Abstract
Chief executive officer (CEO) narcissism is an important area of research due to the strategic implications of how this multifaceted personality trait affects CEO behavior. This article presents a combined meta-analytic and narrative review of CEO narcissism and makes future research recommendations. Our review and meta-analytic findings lead to the creation of a framework for CEO narcissism research focused on narcissistic CEO supply, demand, behavior, and consequences. Additionally, our review identifies five methods of measuring CEO narcissism, each with strengths and weaknesses. We find that while extant findings exhibit common themes, such findings remain mixed and potentially dependent upon methods. We recommend that future research expand beyond the strategic consequences of CEO narcissism to consider additional foci of the research framework and its moderators. Additionally, we suggest that research can benefit from moving beyond the predominant theoretical lenses of upper echelons theory and leadership theory to the lenses of the extended agency model of narcissism, the admiration-versus-rivalry perspective of narcissism, and tournament theory.
Keywords
While narcissism has been long studied in psychology (Levy, Ellison, & Reynoso, 2011), research on narcissism in chief executive officers (CEOs) and its effects on firm outcomes has only recently emerged (e.g., Buyl, Boone, & Wade, 2017; Patel & Cooper, 2014). This stream of literature was spurred by the development of narcissism measures that do not rely on CEOs to provide data and has led to a growing body of empirical findings (e.g., Chatterjee & Hambrick, 2007; Resick, Whitman, Weingarden, & Hiller, 2009). These findings encompass a variety of strategic outcomes that at times have displayed seemingly contradictory results that are yet to be reconciled (Chatterjee & Hambrick, 2007; Ham, Seybert, & Wang, 2018; Olsen, Dworkis, & Young, 2014). Without such reconciliation, the field might continue down a variety of either parallel or divergent research streams with increasing difficulty in integrating them.
This review seeks not only to understand and reconcile extant findings on CEO narcissism but also to identify future research opportunities. To achieve these objectives, we present a combined meta-analytic and narrative review. In doing so, this review makes four main contributions. First, we review extant measures of CEO narcissism. We provide guidance in measurement choice, make recommendations for improving existing methods, and suggest possible new methods. Second, we aggregate empirical findings and identify contingencies of CEO narcissism’s effects. In so doing, we clarify consistent findings and identify possible reasons for inconsistent findings. This effort highlights how methods may, in fact, be contributing to divergent results. Third, we provide a framework for future research exploring CEO narcissism. This framework adds depth to the current emphasis on strategic firm outcomes, drawing attention to the supply and demand of narcissistic CEOs and the manifestation of CEO narcissism. Our intent is that this framework will create the possibility of a more balanced and theoretically rich research stream. Fourth, we identify alternative theoretical lenses that may provide insightful perspectives to explain unexplored questions about CEO narcissism.
CEO Narcissism Definition and Methods
We begin our review by defining narcissism according to the guidelines of the American Psychiatric Association (APA; 2013), as contained in the Diagnostic and Statistical Manual for Mental Disorders—the DSM. The DSM defines narcissism as a multifaceted personality trait that combines grandiosity, attention seeking, an unrealistically inflated self-view, a need for that self-view to be continuously reinforced through self-regulation, and a general lack of regard for others. Grandiosity suggests entitlement, self-centeredness, and a belief that one is better than others. Attention seeking implies that one exerts significant effort to become the focus of attention. An unrealistic self-view suggests that central to the individual’s identity is an overinflated, inaccurate picture of the self. Self-regulation refers to “all the processes, mechanisms, tactics, and strategies people engage in to regulate and shape their self-images” (Morf, Torchetti, & Schürch, 2011: 57) and is critical to understanding an individual’s motivations and thought patterns. Finally, a general lack of regard for others refers to a lack of empathy toward others and a tendency to exploit situations and persons for personal gain.
After identifying and comparing definitions used in prominent CEO narcissism articles, we found that most definitions of narcissism could be traced back to the DSM. In fact, the DSM is the common denominator across most studies. Because of this, we grounded our article in the DSM definition. Despite its origin in clinical psychology, the DSM’s definition of narcissism is also widely accepted for defining narcissism in its nonclinical form (Judge, LePine, & Rich, 2006; Raskin & Terry, 1988). In addition, the APA continually updates the DSM with the latest information, thereby making its definition advantageous for future research purposes (for a detailed history, see Levy et al., 2011).
To gather articles for our review, we searched within major management, accounting, and finance journals with the prominent databases Google Scholar and EBSCO Complete. We searched for the words “narcissism” and “CEO” in the article titles and conducted keyword searches for the words “narcissism” and “CEO.” We searched published, online, and unpublished articles. Any article not available online or through the journal we solicited directly from the author. If we found sources that were not originally identified, we continued to add them to the review throughout the entire research project. Our efforts yielded 62 peer-reviewed articles, books, academic-oriented articles, unpublished articles, or dissertations; however, for the narrative review itself, we relied primarily on peer-reviewed articles.
For our meta-analysis, we first identified 37 articles from peer-reviewed and non-peer-reviewed sources reporting effect sizes and then followed the methods of Hunter and Schmidt (2004). Effect sizes were coded by sample, article, and measure to ensure that no effect sizes from the same sample were double counted. Due to the high proportion of samples from databases using U.S.-based Securities and Exchange Commission (SEC) filings (e.g., Execucomp), we conducted additional tests for robustness looking for bias due to nonindependence (van Essen, Otten, & Carberry, 2015). We found our findings to be robust (see Appendix).
Rather than dedicate a separate section to our meta-analytic findings, we intersperse those findings throughout our review. We organize our review into the following sections. First, we focus on the differing measurements of CEO narcissism since they are the foundation of emerging research. Understanding how the key independent variable is measured provides insight into what can be researched, as well as the limitations of extant findings. Second, we review extant findings on CEO narcissism. Third, we draw on broader management and narcissism research to organize our findings into a framework for future research. Finally, we explore the theoretical lenses used in CEO narcissism research.
Measuring CEO Narcissism
CEO narcissism has been measured in a variety of ways across multiple disciplines. In our review, we categorized differing measures of CEO narcissism and identified potential methodological improvements. We found that extant research has used five measurement types: CEO narcissism index, psychometric self-report, psychometric third party, pronoun use, and signature size. Central to the evolution of these measures is the innovation of new, unobtrusive methods (Webb & Weick, 1979) to reduce reliance on CEOs completing psychological scales. We consider each of these five measurement types in turn before discussing emerging methods that could be adapted to CEO narcissism.
