Chat with us, powered by LiveChat In this assignment, you are to critically - Study Help

 In this assignment, you are to critically read and evaluate a scholarly article’s strengths, weaknesses, and contributions to the study field. Learning how to critique a journal article has several benefits, including preparing you for publishing in the future and keeping you current on the literature in your field of study. The practical application is developing the ability to look at research within your organization and industry with a knowledgeable, critical eye. 

 Gupta, V. K., Mortal, S. C., Silveri, S., Sun, M., & Turban, D. B. (2020). You’re fired! Gender disparities in CEO dismissal. Journal of Management, 46(4), 560-582. 

 Following your review, choose one article from this list; critically evaluate the article’s strengths, weaknesses, and contribution to the study field using the outline below as a guide:

 Cover page

  • The cover page will include:
    • Articles Title and Author (s)
    • Name of Journal (s)
    • Date of publication
    • Your name

Executive Summary

  • Summarize the significant aspects of the entire article, including:
    • The overall purpose and general area of study of the article.
    • The specific problem being addressed in the study.
    • The main findings of the article.

Literature Review

  • Briefly summarize the overall themes presented in the Literature Review.
    • Was the literature review applicable to the study, current and thorough?
    • Were there gaps in the literature review?

Data Analysis

  • Identify the methodology used: qualitative, quantitative, mixed? Was the chosen methodology appropriate for the study? Why or why not?
  • Did the data analysis prove or disprove the research questions? Explain.


  • In this section, you will address the following:
    • Describe the article’s relevance to the field of knowledge.
    • Outline the strengths and weaknesses of the article. Be specific.
    • Based on the article, what future research do you think needs to be accomplished in this area?
    • What are your key points and takeaways after analyzing the article?

Proper APA in-text citation must be used. The review is to be word-processed double spaced, not less than two pages, and no more than five pages in length. Paper length does not include the cover page, abstract, or references page(s). 


Journal of Management
Vol. 46 No. 4, April 2020 560 –582
DOI: 10.1177/0149206318810415

© The Author(s) 2018
Article reuse guidelines:

You’re Fired! Gender Disparities
in CEO Dismissal

Vishal K. Gupta
Sandra C. Mortal
The University of Alabama

Sabatino Silveri
University of Memphis

Minxing Sun
Clemson University

Daniel B. Turban
University of Missouri

CEO dismissals attract considerable attention, presumably because of the visibility, publicity,
and intrigue that often surrounds the decision to fire the CEO. With the goal of advancing schol-
arly understanding of CEO dismissals, we examine whether CEO gender influences the likeli-
hood of dismissal. We theorize and find that ceteris paribus, female CEOs are significantly more
likely to be dismissed than male CEOs. Perhaps even more importantly, we find a CEO gender
by firm performance interaction such that male CEOs are less likely to be dismissed when firm
performance is high (compared to when it is low), whereas female CEOs have a similar level of
dismissal likelihood regardless of firm performance. Notably, our results are robust to multiple
analytical techniques and various econometric specifications, bringing greater credence to the
validity of our findings. Implications and directions for future research are also discussed.

Keywords: CEO dismissal; gender; firm performance

Acknowledgments: Previous versions of this research were presented at the Southern Management Association
Conference and discussed at the Indian Institute of Management at Shillong, where we benefited from good feed-
back and constructive comments. Conversations with Alka Gupta, Junsoo Lee, and Steven Michael were useful in
the development of ideas discussed here. Erik Markin and Joshua White provided editorial assistance at different
stages during the writing of this article. We are very grateful for the helpful insights and suggestions from action
editor Karen Schnatterly and three anonymous reviewers, all of whom played an important role in strengthening
this manuscript. Of course, all omissions and errors remain our own.

Corresponding author: Vishal K. Gupta, The University of Alabama, Tuscaloosa, AL 35487, USA.

