Research Assignment

r Academy of Management Journal 2018, Vol. 61, No. 2, 379–404. https://doi.org/10.5465/amj.2015.1238

GENDER-INCLUSIVE GATEKEEPING: HOW (MOSTLY MALE) PREDECESSORS INFLUENCE THE SUCCESS OF

FEMALE CEOS

PRIYANKA DWIVEDI Texas A&M University

APARNA JOSHI VILMOS F. MISANGYI

Pennsylvania State University

Male-typed leadership schemas have been widely acknowledged as barriers to women’s success in leadership roles. We explore how local organizational agents and contexts enable women leaders to overcome these barriers and achieve success at the highest levels in firms. Specifically, we focus on chief executive officer (CEO) suc- cession events and study how several facets of predecessor CEOs and the succession context combine to influence incoming women’s post-succession performance. We conduct a qualitative comparative case study of all CEO successions that involved female successors between 1989 and 2009 across the largest corporations in the United States. Our findings suggest that women’s success occurred when a confluence of local firm-level factors and attributes of the (mostly) male predecessors promoted gender- inclusive gatekeeping during succession. Our qualitative comparative analysis ap- proach reveal three recipes for female success: “handing over the legacy,” “partnering the legacy,” and “turning around the legacy.” Moreover, a comparison to a matched- sample of men CEO succession events shows that these three recipes for success are unique to women. Based upon our findings, we propose that male predecessors’ gender-inclusive gatekeeping facilitates female leaders’ success and occurs when local enabling conditions and the embedded context enact agentic and structural mecha- nisms to alter leadership schemas.

Across the largest corporations in the United States, over the past two decades, just under a hun- dred women have been appointed to the position of chief executive officer (CEO). Given the historic un- derrepresentation of women in top leadership roles, it is perhaps not surprising that management re- search has been devoted to identifying the obstacles that women face as they make their way into the upper echelons. Previous research in this domain has put forth three interrelated arguments: women face barriersbecausetheyarenumericallyunderrepresented

(Kanter, 1977); women are likely to fail in CEO roles because they do not fit the stereotypic expectations of what it means to be an effective leader (Eagly & Karau, 2002; Zhang & Qu, 2016); and, women tend to be appointed to leadership positions in the worst of times, which sets them up for failure (e.g., the “glass cliff”) (Haslam&Ryan,2008;Ryan&Haslam,2005).Themajor theme across all of these perspectives is that gender biases,stereotypes,andmale-typedleadershipschemas shape global normative expectations about women’s ability to lead and undermine the success of women in leadership roles. Based on this underlying theory, studies so far have largely focused on perceptions of women leaders and provided evidence that incoming women CEOs experience negative reactions from in- vestors, media, and other external stakeholders (Dezsö &Ross,2012;Dixon-Fowler,Ellstrand,&Johnson,2013; Lee & James, 2007; Park & Westphal, 2013). However, whether or not firms led by women underperform is not at all clear. Indeed, research has yielded mixed results for the effect of CEO gender on firm

The authors contributed equally to this research and are listed alphabetically. We thank Editor-in-Chief Gerry George and three anonymous reviewers for their thought- ful guidance throughout the review process. We would like to thank Forrest Briscoe and Don Hambrick for their feed- back on early drafts of this paper. We also thank seminar participants at the Pennsylvania State University, Uni- versity of Maryland, University of Delaware, and Univer- sity of Illinois.

379

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission. Users may print, download, or email articles for individual use only.

performance (e.g., Davis, Babakus, Englis, & Pett, 2010; Hoobler, Masterson, Nkomo, & Michel, 2016; Jeong & Harrison, 2016; Khan & Vieito, 2013; Kolev, 2012; Zhang & Qu, 2016).

In a substantive departure from the past accounts of women’s performance in executive roles, we focus on the potential role that key organizational agents and local contexts may play in driving variability in the application of gender biases and stereotypes that canhinderwomen’ssuccessasleaders.Theprevailing emphasis in past research has been on organizational stakeholders’ negative responses to female leaders emanating from pervasive male-typed schemas of leadership; a perspective which presumes that such schemas uniformly harm women and uniformly ben- efit men. This past lens therefore implicitly lends itself to a focus on differences between men and women in leadership effectiveness and deflects scholarly atten- tion away from examining the factors that create vari- ability among women who occupy executive positions in firms. Indeed, as is reflected in the mixed findings in thisdomain,itisnotyetfullyunderstoodwhyandhow some female leaders are able to overcome male-typed leadership schemas to become successful as CEOs, while others fail to do so. This has led some scholars to recommend that, if actionable theories of gender in- tegration are to be developed, researchers must refrain from taking a “sex differences” approach and focus instead on identifying the structural and contextual factors that shape variability within each gender (e.g., Ely & Padavic, 2007). To this end, we integrate past theory and research on female leadership, gate- keeping, and executive successions to examine con- ditions under which predecessor CEOs—who in the upper echelons of firms are mostly male—may func- tionas“gatekeepers”andwhetherthisgatekeepinghas potentially beneficial or pernicious effects on the per- formance of female CEO successors.

We consider two ways in which male predecessor CEOs may shape success among female CEOs. First, we consider the gatekeeping role played by the pre- decessor CEOs through their involvement in hiring, selecting, and socializing incoming female executives. The predecessor CEO’s impact is particularly poten- tiallyprominentinthecontextofsuccessionevents—a sensitiveperiodfortheCEOrole—wherealegacybuilt by one organizational leader is bequeathed to the next (Higgins, 2005; Joshi, Dencker, Franz, & Martocchio, 2010; Marquis & Tilcsik, 2013). The executive suc- cession literature suggests that whether predecessors are supportive or meddlesome, inconspicuous, or charismatic, they have a huge influence on how suc- cessors are hired and groomed for the CEO position

(e.g., Zajac & Westphal, 1996) and oftentimes remain involved in firm governance to directly influence the successors’ subsequent success in the role (Chen & Hambrick, 2012; Quigley & Hambrick, 2012). Fur- thermore, research on women’s entry into tradition- ally male-dominated contexts such as academia, engineering, law, science, and supervisory ranks is instructive for understanding the predecessor’s po- tential gatekeeping role. Gatekeepers have been tra- ditionally defined as individuals (e.g., supervisors) or groups (e.g., professional associations or boards) who restrict (or enable) entry into highly prestigious occu- pations through their direct involvement in recruit- ment or the setting of occupational standards and norms(Merton, 1957). Past research suggeststhatmale gatekeepers influence the maintenance, reproduction, or mitigation of gendered expectations and hierar- chies and can potentially hinder or enhance gender parity in male-dominated settings: while some male gatekeepers may be exclusionary and withhold re- sources from incoming women, others may be inclusionary and facilitate women’s success by opening avenues for gender integration (Acker, 1992; Briscoe & Joshi, 2016; Connell, 2005; Reskin & Padavic, 1988).

Second, we posit that gatekeeping also occurs through the setting and managing of the local nor- mative expectations associated with the particular CEO role. These local expectations can either mitigate or exacerbate the effects of global male-typed schema about women’s ability to lead a firm and in this way influence the success or failure of their female suc- cessors. Past research shows that predecessors are integral in such a process as they may be highly in- fluential in shaping idiosyncratic firm-specific skill expectations associated with the executive role that the successor must live up to in order to be successful (Burton & Beckman,2007).Priorresearchon the inter- generational transferoflegacies(e.g.,Joshi etal.,2010; Wade-Benzoni, 2002) also highlights the role that predecessors play in shaping the resource-based benefits or burdens associated with leader positions and how these have implications for the performance of incumbents in their new roles (Wade-Benzoni, Hernandez, Medvec, & Messick, 2008). Although these“local”expectationshavetypicallybeenviewed as constraints on successors, we consider whether they might also alter the stereotypic role expectations driven by global male-typed schema of leadership.

We draw on these disparate streams of research to proposethatpredecessorsandthelocalcontextinwhich they are embedded are an important missing link in extant theorizing about women’s successes and failures

380 AprilAcademy of Management Journal

in CEO roles. We explore the predecessors’ influence in female CEOs’ success through a comparative case anal- ysis of all of the CEO successions that involved female successorsbetween1989through2009acrossthelargest firms in the United States. Although the appointment of women as CEOs in large public corporations remains a relatively rare event, a small yet critical mass of women has ascended to this highest level of firms over the past couple of decades. This provides a unique opportunitytobegintodeveloptheorythataccountsfor women’ssuccessesintopleadershiproles.Todoso,we used the qualitative comparative analysis (QCA) (Ragin, 2000, 2008) case study approach to system- atically examine the entire set of female CEO cases—our final sample consists of 84 female suc- cession events—and to build theory based upon the underlying cases (e.g., Greckhamer, Misangyi, & Fiss, 2013).

As is conventional in case study research using QCA, we start by drawing on existing theory to identify various attributes of predecessor CEOs and the suc- cession context that seem relevant for shaping the successes or failures of incoming women CEOs. Like prior research on female executives, we begin with the premise that all female CEOs are subject to global normative male-typed expectations about leadership, but we surmise that variability in the post-succession performance among women CEOs can be in part at- tributed to the more proximal influence of certain predecessors operating in conjunction with other key attributes of the succession context (Finkelstein, Hambrick, & Cannella, 2009; Marquis & Tilcsik, 2013). Our qualitative comparative case study reveals that a complex confluence of predecessor attributes and local contextual conditions combined to form three gender-inclusive gatekeeping recipes through which predecessors facilitated the success of women CEOs in these organizational settings. Based upon our findings, we offer a midrange theory of gender- inclusive gatekeeping for women’s effectiveness in the upper echelons, and we discuss the implications ofourtheoryforfutureresearchonwomen’stransition into leadership roles more generally.

WOMEN IN EXECUTIVE ROLES: THEORY AND EVIDENCE

The predominant theme in extant research on women in executive roles is that pervasive male-typed leadership schemas or stereotypes create barriers for women’s success as leaders (e.g., Eagly & Karau, 2002; Schein, 1977). This view largely rests on role congruity theory (Eagly, Wood, & Diekman, 2000) and suggests

that leadership roles implicitly invoke stereotypically male traits—such as competitiveness, aggressiveness, independence, forcefulness, and decisiveness. Given women’s gender-based social roles, they are stereo- typically considered to possess communal and nur- turing qualities like cooperativeness, compassion, thoughtfulness, and warmth, that are believed to be more apt for caretaking roles rather than leadership roles that require dominance (Eagly & Karau, 2002). Thisincongruencebetweenfemalestereotypesandthe male-typed traits associated with leaders leads to per- ceptions that women are ill-equipped for leadership rolesandresultsinlessfavorableandevenantagonistic evaluations of women’s leadership effectiveness (Eagly & Karau, 2002; Joshi, 2014; Koenig, Eagly, Mitchell, & Ristikari, 2011; Nelson & Quick, 1985). Schein’s(2007)“thinkmanager–thinkmale”paradigm adds to this perspective, highlighting how traditional stereotypes of leaders involve agentic characteristics that tend to be associated with men (Schein, 1973, 2007). In short, this past research suggests that male- typed schema guide assessments of the characteristics required for success in managerial roles, thereby fos- tering bias against women in managerial selection, placement, promotion, and training.

Broad support for this perspective has come from studies that have examined organizational stake- holders’ perceptions of women’s ability to perform (e.g., Dixon-Fowler et al., 2013; Lee &James,2007; Park & Westphal, 2013). For example, announcements of female CEO appointments receive negative reactions from shareholders (Lee & James, 2007), as well as have negative spillover effects to other firms with female CEOs(Dixon-Fowleretal., 2013).And,mediacoverage on women CEOs is far more likely to prime gender roles by alluding to the executive’s children and mar- ital status, ultimately portraying female CEOs based on gendered norms rather than as viable leaders (Lee & James, 2007). There is also evidence to suggest that CEOs at other firms are more likely to attribute low performance to a peer CEO (i.e., rather than to envi- ronmental factors) when the peer is a female rather than a male. This triggers further negative media cov- erage for female-led firms and harms the reputations of women CEOs (Park & Westphal, 2013). Moreover, there is some evidence to suggest that when failure occurs, women are more likely to be singled out than men for criticism and blame (Haslam, Ryan, Kulich, Trojanowski, & Atkins, 2010). A related stream of research suggests that women are likely to be appointed to leadership roles in times of crisis and that their appointments are thus “more precarious” and set up for failure (also known as the “glass cliff”

2018 381Dwivedi, Joshi, and Misangyi

phenomenon; Ryan & Haslam, 2007; Ryan, Haslam, Hersby, & Bongiorno, 2011). However, research evidence testing this “glass cliff” perspective has been mixed (e.g., Paustian-Underdahl, Walker, & Woehr, 2014; Rosette & Tost, 2010).