CEO Narcissism Index
Chatterjee and Hambrick (2007, 2011) pioneered the use of the CEO narcissism index as a composite measure of several narcissism indicators observed in firm reports and publicly available documents. The CEO narcissism index has since become the most frequently used method (23 of 42 articles; Bianchi, 2014; Buchholz, Jaeschke, Lopatta, & Maas, 2018; Buyl et al., 2017; Chatterjee, 2009; Chatterjee & Hambrick, 2007, 2011; Engelen, Neumann, & Schmidt, 2016; Gerstner, König, Enders, & Hambrick, 2013; Ingersoll, Glass, Cook, & Olsen, 2017; Judd, Olsen, & Stekelberg, 2017; Kashmiri, Nicol, & Arora, 2017; Liu, 2009; Marquez-Illescas, Zebedee, & Zhou, 2018; Oesterle, Elosge, & Elosge, 2016; Olsen et al., 2014; Olsen & Stekelberg, 2016; Patel & Cooper, 2014; Rijsenbilt, 2011; Rijsenbilt & Commandeur, 2013; Schrand & Zechman, 2012; Tang, Mack, & Chen, 2018; Zhu & Chen, 2014, 2015). Chatterjee and Hambrick’s (2007) original index included five components: (1) the relative cash pay of the CEO to the next-highest paid executive, (2) the relative noncash pay of the CEO to the next-highest paid executive, (3) the size of the CEO’s picture in the annual report, (4) the number of CEO mentions in company press releases, and (5) the number of first-person singular pronouns used by the CEO during interviews.
Chatterjee and Hambrick (2007) demonstrated internal statistical consistency of the measure by providing evidence that these measures were more related to the CEO than the CEO’s firm (r = 0.90, p value is unreported). Qualified security analysts also rated the CEOs as narcissistic (r = 0.82, p < .01). Other researchers using the Chatterjee and Hambrick (2007) index have used third-party evaluators to further validate the measure. For instance, Patel and Cooper (2014) had qualified security analysts (r = 0.74, p < .001) and psychology researchers (r = 0.70, p < .001) independently rate their sample, while Zhu and Chen (2014) used psychologists who provided CEO narcissism ratings after watching short videos of the CEOs talking (r = 0.84, p value is unreported). We note that while Chatterjee and Hambrick (2007) used five components, Chatterjee and Hambrick (2011) found that the first-person pronoun component was unreliable post the Sarbanes Oxley Act of 2002 and thereafter used only the remaining four components.
Despite its common use and validation, the CEO narcissism index faces two major limitations. First, Brown (2016) raises the concern that the CEO narcissism index does not directly link to the multifaceted subdimensions of narcissism (e.g., grandiosity). Second, several CEO narcissism index items could be influenced by factors outside the CEO’s control. For example, two of the index’s components are based on CEO compensation, and firm size alone may explain as much as 40% of a CEO’s pay (Tosi, Werner, Katz, & Gomez-Mejia, 2000). Due to data availability, some research employs only three survey items (Ingersoll et al., 2017; Schrand & Zechman, 2012), while others merely employ the compensation items (Bianchi, 2014), thereby placing even more weight on compensation measures.
Our meta-analysis suggests support for the criticism regarding the influence of firm size on the CEO narcissism index. We found that firm size had a positive and statistically significant correlation with CEO narcissism for all measures of CEO narcissism (ρ = .08, SDρ = .02, p < .001; see Table 1). This effect increased when we considered only studies that employed the CEO narcissism index (ρ = .15, SDρ = .03, p < .001; see Table 2). This suggests that the CEO narcissism index measure has a higher correlation with firm size than other measures. These results do not necessarily mean that firm size contaminates the CEO narcissism index measure, as it could be that firm size enhances narcissistic tendencies or that narcissistic CEOs tend to gravitate to larger firms. However, these results do imply caution in the use of the index.
Meta-Analytic Results of CEO Narcissism and Firm Attributes and Outcomes
Note: Mean ρ = average corrected correlation; SDρ standard deviation of corrected correlation; LL = the lower limit of the 95% confidence interval; UL = the upper limit of the 95% confidence interval; Q = Q statistic to test for homogeneity in the corrected correlations across studies; variance = percentage of the variance in correlations explained by statistical artifacts.
p < .05.
p < .01.
p < .001.
The total number of studies using each CEO narcissism measurement method are as follows: CEO narcissism index, n = 23; psychometric self-report, n = 6; psychometric third-party rating, n = 7; pronoun use, n = 4; signature size, n = 2. Each of these articles is cited in the Measuring CEO Narcissism section. However, only studies with correlations between the firm-level measure and the CEO narcissism measure were analyzed, and only analysis with two or more correlation pairs are reported in the Table 2 meta-analysis results.
Meta-Analytic Results of CEO Narcissism Based on Method of Measurement
Note: Mean ρ = average corrected correlation; SDρ standard deviation of corrected correlation; LL = the lower limit of the 95% confidence interval; UL = the upper limit of the 95% confidence interval; Q = Q statistic to test for homogeneity in the corrected correlations across studies; variance = percentage of the variance in correlations explained by statistical artifacts.
p < .01.
p < .001.
As an alternative, Rijsenbilt and Commandeur (2013) developed an index of 15 items (e.g., private jet use, publicity, acquisition activity). The increased number of items could provide enhanced measurement accuracy. Nevertheless, this expanded measure has not been widely used due to limited data availability (Buchholz et al., 2018; Rijsenbilt, 2011; Rijsenbilt & Commandeur, 2013). Despite the potential limitations of the Chatterjee and Hambrick (2007) CEO narcissism index, it maintains the advantage of unobtrusive data that are publicly available for a large sample of CEOs across multiple industries. Thus, it is not surprising that it has been the most commonly used measure and is currently the most validated unobtrusive measure for researchers to consider for future research.
Psychometric Self-Report
Psychometric self-report measures are gathered when CEOs complete an assessment (6 of 42 articles; Beauchesne, 2014; Peterson, Galvin, & Lange, 2012; Reina, Zhang, & Peterson, 2014; Wales, Patel, & Lumpkin, 2013; Wang, 2016; Zhang, Ou, Tsui, Wang, & Hui, 2017), though such measures are rarely used due to difficulty in obtaining the data. The Narcissism Personality Inventory (NPI) has been the only assessment used in CEO research. The NPI is the most widely used assessment of narcissism as a personality trait and is considered accurate and valid (Gentile et al., 2013; Raskin & Terry, 1988). While the 40-question NPI best captures the multifaceted nature of narcissism, shorter versions have been used (e.g., NPI-16; Gentile et al., 2013). The NPI has been used as the basis for self-report and third-party psychometric measurement of CEO narcissism and as conceptual validation of the CEO narcissism index and non-NPI-based third-party psychometric measurements (Chatterjee & Hambrick, 2007; Resick et al., 2009; Rijsenbilt & Commandeur, 2013). In a perfect world, this direct measure would be the first choice for researchers to consider for future research.
Psychometric Third Party
To take advantage of a validated psychometric scale while attempting to overcome the challenge of CEO access, a growing number of researchers have used third parties to score assessments measuring the CEO’s narcissism (7 of 42 articles; Cragun, 2018; Gupta & Misangyi, 2018; O’Reilly, Doerr, Caldwell, & Chatman, 2014; O’Reilly, Doerr, & Chatman, 2017; Petrenko, Aime, Ridge, & Hill, 2016; Regnaud, 2014; Resick et al., 2009). For example, Resick et al. (2009) identified eight adjectives aligned with the arrogance, grandiosity, and self-promoting behavior outlined in the DSM definition of narcissism (APA, 2013) and then asked third-party assessors to use these adjectives to rate CEOs based on their biographies.