E-mail: [email protected]

810415 JOMXXX10.1177/0149206318810415Journal of ManagementGupta et al.

Gupta et al. / Gender and CEO Dismissal 561

The increasing presence of women in chief executive roles motivates considerable interest
in understanding what happens to women after they reach the CEO position (Oliver, Krause,
Busenbark, & Kalm, 2018; Zhang & Qu, 2016). Some scholars suggest that women receive
preferential treatment compared to men for having reached previously inaccessible leader-
ship roles (“female leadership advantage” logic; Underdahl, Walker, & Woehr, 2014). Others
contend that women continue to be disadvantaged even after they attain the highest position
in the organizational hierarchy (Glass & Cook, 2016). Specifically, using the metaphor of the
“glass cliff,” researchers argue that women in leadership positions face more perils and risks
compared to their male counterparts (Ryan & Haslam, 2007). The bias against women in
leadership roles is believed to be rooted in widespread stereotypical beliefs that associate the
characteristics needed for success as a leader with men but not with women (“think manager–
think male” effect; Eagly & Karau, 2002; Schein, 2001). With the goal of further advancing
this line of inquiry, and providing a strong test of the greater precariousness of women’s
leadership position vis-à-vis men (Ryan & Haslam, 2007), we investigate whether female
CEOs face differential dismissal risk compared to male CEOs.

CEO dismissal refers to the forced departure of the chief executive from the firm (Haleblian
& Rajagopalan, 2006; Huson, Parrino, & Starks, 2001). Three decades ago, Fredrickson,
Hambrick, and Baumrin defined CEO dismissal as “a situation in which the CEO’s departure
is ad-hoc (e.g., not part of mandatory retirement policy) and against his or her will” (1988:
255), a definition that continues to resonate with researchers in this area. While CEOs depart
from their firms for many reasons, dismissal—sometimes also referred to as forced turnover
(Farrell & Whidbee, 2002) and involuntary exit (Alexander, Fennell, & Halpern, 1993)—has
long been considered the most theoretically interesting form of CEO departure (Finkelstein,
Hambrick, & Cannella, 2009). Scholars posit, with considerable supporting evidence, that
firm performance is the primary metric by which CEOs are assessed and, thus, an important
predictor of CEO dismissal (Hilger, Mankel, & Richter, 2013). Consequently, researchers
generally view CEO dismissal as “one of the main corporate governance instruments”
(Fiordelisi & Ricci, 2014: 66), so that the threat of dismissal is considered a powerful tool to
pressure managers to lead their firms better and pursue value-enhancing policies (Lehn &
Zhao, 2006).

Unexpected departures of CEOs from public corporations receive considerable media
scrutiny (Li, Lu, Makino, & Lau, 2017). The growing number of women in the C-suite,
coupled with the current zeitgeist of gender equality in society (Sandberg, 2013), motivates
interest in possible gender differences in CEO dismissal. Contemporary media articles con-
tend that female CEOs face greater threat of dismissal compared to male CEOs because the
former get blamed disproportionately more for the problems and issues facing the company
(Leung, 2014; Reingold, 2016). A recent PwC investigation noted that among CEOs leaving
office in large public corporations between 2003 and 2013, 38% of women were forced out
compared to only 27% of men (Favaro, Karlsson, & Nielson, 2014). While these reports sug-
gest gender biases in CEO dismissal, it is worth remembering that media accounts are “not
held to high methodological and peer review standards that academic work is subject to” (R.
B. Adams, 2016: 373) and, thus, do not advance ongoing discussions about challenges faced
by female leaders in a rigorous fashion (Ryan, Haslam, Morgenroth, Rink, Stoker, & Peters,
2016). The CEO position is the highest corporate leadership role (Zhang & Qu, 2016), and
CEO dismissals are often clouded in controversy and confusion, so that systematic research

562 Journal of Management / April 2020

is needed to examine whether there are gender differences in CEO dismissals. In particular,
it is important to know, using rigorous statistical tools, whether CEO dismissals actually
show significant gender bias, which may inadvertently perpetuate social stereotypes suggest-
ing that men, but not women, have the attributes required for successful leadership.

On the basis of 641 dismissals from 2000 to 2014, we investigate whether female CEOs
are more likely to be dismissed than male CEOs and whether firm performance is differen-
tially related to dismissal for male and female CEOs. As such, we examine and illuminate
the gendered nature of CEO dismissal, an important issue that is a matter of scholarly inter-
est and also has substantial practical implications for sound corporate governance (Hilger
et al., 2013). Our research extends prior work that examines whether female, compared to
male, executives face greater challenges and threats (Ryan & Haslam, 2007). Rigorous
scholarship on possible gender bias at the apex of the organizational hierarchy can provide
valuable insights into the similarities and differences in career derailment between high-
potential men and women (Krishnan, 2009; J. B. Leslie & Van Velsor, 1996). More broadly,
we examine the provocative claim that the rise of women to top executive positions in large
corporations suggests that gender is no longer a relevant factor in one’s career trajectory
(Elsesser, 2016), presumably because gender bias is adequately addressed by corporate
policies and procedures that seek to level the playing field for men and women (Barrett &
Morris, 1993; Landy, 2008).