Whether these perceptions and resulting disadvan- tages affect women CEOs’ actual successes—e.g., post- successionfirmperformance—remainsunclear.While some studies have found marginal performance benefits for firms with female CEOs (Davis et al., 2010; Khan & Vieito, 2013), others have found no significant difference in the shareholder returns of firms led by male versus female CEOs (Kolev, 2012). These equivocal findings have prompted researchers to look for contingencies that may shape the effects of CEO sex on firm performance. In a recent study set in the context of publicly-traded Chinese firms, Zhang and Qu (2016) found support for their suggestions that “gender-change” among CEOs (i.e., a woman succeeding a man) would have a “disruptive” effect on post-succession firm performance because this gender change makes the female successor’s per- ceived “out-group status” especially salient. More- over, they found that the successor’s insider origin and the presence of other women on the board of directors lessened this negative effect, which they suggested is due to these conditions reducing “the salience of the female successor’s ‘out-group’ status based upon gender” (2016: 1850).

Although these latter findings offer an important initial insight into how factors of the succession con- text may play in shaping women’s performance out- comes in CEO roles, they do not go far enough in engaging the complexities of successions wherein attributes of both the predecessors and successors combine with the succession context to influence success (Karaevli & Zajac, 2013). Indeed, past re- search suggests that succession events represent a transitory period in which elements of the past—the historic legacy associated with the role and the past experiences of the successor (Burton & Beckman, 2007)—aswellasthepresent—thesuccessioncontext and the involvement of the predecessor in and after the transition (Quigley & Hambrick, 2012; Zajac & Westphal, 2004)—jointly operate in complex ways to shape subsequent performance in the role (Finkelstein et al., 2009; Marquis & Tilcsik, 2013).

The Influence of Predecessors and the Succession Context

Our thesis is that certain attributes of the prede- cessor combine with several key facets of the

succession context to facilitate inclusionary (exclu- sionary) gatekeeping that might mitigate (amplify) the broader gendered stereotypical expectations for leadership roles. We first consider three key prede- cessor attributes that may contribute to predecessors’ influence on the subsequent performance of the in- coming female successors.

The predecessor’s founding status and tenure. The predecessor CEO’s status as a founder or a member of the founding family is likely to have con- siderable influence on stakeholders’ expectations about the role—and hence on the success of the in- coming successor—primarily because founders cre- ated the role to begin with. Indeed, Burton and Beckman (2007) found that the backgrounds, skills, and expertise of founding executives left a lasting “blueprint” for future incumbents in that role: the extent to which successors fit the expectations set by this blueprint predicted their own longevity in the role. Apart from shaping expectations about the skills and expertise associated with the role, found- ing CEOs may have more direct involvement in the incoming CEOs’ selection and socialization. In their review of organizationally-based generations, Joshi and colleagues highlighted how idiosyncratic skills developed in a role over a period of time form a part of a unique role-based generational legacy which has long term implications for the performance of future incumbents (Joshi et al., 2010; Joshi, Dencker, & Franz, 2011). Therefore, even when the predecessor is not a founder, a long tenure in the role would also lead to a specific legacy or blueprint for the expertise and skills needed in the role itself that may set local normative expectations that must be met by the successor in order to be successful (see also Marquis & Tilcsik, 2013). To the extent that a longer tenure in the role allows the CEO to build social relationships with key stakeholders, it also affords the predecessor the ability and legitimacy to manage stakeholder expectations about the incoming CEO. Moreover, based on the gravitas conferred by their long expe- rience in the role, long-tenured predecessor CEOs are also likely to be directly engaged in the incoming executives’ selection and socialization. As such, founder status or a long tenure would make the predecessor uniquely positioned to affect global schemas or stereotypic expectations about the local leader role.

These perspectives suggest that founding status or tenure is likely to enhance the male predecessors’ gatekeeping function during succession events in- volving female successors in two ways: by enhanc- ing the potential for the predecessor’s direct

382 AprilAcademy of Management Journal

involvement in selection and socialization of female successors and by indirectly forging local expecta- tions for the skills needed in the role. However, since past gatekeeping research has not focused on the specific context of executive successions, how the predecessor’s founding status or tenure will affect a female CEO’s subsequent success is unclear. On the one hand, this literature suggests that to the extent that CEOs who are long-tenured or founders, they may be in a position to buffer the incoming executive from global stereotypic expectations of the CEO role and also be a source of mentoring and socialization in the role. On the other hand, we might expect that since predecessor CEOs are mostly male, tenure or founding status might confer a “larger than life” persona for the departing male executive that simply mirrors or further strengthens societal male-typed implicit leadership schemas highlighting the female successors’ atypicality in the new role.

The predecessor’s pre-succession power. The power enjoyed by the predecessor CEO prior to succession, a key attribute to be considered in any consideration of succession events (e.g., Finkelstein et al., 2009), stands as a second potentially critical influence on the successor female CEO. In particular, CEO duality—i.e., when the CEO is also the chair of the board of directors—stands as an unambiguous marker of the CEO’s power and leadership (e.g., see Finkelstein & D’Aveni, 1994). Powerful predecessor CEOs are likely to have a direct say in the hiring or selection of the incoming female prior to the suc- cession (e.g., Zajac & Westphal, 1996). Further, like founding CEOs, powerful CEOs may also be in a po- sition to forge a blueprint on the role itself that might serve as the standard based upon which the in- coming CEO is evaluated (see Burton & Beckman, 2007). Indeed, it seems reasonable to expect that the predecessor’s duality (or lack thereof) might itself be part of that blueprint, such that the incoming exec- utive’s duality (or lack thereof) might serve as a sig- nal of their fit with the role. While the predecessor’s power is clearly a crucial attribute to be considered in developing an understanding of female success, it is unclear whether it will magnify or alter global normative expectations of women’s ability to suc- ceed as the CEO. On the one hand, a powerful pre- decessor CEO can serve as a beneficial gatekeeper, if for instance he uses it to provide potential female successors with access to the technical skills and social capital requisite to the job. On the other hand, if the male predecessor’s power triggers male-typed schema, or allows him to withhold access to technical and social resources from the incoming executive,

this might have pernicious effects for the female suc- cessor’s performance.

The predecessor’s post-succession presence on the board. The CEO succession literature further highlights that predecessors can also directly influ- ence their successors after the succession event (e.g., Fredrickson, Hambrick, & Baumrin, 1988; Quigley & Hambrick, 2012). For instance, Quigley and Hambrick (2012) found that predecessor CEOs may linger on as a member of the board of directors, often as the chairman of the board of directors, and that this presence can limit the successor CEO’s ability to make strategic changes and impact firm performance. Whether this continued presence has a beneficial or detrimental effect on the female CEO’s subsequent performance is unclear, however, as it also hinges on the way in which the predecessor serves as a gatekeeper. For instance, by remaining active on the board, a predecessor who is committed to gender diversity and inclusion at the highest levels may continue to provide access to social capital and other resources that are crucial for meeting the local normative expectations associated with the CEO role. On the other hand, this lingering presencecould also serve as an active meddlesome influence that severely hinders the female successor’s legitimacy and ability to act in the role.

In sum, whether these three attributes of a pre- decessor’s influence—which may combine in several ways to involve active engagement pre- and post- succession as well as set expectations associated with therole—haveadetrimentalorbeneficialimpactonthe female successor’s fate remains an open question. We now turn to considering several aspects of the succes- sion context which may combine with the predecessor influences to shape post-performance success.

The Successor and Succession Context

The CEO succession literature points to several key facets of the succession context—the female succes- sor’s origin, the inherited performance, and the power of the successor—that may potentially combine with the predecessor’s attributes in setting local expecta- tions that either mirror or buffer global male stereo- typic expectations about the CEO role as well as affect the agency of the successor (e.g., Finkelstein et al., 2009; Karaevli & Zajac, 2013).

Successor origin. The origin of the female successor— whether the incoming female CEO was promoted from inside (insider) or hired from outside (out- sider) the firm—is widely acknowledged by extant research as a key facet of executive successions

2018 383Dwivedi, Joshi, and Misangyi

(e.g., Castanias & Helfat, 2001; Zhang & Qu, 2016; Zhang & Rajagopalan, 2004). Furthermore, Rosette and Tost (2010) found in a series of experiments that female leaders are ascribed as being more agentic and rated more highly in general leadership effec- tiveness when the firms’ successes are internally attributed directly to their efforts. These past find- ings imply that female insiders with a proven track record of success within their firms could forge lo- cal expectations about their competence and lead- ership ability that replace global gendered expectations by building relationships with key stakeholders in the organizations even prior to en- tering the role. How the successor’s origin combines with the several foregoing predecessor influences is an open question, however. For instance, the po- tential just described benefits of being an insider female may be enhanced or diminished by a long- tenured or founder predecessor. On the one hand, such a predecessor could play an active role in reinforcing the insider female CEO’s “in-group” status and proven track record. On the other hand, as we outlined above, the “larger than life” persona of a long-tenured or founding predecessor might be burdensome for the insider female who may be per- ceived as the predecessor’s acolyte thereby dimin- ishing her own effectiveness in the role.

Successor’s inherited performance. Adding a further layer of complexity to the combinatorial nature of succession events, a successor’s insider (outsider) origin is closely intertwined with the performance conditions that are inherited by the successor from the predecessor (e.g., Finkelstein et al., 2009; Karaevli, 2007). That is, outsiders are typically hired in contexts where there is at least the perception that change is needed—i.e., by firms who are struggling in their performance. Outsider ap- pointments are valued by boards both for the new skill sets that outsider CEOs bring, which may lead to performance turnarounds for struggling firms (Chen & Hambrick, 2012), and for the clear signal of change they provide that helps to assuage stakeholder ex- pectations (Finkelstein & D’Aveni, 1994). In short, the prior performance conditions clearly affect “lo- cal” expectations in ways that may or may not am- plify global stereotypes. Indeed, the “glass cliff” arguments proposed in past research for the hiring of outsider female CEOs in distressed firm situations have received mixed support. One reason for these past equivocal findings is that prior firm performance may not operate as an independent mechanism shap- ingwomen’spost-successionperformance.Rather,the performance legacy operates in conjunction with the

influences of the predecessor and the other facets of the succession context (Karaevli & Zajac, 2013). For instance, a positive (negative) performance inheri- tance could enhance (minimize) the combinatorial effect of the predecessor’s influence and the succes- sor’s origin on the successor female CEO’s subsequent successes.

Successor’s power. Somewhat mirroring the im- portance of the predecessor’s power while he is CEO, whether the female successor is appointed as the chairperson when the CEO (i.e., CEO duality) also stands as a critical element that potentially contrib- utes to the female successor’s success. As noted above, duality itself may be part of the blueprint or expectations for a role: when a predecessor was powerful so too should be the successor—in order to fit local firm-level expectations (see Burton & Beckman, 2007). Conversely, when a powerful or long-tenured predecessor has already managed to weaken the influence of global stereotypic expecta- tions through their gatekeeping function, it may be that the female successor may not need duality after all. In this situation, the insider female successor may be seen as meeting the local expectations of the role already (for reasons already discussed above) and thus might receive the support and endorsement from stakeholders who may eventually enable her success in the role. On the other hand, given that duality is a clear symbol of power, this position may enhance local perceptions of the female successor as meeting more global expectations. In any case, holding such a position of power provides the female CEO with more discretion, and thus agency, than if she didn’t hold this dual role (Hambrick & Finkelstein, 1987).

In sum, how the predecessor’s founding status or tenure, past power, or continued presence on the board combine with the female successor’s origin, inherited performance conditions, and power to shape subsequent outcomes is a vital open question. Below we describe the QCA approach we undertook to systematically investigate whether and how these attributes and facets combine to affect female CEO success.