Other studies have employed similar approaches with the NPI based on direct observation of CEOs (e.g., Cragun, 2018) or after watching or reading a CEO’s speech (e.g., Gupta & Misangyi, 2018; Petrenko et al., 2016). These direct observation methods can be valid uses of scales developed as self-report measures (Boldero, Bell, & Davies, 2015). While third-party assessments require access to qualified observers to provide ratings, such ratings are unobtrusive to the CEO and grounded in validated measures. Where possible, obtaining data from third-party informants who have direct, prolonged, day-to-day exposure to the CEO provides intimate observation of CEO behavior. For example, chief human resource officers report that they are confidants to their CEOs and spend a significant portion of their time directly with their CEOs (Wright, Cragun, et al., 2016; Wright, Nyberg et al., 2016). While widespread adoption of direct third-party observation is unlikely, due to the difficulty in gaining access to qualified observers, it is a desirable choice for researchers to consider where feasible.
Pronoun Use
While not as commonly used as a stand-alone measure of CEO narcissism, research has employed personal pronoun usage as a measure of narcissism in general (DeWall, Buffardi, Bonser, & Campbell, 2011; Raskin & Shaw, 1988) and CEO narcissism specifically (4 of 42 articles; Aktas, de Bodt, Bollaert, & Roll, 2016; Capalbo, Frino, Lim, Mollica, & Palumbo, 2017; de Bodt, Bollaert, Grandin, & Roll, 2015; Knights, 2014). This measure draws on speech styles and the narcissists’ self-focus and is calculated by contrasting the use of singular pronouns (e.g., I, me, mine) versus plural pronouns (e.g., we, us, ours) in regular dialogue (Raskin & Shaw, 1988).
Pronoun usage as a measure of narcissism has received criticism (Carey et al., 2015), and obtaining sufficient video or interview transcripts of CEOs speaking in everyday conversation is challenging. Furthermore, CEOs have become more restrained in their discussions since the passage of the Sarbanes-Oxley Act of 2002 (Chatterjee & Hambrick, 2011). Due to the lack of consistent data, questions regarding which construct the measure reflects and changes in CEO behavior, the use of pronouns as a measure may require additional validation.
Signature Size
Ham et al. (2018) and Ham, Lang, Seybert, and Wang (2017) employed an alternative unobtrusive measure of CEO narcissism by measuring the size and contents of a CEO’s signature in SEC filings (2 of 42 articles). Their rationale posits that a larger signature represents the grandiose nature of a narcissist. To validate the measure, Ham et al. used student samples in a laboratory setting. While the main advantage of the measure is that it captures a behavior under the direct control of the CEO (i.e., his or her signature), the measure may not fully capture narcissism’s multifaceted nature. Ham et al. also provided external validation by correlating the measure with employee ratings of CEO narcissism, as obtained by O’Reilly et al. (2014). As this is a newer measure, it has seen limited use to date.
Measures of Narcissism Not Yet Found in CEO Narcissism Research
To the best of our knowledge, two promising psycholinguistic methods of measuring CEO personality have not yet been adapted for use in CEO narcissism research: first, a psycholinguistic dictionary–based evaluation based on a closed list of terms (e.g., Nadkarni & Chen, 2014); second, open language–based methods. For example, Harrison, Thurgood, Boivie, and Pfarrer (2019) developed an open language method of evaluating text to determine CEO personality based on the Big Five personality traits. Given a high level of interest in connecting CEO language to CEO narcissism (e.g., Amernic & Craig, 2007; Craig & Amernic, 2011) and the innovation of applying key language in the evaluation of CEO speech by trained evaluators (Resick et al., 2009), creating a set of closed language around narcissism seems plausible. Extending that to an open language approach also seems possible and could yield a useful measure of CEO narcissism.
That said, open language approaches distilled through machine learning are still emerging and have been challenged theoretically and empirically. Care should be taken not to overfit the model, as open language models are susceptible to issues of multicollinearity (Kern et al., 2016). Another significant issue with open language analysis is that samples are nonrandom (Kern et al., 2016). This affects the structural validity and generalizability of the analysis, as the analysis uses a subset of the individual’s overall language within the context of a particular situation (Bleidorn & Hopwood, 2019). Another significant criticism of machine learning methods is that, despite their predictive capability, it is unclear to what extent they measure “relatively stable patterns of thoughts, feelings, and behavior (i.e., personality traits) versus related psychological characteristics, such as preferences, interests, attitudes, motives, or beliefs” (Bleidorn & Hopwood, 2019: 195). When CEO language is gathered from public transcripts of interviews, it must remembered that CEOs prepare for these events and are influenced by social desirability and legal and regulatory considerations when speaking (Buchholz et al., 2018; Chatterjee & Hambrick, 2011). Thus, while offering a potential avenue for measuring CEO narcissism, psycholinguistics methods will need validation in future research.
Summary of CEO Narcissism Measurement
To better understand how CEO narcissism influences a variety of outcomes, researchers have identified multiple ways to measure it. However, difficulty in obtaining validated and reliable self-report measures from CEOs limits CEO narcissism research and has caused the emergence of new methods, such as the unobtrusively obtained CEO narcissism index (Chatterjee & Hambrick, 2007). Initial success and insights provided by the CEO narcissism index demonstrate serviceability and usefulness. Yet, while amenable to analyzing large firms for which data are available, it does not provide the rich, multifaceted psychological evaluation that validated scales could provide. Alternative unobtrusive measures, such as third-party psychometric ratings (Petrenko et al., 2016; Resick et al., 2009) or broader behavioral indicator indices, could also be employed (Rijsenbilt & Commandeur, 2013). Emerging measures of personality, such as closed and open language approaches, could be adapted to CEO narcissism research (e.g., Harrison et al., 2019), though theoretical and empirical challenges remain. Thus, although psychological scale-based measures are most ideal, due to their validity and general reliability (Gentile et al., 2013; Raskin & Terry, 1988), researchers must balance the benefits and disadvantages of available measures based on their sample of interest.
CEO Narcissism Research Findings
We next review how narcissistic CEOs influence their firms. Specifically, we identified five ways in which narcissistic CEOs significantly affect their firms: firm performance, innovation and growth, policies and risk, financial leverage, and questionable behavior. We include findings based primarily on our narrative review and augment those findings with meta-analysis findings where possible.
Firm Performance
Several studies have examined the effect of CEO narcissism on firm performance; however, the evidence has not yielded consistent findings. Chatterjee and Hambrick (2007) found that CEO narcissism positively affects performance extremes, measured as deviations from average industry stock returns or return on assets (ROA) and performance fluctuation (the change in ROA year over year). Reina et al. (2014) found evidence that CEO narcissism is related to firm performance (measured by ROA) and moderated by organizational identification. Olsen et al. (2014) found that CEO narcissism positively affects a firm’s earnings per share and stock price. In contrast, Ham et al. (2018) found that CEO narcissism negatively affects a firm’s profitability, as measured by ROA and operating cash flows.