Theory and Hypotheses

Considerable evidence indicates that gender-role stereotypes, socially shared expectations
about the characteristics and behaviors of men and women (Eagly & Mladinic, 1989), have
limited the ascension of women to the highest level in organizations (Heilman, 2001). Gender
stereotypes typically associate men with agentic characteristics, which capture achievement-
oriented tendencies (e.g., aggressive), whereas women are associated with communal attri-
butes, which entail concern for the welfare of others (e.g., caring; Haines, Deaux, & Lofaro,
2016). Because leadership roles are often described in masculine (agentic) terms (Schein,
2001), and women are viewed as deficient in such qualities (Heilman, 1983), gender stereo-
types constrain women’s advancement to leadership positions (Koenig, Eagly, Mitchell, &
Ristikari, 2011).

Despite substantial barriers, some women have ascended into top leadership ranks in the
corporate world, including at very prominent firms such as GM (Mary Barra), Pepsico (Indra
Nooyi), and IBM (Virginia Rometty). Unfortunately, however, there is limited understanding
of whether men and women have different experiences when they occupy similar leadership
roles. Ryan and Haslam (2005) introduced the idea of the glass cliff to highlight that women
who break through the “glass ceiling” will find themselves in perilous situations as they will
be promoted to higher risk leadership positions and, thus, face more difficulties once they are
in leadership positions. The glass cliff metaphor captures the precariousness of such posi-
tions due to invisible dangers of falling from the heights of corporate leadership (Ryan,
Haslam, & Kulich, 2010). Evidence is mixed, however, regarding whether women are more
likely to be promoted to risky top leadership positions, with some studies finding supporting
results (e.g., Cook & Glass, 2014a; Ryan, Haslam, Hersby, & Bongiorno, 2011), while others
do not (S. M. Adams, Gupta, & Leeth, 2009; Cook & Glass, 2014b). However, the equally

Gupta et al. / Gender and CEO Dismissal 563

important question of whether men and women are treated differently in comparable leader-
ship positions has largely gone unanswered. As Oliver et al. observed, “Though research has
focused on the ascent and acceptance of female CEOs, the post-promotion circumstances
female CEOs face remain unclear” (2018: 113). This is a notable omission because from an
equal opportunity perspective, what happens to women in leadership positions is at least as
critical as the conditions under which they made it to these positions (Zhang & Qu, 2016).
Thus, we examine the important research question of whether women who reach the CEO
position are treated differently than men in such positions.

Two well-regarded conceptual frameworks—token theory (Kanter, 1977) and role con-
gruity logic (Eagly & Karau, 2002)—cast light on the potential challenges that female lead-
ers, but not male leaders, face after they make it to coveted leadership positions like the CEO
role. Token theory, sometimes labeled “critical mass theory” (Bratton, 2005), suggests that
numerical minorities (such as female CEOs) often experience enhanced visibility and atten-
tion, exaggerated stereotypes, and heightened monitoring and scrutiny (E. B. King, Hebl,
George, & Matusik, 2010). Furthermore, role congruity theory suggests that cultural stereo-
types associating leadership with masculinity can undermine evaluations of women’s com-
petence and ability to lead (Kark & Eagly, 2010). Taken together, token theory and role
congruity theory suggest that female leaders are more likely to be the target of others’ unfa-
vorable perceptions about their ability and competence. Consistent with this, investor reac-
tion to announcements of female CEO appointments is significantly more unfavorable than
of male CEO appointments, presumably because female CEOs are perceived as less compe-
tent than male CEOs (Lee & James, 2007). Additionally, research, using an experimental
design, finds that firms having an initial public offering are perceived as less likely to suc-
ceed when a woman, rather than a man, is at the helm, which then lowers investments directed
at female-led firms compared to male-led firms (Bigelow, Lundmark, Parks, & Wuebker,

However, the greater salience of female CEOs may also provide them with certain bene-
fits (L. M. Leslie, Manchester, & Dahm, 2017). Some scholars contend that there is a “female
advantage” for women who make it to the top leadership positions (Yukl & Chavez, 2002).
Gender stereotypes describe women as more skilled at inclusiveness, interpersonal relations,
power sharing, and nurturing others—characteristics considered essential for leadership in
modern organizations—and as a consequence, women should be superior leaders (Grant,
1988; Loden, 1985). In this vein, Rosette and Tost argue that “because a feminized approach
to managing others is increasingly viewed as a strength” (2010: 222), female leaders are seen
as more valuable for the firm than male leaders. Furthermore, other scholars argue that
because women surmount more barriers than men on their way up, women who make it to
the CEO position may be “particularly gifted and/or especially good at learning and/or deal-
ing with adversity” (Gupta, Mortal, & Guo, 2018: 2039). For these reasons, some scholars
suggest a “female leadership advantage” exists for women who make it to the CEO position
(Eagly, 2007), a sentiment that also resonates well with the popular press and the mass media
(Folkman, 2012).