METHOD

We used the fuzzy-set approach to qualitative comparative analysis (fsQCA; Ragin, 2000, 2008) to examine how predecessor attributes and succes- sion contextual elements combine to affect succes- sors’ firm performance among incoming female CEOs. This comparative case method affords

384 AprilAcademy of Management Journal

a configurational approach (e.g., see Crilly, Zollo, & Hansen, 2012; Fiss, 2011; Misangyi & Acharya, 2014) that is especially well suited for our inquiry because it affords both a systematic and rich comparative analysis—through its use of Boolean logic and algebra—across a relatively larger number of cases than is conventionally possible through other cross- case comparative techniques (e.g., see Greckhamer et al., 2013). Detailed explanations of this method- ology are presented elsewhere (e.g., Greckhamer, Misangyi, Elms, & Lacey, 2008; Ragin, 2000, 2008). We first describe our selection of cases, followed by an explanation of our coding of the cases’ member- ships in the outcome and each of the theoretically relevant attributes. We then briefly explain our pri- mary and supplemental analytical approaches.

Data and Sample

Since generalizing the findings of fsQCA has to be done with care (Greckhamer et al., 2013)—as is true with any qualitative case-oriented analytical approach—we looked for all CEO succession events involving women that occurred over the past three decades among the S&P 1500 and Fortune 500 firms, given their prominence and representativeness of firm size (e.g., see Misangyi & Acharya, 2014). Our initial search resulted in a total of 98 female CEO succession events, from which we found complete data for 84 cases that span the years 1992–2009. In constructing our sample, we treated the year of succession as year t, and collected the relevant pre- succession attributes in years t – 1 through t – 3 and post-succession attributes in years t 1 1 through t 1 3. All the data come from archival sources in- cluding firms’ annual reports and proxy statements, press releases on succession announcements, the Business Week and Forbes executive database, Com- pustat’s Annual and Execucomp files, and the Risk Metrics database. Online Supplemental Appendix 11

presents the cases of female CEOs we studied.

Coding Cases’ Set Memberships

In fsQCA, the theoretical attributes (predecessor influences, successor characteristics, contextual conditions) and outcome (firm performance) are each viewed as a set in which each case may have membership. Consistent with previous research, we

coded membership in binary attributes using crisp sets (i.e., “fully in” or “fully out”), we used four-value fuzzy-sets to code attributes derived from qualitative data (i.e., fully in, more in than out, more out than in, fully out; Misangyi & Acharya, 2014), and we used the direct method of calibration in the fsQCA soft- ware to transform continuous quantitative data into fuzzy-set memberships(based upon three qualitative thresholds: fully in, the crossover point, fully out; Ragin, 2008). We relied on existing theory and substantive knowledge to establish all calibration thresholds. Table 1 summarizes the calibration thresh- olds for each attribute under study, as well as the pertinent descriptive statistics of the underlying measures upon which the coding of set memberships was based.

Outcome

Post-succession high firm performance. Follow- ing the extant CEO succession literature, we used the firm’s performance over three years following the succession event—measured as average firm return on assets (ROA; i.e., net income divided by total as- sets) ofyears t 1 1, t 1 2, and t 1 3 (e.g.,Cannella,Park, & Lee, 2008; Cannella & Shen, 2001; Shen & Cannella, 2002). Consistent with previous studies that have examined firm performance as an outcome using QCA (Fiss, 2011; Misangyi & Acharya, 2014), we used this measure to code each case’s membership in the set of firms with high post-succession ROA—i.e., a firm has to be an above-average performer within its industry to be in this set—based upon the following thresholds: firms that were at or below their industry median ROA were coded as fully out of this set, firms with ROA in the upper quartile of their industry (i.e., $ 75th percentile) were coded as fully in, and the crossover point was the halfway mark between these thresholds. Industry ROA median and upper quartile scores were calculated using the three-year average post-succession ROA of all firms in the Compustat annual file in the same four-digit SIC as the focal company.

Theoretical Attributes

Predecessor tenure_or_founder. We measured predecessor tenure as the number of years the pre- decessor CEO served prior to the succession event. Given that the average tenure of large company CEOs is 9.7 years (The Conference Board, 2014) and that it takes roughly three years for CEOs to fully make their mark on a company (Hambrick & Fukutomi, 1991;

1 Link to online supplemental appendices(s): https:// sites.psu.edu/vilmos/files/2018/01/Online-Supplemental- Appendix-194243a.docx

2018 385Dwivedi, Joshi, and Misangyi

T A B L E 1

F u zz

y -S et

M em

b er sh

ip C a li b ra ti o n s a n d M ea

su re

D es cr ip ti v e S ta ti st ic s

F u zz y S e t C a li b ra

ti o n s

M ea

su re

D es cr ip ti v es

a

A tt ri b u te s

F u zz y -s e t /M

ea su

re F u ll y in

C ro ss o v e r

F u ll y o u t

M ea

n S D

M a x

M in

1 .

P re d ec

es so r’ s P re -

S u cc es si o n T en

u re

S u cc

es si o n ev

en ts

w h er e th e

p re d ec es so r C E O

h a d a lo n g

te n u re

/ n u m b er

o f y ea

rs th e

p re d ec

es so r C E O

se rv ed

p re -

su cc es si o n

1 0 ye

a rs

3 ye

a rs

1 ye

a r

(F ) 9 .3

6 .9

3 4

1 .0

(M ) 8 .5

7 .5

3 8

0 .2 5

2 .

P re d ec

es so r w as

F o u n d er

b S u cc

es si o n ev

en ts

w h er e th e

p re d ec

es so r C E O w a s a fo u n d er

/ p re d ec

es so rs

w h o w er e

fo u n d er s, co

-f o u n d er s, o r

fo u n d in g fa m il y

P re d ec

es so r is

fo u n d er

P re d ec

es so r is co

- fo u n d er / fo u n d in g

fa m il y (0 .6 6 )

P re d ec

es so r n o t

fo u n d er

(F ) 0 .2 2

0 .4 0

1 0

(M ) 0 .2 1

0 .4 2

3 .

P re d ec

es so r P o w er

c S u cc

es si o n ev

en ts

w h er e th e

p re d ec

es so r C E O

al so

se rv ed

as th e ch

ai r o f th e b o ar d p ri o r

to su

cc es si o n

P re d ec

es so r C E O

w a s

ch a ir

P re d ec

es so r C E O

w a s n o t ch

a ir

(F ) 0 .6 9

0 .4 6

1 0

(M ) 0 .7 0

0 .4 6

4 .

P re d ec

es so r’ s P o st -

S u cc es si o n

P re se n ce

o n B o ar d d

S u cc

es si o n ev

en ts

w h er e th e

p re d ec

es so r re m a in s o n th e

b o a rd

fo r a t le a st 2 4 m o n th s p o st

su cc es si o n as

ch ai rm

an ,v

ic e-

ch ai rm

an o r d ir ec

to r

P re d ec

es so r C E O

re m a in s a s ch

a ir

P re d ec

es so r C E O

re m a in s a s vi ce

-c h a ir /

d ir ec

to r (0 .6 6 )

P re d ec

es so r C E O

d o es

n o tr em

a in

o n b o a rd

(F ) 0 .2 5

0 .4 2

1 0

(M ) 0 .4 1

0 .4 8

5 .

S u cc

es so r is

In si d er

S u cc

es si o n ev

en ts

w h er e th e

su cc es so r C E O

w a s a n in si d er

/ n u m b er

o f y ea rs

as an

ex ec

u ti v e

o f fo ca

l fi rm

p ri o r to

b ec o m in g

C E O

5 ye

a rs

2 ye

a rs

1 ye

a r

(F ) 7 .7

8 .8

3 3

0 (M

) 5 .4

7 .6

3 1

0

6 .

S u cc

es so r P o w er

e S u cc

es si o n ev

en ts

w h er e th e

su cc es so r C E O w a s a ls o th e ch

a ir

b y th e tw

el ft h m o n th

o f th ei r

p o st -s u cc

es si o n

S u cc es so r C E O

is ch

a ir

S u cc es so r C E O

is n o t ch

a ir

(F ) 0 .3 3

0 .4 7

1 0

(M ) 0 .2 1

0 .4 1

7 .

P re d ec

es so r P o o r

P er fo rm

an ce

f S u cc es si o n ev en

ts w h er e th e

p re d ec es so r C E O

h a d p o o r p re -

su cc es si o n p er fo rm

a n ce

/ th re e-

y ea r p re -s u cc es si o n av

er ag e fi rm

R O A

R O A #

In d u st ry

M ed

ia n R O A

H a lf w a y

R O A

$ In d u st ry

Q 3 R O A

(F ) 0 .2 5

0 .4 4

(M ) 0 .3 6

0 .4 8

8 .

P re d ec

es so r F ir ed

g S u cc es si o n ev en

ts w h er e th e

p re d ec es so r C E O w as

fi re d /

p re d ec es so r w as

fi re d o r le ft

vo lu n ta ri ly

(p la n n ed

re ti re m en

t; d u e to

h ea lt h o r p ro fe ss io n al

re as on

s)

P re d ec

es so r C E O

d is m is se d

_ P re d ec

es so r C E O

vo lu n ta ry

ex it

(F ) 0 .1 6

0 .3 6

1 0

(M ) 0 .1 5

0 .3 6

a W

e in cl u d e h er e th e m ea su

re d es cr ip ti v es

fo r th e m at ch

ed sa m p le

o f m en

fo r co

m p ar at iv e p u rp o se s.

b D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e p re d ec

es so r C E O s w er e fo u n d er s, co

-f o u n d er s o r fo u n d in g fa m il y .

c D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e p re d ec

es so r C E O s al so

se rv ed

as ch

ai r o f th e b o ar d p ri o r to

su cc es si o n .

d D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e p re d ec

es so r C E O s re m ai n ed

o n th e b o ar d p o st -s u cc

es si o n as

ch ai rm

an /v ic e- ch

ai rm

an /d ir ec

to r.

e D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e su

cc es so r C E O s se rv ed

as ch

ai r o f th e b o ar d b y th e tw

el ft h m o n th

p o st -s u cc

es si o n .

f D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e p re d ec es so r C E O

h ad

p o o r p er fo rm

an ce .

g D es cr ip ti v e st at is ti cs

ar e th e p ro p o rt io n o f fi rm

s in

th e sa m p le

w h er e th e p re d ec

es so r C E O

w as

d is m is se d .

386 AprilAcademy of Management Journal

Miller, 1991), we coded a fuzzy-set to capture each case’s membership in the set of succession events with a long-tenured predecessor CEO based on the following thresholds: cases where the predecessor’s tenure was $ 10 years were coded as fully in; pre- decessor tenures # 1 year were coded as fully out; and a tenure of three years was treated as the cross- over point. We assessed each case’s membership in the set of succession events where the predecessor CEO was a founder using a four-value fuzzy-set: cases where the predecessor was the founder were coded as fully in; co-founders and members of the founding family were coded as more in than out; and cases where the predecessor was not any type of founder were coded as fully out. In cases where the predeces- sor was an interim CEO—as either described as such inmediareportsorwhoservedforlessthan12months (Quigley & Hambrick, 2012)—we considered the CEO prior to the interim CEO appointment as the pre- decessor. Because either of these attributes should theoretically have a similar effect (e.g., forge a blue- print on the role), we thus used a meta-set in our suf- ficiency analyses consisting of the union of these two fuzzy-sets which captured each case’s membership in thesetofsuccessioneventswherethepredecessorhad a long tenure or was a founder of the firm (via the “fuzzy or” function in the fsQCA software, which constructs a fuzzy-setbasedupon themaximum value for each case of the two examined sets; Ragin, 2008).

Predecessor duality. Predecessor duality was assessed by examining whether the predecessors also served as the chairperson of theboardwhile they were the CEO (i.e., predecessor CEO duality prior to the succession event). Each case’s membership in the set of succession events where the predecessor CEO was also the chair was assessed with a crisp set such that cases where the predecessor was also the chair prior to the succession event were coded as fully in; otherwise, cases were coded as fully out.

Predecessor presence on board. Because out- going CEOs may remain on the board for at least part of the first year after the succession event (Quigley & Hambrick, 2012), we didn’t consider a predecessor’s presence on the board to be meaningful unless they were on the board for at least two years post succes- sion. Thus, we assessed each case’s membership in the set of succession events where the predecessor remains on the board using a four-value fuzzy-set as follows: firms where the predecessor remains on the board as chairperson for at least 24 months after the succession date were coded as fully in; those where the predecessor continued as a board director or as vice-chairperson for at least 24 months were coded

asmore in than out; and caseswhere the predecessors were not on the board post-succession were coded as fully out.