Our own meta-analysis shows that the overall relationship between CEO narcissism and firm financial performance (e.g., ROA, total shareholder returns, operating profit, Tobin’s Q, earnings per share) is positive and statistically significant (ρ = .06, SDρ = .02, p < .001), yet the statistically significant Q value (Q = 32.66, p < .05) indicates additional moderators (see Table 1). Further breakdown for measures with sufficient effect sizes is possible for ROA and Tobin’s Q, yet neither shows statistically significant results. Instead, the data show an aggregation of performance, as represented by a variety of performance measures (operating profit, total shareholder returns, return on sales, earnings per share, cash, market value added, win percentage, and market-to-book ratio) as having a statistically significant effect size (ρ = .10, SDρ = .03, p < .001), yet the statistically significant Q value (Q = 23.40, p < .05) again indicates additional moderators (see Table 1).
We next explored the effects of CEO narcissism on performance by varying measures of CEO narcissism. We found that the CEO narcissism index had a statistically significant effect size (ρ = .08, SDρ = .03, p < .01), yet the statistically significant Q value (Q = 21.21, p < .01) indicates the presence of additional moderators (see Table 2). No other measures of narcissism had enough effect sizes to aggregate. Thus, while the results of aggregating the effect sizes of the narcissism-to-performance relationship—either by performance measure or narcissism measure—provide preliminary evidence of the nature of a CEO’s narcissism effect on firm performance, there is still room for further analysis. More studies are needed to reach a more definitive conclusion.
Innovation and Growth
A number of studies have explored the relationship between CEO narcissism and firm innovation and growth. Kashmiri et al. (2017) found evidence that CEO narcissism is related to new product innovation and radical innovations in product portfolios, and Zhang et al. (2017) found a correlation between narcissistic CEOs who also show humility and firm innovative performance. Evidence also suggests that narcissistic CEOs spend more on research and development (R&D) and acquire new companies more aggressively (Chatterjee & Hambrick, 2011; Ingersoll et al., 2017).
To explore this relationship further, we calculated the effect sizes between CEO narcissism and measures of firm innovation and growth using meta-analysis. We found mixed statistical evidence of the relationship between CEO narcissism and innovation and growth. Like firm performance outcomes, the measurement of innovation and growth constructs may affect the analysis. At the highest level, the relationship between CEO narcissism and innovation and growth was statistically significant (ρ = .09, SDρ = .02, p < .001; Q = 58.61, p < .001; see Table 1). However, at the second level, the results were not as consistent. That is, our meta-analysis showed a lack of statistical significance of the relationship between CEO narcissism and either growth or mergers and acquisitions. Growth was measured as sales growth or new market entry, and mergers and acquisitions were measured as the number and size of acquisitions. However, we found a positive and statistically significant correlation between CEO narcissism and innovation when measured as new product introductions or entrepreneurial orientation (ρ = .14, SDρ = .05, p < .01); we also found a positive and statistically significant correlation between CEO narcissism and innovation when measured by R&D spending (ρ = .11, SDρ = .03, p < .001). In both cases, the statistically significant Q value (innovation: Q = 11.17, p < .05; R&D: Q = 26.63, p < .001) indicated additional moderators (see Table 1).
We also investigated the relationship between CEO narcissism and measures of firm innovation and growth, based on how studies measured narcissism. When based on the CEO narcissism index only, the aggregated effect size of narcissism to innovation and growth was positive and statistically significant (ρ = .09, SDρ = .02, p < .001), yet the statistically significant Q value (Q = 31.10, p < .01) indicated additional moderators (see Table 2). Other measures of CEO narcissism did not have enough effect sizes to aggregate, or they were statistically nonsignificant. Therefore, how CEO narcissism affects firms’ innovation and growth seems context and measurement dependent.
Policies and Risk
An additional outcome that research has considered is the link between narcissism and increased risk taking (Campbell, Goodie, & Foster, 2004). Our review found that this link typically manifests as CEOs making riskier decisions for recognition (Chatterjee & Hambrick, 2011) or manipulating policies to achieve desired results (Buyl et al., 2017; Olsen et al., 2014). Our meta-analysis, however, failed to find a statistically significant positive relationship between CEO narcissism and risk taking, with risk taking measured as effectiveness of internal controls, high-risk investment strategies, aggressive financial reporting, or risky spending (see Table 1).
While some empirical research finds evidence that CEO narcissism influences risk taking, the meta-analysis results are affected by construct definition and measurement of “risk” across studies (and disciplines). For example, Ingersoll et al. (2017) investigated CEO narcissism within the ethics domain, defining risk taking as a composite measure of R&D and capital expenditures and acquisitions, while others have used R&D spending as evidence of growth and innovation (Chatterjee & Hambrick, 2007). In our meta-analysis, we coded any measure, including R&D spending, as growth and innovation rather than risk taking. Since risk is fundamental to the study of CEO narcissism, more research is warranted to clarify the construct and measurement of risk.
Financial Leverage
Narcissistic CEOs may increase their firm’s financial leverage to improve financial results, due to their tendency for overconfidence, risky decision making, and influencing or manipulating accounting numbers (Buyl et al., 2017; Capalbo et al., 2017). We examined the relationship between CEO narcissism and leverage ratios, including debt to capital, debt to equity, and current ratio. The meta-analysis shows a lack of statistical significance regarding the overall relationship between CEO narcissism and financial leverage (see Table 1). This suggests that there is no such relationship or that further study is needed to understand how CEO narcissism might affect a firm’s leverage.
Questionable Behavior
Finally, an underlying theme of CEO narcissism research is that narcissistic CEOs engage in questionable behavior often characterized as unethical, deceitful, illegal, reckless, or purely selfish. Such questionable behavior includes bullying (Regnaud, 2014), fraud (Rijsenbilt & Commandeur, 2013), distorting financial information (Capalbo et al., 2017), increased likelihood of using corporate tax shelters (Olsen & Stekelberg, 2016), and fraudulent financial reporting (Risjenbilt & Commandeur, 2013). However, there were not enough effect sizes to aggregate findings based on questionable behavior.
Behaviors such as increased corporate social responsibility could also be construed as questionable because they may be enacted for purely self-serving purposes—to attract attention and praise. In fact, Petrenko et al. (2016) found that although narcissistic CEOs spend significant money on corporate social responsibility, the impact of so doing diminishes with greater narcissism of the CEO. A potential contributing factor to a narcissist’s ability to justify using company resources for self-serving behavior is the narcissist’s level of organizational identification: The more a narcissistic CEO identifies with the organization, the more that CEO sees the organization as a personal extension (Fuller, Galvin, & Ashforth, 2017; Galvin, Lange, & Ashforth, 2015).
Ethics researchers argue that environment is a legitimate and meaningful contributor to CEO narcissistic unethical behavior (e.g., Chen, 2010) and that CEO wrongdoing increases when there is pressure and opportunity with rationalization (Schnatterly, Gangloff, & Tuschke, 2018). The implication for researchers and companies is that if controls are not applied to mitigate risk factors, narcissistic CEOs can cause great harm. Future research could examine how CEO narcissism might contribute to unethical behaviors and which factors might amplify or mitigate these outcomes.