The literatures reviewed above offer competing predictions for men and women in CEO
positions. Empirical research also paints a mixed picture. Some studies find that female
CEOs have a lower likelihood of departing from the firm than male CEOs (Elsaid & Ursel,
2018), whereas other evidence suggests that female CEOs have shorter median tenure than

564 Journal of Management / April 2020

male CEOs (Glass & Cook, 2016). Importantly, however, none of these studies distinguished
between dismissals and voluntary exits, thus failing to offer a direct test of whether there are
gender differences in CEO dismissal. From the perspective of token theory and role congru-
ity theory, female CEOs are seen as having attributes inconsistent with leadership roles,
which causes female CEOs to be scrutinized more critically and evaluated more negatively
than male CEOs. The growing literature on the glass cliff phenomena also recognizes that
“women managers tend to receive greater scrutiny and criticism than men, and they tend to
be evaluated less favorably, even when performing exactly the same leadership roles as men”
(Ryan & Haslam, 2007: 550). However, from the “female advantage” perspective, female
CEOs are seen as more competent than male CEOs, which might be associated with lower
levels of dismissal. We theorize, however, that the heightened scrutiny of female CEOs
increases the salience of social expectations about what men and women usually do or ideally
should do (Rudman & Glick, 2001), which generates less favorable evaluation of women’s
leadership compared to men’s. Furthermore, the perceived cultural mismatch between
women and demands of leadership roles is “likely to be the most extreme at the highest levels
of leadership” (Eagly & Karau, 2002: 577), such that women “often seem inappropriate or
presumptuous” when they occupy senior leadership positions (Wynen, Beeck, & Ruebens,
2015: 379). In effect, we expect the lack of fit between the feminine stereotype and the CEO
role causes female managers to be scrutinized more heavily and criticized more often.
Consequently, all else being equal, we expect women will seem less able and competent than
men to effectively lead the firm, even when they already occupy the CEO position.

Thus, based on the conceptual logic articulated above, we hypothesize that:

Hypothesis 1: The likelihood of dismissal will be higher for female CEOs than for male CEOs.

Much of the work on CEO dismissal is grounded in agency theory (Crossland & Chen,
2013), which posits that when firms do not perform well, an effective internal governance
practice is to sanction the CEO (Kato & Long, 2006). For researchers working in this vein
(e.g., Murphy, 1999), poor firm performance is an ipso facto case of poor CEO performance.
Consistent with this idea, firm financial performance is considered the primary predictor of
CEO dismissal (Hubbard, Christensen, & Graffin, 2017). As Fredrickson et al. note, the
“most obvious answer” to the question of why CEOs get dismissed, and “one for which there
is [considerable] empirical support” (1988: 255), is that the firm is not performing well and
is not expected to perform well in the future. While much research in this area focuses on the
role of the performance of the firm itself (Hilger et al., 2013), some scholars argue that the
board’s expectation about future performance—based on comparisons with industry and
market performance—is a critical factor in CEO dismissal (Haleblian & Rajagopalan, 2006).

Although it is unsurprising that poor firm performance is considered a key determinant of
CEO dismissal (Wiersema & Zhang, 2011), some research suggests that CEOs get fired even
when firm performance is good (Martin & Combs, 2011). Ertugrul and Krishnan (2011), for
example, find that approximately half the dismissed CEOs in their sample were forced out
when the firm was performing well. When a firm is performing poorly, implying that the
CEO is not leading the firm properly, there may be a normative expectation that the board
should take action and dismiss the CEO so that a new leader can take the firm in a different
direction (Hilger et al., 2013). Indeed, the behavioral theory of the firm proposes, with

Gupta et al. / Gender and CEO Dismissal 565

considerable supporting evidence, that when firm performance is below expectations, based
on the firm’s prior performance and/or competitors’ performance, organizational changes are
likely (Argote & Greve, 2007). Thus, when firms are performing poorly, we expect that an
appropriate action is to dismiss the CEO, and in making such a decision, the board demon-
strates its vigilance in exercising its statutory monitoring function. However, when the firm
is performing well, there is considerable ambiguity about the CEO’s leadership of the firm
and no clear script for the board to follow. In such situations we expect that gender-role ste-
reotypes will influence dismissal decisions.