Successor is insider. We measured the number of years each incoming CEO served as an executive of the focal firm prior to her becoming CEO and trans- formed this measure into a continuous fuzzy-set capturing each case’s membership in the set of suc- cession events where the successor CEO was an insider using the following calibration thresholds. Conventionally, an incoming CEO is considered to be an insider once they have spent at least two years in the company pre-succession (Shen & Cannella, 2002). We thus used this theoretical precedent as the crossover point and set the upper (pre-succession tenure of $ 5 years, fully in) and lower (# 1 year, fully out) thresholds in a manner that allowed us to cap- ture the conceptual underpinnings of this attribute. That is, spending less than a year in the firm pre- successionwouldnotaffordtheopportunitytodevelop a proven track record or the building of meaningful relationships; a pre-succession tenure of five years would clearly provide ample time to do so.

Successor duality. Successor duality was assessed by examining whether the incoming female CEO was also appointed as the chairperson of the board. Given that it is common for predecessors to remain chair for at least some of the first year post-succession (Quigley & Hambrick, 2012), we considered a succes- sor to have duality if she was appointed as the chair by the end of the first year post-succession. Thus, we coded each case’s membership in the set of succes- sion events where the successor CEO had duality with a crisp set such that cases where the successor was also the chair by the twelfth month of their post-succession tenure were coded as fully in; other- wise, cases where coded as fully out.

Successor inherits poor performance. The final succession contextual element we examined is the performance conditions left behind by the predeces- sor, and past research suggests that there are two crit- ical factors by which such conditions can be assessed (e.g., see Finkelstein et al., 2009): pre-succession firm performance(wasitpoorornot?)andthepredecessor’s exit (was it voluntary or not?). Given the conventional interest in the CEO succession literature around poor starting conditions, we captured poor pre-succession firm performance in a similar manner as described above for post-succession performance (i.e., we used the average ROA in the three years prior to the suc- cessionevent)andassessedcasesfortheirmembership in theset of firmswith poorpriorperformance byusing the converse of the above-described measure: cases

2018 387Dwivedi, Joshi, and Misangyi

with three-year pre-succession average firm ROA # 50th percentile of their four digit SIC industry ROA werecodedasfullyin;firmROA$ 75th percentilewere coded as fully out; crossover point 5 the halfway mark between these thresholds. In coding the predecessor’s exit, we were interested in whether the predecessor CEO left involuntarily (i.e., was fired) and did so by examining annual reports, proxy statements, company announcements, and news and media reports around the succession event. We also followed previous stud- ies in using both predecessor age and their continued board presence to corroborate our assessments (Shen & Cannella, 2002). We then used these data to assess each case’s membership in the set of succession events where the successor CEO is dismissed using a crisp set: cases where the predecessor CEO was dismissed were coded as fully in; otherwise, as fully out. Because the successor clearly inherits much less favorable perfor- mance legacy conditions when either performance was low or the predecessor left involuntarily, we used a meta-set to capture each case’s membership in the set of succession events where the predecessor had poor performance OR left involuntarily (i.e., the union of these two sets via the “fuzzy or” function).

Analytical Technique

Sufficiency analyses. Primary analysis. Theforegoingcodingofcases’set

memberships in the outcome and each of the theoret- ically relevant attributes sets the stage for conducting several types of analyses. Our primary analysis of in- terest involved examining whether the theoretical at- tributes specified above are sufficient for observing high post-succession performance, an analysis that relies upon an examination of subset relations (Ragin, 2000, 2008). Briefly, for any of the above attributes, or a combination of them, to be sufficient for observing high performance, their presence (or absence) will al- ways be accompanied by the presence of high perfor- mance. As with past exploratory studies using QCA, wefirstusedthesufficiencyanalysestoidentifywhich, if any, of the combinations of the specified theoretical attributes were sufficient for post-succession firm per- formance, and we then further examined the cases underlying the sufficient configurations through more in-depth qualitative analysis to gain additional insight into how and why the particular combinations led to success (Aversa, Furnari, & Haefliger, 2015; Crilly, 2011; Misangyi & Acharya, 2014). In describing our results, we first report the configurations found in the sufficiency analysis and then discuss the findings from our subsequent qualitative analysis.

Technically speaking, the sufficiency of an attribute (or a combination of attributes) for observing the out- come is shown if membership scores in the attribute (combination) are consistently less than or equal to the membership scores in the outcome, where consistency “indicates how closely a perfect subset relation is ap- proximated” (Ragin, 2008: 44). Subset relation consis- tency is therefore the metric by which sufficiency is evaluated. Because perfect subset relations (i.e., the particular explanatory condition is always present when the outcome is also present) rarely, if ever, exist, researchers determine a priori the minimum consis- tency threshold they deem acceptable (i.e., they settle for “quasi-sufficiency”; Ragin, 2000). In conducting our sufficiency analyses, we followed prior research (Misangyi & Acharya, 2014; Ragin, 2008) and (1) set the minimum raw and proportional reduction in in- consistency (PRI) consistency thresholds to . .75 and, (2) utilized the natural break in raw consistency scores during the analysis as the threshold consistency. We report the actual levels of both types of consistencies used for each analysis in the footnotes of the results ta- bles. With respect to solution consistency, we used a minimum acceptable overall solution consistency of .80 (Ragin, 2008). We set the minimum frequency in the analyses to two cases per configuration as this was re- quired to have an inclusion rate of at least 80% of the cases (see Ragin, 2008; Rihoux & Ragin, 2009).

Contingency analyses. Although control variables are neither required nor possible in QCA (e.g., see Misangyi, Greckhamer, Furnari, Fiss, Crilly, & Aguilera, 2017), we sought to examine whether sev- eral additional conditions affected how the foregoing theoreticalattributescombinedforthesuccessfulcases. Though not specific to succession contexts generally, these conditions have nevertheless been shown by previous research to be important to the phenomena under study. In particular, the degree to which the in- dustry is male or female dominated (e.g., Joshi, Son, & Roh, 2015) and the presence of women on the board of directors has been found to be important to female ex- ecutive success (e.g., Zhang & Qu, 2016). Previous re- search also suggests that firm size has an effect on firm ROA (e.g., Hansen & Wernerfelt, 1989). Therefore, to examine what, if any, effect these conditions may have on our main findings, we performed three additional sufficiency analyses in which each of these conditions were examined as contingency conditions.

We coded each case for their membership in each of these sets as follows. Cases with one or more female directors on the board other than the female successor CEO were coded as fully in the set of succession events that occurred in firms with women on the board of

388 AprilAcademy of Management Journal

directors; otherwise they were coded as fully out. We captured succession events that occurred in male- dominated industries using a continuous fuzzy-set: cases in industries where the ratio of male senior exec- utives to total senior executives $ .85 were coded as fully in; cases in industries with a proportion of male senior executives to total senior executives # .50 were coded as fully out; the cross-over was coded as a pro- portion of .70. Examples of male-dominated industries includecomputerequipmentmanufacturing(SIC3577), data processing and computer programing services (SIC 7374), beverages manufacturing (SIC 2086), and chemicals and chemical preparations (SIC 2899). Large firm size was measured using the net sales in the year of succession for each firm obtained from the Compustat database.Extantresearchgenerallydefineslargefirmsas havingsalesinexcessof$100millionandthuswecoded firms as being fully out of this set if their net sales were $100 million or less. We then used the distribution of the data and used the median (sales 5 $951.2 million) asthecrossoverpoint,andthe75thpercentile(quartile3) as the fully in threshold (sales 5 $2817.7 million).

Supplemental analyses. We performed two addi- tional sufficiency analyses. First, we followed the conventional practice of examining the sufficiency of the specified theoretical attributes for the absence of the outcome—i.e., for not-high post-succession firm performance (e.g., Greckhamer et al., 2008; Misangyi & Acharya, 2014; Ragin, 2008). Second, we also sought to understand how the findings with re- spect to women CEOs compare to those of men CEOs. To do so, we examined the sufficiency of the studied theoretical attributes for both high and not-high post- succession firm performance on a matched sample of male CEO successions.

Robustness analyses. We also further investigated the robustness of the main results through three addi- tional analyses. First, we examined the necessity of the various attributes under study for achieving high per- formance. Second, we conducted a diversity analysis to examine the configurations that existed empirically but were not sufficient for performance (Greckhamer et al., 2008). Finally, given that we are interested in understanding women’s success, and that tenure in the new role has been used as a measure of success in past research (e.g., Burton & Beckman, 2007), we also sought to examine whether our findings were robust to this alternative measure.

FINDINGS

We first provide key sample descriptives. With re- spect to the outcome of interest—post-succession firm

performance—59ofthe84femaleCEOshadanon-zero fuzzy-set membership score (i.e., ROA above their in- dustry median), and 44 cases had a membership score above the crossover point of the high performance fuzzy-set (i.e., were “more in than out”to “fully in” this set). All but one of the predecessors to the female suc- cessors were male CEOs (Anne Mulcahy was the only female predecessor CEO to Xerox CEO Ursula Burns). The average tenure of the predecessors was 9.3 years (see Table 1), and 39% of the cases involved prede- cessors with 10 or more years of tenure. Twenty-one (25%)ofthepredecessorswerefoundersorapartofthe founding family of their firms. Fifty-seven (69%) of the predecessors were dual CEOs prior to the succession. Outofthese,15remainedontheboardpost-succession (14 as chairperson and one as a director). Four pre- decessors were not chairpersons when CEO but were then appointed as chairperson post-succession, and another three remained on the board as a director (in total, 22 predecessors remained on the board). In terms of the female successors, 51 of the 84 (61%) were in- siders (average pre-succession tenure was 7.7 years; see Table 1). Within a year of assuming the CEO posi- tion, 28 (33%) of the successors were appointed as the chairperson of the board (i.e., dual CEOs). Only 21 (25%) of the female successors inherited poor firm performance from their predecessor, and in 13 cases, the predecessor left involuntarily.

Configurations Sufficient for Women CEOs’ High Performance

The primary analysis involved examining the suffi- ciency of the studied theoretical attributes for observ- ing female CEOs who achieved high post-succession firm performance. The results of this analysis are re- ported in panel 1 of Table 2. As the table shows, there are three different recipes for women’s success as a CEO, which cover 24 of the 44 cases in our sample (55%) that are above the crossover point of the set of high performing firms.2 The results of each of the contingency analyses which examined the effect that

2 Coverage is a measure of empirical relevance in QCA— akin to the R2 in regression (Fiss, 2011)—but does not equate to the proportion of cases covered by the configuration or solution (see Ragin [2006] for a detailed discussion of how coverage is calculated in QCA). Thus, in addition to the conventional reporting of the raw and unique coverage of each configuration as well as the overall solution coverage, we also report the actual number of cases in each configura- tion as well as the proportion this represents of the relevant cases (here the high performing cases).

2018 389Dwivedi, Joshi, and Misangyi

women on the board, male-dominated industries, and firm size have on the main findings are reported in panels1 through 3 of Table 3, respectively. Inessence, and as we describe below, these contingency analyses show the contexts in which the three different recipes found in the main analysis tended to occur.

Before further describing the results, note that we interpret and report the intermediate solution pro- duced by the fsQCA software in the configuration tables (e.g., Fiss, 2011; Misangyi & Acharya, 2014) and denote the presence and absence of attributes within each configuration as follows: central condi- tions are denoted by (present) and (absent) while contributing conditions are represented by d (present) and (absent). Core conditions are decisively suffi- cient given that they rest upon the existing data among the studied cases rather than on counterfactual anal- ysis, while the inclusion of contributing conditions in the solution is derived from counterfactual anal- ysis (see Ragin, 2008). Thus, although reporting this distinction is important for transparency reasons, an

interpretation of core conditions as being theoretically more important than contributing conditions is only rel- evantwhenoneaprioritheorizesaboutsuchadistinction (Misangyi et al., 2017; e.g., see Fiss, 2011; Grandori & Furnari, 2008). Therefore, we denote this distinction for transparency, but do not distinguish between the con- ditions in our theoretical interpretations.

Handing over the legacy. Configuration 1 of Table 2 shows that when powerful long-tenured predecessors completely handed over the reins of power to insider female successors of favorably performing firms, this was sufficient for observing high post-succession firm performance. In this con- figuration, female successors were long time insiders (13 years on average in the company before becom- ing a CEO) who followed long-tenured (14 years as CEO on average; four were firm founders) predeces- sors. All of the predecessors enjoyed the power of board chairpersonship while they were CEOs and then fully relinquished this power to their female successors (all predecessors retired): all of the female

TABLE 2 Configurations Sufficient for Women’s Post-Succession Performance

Notes: Central conditions are represented by (presence) and (absence); contributing conditions by (presence) and (absence). Configurations 1a and 1b in Panel 2 are “neutral permutations”; they share the same central conditions but differ in their contributing conditions. Actual minimum thresholds used in the analyses, respectively: raw consistency 5 .77; .77; PRI consistency 5 .76, .75; a minimum frequency of two cases/configuration was used in both analyses.

