between views that executives determine outcomes and views that they are largely constrained; discretion is argued to derive from the task environment, internal organization, and the executive's own characteristics, and to determine how much executives matter. Wangrow et al. review the empirical literature and conclude that the theory remains incompletely validated: most studies operationalize discretion via industry/task-environment factors, with far less attention to organizational and individual (managerial) sources; measurement is inconsistent; and constraints on discretion may act as complements or substitutes—an agenda for future work.

1987 is conceptual (no data). 2015 is a narrative empirical review; conclusions depend on the studies available and highlight that discretion is frequently proxied rather than measured directly.

Founder-CEO succession differs from later CEO successions because of founders' emotional attachment and large equity stakes. Two milestones sharply raise the hazard of founder replacement: completing initial product development and raising each round of outside financing. Hence the "paradox": the founder's success at hitting milestones increases, rather than decreases, the probability that they are replaced.

Event-history (hazard) analysis of 202 Internet start-ups, late 1990s/early 2000s. Single sector, dot-com era, U.S. venture-backed firms; findings on succession timing may not transfer to bootstrapped or non-tech firms.

Module 1, Module 9, Module 10, Module 13

Founders face a trade-off between being "rich" (maximising firm value by giving up equity and control to co-founders, hires and investors) and being "king" (keeping control). Drawing on his data on 212 U.S. start-ups, Wasserman reports that founders who gave up more equity built more valuable companies, and that most founders are eventually replaced as CEO, often against their wishes. He advises founders to decide explicitly which goal they prioritise.

HBR synthesis of the author's descriptive research; not a peer-reviewed causal study. The rich-vs-king correlation is confounded by venture quality (better ventures attract more investors and dilution).

The level of strategic change (shifts in resource-allocation patterns) has an inverted-U relationship with firm performance—moderate change helps, excessive change hurts. Both the upside of moderate change and the downside of excessive change are more pronounced for outside CEOs than for inside CEOs, and supplementary analysis shows the pattern is driven by later tenure years rather than the initial post-appointment period.

Archival data on 193 U.S. CEOs who departed between 1993 and 1998. Modest sample; strategic change proxied by financial resource-allocation indicators; performance is accounting-based; observational design.

Module 6, Module 7, Module 10, Module 13

Taking a paradox perspective, the authors find that humility and narcissism are not mutually exclusive in CEOs and that the combination of high humility and high narcissism is associated with the strongest firm innovation outcomes (innovative culture and innovation performance). The interaction effect is mediated by "socialized charisma" (charisma directed toward collective rather than self-serving ends).

Two studies of Chinese CEOs: Study 1 (63 CEOs, longitudinal multi-source design) and Study 2 (143 CEOs, cross-sectional), with humility rated via Owens et al.'s (2013) scale and narcissism via the NPI-16 (shortened in Study 2); data from ~1,163 managers overall. Chinese context; self-reported narcissism and other-rated humility; innovation measured partly perceptually; correlational.

Module 4, Module 5