Considerable evidence indicates that stereotypes have greater influence in more ambiguous
situations (Heilman, 1997; Heilman & Haynes, 2008). Therefore, we expect gender-role ste-
reotypes, and biases associated with female leaders, to have more influence when firms are
performing well because of the inherent uncertainty and ambiguity in connecting CEOs’ deci-
sions and actions with specific firm outcomes (Auster & Prasad, 2016). Because of their token
status and role incongruity (discussed earlier), female CEOs are more likely to be perceived as
having less ability and competence than male CEOs. Indeed, Ertugrul and Krishnan (2011)
observed that CEOs in well-performing firms were more likely to be dismissed when they are
perceived as deficient in ability or competence. By extension, we argue that because of ambi-
guity and vagueness in determining the CEO’s contributions to firm performance, CEO gen-
der will influence the extent to which the CEO is given credit for good firm performance. We
contend that when the firm is performing well, female CEOs are less likely than male CEOs
to be considered good stewards of the firm and therefore are more likely to be dismissed.

In summary, we expect that CEO gender will affect performance-dismissal sensitivity,
such that dismissal will be more sensitive to performance—both absolute and relative—for
male CEOs compared to female CEOs. More specifically, we expect no gender differences
in dismissal for poorly performing firms, as we expect boards to feel pressured to take action
regardless of CEO gender. However, we expect that as firm performance improves, the prob-
ability of dismissal for male CEOs decreases but does not protect female CEOs from dis-
missal. Thus, we hypothesize:

Hypothesis 2: CEO gender will moderate the performance-dismissal relationship, such that perfor-
mance will be more strongly related to dismissal for male versus female CEOs. More specifi-
cally, we expect that male and female CEOs are equally likely to be fired in underperforming
firms, but female CEOs are more likely to be fired than male CEOs in well-performing firms.



Large public firms in the United States compose the sample for this study. We start with
the ExecuComp database, which contains information on top executives for large firms. We
obtain board variable data from BoardEx, which has consistent coverage starting in 2000.
Accounting information is from the Compustat Industrial Annual files, and stock return
information is from the monthly CRSP tapes. Our final sample comprises 21,772 firm-year
observations spanning 2,390 unique firms from 2000 to 2014. CEOs who have been in office
less than 1 year are not included in our sample because dismissal is highly unlikely within the
1st year.

566 Journal of Management / April 2020


Our dependent variable, CEO Dismissal, equals 1 if the CEO is fired and 0 if there is no
turnover.1 Identifying CEO dismissals is difficult because firms rarely explicitly state that the
departing CEO is being fired (Dedman & Lin, 2002; Shen & Cannella, 2002). Therefore, we
follow Parrino (1997) to identify CEO dismissal by combining content analysis of public
news reports with information about CEO age and continued affiliation with the firm.
Although this approach is labor intensive, it has been employed in recent studies because it
is considered quite effective in capturing CEO dismissal (Jenter & Kanaan, 2015; Zhang,
2008). On the basis of a Factiva news search of press reports, we classify CEO departure as
a dismissal when it is reported that the CEO is fired or forced out (which is quite uncommon)
or resigns or leaves as a result of policy differences or pressure (Jenter & Kanaan, 2015).2
Furthermore, exits for CEOs below the age of 60 are classified as dismissal if either the press
does not report the reason as death, poor health, or acceptance of another position or the press
notes that the CEO is retiring but does not announce the retirement at least 6 months in
advance (Parrino, 1997). If news reports explain the departure as due to reasons unrelated to
the firm’s activities, it is not classified as a dismissal. This method identified 641 dismissals
from a total of 2,416 departures in the sample (there were 1,769 voluntary departures and 6
that could not be classified).

Although coding of press reports is the most prevalent method to code CEO dismissal
(Gao, Harford, & Li, 2012; Hubbard et al., 2017), on the basis of an anonymous reviewer’s
comment, we conducted robustness tests using the Peters and Wagner (2014) age-based indi-
cator for CEO dismissal. We code all CEO departures below the age of 56 as dismissals and
the rest as voluntary exits (Peters & Wagner, 2014). The press- and age-based measures of
CEO dismissal are strongly correlated (r = .75, p < .01), and both measures yielded similar
support for our hypothesized relations. Because of concerns about using only age as an indi-
cator of departure (Wiersema & Zhang, 2011), and to be consistent with much of the prior
research in this area (Hilger et al., 2013), we report results using the press-based measure.