Summary of Research Findings
Insightful research streams investigating the effects of CEO narcissism are emerging. Research on CEO narcissism and firm performance has yielded varied results, depending on study design and methodology (Chatterjee & Hambrick, 2007; Ham et al., 2018; Olsen et al., 2014), while its influence on firm innovation and growth indicates that narcissists spend more on R&D and acquisitions (Chatterjee & Hambrick, 2011; Ingersoll et al., 2017). Research on CEO narcissism with regard to policies and risk has provided initial evidence that CEOs make riskier decisions (Chatterjee & Hambrick, 2011) and manipulate policies to achieve desired results (Buyl et al., 2017; Olsen et al., 2014). Finally, evidence of the detrimental effects of narcissistic CEOs’ questionable behavior is accumulating. This behavior includes bullying (Regnaud, 2014), fraud (Rijsenbilt & Commandeur, 2013), and distorting financial information (Capalbo et al., 2017), among others.
Key Questions and Future Directions
The majority of CEO narcissism research examines the strategic consequences of CEO narcissism (e.g., firm performance and risk taking). However, this focus is limiting. Future research could delve deeper into the causal chain to understand the what, how, why, who, where, and when of CEO narcissism (Whetten, 1989). For a framework of future CEO narcissism research, we draw on narcissism and strategic leadership literatures (Campbell & Miller, 2011; Finkelstein, Hambrick, & Cannella, 2009; Rosenthal & Pittinsky, 2006) and propose four overarching research avenues: What increases the supply of narcissistic CEO candidates? What creates demand for narcissistic CEOs? When will a narcissistic CEO’s behavior manifest? What are the consequences of CEO narcissism? This framework outlines the potential for a balanced and theoretically rich research stream of literature. We present extant evidence in each of these categories and make recommendations for future research. The framework and its elements are summarized in Figure 1.
Framework for Future CEO Narcissism Research
What Increases the Supply of Narcissistic CEO Candidates?
While it is beyond the scope of the CEO narcissism literature to establish the psychological origins of narcissism, future research could provide richer understanding regarding the life and career experiences that lead to a narcissistic individual advancing through the ranks to become a CEO. For example, evidence suggests that narcissists sort not only into business schools but also into CEO-oriented career tracks (Brown, Akers, & Giacomino, 2013; Young & Pinsky, 2006). Prior research has also linked CEO narcissism to entrepreneurship (e.g., Engelen et al., 2016; Navis & Ozbek, 2016). Rather than climbing the corporate ladder, narcissists might be inclined to start their own firms. Indeed, Wales et al. (2013) provided evidence that narcissism and entrepreneurial orientation are correlated. Anecdotal evidence of entrepreneurs suggests a link between narcissistic CEOs and entrepreneurial founders (e.g., Steve Jobs, Elon Musk, Elizabeth Holmes). Nevertheless, this research stream is underdeveloped and could still yield insights into the career path of narcissistic individuals who become CEOs.
Alternatively, if a narcissistic individual enters the workforce, narcissism has the potential to be a key driver in the corporate ladder sorting process (Kets de Vris & Miller, 1985; Maccoby, 2000). Evidence suggests that narcissism is associated with leadership emergence (Grijalva, Harms, Newman, Gaddis, & Fraley, 2015) and managerial potential (Furnham, Trickey, & Hyde, 2012). However, there is little evidence regarding how a narcissist advances in a professional career to become a viable CEO candidate. Research could also identify and describe how company culture might foster narcissism (Schwartz, 1991). CEO career research has identified how a particular experience set leads to a person rising to the office of CEO (Crossland, Zyung, Hiller, & Hambrick, 2014). Perhaps, a narcissist’s rise to the CEO position results from high-risk, high-payoff behavior (Goel & Thakor, 2008). Alternatively, a set of career experiences or certain corporate cultures exist that reward and reinforce narcissistic behaviors.
Significant workplace experiences can also affect narcissistic development. Bianchi (2014) found that CEOs who entered the workforce in less prosperous economic times demonstrated fewer narcissistic behaviors when they later became CEOs. The theoretical corollary would state that individuals entering the workforce and significant developmental experiences in successful economic times marked by opulent rewards would exhibit more narcissistic behavior when they later became CEOs. Additionally, an examination of executives’ careers indicates that with great success comes significant pay increases, perks, attention, publicity, and recognition, all of which have potential to exacerbate narcissistic tendencies (Campbell & Foster, 2007; Finkelstein et al., 2009; Gentile, 2011). The little we do know about the supply of narcissistic CEO candidates suggests that there is much more to learn, thereby providing an opportunity for future research to address a variety of questions with a variety of research methods (e.g., field studies, case studies, surveys, and ethnographies).
What Creates Demand for Narcissistic CEOs?
While research has explored the organizational outcomes of CEO narcissism, often focusing on the downsides of narcissism, there has been little exploration into factors that lead boards of directors and their firms to hire narcissistic CEOs. Given that narcissistic behavior is a predictable, pronounced trait that endures over time (APA, 2013) and is demonstratively apparent to third-party raters (Boldero et al., 2015), one could presume that boards that hire narcissistic CEOs do so knowingly. For example, it is possible that a board focused on innovation in a period of poor firm performance feels that it has little to lose, thus leading to the selection of a narcissistic CEO. It is also possible that narcissistic boards prefer narcissistic CEOs. Zhu and Chen (2014) provide evidence that narcissistic CEOs are associated with the appointment of narcissistic directors. It is plausible that the reverse is also true.
The leadership of narcissistic CEOs can be associated with benefits and harm. In times of crisis, instability, or lack of direction, a candidate’s display of self-confidence and reassuring rhetoric can calm anxious organizational members and create a path for the emergence of a narcissistic leader (Rosenthal & Pittinsky, 2006). Nevertheless, the situational advantages that give rise to the narcissist leader can fade quickly (King, 2007). While narcissistic leaders may thrive in chaotic times, they may also seem out of place in more tranquil times (Maccoby, 2000). Future research could provide greater insight into the puzzle of which circumstances lead to the appointment and success of a narcissistic CEO.
When Will a Narcissistic CEO’s Behavior Manifest?
In the extant CEO narcissism literature, narcissism is treated as a personality trait. Trait theorists believe that an individual’s personality is stable across time and situations, which results in predictable behaviors and motivations (Miller & Campbell, 2008). Treating narcissism as a trait allows researchers to measure narcissism at any point in time and then assume or empirically specify that it does not change. For example, numerous studies measure CEO narcissism in the early years of a CEO’s tenure, assume a fixed value for the narcissism measure, and then examine its effect across multiple years in panel data form (e.g., Chatterjee & Hambrick 2007, 2011; Olsen et al., 2014; Zhu & Chen, 2014).
However, behaviors and motivations may also result from a combination of personality- and situation-based characteristics (Furr & Funder, in press; Pervin & John, 1999). This suggests that the effect of CEO narcissism could differ by context (Amernic & Craig, 2010; Gamache, McNamara, Mannor, & Johnson, 2015; Mannor, Wowak, Bartkus, & Gomez-Mejia, 2016). Maccoby (2003, 2007) suggests that successful narcissistic leaders—those who find productive manifestations of their narcissistic personality traits—are constrained within an organizational context by their own self-knowledge and restraining situational and environmental factors. Therefore, a more nuanced approach to CEO narcissism might include drawing on trait activation theory to understand narcissistic behavior triggers, rather than assuming that they manifest in narcissistic CEOs (Judge & Zapata, 2015).