390 AprilAcademy of Management Journal

CEO successors in this configuration were appointed chairperson of the board within a year of their suc- cession. A total of 12 women (or 27% of the 44 high performing cases) followed this recipe for success, including DuPont’s Ellen Kullman, Mattel’s Jill Barad,Pepsi’sIndraNooyi,andXerox’sAnneMulchahy and Ursula Burns.

To gain a deeper insight into the workings of this combination of attributes, we further analyzed these 12 cases comprising this configuration by examining the company websites, annual reports, press releases, and succession announcements as well as news arti- cles, business reports, and books on the cases. Our deeper qualitative analysis identified three additional mechanisms as underlying how the studied theoreti- cal attributes combined for success.

The first mechanism, mentoring and sponsorship by the predecessor, was a common theme across all of the 12 cases in this configuration. The prede- cessors handpicked the female successors for

positions of responsibility and mentored them to take challenging roles that not only increased the breadth of their functional experiences, but also helped them to gain visibility in the organization as successful leaders. Indra Nooyi’s rise at Pepsi is il- lustrative: when Steve Reinemund became the CEO of Pepsi Co., he made Indra Nooyi his second in command (she was chief financial officer [CFO] and President). He is famously quoted as having told Nooyi “I can’t do it unless I have you with me” (Smith, 2015). CEO Linda Lang of Jack in the Box presents another good example, as captured in a post-succession reflection given by her predecessor CEO, Robert Nugent, on his influence on Lang’s ca- reer progress: “I’d ID’d Linda early on. . ..It quickly became clear to me that she was very astute about fi- nancial matters. . ..I told her that before she could move up any further, she’d have to serve in operations” (Green, 2006). Nugent put Lang in charge of the Southern California region, which provided her with

TABLE 3 Configurations Sufficient for Women’s High Post-Succession Performance: The Contingencies of Women on the Board,

Male-Dominated Industries, and Firm Size

Notes: Central conditions are represented by (presence) and (absence); contributing conditions by (presence) and (absence). Configurations 2a and 2b in Panel 1 are “neutral permutations”; they share the same central conditions but differ in their contributing conditions. Actual minimumthresholds for the analyses, respectively: raw consistency 5 .82, .83, .78; PRI consistency 5 .75; .75, .75; minimum frequency of two cases/configuration in all three analyses.

2018 391Dwivedi, Joshi, and Misangyi

the operational experience she needed to become president, chief operating officer (COO) and later CEO when Nugent retiredin 2005.Similarly,Paul A. Allaire, former Chairman and CEO of Xerox, not only mentored his successor, Anne Mulchahy, but also had a hand in sponsoring her successor, Ursula Burns. CEO Ellen Kullman (DuPont) and Jill Barad (Mattel) also have explicitly credited their predecessor CEOs, Charles Holliday and John Amerman, respectively, as being pivotal to their development as the future leader of their firms. Moreover, in three cases, the sponsoring predecessors were also the founder of the company and in two of these the female successors were daughters of the founders (Claire C. Skinner, CEO of All American Group; Marla Schaefer, Co-CEO of Claire’s Stores).

The second set of mechanisms underlying this configuration were more structural in nature as the organizations either had founder(s)that supported gender inclusion or organizational programs that promoteddiversitywithin the company. Ouranalyses revealed that in five cases (All American Group Inc., Bombay Company Inc., Claire’s Stores, Edgewater Technology Inc., Lenox Group) gender inclusion at the very top was initiated in the organizations by the founding top management, while seven cases (Avon, DuPont, Jack in the Box, Mattel, Pepsi, Xerox) had cultural practices aimed at gender inclusion. For example, Xerox Corporation, where two female executives—Anne Mulchahy and Ursula Burns— rose to the position of CEO, has a long history of “promoting inclusion and understanding” and cul- tivating a “pluralistic environment” (Allen, 1998) where women are supported and encouraged to rise through the management ranks to hold powerful top positions. PepsiCo’s “strong legacy of leading in di- versity practices” (Pepsi, 2016) is another exemplar; women have been on its board since the 1950s. Moreover, these inclusive organizational practices were actively promoted by the long-tenured and powerful predecessors who were clearly committed to fostering inclusive cultures in all seven of these companies. For example, Steve Reinemund, Indra Nooyi’s predecessor at PepsiCo, acknowledged his commitment to diversity in a post-succession interview:

As CEO, I spent a lot of my time on diversity. It’s a business opportunity and the right thing to do. Di- versity, for a consumer-products company, is obvious to me. We can’t grow as a company if we don’t rep- resent our consumers from the frontline to the boardroom. (Erisman & Beenan, 2007)

The results of the contingency analyses are re- ported in Table 3 and show that the “handing over the legacy” recipe for high post-succession perfor- mance tended to occur in firms with women repre- sented on the board of directors (configuration 1, panel 1), in male-dominated industries (configura- tion 1, panel 2), and among large firms (configuration 1, panel 3). Perhaps most interesting are the results with respect to male-dominated industries: panel 2 shows that the “handing over the legacy” recipe (configuration 1)—the one configuration in which power is passed from the predecessor to the successor—is the only configuration that occurred in male-dominated industries. Moreover, the results sug- gest that structural mechanisms (i.e., founders or orga- nizational programs that promote gender inclusion) are particularly important for buffering the effects of global male-typed schema in male-dominated industry set- tings. Furthermore, which of these mechanisms were at work was contingent on firm size: organizational programs promoting gender inclusion occurred in large firms (configuration 1, panel 3), whereas all but one of the five cases in which founders supported gender inclusion were smaller firms. The findings with respect to women on the board (configuration 1, panel 1) revealed that seven large firms with gender inclusive programs had at least one additional woman beyond the female CEO, while three of the organizations with inclusive founders did.

Partnering the legacy. Configuration 2 (panel 1, Table 2) captures a second path to women CEO success: while here again the women successors were long-term insiders (13 years on average) who inherited high performing firms from long-tenured predecessors (eight years as CEO on average; three of the seven cases were firm founders), this second scenario differed from the first configuration with respect to power. Here, neither the male predeces- sors nor the successors were chairpersons of the board while they served as the CEO. Nevertheless, this suggests that a match in the predecessors’ and successors’ dual statuses (i.e., the lack of) was again an ingredient for success.

Our deeper qualitative analysis on the cases mak- ing up this configuration was again revealing. First, similar to configuration 1, a common mechanism found to be underlying all of the “partnering the legacy” cases was the mentoring and sponsorship by the predecessor of the female executives prior to their ascendance to the CEO role. To pick just one example, Kimberly McWaters, the CEO of Universal Technical Institute (UTI), credits her predecessor, Robert Hartman, as playing a critical role in her

392 AprilAcademy of Management Journal

development as a leader. Early in her career, Hartman actively advised McWaters to develop her management and leadership skills, including encouraging her to pursue an undergraduate de- gree in business.

Second, and unlike in configuration 1, we found an experience-profile match between the predecessor and successorand a similarity in the career pathways as a common structural mechanism across these cases. As just described above, Kimberly McWaters of UTI held several different cross-functional roles, within UTI before leading the firm, that were a mirror image of her predecessor’s rise to CEO. Similarly, Tamara Lundgren, CEO of Schnitzer Steel Industries (SSI), matched predecessor John D. Carter’s profile: both Lundgren and Carter had legal educational backgrounds from elite universities, were partners in law firms in the initial phases of their careers, and had eventually accumulated functional experience in consulting, finance, and management. Given that the predecessor CEOs were long-tenured, this finding is consistent with the notion that the pre- decessors forged expertise and skills based blue- prints on the role, and that the female successors’ matching these role imprints was an important structural mechanism underlying the subsequent success of their female successors.

Our qualitative analyses also uncovered an addi- tional mechanism that differed from the clear absence of the predecessors’ post succession influence in configuration 1: the female successors in this second configuration led their firms in a post-succession partnership (hence “partnering the legacy”), a part- nership which occurred in large part with the pre- decessors. Indeed, the data revealed that in five of the seven cases that made up this configuration, the pre- decessor stayed on the board and played a guiding role post-succession (three became chairman of the board post-succession; two were firm founding members and remained as directors). For exam- ple, when Tamara L. Lundgren succeeded John D. Carter as CEO of SSI, the succession announce- ment emphasized “the complementary combination of Mr. Carter and Ms. Lundgren.” Similarly, in the press release announcing the succession of AMN Healthcare CEO Susan Salka, co-founder and predecessor Steven C. Francis said, “I am very confident in Susan’s ability to lead AMN Healthcare with its growth healthcare staffing industry. I look forward to my ongoing role in providing guidance and oversight to the company.” In the remaining two cases where the predecessors completely exited the firm in the post-succession period, the female successor CEOs—Linda Huett at

Weight Watchers and Stephanie DiMarco at Advent Software—led their firms in partnership with their powerful board chair counterparts.

The results of the contingency analyses reported in Table 3 again shed light on the contexts in which this configuration was embedded. Specifically, the “partnering the legacy” scenario occurred in non- male-dominated industries (configuration 2, panel 2), among smaller firms (configuration 3, panel 3), and regardless of whether there were women on the board of directors (configurations 2a and 2b, panel 1). Thus, the contingency results suggest that while the mentoring mechanism is not contingent upon the embedded context, the predecessor’s influence on female success in smaller firms and in relatively gender-integrated industry settings (such as apparel and accessories, educational services) where global male-typed schema are less likely to be prevalent involved an additional agentic (partnering post- succession) and a different structural (expertise and skills blueprint for the role) mechanism. Moreover, while the “power” role imprint still operated here, the embedding context affected its operation: in gender-integrated industries, this structural mecha- nism involved not having the power of the chair.

Turning around the legacy. The third combina- tion (configuration 3, panel 1, Table 2) for high post-succession firm performance involved powerful long-tenured predecessors (all dual CEOs, tenures of 12 years on average, three of the five predecessors were firm founders or founding family members) who turned over poorly performing firms to their insider female successors (pre-succession tenures of seven years on average), and all of the predecessors remained as chairs of their boards post-succession.

Ourdeeperanalysisof thecasesin thisconfiguration suggested that two mechanisms drove the pre- decessors’ influence on the incoming female executives’ successes: predecessor pre-succession mentoring and post succession partnering. Indeed, our analyses revealed that the female successors were essentially protégés of their predecessors both before and after succession. Two cases exemplify this sce- nario. Rite Aid’s Mary Sammons, whose predecessor, RobertMiller,washerbossandmentorbothatRiteAid and at Fred Meyer beforehand; Miller had served as Sammons’s mentor for over a decade altogether. When Miller handed over the CEO post at Rite Aid to Sam- mons, he proclaimed: “I intend to be an active chair- man, assisting with strategic planning and working with Mary and the senior management team.” Another example is Angela Braly who, before taking over the CEO post, served as president of WellPoint Inc. under

2018 393Dwivedi, Joshi, and Misangyi

her predecessor, Larry C. Glasscock. In the succession announcement, Glasscock said:

Angela has been one of my most trusted and valued colleagues, partnering with me on literally every major strategic initiative undertaken by the company [. . .] In my ongoing role as WellPoint’s Chairman, I look forward to continuing to work closely with Angela. (WellPoint, 2007)

The contingency analyses then suggest that the only contextual condition applicable to the “turning around the legacy” scenario has to do with the size of the firm—that is, this configuration occurred in larger firms (see Table 3). This scenario did not ap- pear among the sufficient configurations with re- spect to women on the board or male-dominated industry.

Supplemental Analysis 1: Women CEOs’ Not-High Performance

We examined what, if any, configurations of the main attributes were sufficient for women’s “not- high” performance. The findings are reported in panel 2 of Table 2. Configurations 1a and 1b—which are “neutral permutations” of each other given that they share the same central conditions and differ in their contributing conditions (Fiss, 2011)—centrally involved outsider female CEO successors who inherited poor performance from predecessors that exited the firm and the board post-succession— indeed, half of the predecessors were fired. In 10 of the 12 cases across these two configurations, the predecessors were long-tenured. Interestingly, config- uration 1a provides a direct contrast to the “turning around the legacy” scenario for success found in our main analysis as it involves female CEOs who inherited unfavorable performance. Moreover, it is worth noting that the findings also show that this failure condition occurs regardless of the female successors being appointed as the board chair (five of the eight cases in configuration 1a held such power). Configuration 2 (panel 2, Table 2) further shows that outsider female successors (who lacked mentoring or partnering) failed regardless of the inherited performance conditions. In three of the five cases, the female successor inherited unfavorable performance (and in two of these the pre- decessor was fired), in the other two cases the successor inherited favorable performance.

