Measures of CEO narcissism

Introduction

In the early 1980s, Hambrick and Mason (1984), in a seminal publication, proposed the upper echelons perspective.  According to this theory, the personality of executive managers, especially CEOs, affects the choices and ultimately the performance of organisations.  Since this time, hundreds of studies have explored the association between the personality of CEOs and the performance of organisations.  One personality trait in particular has been studied extensively: CEO narcissism. These studies have revealed that narcissistic CEOs can occasionally enhance some measures of financial performance or innovation but also often pursue risky, unethical, or even unlawful courses of action (for a review, see Cragun et al., 2020).

To explore the consequences of CEO narcissism, researchers need to measure this trait accurately.  To achieve this goal, some researchers invite CEOs to complete a measure of narcissism, typically the Narcissism Personality Inventory.  To illustrate, Zhang et al. (2017) explored whether CEO narcissism and humility affect the innovation of firms.  To explore this matter, the researcher invited 63 CEOs to complete the shortened version of the Narcissistic Personality Inventory, comprising 16 items.  A typical item is “I always know what I am doing”.  Other senior managers assessed the humility of these CEOs, and middle managers evaluated the degree to which the organisation embraces innovation.  When the CEOs exhibited both narcissism and humility, the organisation was especially innovative.

Nevertheless, in practice, as Cragun et al. (2020) acknowledged, researchers who explore the consequences of CEO narcissism seldom ask these CEOs to complete personality measures.  CEOs are often too busy to complete these instruments.  And narcissistic CEOs may be especially reluctant to complete these measures, potentially biasing the results.  Accordingly, researchers have developed, and to some extent validated, a range of other methods to gauge CEO narcissism.

The CEO narcissism index and variations

Because CEOs are often unable or unwilling to complete measures of narcissism, researchers instead often attempt to assess this trait unobtrusively.  That is, these researchers distil public information about CEOs to estimate levels of narcissism.  One of the most influential examples of this approach is called the CEO Narcissism Index (Chatterjee & Hambrick, 2007; Chatterjee & Hambrick, 2011).  To illustrate, Chatterjee and Hambrick (2007) gleaned five metrics from public databases:

• the salary of the CEO relative to the salary of the executive who receives the second highest income,
• other benefits the CEO receives, such as stock options, relative to the other benefits the next most remunerated executive receives,
• the size or prominence of the CEO picture in the annual report, ranging from no photograph of the CEO to a photograph of the CEO alone that covers a whole page,
• the number of times the CEO is mentioned in company press releases,
• the number of times the CEO uses first-person singular pronouns, such as my or I, compared to first person pronouns overall, during interviews—a metric that Chatterjee and Hambrick (2011) later decided to exclude.

To generate an overall measure of narcissism, researchers tend to standardise and then average these metrics.  As evidence of validity, researchers have shown that other assessments of narcissism do correlate highly with this index.  For example

• as one study revealed, the correlation between this index and ratings of CEOs from qualified security analysts was .74 (Patel & Cooper, 2014),
• as another study showed, the correlation between this index and ratings of CEOs from psychologists who watched short videos of these individuals speaking was .84 (Zhu & Chen, 2014).

This index tends to predict a range of hypothesised outcome, offering some evidence of validity.  Yet, although a seductively elegant method, concerns have been raised about this index (for a brief review, see Cragun et al., 2020).  Specifically

• the index does not differentiate grandiose narcissism from vulnerable narcissism,
• the evidence of validity is not watertight, because the various measures—such as ratings from psychologists watching videos—may not be valid either,
• instead, for example, all the various metrics could be measuring the degree to which CEOs like to appear confident rather than narcissism.

Since the work of Chatterjee and Hambrick (2007, 2011), many researchers have utilised and adapted this index.  Researchers often utilise only a subset of the five metrics.  Alternatively, some researchers have introduced additional metrics.  To illustrate, Rijsenbilt and Commandeur (2013) proposed an index that integrates 15 items including

• whether the CEO uses a private jet,
• number of times the CEO has appeared in major publications,
• number of acquisitions while the individual was CEO,
• number of lines in the Marquis Who’s Who database,
• number of awards received,
• governance index that measures the rights of managements compared to the rights of shareholders,
• whether the CEO is also the chairperson,
• number of role titles in addition to CEO,
• as well as the metrics that Chatterjee and Hambrick (2011) recommended.

Pronoun use

One metric that is occasionally included in these narcissism indices, but sometimes deliberately excluded, revolves around pronoun use. That is, researchers assume that narcissistic individuals, including narcissistic CEOs, will tend to use first-person singular pronouns such as “I” and “me”, more often than first-person plural pronouns, such as “we” or “us”. 

Indeed, this possibility can be traced to the work of Raskin and Shaw (1988).  In this study, 48 participants, all undergraduate students enrolled at the University of California at Santa Cruz, discussed a topic of their choice for five minutes. In addition, these participants completed the Narcissistic Personality Inventory. In general, the students who scored high on narcissism were more inclined to use first-person singular pronouns such as I, me, mine, my, and myself rather than first-person plural pronouns such as we, us, our, ours, and ourselves.  

Since this time, researchers have utilised this metric to gauge the narcissism of CEOs.  For example, Aktas et al. (2016) analysed the speeches of CEOs in response to fair disclosures.  To estimate narcissism, these researchers calculated the number of times these CEOs utilised first-person singular pronouns divided by the number of times these CEOs utilised first-person pronouns overall.  This measure was positively associated with the tendency of CEOs to initiate deals around acquisition but inversely associated with the tendency of CEOs to finalise deals.