Our main predictor variables are Female, an indicator variable for the CEO being female,
and prior firm performance. We present all analyses with only market-adjusted returns (Ret
Mkt Adj, which is the difference between the firm’s return and the market’s return over the
year preceding the departure date), as standard economic theory suggests that exogenous
market performance gets filtered out in evaluating firm performance attributed to the CEO.
Notably, results are quite similar with alternative measures of performance: raw returns (Ret
1 Year: the firm’s stock return during the year preceding the departure date) and industry-
adjusted returns (Ret Ind Adj: the difference between the firm’s return and the industry’s
return during the year preceding the departure date).3 These performance measures capture
both absolute performance (Ret 1 Year) and relative performance (Ret Mkt Adj and Ret Ind
Adj). As reported in the Supplementary Analyses section, results are robust to employing
self-relative performance (which captures historical aspiration level; Harris & Bromiley,
2007) and accounting-based firm performance measures instead. We classify each firm’s
industry according to the 12-industry classification in Fama and French (1997), and we
obtain industry returns from Ken French’s data library.4

We include several control variables to capture the various factors that can influence the
likelihood of CEO dismissal.5 First, we control for Firm Size, measured as the natural log of
a firm’s total assets for the year prior to the departure date, to account for greater expectations

Gupta et al. / Gender and CEO Dismissal 567

of CEOs at bigger firms (Shen & Cannella, 2002). Second, we control for eight specific CEO
attributes that could influence the likelihood of dismissal (Hubbard et al., 2017): CEO Duality
(indicator variable for CEO also being chair of the board), CEO Ownership (the shares
owned by the CEO scaled by total shares outstanding), CEO Origin (dummy variable that
equals 1 if the CEO comes from inside the firm and 0 otherwise; Zhang & Rajagopalan,
2010), CEO Social Status (the number of other boards of listed firms the CEO serves on rela-
tive to board size of focal firm; Flickinger, Wrage, Tuschke, & Bresser, 2016), CEO Age,6
CEO Functional Experience (dummy variable that equals 1 if the CEO has finance experi-
ence and 0 otherwise; Gomulya & Boeker, 2014), Other-CEO Candidates (dummy variable
that equals 1 if the firm also has a separate person in the COO or president position and 0
otherwise; Zhang, 2006), and CEO Ability (dummy variable that equals 1 if the CEO obtained
a degree from a top 20 SAT school; T. King, Srivastav, & Williams, 2016). Third, we control
for three specific board characteristics since the board hires and fires the CEO (Hilger et al.,
2013): Board Size (the number of directors on the board), Independent Directors (the propor-
tion of board members who are not executives of the firm), and Female Directors (the pro-
portion of female board members excluding the CEO). In all, we include 12 covariates, plus
firm performance, as controls in the reported regressions. All continuous variables are win-
sorized at the 1% and 99% levels.


We use a probit model to test the two hypotheses: (1) CEO dismissal is more likely for
female CEOs than for male CEOs, and (2) dismissal among female CEOs is less influenced
by firm performance relative to male CEOs. A probit model is appropriate here because the
dependent variable, CEO Dismissal, is binary. Our results are robust to using a logit model,
as we report in the Supplementary Analyses section. We cluster standard errors at the firm
level, although as we discuss later, our results are robust to clustering at the industry or CEO
level. Our analyses use both industry and year fixed-effects. STATA version 14 was used for
all analyses.

Recent popular and academic discussions suggest that firms with female CEOs will dif-
fer systematically from firms where the CEO position remains a “male bastion” (Ryan et al.,
2016: 447). If correct, this could mean that the results from the full sample probit model
discussed above are affected by selection bias. To alleviate self-selection concerns, we fol-
low prior research (e.g., Bugeja, Matolcsy, & Spiropoulos, 2012; Geiler & Renneboog,
2015) in generating a propensity-matched sample of male- and female-led firms to test our
predictions. These results, which also validate our predictions, are presented in alternative
analyses below.


Tables 1 and 2 present descriptive statistics and correlations for our variables. Table 1
presents descriptive statistics for the sample as a whole and for male- and female-led firms
separately. Table 2 presents correlation coefficients for the sample as a whole. There are
21,772 firm-year observations in total, of which 617 are female-led and …

error: Content is protected !!