In short, the findings of this first supplemental analysis show that female CEOs did not achieve high performance when they were outsiders who were not afforded the mentoring or partnering

relationship with their predecessors before or after becoming CEO.

Supplemental Analysis 2: Configurations Sufficient for Men CEOs’ Performance

We sought to understand how the recipes for success among women compare to those for men. Therefore, we conducted sufficiency analyses of the main attributes under study for both high and not- high performance on a matched sample of men.3

Descriptively, the male and female successions are similar in most aspects: 56 out of the 80 male CEOs had a non-zero fuzzy-set membership score (as compared to 59 of the 84 women), 40 of which had a membership score above the crossover point (44 for women). In the male sample, the average predeces- sor tenure was 8.5 years (9.3 years for women), 21% of the predecessors were founders or a part of the founding family of their firms (25% for women), 70% of the predecessors were dual CEOs (vs. 69%), and 15% of the male successor CEOs followed a prede- cessor who was fired (vs. 15%). There were, how- ever, some interesting differences between male and female CEOs: 13% of the male cases (39% among the women) involved predecessors with 10 or more years of tenure, 31% of the male cases (vs. 26%) had predecessors that stayed on the board post- succession, 49% (vs. 61%) of the male successors were insiders with an average tenure of 5.4 years (vs. 7.7 years), 21% (vs. 33%) of the male successors were appointed chair, and 36% (vs. 25%) of the male successors inherited poor performance. This last descriptive dispels the notion that women CEOs were generally set up to fail—evidently, men were more vulnerable to this in our sample.

The sufficiency analyses results are reported in Table 4; panel 1 shows the configurations sufficient for high performance and panel 2 reports the solution fornot-highperformance.Startingwiththesolutionfor men’s high performance (panel 1; Table 4), configu- ration 1 (panel 1, Table 4) shows a combination that is essentially the same scenario as the “partnering the legacy” scenario found for women (see configuration 2, panel 1, Table 2), but here the male successors were

3 We identified a matching sample of succession events involving male CEOs which occurred in the same year (and when none occurred, the year before), the same four-digit SIC industry code (and where none, the same three-digit SIC), and in a firm of similar size (as measured by market capitalization). Based upon this matching process, we were able to find complete data for 80 matching men cases.

394 AprilAcademy of Management Journal

outsiders who had stellar reputations for their past leadership experience (and two out of the five cases had prior experience as CEOs) at other companies who replaced highly successful predecessors that remained actively involved in strategic decision making in the post-succession period. Thus, in con- trast to the deep mentoring relationships that con- tributed to women’s success in this scenario, successful men were outsiders who were hand-picked by their predecessors and brought into the firm in high-level executive positions for a relatively short grooming pe- riod prior to succession (i.e., a year or less). Configura- tion 2 (panel 1, Table 4) shows a recipe for men’s success similar to the “handing over the legacy” recipe for women—a powerful long-tenured predecessor handed over board chairmanship and favorable per- formance to the male successor CEO—but here again successamongmenoccurredwhenthemalesuccessors were outsiders. Configuration 3 for men’s success does not have a direct parallel among those configurations found for women, and clearly does not afford the op- portunity for mentoring by the predecessor (given that the predecessor was short-tenured).

The solution for men’s not-high performance further illuminates the differences between male and female CEOs. Configuration 1 of this solution (panel 2, Table 4) suggests that men failed when they inherited a poorly performing firm, and while women were outsiders when failing in this same scenario, both insider and outsidermenfail;indeed,onlythreeofthe10malecases came from the outside. The differences between male and female CEOs is further highlighted by configura- tions 3–4 of the men’s not-high solution: they show that insider male successors also fail to turn their firms around when handed poor performance from long- tenured (and often powerful) predecessors (who often stayed on the board post-succession).

In summary, whereas insider women—who were either handed over the legacy or who partnered the legacy with their predecessors—succeeded when they inherited positive performance legacies from long-tenured predecessors, it was outsider men who succeeded under the same conditions. When it comes to inheriting poor performance, while insider female protégés succeeded in turning around their companies in partnership with their powerful predecessors who

TABLE 4 Configurations Sufficient for Men’s Post-Succession Performance

Notes: Centralconditionsare representedby (presence)and (absence); contributingconditionsby (presence)and (absence). Actual minimum thresholds used in the analyses, respectively: raw consistency 5 .77; .76; PRI consistency 5 .76, .75; a minimum frequency of two cases/configuration was used in both analyses.

2018 395Dwivedi, Joshi, and Misangyi

stayed on the board (“turning around the legacy”), the results show that insider men were unable to turn around poorly performing firms under the same con- ditions. In short, the main mechanisms that appear to underlie the workings of the combinations sufficient for female success were not found to afford high post- succession firm performance to male successors.

Robustness Analyses

Analysis of necessity. Given the foregoing find- ings, we had a particular interest in examining the necessity of the predecessors’ long tenures, pres- ences on boards, or the successors’ origins for ob- serving high post-succession performance. That is, we examined whether one or the other of these at- tributes “must be present for the outcome in question to occur” (Ragin, 2000: 91). Table5 reports the results of this analysis for both women’s and men’s high performance. The results show that none of the attributes under study were necessary alone for post-succession performance, assuming a benchmark consistency of .80 (see the footnote to Table 5 for a more detailed explanation). As shown at the bot- tom of Table 5, however, a necessary condition for women’s success that occurred across all of the

studied cases (i.e., including also those cases that were not part of the sufficiency solutions) was that the predecessor was either long-tenured or stayed on the board post-succession. The results also show that fe- male success amongallof the studied cases depended upon having either a long-tenured predecessor or afemaleinsider.Moreover,theresultsshow thatthese same attributes were not necessary for men’s success.

Analysis of diversity. The sufficiency analysis covered approximately half of the successful female cases, and we therefore sought to investigate the remaining non-sufficient configurations (i.e., those that were not sufficient for high performance) through a di- versity analysis (e.g., see Greckhamer et al., 2008). We report the full details of this analysis in Online Sup- plemental Appendix 2. Briefly, the analysis allowed us to examine the most frequently occurring non-sufficient configurations and we found that there were six such configurations, which differed by just one attribute from one of the three success recipes found in the sufficiency analysis reported above. This analysis thereby showed the impor- tance of each particular ingredient to the respective recipe for success and thus further corroborated the three main recipes of female success found in the sufficiency analyses.

TABLE 5 The Necessity of the Attributes for Post-Succession High ROA

Women High Performance Men High Performance

Consistency z-scorea Consistency z-scorea

Predecessor Tenure_or_Founder 0.80 20.16 0.66 22.79 Predecessor Duality 0.68 22.47 0.72 21.66 Predecessor Presence on Board 0.33 29.12 0.48 26.15 Successor is Insider 0.72 21.70 0.47 26.34 Successor Duality 0.31 29.58 0.19 211.58 Successor Inherited Poor Performance 0.32 29.38 0.41 27.46 Predecessor Tenure_or_Founder _or_ Predecessor

Presence on Board 0.91 1.95* 0.48 26.15

Predecessor Tenure_or_Founder _or_ Successor is Insider

0.93 2.33** 0.82 0.21

*p , .05 **p , .01, one-tailed tests

a We assessed whether the observed consistencies are significantly greater than a pre-specified “benchmark” consistency using proba- bilistic criteria; we used a benchmark of .80. This is accomplished via a z test using the following formula (see Ragin, 2000: 109–115, 227–229):

z 5 ((CO – CB) – 1 / 2n) / sqrt ((CB * (1 – CB)) / n) where CO is the observed consistency, CB is the benchmark consistency (here .80), and n is the number of cases with nonzero membership in

the set of firms with high firm performance (women, n 5 59; men, n 5 56). Thus, this assesses the difference between the observed consistency and the benchmark consistency relative to the standard error of the benchmark (i.e., the latter is represented by the formula’s denominator). The z score can then be evaluatedusing the standardnormal distribution table (and following convention, we use an a of .05 for significance; i.e., z $ 1.65 as a one tailed test is appropriate here). In essence, this use of probabilistic criteria allows for inferences regarding the “quasi-necessity” of the attributes under study (i.e., the particular finding is not due to chance); a .80 benchmark means that the attribute(s) in question is (are) “almost always necessary” for performance (Ragin, 2000: 109).

396 AprilAcademy of Management Journal

Alternative outcome: Women’s post-succession tenure. As noted above, post-succession tenure length in the new role has been considered as a meter of suc- cess in studies examining the influence of predecessors on successors (e.g., Burton & Beckman, 2007) and we thusalsoexaminedtherobustnessofourfindingstothis alternativemeasure.Todoso,wecapturedthetenureof the female CEOs using a fuzzy-set calibrated similar to that used to capture predecessors’ pre-succession ten- ures: successors with post-succession tenures $ 10 years were coded as fully in, tenures of # 1 year were coded as fully out, and the crossover point was three years. We then examined the sufficiency of the main attributes for observing longer post-succession tenures among women. The findings are discussed in Online Supplemental Appendix 3. In brief, the findings further corroborated our firm performance results and pro- vided additional insights into female CEOs’ success. The configurations found to be sufficient for observing longer tenures among women represent the “handing over the legacy,” “partnering the legacy,” and “turning around the legacy” recipes and thereby further cement these success scenarios. Further, our findings suggest that while mentoring was a crucial mechanism for fe- male CEOs’ post-succession performance success, it was not integral to their longer tenures.

DISCUSSION

Past research on female executives has often relied on the assumption that global male-typed leadership schemas uniformly impose barriers on the success of women entering into leadership roles. While immensely valuable, this past lens has deflected scholarly attention away from examining the proxi- mal influence of organizational agents and the local context on either reinforcing or mitigating the effect of male-typed schema on women leaders’ effective- ness. We integrated past research on female leader- ship, executive succession, and male gatekeeping to investigate how several attributes of the (mostly) male predecessor CEOs combined with key facets of the succession context to jointly determine women CEOs’ subsequent performance. Our exploratory qualitative comparative case study of all the female CEO succession events across the largest U.S. firms over 20 years uncovered three different combina- tions of the studied theoretical attributes sufficient for the success of female CEO successors. In essence, our findings show that certain enabling factors and the local context set the conditions for gender- inclusive gatekeeping wherein male predecessors facilitated female leadership effectiveness.

Overall, our findings clearly suggest that the long tenure of the male predecessor and the insider origin of the female successor are both essential enablers of female success—not only are both of these condi- tions part of all three sufficient recipes for female success, but our analyses revealed their necessity: one or the other of these conditions was always present across all of the empirically-occurring con- figurations (sufficient or not) of women’s success. Our results clearly show that these two essential conditions enabled gender-inclusive gatekeeping by combining with the local governance structure and the embedding context to form three different rec- ipes for success. Two of these success recipes in- volved a favorable performance legacy in which the long-tenured predecessor either handed over to, or partnered with, an insider female successor. In the third scenario, the long-tenured predecessor guided the insider female successor to turnaround an unfavor- able performance legacy. Moreover, our findings re- garding female “not-high” performance reinforced the importance of women’s insider origins: failure occurred when women where outsiders who inherited unfavor- able performance from long-tenured predecessors.

Our deeper analyses of the sufficient cases revealed several mechanisms underlying these recipes for success. One mechanism was common across all three recipes: the pre-succession mentoring and sponsorship of the insider female successors by the long-term predecessors. In addition to this rather agenticmechanism,the successfulcasesrevealed that the local governance structure—the predecessor’s and successor’s power while CEO as well as the pre- decessor’s post-succession board presence—was an- other key enabling condition. In the “handing over the legacy” recipe, the mentoring and sponsorship mechanism appeared to work in conjunction with two other mechanisms that were more structural in nature: the successors’ appointments as chair of the board fulfilled the predecessor’s “power” blueprint on the CEO role, and they rose in firms founded by leaders that promoted inclusiveness or that had well- established diversity programs.