Despite some success, researchers have raised concerns about this metric.  First, a more comprehensive study, conducted by Carey et al. (2015), challenged the validity of this measure.  Derived from over 15 samples, these researchers discovered the correlation between narcissism, as gauged from valid instruments such as the Narcissistic Personality Inventory, and pronoun use was close to zero.

Second, many other considerations may constrain the words that CEOs write or speak in public.  For example, the Sarbanes-Oxley Act (SOX) of 2002—a US federal law that mandates specific practices around financial reporting—may limit the discretion of CEOs when they discuss finances.  Indeed, Chatterjee and Hambrick (2011) referred to this Act to justify their exclusion of pronouns from their narcissism index.

Word use

Rather than first-person singular pronouns, narcissists might use other terms more frequently.  Anglin et al. (2018) , for example, argued that narcissists may often express words that coincide with authority, superiority, exhibitionism, vanity, and self-sufficiency—some of the dimension or facets of the Narcissistic Personality Inventory.   Examples of these words may include

• accomplished, mastermind, and virtuoso to represent authority,
• fiercest, shrewdest, or outstanding to represent superiority,
• fame, notable, or prominence to represent exhibitionism,
• grandest, pretty, or praiseworthy to represent vanity, and
• mastery, resilience, unfaltering to represent self-sufficiency.

Anglin et al. (2018) utilised this reasoning to develop a measure of narcissism.  To achieve this goal, they applies a procedure, utilising Computer-Aided Text Analysis, that Short et al. (2010) outlined and popularised.  In essence, Anglin et al. (2018)

• utilised the Narcissistic Personality Inventory to extract key words that represent each facet,
• uncovered synonyms to extend this dictionary of words,
• attempted to identify other possible words from a sample of relevant extracts—in this instance, Kickstarter campaigns, the focus of this study,
• undertook factor analysis to validate these word lists.  

This procedure identified a few 100 words that tend to epitomise narcissism.  The researchers discovered an inverted U shape relationship between the use of these words and performance in Kickstarter campaigns.  That is, moderate use of these words in Kickstarter campaigns, indicative of some narcissism, attracted significant funds.  Very low or high use of these words diminished funding.   

Other researchers have also used this measure narcissism.  For example, Kruse et al. (2025) analysed the answers of CEOs to questions during earnings calls: conference calls between the management and stakeholders of companies.  CEOs who used these words frequently were more likely to choose executive managers whose experience was limited—managers who are perhaps not as able to question decisions. 

Size of signatures

Researchers have also uncovered another unobtrusive metric that can be derived from public databases but may demand some more analysis: the size of their signatures.  That is, as Ham et al. (2017, 2018) proposed, CEOs who write larger signatures are more likely to exhibit the signs of narcissism. 

Research has indeed corroborated this possibility, at least to some degree.  In one study, published by Mailhos et al. (2016), 340 university students, studying at the Universidad de la República in Uruguay, completed the Narcissistic Personality Inventory—a measure of grandiose narcissism—as well as measures of self-esteem and dominance.  In addition, participants signed the consent form.  The total area the signature covered, when surrounded by a rectangle, was calculated.  The findings revealed that

• signature size was positively associated with dominance,
• signature size was positively associated with narcissism, but only in females.

These results, therefore, can be regarded as only qualified support of the notion that signature size epitomises narcissism.  The relationship was not significant in males.  And findings that apply to university students may not extend to CEOs, who presumably sign more frequently.  In addition, other characteristics, such as whether individuals sign only with their initials or include other stylistic features, also affects signature size but not relate to narcissism. 

Despite these concerns, many studies have shown that CEO narcissism, as gauged from signature size, does indeed predict a range of workplace outcomes.  For example,

• as Li et al. (2025) showed, when the signature of CEOs is large, the costs of debt tends to be higher, partly because the quality of financial reporting is lower,
• as Hou et al. (2025) demonstrated, when the signature of CEOs is large, credit ratings tend to be lower.

Assessments by third parties

Rather than unobtrusive measures, some researchers invite other individuals to evaluate the narcissism of CEOs.  These individuals vary from people who work closely with the CEO to people who merely observed extensive video footage or read biographical material about the CEO. These individuals might evaluate these CEOs on a few adjectives, derived from the DSM V, or on a comprehensive measure of narcissism. 

To illustrate, Resick et al. (2009) derived extensive biographic information about 75 CEOs of baseball clubs, from a range of sources, such as encyclopedias, local newspapers, and relevant periodicals. Trained assessors then utilised this information to evaluate the CEOs on a range of adjectives, each of which corresponded to established measures of narcissism.  CEOs who were deemed as narcissistic on this measure were more inclined to apply a leadership style called contingent reward rather than transformational leadership. In contrast, Gupta and Misangyi (2018) derived measures of CEO narcissism from brief video footage of these individuals.  These videos were extracted from the internet but edited to mask the identity of these CEOs, such as obscured the face. The videos were usually a few minutes long.   Three trained assessors then utilised the Narcissism Personality Inventory, comprising 16 items, to evaluate each CEO.   As the study revealed, when CEOs were deemed as narcissistic, their firm was less likely to be emulated by other rival companies.  That is, other companies did not tend to trust these leaders and were thus disinclined to emulate their behaviour.