For example, narcissists crave attention and approbation (APA, 2013); therefore, situations that provide attention and approbation may trigger narcissistic behavior. Prior CEO narcissism research has found consistent support for this notion. Petrenko et al. (2016) found that narcissistic CEOs are more likely to engage in corporate social responsibility efforts to draw attention and reinforce image. Chatterjee and Hambrick (2011) found that narcissistic CEOs are emboldened in their risk taking by social praise, while Olsen et al. (2014) found that narcissistic CEOs will engage in greater real-earnings management to improve their reported earnings per share. Thus, the institutional contexts in which a CEO operates can elicit narcissistic behavior by providing avenues for narcissistic CEOs to garner praise, attention, and image building.
Situational moderators
To tease out factors that enable the manifestation of narcissistic CEO behavior, researchers could consider situational factors that moderate this relationship. For example, boards of directors may communicate to a CEO through language, governance choices, or by example that narcissistic behavior is tolerated, valued, or discouraged. Compensation design and performance monitoring are other mechanisms by which boards can curtail negative behavior and encourage positive behavior. Future research could examine the role and influence of corporate governance on the behavior of narcissistic CEOs.
Potentially complicating the examination of a board’s governing role over narcissistic CEOs is the fact that the CEO may also be chair of the board. To this point, our meta-analysis suggests a positive and statistically significant relationship between CEO narcissism and chair-of-the-board positions (i.e., CEO duality: ρ = .09, SDρ = .03, p < .01; see Table 1). Narcissistic CEOs gain power and prestige in this dual role, which may enable the CEO’s narcissistic behavior. CEO succession literature would support such a view, as CEOs with more power are less likely to be dismissed in times of poor performance (Boeker, 1992). Future research might examine how CEO duality could facilitate narcissistic behavior among CEOs.
Situational factors may also lead a CEO to behave narcissistically, even when the CEO does not demonstrate high levels of narcissism under normal circumstances. This acquired situational narcissism may affect CEOs in ways similar to the way that it is theorized to affect celebrities, athletes, and politicians. Robert B. Millman coined the term acquired situational narcissism, in which movie stars, politicians, and professional athletes develop narcissistic traits as a result of their fame (Sherril, 2001). In other words, while narcissism may lead people to seek out fame, it can also develop after people achieve success and prolonged concentrated attention (Gentile, 2011). This pattern could occur in CEOs who achieve high levels of wealth, power, prestige, or celebrity. Future research might identify situations that cause or reduce the development of situational narcissism.
CEO attributes as moderators
Other characteristics of a CEO could influence the manifestation of narcissism. Evidence suggests that, at times, narcissists demonstrate counterintuitive behavior, such as humility. Zhang et al. (2017) found that humility and narcissism are not opposite ends of the same construct but rather orthogonal constructs wherein a more narcissistic CEO can also behave humbly. Other constructs that have potential to magnify or diminish narcissistic behavior include CEO celebrity (Hayward, Rindova, & Pollock, 2004), CEO materialism (Davidson, Dey, & Smith, 2018), entrepreneurial orientation (Wales et al., 2013), age (Marquez-Illescas et al., 2018), and gender (Ingersoll et al., 2017).
First, CEO celebrity—when journalists publicly attribute a firm’s positive performance to its CEO’s actions (Hayward et al., 2004), it could encourage narcissistic behavior due to the external attention to the CEO. Second, CEO materialism—the CEO’s number of luxury assets (Davidson et al., 2018) is derived from the psychological concept of materialism, wherein an individual displays an attachment to worldly possessions and material needs and desires (Richins & Rudmin, 1994). Narcissistic CEOs high in materialism may give additional emphasis to pay and incentives. Third, entrepreneurial orientation occurs when a CEO engages in product-market innovation, undertakes somewhat risky ventures, and is first to come up with “proactive” innovations, thereby beating competitors to the punch (Wales et al., 2013). While narcissists who are high in entrepreneurial orientation may fuel their narcissistic esteem by developing new products, those low in entrepreneurial orientation may spend more energy protecting existing product portfolios. Fourth, tempering of narcissistic behavior may simply occur with the wisdom of experience. For example, narcissistic behaviors are often considered less pronounced with age (Foster, Campbell, & Twenge, 2003). Marquez-Illescas et al. (2018) found that the overly positive tone of earnings announcements by narcissistic CEOs was moderated by the CEOs’ age. They also theorized that with age, CEOs become more content and humble, which is consistent with findings that CEOs can evince narcissistic and humble behaviors simultaneously (Zhang et al., 2017). Future research might explore other situational factors that moderate narcissistic behavior.
Finally, an important caveat is that studied CEOs are predominantly male (Zarya, 2016), and gender should be accounted for when interpreting findings. In general, men are more likely to be narcissistic (Twenge & Campbell, 2009), which is also the case for CEOs (Ingersoll et al., 2017). Taking gender into account is important because personality traits, including narcissism, are perceived differently for males and females (Brescoll, 2016). Indeed, Ingersoll et al. (2017) found gender to be a moderator between narcissism and risk-taking and questionable behaviors such that narcissistic female CEOs engage in less risk-taking and questionable behaviors than do narcissistic male CEOs.
What Are the Consequences of CEO Narcissism?
Our review has emphasized that CEO narcissism can result in both beneficial and detrimental organizational outcomes with short- and long-term consequences. Indeed, it is perhaps more representative to say that CEO narcissism has multidirectional effects that largely depend on the context and situational features that enhance, repress, and activate narcissistic behaviors. Therefore, beyond the current set of outcomes detailed so far, there is a wide set of potential consequences and moderators of narcissistic CEO behavior that have yet to be investigated. We categorize these as internal outcomes, external outcomes, top management team (TMT) moderators, and mitigating effects of CEO narcissism, which we address in turn.
Internal outcomes
Internal outcomes are those with impact primarily inside the firm. For instance, narcissistic CEOs may have different impacts on organizational culture than nonnarcissistic CEOs. While CEO values can influence organizational culture (Berson, Oreg, & Dvir, 2008), cultural outcomes as a consequence of CEO narcissism could include the adoption of narcissistic norms, leadership, and communication styles throughout the organization. CEOs also influence the CEO succession process (Finkelstein et al., 2009); therefore, CEO narcissism may affect CEO dismissal and departure, the identification and development of CEO candidates, and the probability of successive narcissistic CEOs. CEO narcissism may also affect the processes and structure of the organization and thus affect feedback and reward systems as well as the organizational structure and workforce composition. We believe that many opportunities exist to improve understanding of how CEO narcissism affects intermediate organizational outcomes and choices as well as decision-making policies and practices.
It has also been suggested that CEOs staff their TMTs in such a manner as to enable their narcissistic esteem needs (Chatterjee & Pollock, 2017). The implication is that once the narcissistic CEO feels that the TMT is successfully arranged, the TMT becomes an extension of the narcissistic CEO rather than an independent problem-solving entity. Zhu and Chen (2014) provide evidence that narcissistic CEOs are associated with the appointment of narcissistic directors on the board. Thus, narcissistic CEOs seem to surround themselves with fellow narcissists. Other effects that narcissists may have on the TMT include levels of communication and cohesiveness. TMT and board settings provide possible research opportunities to understand the day-to-day inner workings and management styles of narcissistic CEOs with governance practices that enable or mitigate narcissistic behavior.