The “partnering the legacy” recipe involved the mentoring and sponsorship mechanism in conjunc- tion with another agentic mechanism: the insider fe- male successors worked in partnerships with the chairmen of their boards (who most often, but not al- ways, were the predecessors) post-succession. The structuralmechanismsalsoatworkheredifferentiated this second success recipe from the first: the succes- sors had background experiences that mirrored those of their predecessors and thereby appeared to fulfill

2018 397Dwivedi, Joshi, and Misangyi

an experience blueprint left by the long-tenured pre- decessors on the CEO role. Moreover, the successors’ non-duality matched the predecessors “power” im- print for a non-dual CEO. Finally, the third recipe, “turning around the legacy,” involved two agentic mechanisms: the powerful long-tenured predecessors who remained the chair of the board post-succession mentored their insider successors and partnered with them to help turn around poorly performing firms.

Our analysis of several contingency conditions— the presence of women on the board, the degree to which the industry was male-dominated, and the firm size—provided deeper insight into the three recipes of success by highlighting the important role that the firm’s context plays in facilitating female leaders’ suc- cesses. Most noteworthy is the effect that the local male-dominated nature of the industry in which the firms were embedded had on the recipes for success. Specifically, the “handing over the legacy” recipe oc- curred in male-dominated contexts, wherein global male-typed schema are likely amplified, and involved structural mechanisms that would seem to directly play to or help mitigate global male-typed schemas— the female CEOs’ were given power (to match the predecessor CEOs’ power blueprint) and rose through the ranks of organizations with founders or practices that advanced gender inclusion. The “partnering the legacy” recipe, on the other hand, tended to occur in industries with a relatively higher proportion of women executives—and thus an embedding context inwhichmale-typedschemaarealreadyblunted—and involved structural mechanisms that helped to meet local expectations set by the predecessors’ experience profile (in the form of a role blueprint) and non-dual status as a CEO. Moreover, partnering with the chair- person was a mechanism for success in such contexts.

Of our several supplemental analyses, we highlight here the findings of the sufficiency analysis of a matched sample of male CEO succession events which occurred in firms of similar size in the same industriesand inthe samesuccession year asthe focal female CEO successions. Our findings showed that the configurations of men’s success and the mecha- nisms that underlie them differed markedly from women’s success configurations. Successful male CEOs were outsiders who received a brief period of pre-succession grooming from the predecessor. The “not-high” performance configurations also differed: male successors were not able to turn poorly per- forming companies around when they followed long- tenured predecessors, regardless of the predecessors’ presence on the board and the male successors’ ori- gins (both insiders and outsiders failed). Overall,

across the studied succession events, these compar- ative findings between men and women show that the mechanisms that have allowed women CEOs to suc- cessfully lead S&P 1500 and Fortune 500 firms are not the same for men under similar circumstances.

Toward a Theory of Gender-Inclusive Gatekeeping for Women’s Leadership Effectiveness

Our initial theoretical framing positioned predecessors as key organizational agents who, functioning as inclusionary (exclusionary) gate- keepers, could shape the initial conditions of success (failure) for women entering leadership roles. The role of predecessors in the success of incumbents has been recognized across many disparate streams of research (e.g., Burton & Beckman, 2007; Joshi et al., 2010, 2011; Wade-Benzoni et al., 2008). In these previous accounts, the predecessor has been viewed as a temporally distal albeit important influence on the success of incumbents. Our findings contribute to this domain by highlighting the more proximal effects of predecessors and go much farther in uncovering the conditions under which mostly male predecessors engage in “gender-inclusive gatekeeping” that can buffer the effects of pervasive male-typed leadership schema on women’s success in leadership roles. Whereas past conceptualizations of gatekeeping have typically referred to exclusionary actions taken by in- dividuals or groups in prestigious occupations or the upper ranks of organizations (Merton, 1957), including how men have restricted women’s access to rewards in male-dominated professions (e.g., Briscoe & Joshi, 2016; Reskin & Padavic, 1988), our research high- lights conditions under which male-gatekeepers can also play an inclusive role.

Male predecessors in our sample did not always play an inclusive gate-keeping role. Indeed, our in- quiry revealed that inclusive gatekeeping occurred only when a confluence of local enabling conditions allowed the predecessor to support the female ex- ecutive’s entry into the CEO role, forming multiple recipes for female success. Based on our findings, we offer a mid-range theory of gender-inclusive gate- keeping for female successors’ effectiveness in the upper echelons: male predecessors serve as critical gatekeepers when particular local enabling condi- tions and the embedding context facilitate the suc- cess of female leaders through both agentic and structural mechanisms that operate to alter local leadership schemas (see Figure 1).

As Figure 1 shows, the predecessor’s legacy— reflected in their long tenure or founding status

398 AprilAcademy of Management Journal

as well as the performance conditions they leave behind—and the successor’s insider origin are two essential ingredients that together enable gender- inclusive gatekeeping in the upper echelons of or- ganizations. As we elaborate below, these conditions together foster the agentic and structural mecha- nisms through which the predecessors may alter lo- cal leadership schemas. The predecessor’s legacy, particularly his long tenure or founding status, al- lows him to imprint normative expectations on the role as well as to shape the inclusive diversity prac- tices of the firm. It also gives the predecessor an opportunity to endorse and mentor the female successor as she rises through the ranks of the orga- nization. The successor’s insiderness enables her to benefit from the predecessor’s endorsements, to gain role-specific and idiosyncratic skills from the pre- decessor as her mentor, and to generally benefit from inclusive organizational practices.

Although our theory is applicable to the highest ranks in firms, we nevertheless expect these two enablers of gender-inclusive gatekeeping—the pre- decessor’s legacy and successor’s insider origin—to generalize to other organizational settings, such as lower level organizational leadership transitions or to other professions where women transition into high visibility leadership roles. In these other con- texts, the predecessor’s legacy, such as the resource benefits or burdens they leave behind, or the successor’s tenure and prior working relationship with the predecessor are likely to influence women’s success in leadership roles as well. Furthermore, we call for future research to explore other local en- abling conditions beyond the factors that we

examined. With respect to the predecessor, it seems reasonable to expect that other attributes (prestige, reputation), individual differences (openness, em- pathy), or leadership types (charismatic or trans- formational) may lend to more (or less) effective gender-inclusive gatekeeping. Similarly, future re- searchcouldconsiderotherattributesof the successor that allow her to take advantage of the inclusive gatekeeping. For instance, in professional settings, does a long-term partnership with the predecessor (in lieu of insider status) through professional net- working associations facilitate gender-inclusive gatekeeping? Moreover, is gender-inclusive gate- keeping more likely infirms that rely on internal labor markets or is it feasible under some conditions in firms that also draw on external hires to increase women’s representation in managerial ranks?

We also propose that other local conditions are critical to gender-inclusive gatekeeping, operating as either enablers or contingencies. Our theorization suggests that, at least in the upper echelons, the governance structure—including at a minimum the power ascribed to the CEO role—is an important enabler of the agentic and structural mechanisms that affect local perceptions of leadership (see Figure 1). Corporate governance structures help to define the normative expectations regarding local expectations of the power held while being CEO as well as determine whether the predecessor is present post-succession to partner with the successor. The beneficent view of predecessors’ post-succession presence uncovered by our findings departs from past research that has viewed this presence as a con- straintonanincomingexecutive(Quigley&Hambrick,

FIGURE 1 A Mid-Range Theory of Gender-Inclusive Gatekeeping

Gender-Inclusive Gatekeeping

Local Enablers

Male Predecessor

Predecessor Legacy

Successor Inside Origin

Governance

Embedding Context

Agentic Mechanisms -Endorsing -Partnering

Strucural Mechanisms

-Role imprints -Inclusive practices

Alter Local

Stakeholders’ Leadership Schemas

Female Successor

Effectiveness

2018 399Dwivedi, Joshi, and Misangyi

2012). In contrast, we posit that for female executives, the predecessors’ continued presence can provide the means for an active partnership that empowers an in- comingfemaleleadertoperform effectivelyintheCEO role and, in some cases, to even turnaround poor per- formance. While we believe that the governance structure is a key construct in a theory of gender- inclusive gatekeeping, this enabling condition may manifest itself in different forms depending upon the organizational context. For instance, future research at lower organizational levels should consider how a firm’s performance management and accountability structures shape the normative expectations associ- ated with the particular role with implications for the successor’s effectiveness. Furthermore, in lower level leader transitions, the predecessor may continue on as the successor’s immediate boss, a relationship which would thereby further enable both the endorsing and partnering agentic mechanisms found to operate in the upper echelons.

Our findings also show that one facet of the embed- ding context of the firm can play a critical role in whether gender-inclusive agentic or structural mecha- nisms enable women leader’s effectiveness: the degree to which the industry is male-dominated. In highly male dominated environments, intervening structural mechanisms—that is, organizational practices that promote inclusion—were influential in altering local firmleadershipschemas,whileinnon-maledominated contextsagenticmechanisms— thatis,post-succession partnering with the predecessor—were integral to women’s success. Given that male dominated settings may amplify male-typed leadership schemas in many differentorganizationalcontexts,includinglower-level leadership transitions, our finding that intervening structural mechanisms that alter male-typed schemas are even more important for women’s success in these local male-dominated settings is striking. We recognize that other facets of the local context beyond male- dominated environments could also shape structural interventions aimed at gender-inclusive gatekeeping in similar ways. For instance, other institutional and regulatory pressures faced by firms within an in- dustry segment may shape the nature of structural interventions they adopt to enable gender inclusion at the highest levels. Thus, we call for further re- search on other attributes of the embedding context of firms that can function as contingencies shaping gender-inclusive gatekeeping by key organizational agents such as predecessors.

Astheforegoingalreadyhighlights,gender-inclusive gatekeeping operates through both agentic and struc- tural mechanisms that alter local male-leadership

schemas and facilitate women’s effectiveness in leadership roles (Figure 1). Again, the combination of male predecessor’s legacy and the female succes- sor’s insider origins allow the predecessor to engage in the mentoring and endorsing (i.e., an agentic mechanism) behaviors that enable women’s success. Furthermore, the local governance structures can enable the agentic mechanism of partnering through the predecessor’s post-succession presence on the board, which allows the predecessor to continue to promote women’s success.

Our theorization also proposes that the structural component of gender-inclusive gatekeeping has two facets: role imprints and inclusive organizational practices. Our theorizing about role imprints builds on past research that has pointed to the lack of the successor’s fit with the expectations set by the pre- decessor as an antecedent of turnover from the role (Burton & Beckman, 2007). Whereas past research has viewed these imprints as constraints on in- cumbents, we propose that the predecessors’ im- prints on the role can also be beneficial for women successors. In particular, our findings suggest that when women fit the “local” expectations set by their male predecessors in terms of their functional back- ground and on other dimensions, their lack of fit with the “global” male-typed leadership schemas appeared to be less salient in the eyes of key organizational stakeholders. Moreover, our finding that “power im- prints” matter to the effectiveness of women succes- sors has important implications for future research: it suggests that researchers need to look beyond task and functional role imprints to other facets of the role that may set the local expectations associated with that particular role and thus become critical for the success of incoming executives.

Finally, although our theorizing focuses on women’s success in the upper echelons, our findings underscore the inherently gendered nature of this rarefied context and behoove future research to further problematize men’s roles as leaders. Consider that the conditions that enabled women’s success— insiderness and predecessors’ long-tenures and in- volvement in post-succession governance—did not enable men’s success. We suggest that researchers apply a critical lens on men’s roles such as the “pre- carious manhood theory” which suggests that, like women, men also attempt to subscribe to socially constructed notions of “being male” with implications for their success (failure) in leadership roles. Through this lens, which views “manhood” as a status that is both “hard won and easily lost,” research could frame leadership transition events in masculine settings as

400 AprilAcademy of Management Journal

arenas for one-upping and dominance displays that can shape outcomes for successors (Schrock & Schwalbe, 2009; Vandello & Bosson, 2013). In the context of the upper echelons, this theory would suggest that the (mostly) male predecessors’ agentic pre-succession endorsing and post-succession part- nering with an incoming male successor might ap- pear to be an affront to the male successor’s efforts to establish dominance—and explain why insider males failed under the same conditions in which female insiders succeeded (e.g., the “turning around the legacy” recipe). In line with this thinking, we attribute the one-upping mechanism as a possible explanation for a crucial ingredient we found for men’s success: being an outsider. As previous exec- utive succession literature suggests, the outsider status of a successor inherently carries expectations for and represents organizational change. It provides the male successor an opportunity to readily distin- guish himself from his male predecessor and allows him to don a dominant leader persona. This provides a potential explanation for why, somewhat surpris- ingly, even when the inherited predecessor legacy conditions are favorable—a situation conventionally thought to call for insiders—we found that male outsiders are more successful. Future research into the consequences of subscribing to male-typed leader schemas to explain variability among males in leadership roles is clearly warranted.