External outcomes
Other outcomes affect the firm externally rather than internally. Potential external outcomes for investigation include mergers and acquisitions, divestitures, growth strategies, international expansion, product innovation, and brand image. As narcissistic CEOs have shown a tendency to manipulate financial earnings (Capalbo et al., 2017), other areas of financial reporting may be affected, such as financial reporting choices, capital structures, and risk management. In addition to what we know about narcissistic CEOs’ questionable behavior (e.g., fraud; Rijsenbilt & Commandeur, 2013), additional outcomes could include product recalls, environmental pollution, and lobbying expenses.
TMT as a moderator
Several TMT-related factors may moderate the effects of CEO narcissism on internal and external organizational outcomes. For example, if the relationship between TMT narcissism and CEO narcissism is independent, there could be a moderating effect of TMT narcissism when it comes to CEO narcissism and performance. In one nonexecutive study, Grijalva, Maynes, Badura, and Whiting (2020) found that NBA teams with higher team levels of narcissism and key players with higher levels of narcissism exhibited poorer organizational performance.
Other TMT attributes could play a role in how the CEO’s narcissism affects organizational outcomes. A TMT’s ability to processes information (as measured by firm size) has been linked to a reduced CEO information-processing load (Henderson & Fredrickson, 1996). Similarly, TMT heterogeneity in terms of education, work experience, or tenure (Carpenter, 2002) may create the same level of capability. Perhaps TMTs with high information-processing capabilities or heterogeneity are better able to handle the demands placed on them by a narcissistic CEO.
Mitigating the effects of CEO narcissism
It is possible that the board or other stakeholders could mitigate the negative consequences of narcissistic CEO behavior. For example, increased levels of oversight, through either administrative committees or increased auditing, could increase awareness of the CEO’s behavior and provide quicker governance reactions, if needed. Other possible avenues to mitigating negative consequences of narcissistic CEO behavior could include but are not limited to the following: removing board responsibilities from the CEO, increasing board involvement in succession planning, and engaging in public relations activity. Future research could examine how effectively these actions and others could mitigate negative behaviors from narcissistic CEOs. Ultimately, if a board can neither prevent narcissistic CEO behavior from manifesting nor mitigate negative consequences of CEO narcissism after the fact, it may need to dismiss the CEO. Future research could identify when such a step is necessary. This line of research could extend into examining how a new CEO is affected in the wake of taking over for a narcissistic CEO.
Summary of Key Questions and Future Directions
The primary focus of CEO narcissism research has been on its immediate strategic consequences. We advocate expanding the investigation into areas of narcissistic CEO supply, demand, and behavior manifestation. The supply of narcissistic CEO candidates may result from some combination of narcissistic drive, career path, and organizational culture. The demand to select a narcissistic CEO may result from a combination of board biases and market pressures. After a narcissistic CEO has been selected, his or her behavior may be more or less manifest under certain situations. While factors such as the opportunity to be the center of attention, increased power, prolonged exposure to concentrated attention, or CEO celebrity may amplify narcissistic behavior, other factors, such as age, may dampen it.
Although strategic consequences have received the most research attention, there are still more strategic consequences to investigate. Internal to the firm are consequences that affect the TMT, organizational culture, succession plans, and policies, among others. External to the firm is a long list of consequences, ranging from mergers and acquisitions to product recalls and financial earnings. Between the narcissistic CEO and downstream consequences is the TMT, who must interpret and execute the CEO’s instructions. The TMT is not only directly affected by CEO behavior but also serves as a buffer to the rest of the organization. Finally, it is possible that the board or other stakeholders may try to mitigate narcissistic CEO behavior by directly intervening in CEO responsibilities or through public relations activity.
Theoretical Basis of CEO Narcissism
The majority of CEO narcissism research is explored through the lens of upper echelons theory and leadership theory. Upper echelons theory is a logical and appropriate framing for CEO narcissism research because it connects the CEO’s motivations and attributes with organizational outcomes (Hambrick & Mason, 1984). While upper echelons theory has yielded fruitful research, it does not necessarily provide an impetus to delve deeper into the underlying mechanisms and processes that drive narcissistic behavior. Furthermore, it may not provide a framework with which to examine potential mitigating factors that might moderate narcissists’ behavior. For example, considering the person, the position, and the environment could help extend upper echelons theory to examine antecedents and moderating factors of CEO narcissism (i.e., the contingency approach; Busenbark, Krause, Boivie, & Graffin, 2016).
Leadership theory (i.e., leader effectiveness due to individual differences; Hoffman, Woehr, Maldagen-Youngjohn, & Lyons, 2011; Lord, Foti, & De Vader, 1984) is useful because it can help clarify similar yet differentiated constructs of narcissism. For example, the leadership lens is used to tease out the relationship between a narcissist’s bold vision and the attribution of charismatic leadership (Galvin, Waldman, & Balthazard, 2010). It is within leadership theory that the differences between such mirrored attributes are articulated and their impacts tested (e.g., Resick et al., 2009). Leadership theory can also serve as a natural starting place to examine not only how narcissists affect their followers but why those followers choose to follow and support a narcissist.
As an alternative, researchers could draw on theoretical lenses from personality theory to dig deeper into the underlying psychology and motivations of narcissists and then apply those theories to CEOs. While these are not the only relevant theories, the agency model of narcissism and the admiration-versus-rivalry perspective of narcissism in particular may provide deeper insight into the motivations of narcissistic CEOs. In addition, tournament theory, an outgrowth of economics literature, may be suited to studying the motivations and strategies used by narcissists as they climb the organizational hierarchy.
The Extended Agency Model of Narcissism
The extended agency model of narcissism is a composite model of the theoretical development and empirical evidence within the narcissism literature (Campbell & Foster, 2007). The extended agency model of narcissism posits that individuals with higher levels of narcissism have a strong need to maintain a positive self-image and think and act in ways to keep these self-views viable. The model adopts the premise that behavior is motivated by a dynamic self-regulatory process, which encompasses a person’s efforts to construct, maintain, defend, and enhance one’s desired self-views (Morf & Rhodewalt, 2001). In this context, the self is seen as a network of cognitive-affective processes in constant transaction with the social environment (Morf & Horvath, 2007). The primary output of such dynamic self-regulatory processes is narcissistic esteem, a sense of self-esteem linked primarily to dominance, rather than closeness or acceptance, and related to the emotion of pride (Campbell & Foster, 2007).
The dynamic self-regulatory process that feeds a narcissist’s narcissistic esteem comprises four mutually reinforcing elements. The first element is a narcissist’s interpersonal skills (e.g., social confidence, charm), which can be used to garner attention or influence. The second element is a narcissist’s intrapsychic self-regulation strategies (e.g., fantasies of power, self-serving bias), which can be used to justify self-serving behavior or outcomes. The third element is a narcissist’s interpersonal strategies (e.g., self-promotion, game-playing), which can be used to achieve goals or control others. The fourth element is a narcissist’s fundamental qualities, those qualities that describe a narcissist’s underlying motivations—agent-versus-communal concern, approach-versus-avoidance orientation, general desire for self-esteem, sense of entitlement, and inflated self-view (Campbell & Foster, 2007). These four elements work together to create a narcissist’s self-regulatory system.
The advantage of the extended agency model of narcissism is that it helps explain the mechanisms and strategies that narcissists use to satisfy their narcissistic esteem. This approach exposes interpersonal skills, intrapsychic strategies, and interpersonal strategies, as well as configurations of all three as mediators between narcissistic motivations and outcomes. Using the extended agency model of narcissism allows future research to consider configurations of motivations and strategies and their results, as opposed to examining just one or two key motivations. This may further expose how CEOs with similarly high levels of narcissism may approach situations differently and thereby experience differing outcomes.
Admiration-Versus-Rivalry Perspective of Narcissism
As an alternative to the extended agency model of narcissism (Campbell & Foster, 2007), Back et al. (2013) introduced the rivalry-versus-admiration perspective of narcissism. This theoretical lens suggests two distinct paths for the manifestation of narcissistic behaviors. The first path, narcissistic admiration, centers on a narcissist’s assertive orientations, such as social potency based on a grandiose self-image and charming, self-assured behaviors. The second path, narcissistic rivalry, centers on antagonistic orientations, such as social conflict where narcissists devalue others, or hostile aggressive behaviors. Narcissistic admiration consists of striving for uniqueness, grandiose fantasies, and charm, whereas narcissistic rivalry consists of striving for supremacy, devaluation of others, and aggressive behavior.
The admiration-versus-rivalry perspective has potential to improve understanding of CEO affect, cognition, and behavior by teasing out specific narcissistic behaviors. Observable behaviors, such as charisma, competitiveness, and confidence, are associated with successful CEOs. Focusing on narcissistic admiration or narcissistic rivalry may provide theoretical mechanisms and constructs to further understand how and why a narcissistic CEO affects strategic outcomes. Another advantage of the admiration-versus-rivalry perspective of narcissism is that it has a validated measure (Back et al., 2013).
Tournament Theory
Tournament theory revolves around the idea that individuals compete for prizes that are designed to incentivize optimal effort and are awarded according to the person’s relative rank (Becker & Huselid, 1992). Narcissists are predisposed to see life as a series of contests (Wallace & Baumeister, 2002) and thus may be more adept at competing in tournaments. The intrinsic satisfaction of competing and the extrinsic reward of status and money are motivating factors to narcissists (Ryan & Deci, 2000; Reidy, Zeichner, Foster, & Martinez, 2008; Twenge, Campbell, Hoffman, & Lance, 2010). Therefore, narcissists often navigate corporate hierarchies to successfully attain executive positions (Brunell et al., 2008; Lubit, 2002; Maccoby, 2007; Rosenthal & Pittinsky, 2006).
Tournament theory could inform how a narcissist rises to the position of CEO and is selected. Researchers often examine the selection of CEOs from the perspective of which candidate is best or optimal (Cragun, Nyberg, & Wright, 2016). However, from a tournament perspective, the defining factor is not about being good but good enough to win the tournament. As certain shareholders appear to favor tournament-style compensation structures (Connelly, Tihanyi, Crook, & Gangloff, 2013), it is reasonable to hypothesize that narcissistic CEOs would emerge from such compensation structures that are more tournament based.
Tournament theory could help investigate possible changes in narcissistic CEO behavior after selection. Before selection, the narcissist competes primarily against other firm executives, whereas after selection, the CEO primarily competes against the performance of other CEOs in the market. Thus, individuals who were once peers and competitors are now subordinates, while board members and CEOs from other companies become peers. As a result, different self-regulation strategies could be used to achieve new objectives, resulting in new behaviors toward shifting peer groups. This could reconcile the conflicting evidence that while narcissism helps individuals rise to the top (Grijalva et al., 2015), a majority of CEOs do not demonstrate high levels of narcissism (Wright, Cragun, et al., 2016).
Summary of Theoretical Basis
From a theoretical perspective, upper echelons theory and leadership theory have historically been the primary lens for investigating CEO narcissism. Future studies based on upper echelons theory may benefit from using a more configural approach (Busenbark et al., 2016). Additional theoretical perspectives that could increase understanding include the extended agency model of narcissism (Campbell & Foster, 2007), the admiration-versus-rivalry perspective of narcissism (Back et al., 2013), and tournament theory (Becker & Huselid, 1992). For a summary of future research questions, see Table 3.
Summary of Future Directions for CEO Narcissism Research
Conclusion
CEO narcissism continues to capture the attention of scholars and practitioners. Despite the move from its nascent stages into a more systematic research stream, there are still opportunities for foundational CEO narcissism research. Unobtrusive measures of narcissism still leave room for innovation and maturity; theoretical approaches beyond upper echelons and leadership theory could be explored; and research outside of CEO narcissism’s immediate strategic consequences has not yet begun in earnest. While we encourage researchers to continue to explore all aspects of CEO narcissism, we note that while narcissism is neither inherently good nor bad, unchecked narcissism can cause great harm. Therefore, we encourage scholars to set a high priority for research efforts that seek to better understand the proper controls and mitigations necessary to avoid CEO narcissism’s potential downsides.
Footnotes
Appendix
Robustness Tests for Meta-Analytic Results of CEO Narcissism and Firm Attributes and Outcomes: Totals
| Correlation relationship | Studies, n | Correlations, n | CEOs, n | Mean ρ | SD ρ | LL | UL | Q | Variance, % |
|---|---|---|---|---|---|---|---|---|---|
| Firm size | 18 | 22 | 6,948 | .09*** | .02 | .04 | .13 | 39.46** | 44 |
| Financial performance | 12 | 13 | 5,629 | .07*** | .02 | .03 | .11 | 20.53* | 57 |
| Innovation and growth | 12 | 16 | 5,264 | .09*** | .02 | .05 | .13 | 20.95* | 56 |
| Risk taking | 4 | 4 | 2,885 | .16** | .05 | .06 | .25 | 23.57*** | 16 |
| Financial leverage | 4 | 5 | 1,011 | .04 | .05 | −.06 | .13 | 8.96* | 44 |
| CEO duality | 6 | 6 | 790 | .12** | .04 | .05 | .19 | 4.60 | 100 |
Note: Mean ρ = average corrected correlation; SDρ standard deviation of corrected correlation; LL = the lower limit of the 95% confidence interval; UL = the upper limit of the 95% confidence interval; Q = Q statistic to test for homogeneity in the corrected correlations across studies; variance = percentage of the variance in correlations explained by statistical artifacts.
p < .05.
p < .01.
p < .001.
Acknowledgements
We thank Sali Li for his encouragement, Donald J. Schepker for his help in formulating our proposal, and Gina C. Grensing for her help in developing this manuscript. We also thank Gerard P. Hodgkinson, the action editor, and two anonymous reviewers for their thoughtful comments and suggestions.