Overall, our study represents an important next step in the development of an actionable theory of women’s success in top leadership roles by shedding new light on how local firm histories and complexities shape the conditions that enable gender-inclusive gatekeeping by key organizational agents such as pre- decessors. Indeed, our study is among the first to high- lighthowvariousaspectsofthefirmcontextcombineto determine variability among women executives. This approach is a substantive departure from past research that has primarily drawn from role congruity theory or implicit theories of leadership to explain performance differences yielding mixed findings (e.g., Hoobler et al., 2016). These past approaches emphasize sex differ- ences in leadership effectiveness, and while analyti- cally expedient, these predominantly regression-based approaches have severely limited researchers’ prowess to detect and fully unpack gender effects when, in fact, gender may have many complex implications in the upper echelons (Ely & Padavic, 2007; Martin, 1994). Given the steady rise of women to the highest levels in firms, our study suggests an urgent need to continue to test the boundary conditions of past theoretical frameworks in the domain of executive successions

and the upper echelons context more broadly. We propose that future research in this domain explicitly recognizethattheseframeworkshavebeendeveloped based on overwhelmingly male samples and may not be transferable to women who are more likely to oc- cupy CEO roles in the future (Strategy &, 2013).

CONCLUSION

In conclusion, our study focuses attention on how organizational agents and local contexts jointly create the conditions for gender-inclusive gatekeeping that enables women’s success in a setting where they have been historically underrepresented. Based on detailed case analyses of all CEO succession events involving women in large U.S. firms over two decades, our findings show the conditions under which (mostly male) predecessor CEOs can be a formidable force that facilitates women CEOs’ subsequent success. Based on the findings, we theorize that not all predecessors are gender inclusive: gender-inclusive gatekeeping by male predecessors occurs only when a confluence of local factors enables both agentic and structural mechanisms that mitigate chronic and pervasive sex- role stereotypes and expectations. Broadly speaking, the mainimplication of ourstudy isthat organizational agents—and, in particular, male predecessors—can potentiallyhelpinalteringlocal contextstomakethem moreinclusive.Tobeclear,thesefindingsdonotimply that women leaders are reliant solely on male pre- decessors for their success. Their success is obviously a function of their own skills, ability, and motivation. The findings do, however, highlight the critical role that male predecessors play in women’s transition into leader roles and place these men front and center in a mandate for greater gender inclusion at the highest levels in organizations.

REFERENCES

Acker, J. 1992. Gendering organizational theory. In A. J. Mills & P. Tancred (Eds.), Gendering organiza- tional theory. Thousand Oaks, CA: Sage.

Allen, S. 1998. Diversity—A passage to the future at Xerox. Retrieved from https://clinton4.nara.gov

Aversa, P., Furnari, S., & Haefliger, S. 2015. Business model configurations and performance: A qualitative comparative analysis inFormula One racing, 2005–2013. Industrial and Corporate Change, 24: 655–676.

Briscoe, F., & Joshi, A. 2016. Bringing the boss’s politics. Supervisor political ideology and the gender gap in earnings. Academy of Management Journal, 60: 1415–1441.

2018 401Dwivedi, Joshi, and Misangyi

Burton, M. D., & Beckman, C. M. 2007. Leaving a legacy: Position imprints and successor turnover in young firms. American Sociological Review, 72: 239– 266.

Cannella, A. A., Park, J. H., & Lee, H. U. 2008. Top man- agement team functional background diversity and firm performance: Examining the roles of team mem- ber colocation and environmental uncertainty. Aca- demy of Management Journal, 51: 768–784.

Cannella, A. A., & Shen, W. 2001. So close and yet so far: Promotion versus exit for CEO heirs apparent. Acad- emy of Management Journal, 44: 252–270.

Castanias, R. P., & Helfat, C. E. 2001. The managerial rents model: Theory and empirical analysis. Journal of Management, 27: 661–678.

Chen, G., & Hambrick, D. C. 2012. CEO replacement in turnaround situations: Executive (mis) fit and its performance implications. Organization Science, 23: 225–243.

Connell, R. 2005. Change among the gatekeepers: Men, masculinities, and gender equality in the global arena. Signs, 30: 1801–1825.

Crilly, D. 2011. Predicting stakeholder orientation in the multinational enterprise: A mid-range theory. Journal of International Business Studies, 42: 694–717.

Crilly, D., Zollo, M., & Hansen, M. 2012. Faking it or muddling through? Understanding decoupling in re- sponse to stakeholder pressures. Academy of Man- agement Journal, 55: 1429–1448.

Davis, P. S., Babakus, E., Englis, P. D., & Pett, T. 2010. The influence of CEO gender on market orientation and performance in service small and medium sized ser- vice businesses. Journal of Small Business Manage- ment, 48: 475–496.

Dezsö, C. L., & Ross, D. G. 2012. Does female representation in top management improve firm performance? A panel data investigation. Strategic Management Jour- nal, 33: 1072–1089.

Dixon Fowler, H. R., Ellstrand, A. E., & Johnson, J. L. 2013. Strength in numbers or guilt by association? Intragroup effects of female chief executive announce- ments. Strategic Management Journal, 34: 1488–1501.

Eagly, A. H., & Karau, S. J. 2002. Role congruity theory of prejudice toward female leaders. Psychological Re- view, 109: 573–598.

Eagly, A.H., Wood, W., & Diekman, A.B., 2000. Social role theory of sex differences and similarities: A current appraisal. In T. Eckes, & H. M. Trautner (Eds.), The developmental social psychology of gender: 123–174. Mahwah, NJ: Erlbaum.

Ely, R., & Padavic, I. 2007. A feminist analysis of organi- zational research on sex differences. Academy of Management Review, 32: 1121–1143.

Erisman, A., & Beenan, G. 2007. Steven S. Reinemund: Leadershipfora21st-centurymultinationalcorporation. Retrieved from http://ethix.org/2007/04/01/leadership- for-a-21st-century-multinational-corporation

Finkelstein, S., & D’Aveni, R. A. 1994. CEO duality as a double-edged sword: How boards of directors’ bal- ance entrenchment avoidance and unity of command. Academy of Management Journal, 37: 1079–1108.

Finkelstein, S., Hambrick, D. C., & Cannella, A. A. 2009. Strategic leadership: Theory and research on ex- ecutives, top management teams, and boards. Ox- ford, U.K.: Oxford University Press.

Fiss, P. C. 2011. Building better causal theories: A fuzzy set approach to typologies in organizational research. Academy of Management Journal, 54: 393–420.

Fredrickson, J. W., Hambrick, D. C., & Baumrin, S. 1988. A model of CEO dismissal. Academy of Management Review, 13: 255–270.

Grandori, A., & Furnari, S. 2008. A chemistry of organiza- tion: Combinatory structural analysis and design. Organization Studies, 29: 315–341.

Greckhamer, T., Misangyi, V. F., Elms, H., & Lacey, R. 2008. Using qualitative comparative analysis in strategic management research: An examination of combina- tions of industry, corporate, and business-unit effects. Organizational Research Methods, 11: 695–726.

Greckhamer, T., Misangyi, V. F., & Fiss, P. C. 2013. The two QCAs: From a small-N to a large-N set theoretic ap- proach. In P. C. Fiss, B. Cambré & A. Marx (Eds.), Configurational theory and methods in organiza- tional research (Research in the sociology of orga- nizations), vol. 38: 49–75. Bingley, U.K.: Emerald Group Publishing.

Green, F. 2006. Jack in the Box is jumpin’. Retrieved from http://legacy.sandiegouniontribune.com

Hambrick, D. C., & Finkelstein, S. 1987. Managerial discretion: Abridgebetweenpolarviewsoforganizationaloutcomes. Research in Organizational Behavior: 9, 369–406.

Hambrick, D. C., & Fukutomi, G. D. 1991. The seasons of a CEO’s tenure. Academy of Management Review, 16: 719–742.

Hansen, G. S., & Wernerfelt, B. 1989. Determinants of firm performance: The relative importance of economic and organizational factors. Strategic Management Journal, 10: 399–411.

Haslam, S. A., & Ryan, M. K. 2008. The road to the glass cliff: Differences in the perceived suitability of men and women for leadership positions in succeeding and failing organizations. The Leadership Quarterly, 19: 530–546.

Haslam, S. A., Ryan, M. K., Kulich, C., Trojanowski, G., & Atkins, C. 2010. Investing with prejudice: The re- lationship between women’s presence on company

402 AprilAcademy of Management Journal

boards and objective and subjective measures of com- pany performance. British Journal of Management, 21: 484–497.

Higgins, M. C. 2005. Career imprints: Creating leaders across an industry, vol. 16. San Francisco, CA: John Wiley and Sons.

Hoobler, J. M., Masterson, C. R., Nkomo, S. M., & Michel, E. J. 2016. The business case for women leaders: Meta-analysis, research critique, and path forward. Journal of Management. Retrieved from https://doi.org.ezproxy.library.tamu.edu/10.1177/ 0149206316628643

Jeong, S. H., & Harrison, D. 2016. Glass breaking, strategy making, and value creating: Meta-analytic outcomes of females as CEOs and TMT members. Academy of Management Journal, 60: 1219–1252.

Joshi, A. 2014. By whom and when is women’s expertise recognized? The interactive effects of gender and ed- ucation in science and engineering teams. Adminis- trative Science Quarterly, 59: 202–239.

Joshi, A., Dencker, J. C., & Franz, G. 2011. Generations in organizations. Researchin Organizational Behavior, 31: 177–205.

Joshi, A., Dencker, J. C., Franz, G., & Martocchio, J. J. 2010. Unpacking generational identities in organizations. Academy of Management Review, 35: 392–414.

Joshi, A., Son, J., & Roh, H. 2015. When can women close the gap? A meta-analytic test of sex differences in performance and rewards. Academy of Management Journal, 58: 1516–1545.

Kanter, R. M. 1977. Men and women of the corporation, vol. 5049. New York, NY: Basic Books.

Karaevli, A. 2007. Performance consequences of new CEO “outsiderness”: Moderating effects of pre and post succession contexts. Strategic Management Journal, 28: 681–706.

Karaevli, A., & Zajac, E. J. 2013. When do outsider CEOs generate strategic change? The enabling role of cor- porate stability. Journal of Management Studies, 50: 1267–1294.

Khan, W. A., & Vieito, J. P. 2013. CEO gender and firm perfor- mance. Journal of Economics and Business, 67: 55–66.

Koenig, A. M., Eagly, A. H., Mitchell, A. A., & Ristikari, T. 2011. Are leader stereotypes masculine? A meta- analysis of three research paradigms. Psychological Bulletin, 137: 616–642.

Kolev,G.I.2012.UnderperformancebyfemaleCEOs:Amore powerful test. Economics Letters, 117: 436–440.

Lee, P. M., & James, E. H. 2007. She’-e-os: gender effects and investor reactions to the announcements of top exec- utive appointments. Strategic Management Journal, 28: 227–241.

Marquis, C., & Tilcsik, A. 2013. Imprinting: Toward a multilevel theory. The Academy of Management Annals, 7: 195–245.

Martin, J. R. 1994. Methodological essentialism, false dif- ference, and other dangerous traps. Signs (Chicago, Ill.), 19: 630–657.

Merton, R. 1957. Science and democratic social structure. In R. Merton (Ed.) Social theory and social structure: 307–316. Glencoe, IL: The Free Press.

Miller, D. 1991. Stale in the saddle: CEO tenure and the match between organization and environment. Man- agement Science, 37: 34–52.

Misangyi, V. F., & Acharya, A. G. 2014. Substitutes or complements? A configurational examination of cor- porate governance mechanisms. Academy of Man- agement Journal, 57: 1681–1705.

Misangyi, V. F., Greckhamer, T., Furnari, S., Fiss, P. C., Crilly, D., & Aguilera, R. 2017. Embracing causal complexity: The emergence of a neo- configurational perspective. Journal of Manage- ment, 43: 255–282.

Nelson, D. L., & Quick, J. C. 1985. Professional women: Are distress and disease inevitable? Academy of Man- agement Review, 10: 206–218.

Park,S.H., &Westphal,J.D.2013. Socialdiscriminationinthe corporate elite: How status affects the propensity for mi- nority CEOs to receive blame for low firm performance. Administrative Science Quarterly, 58: 542–586.

Paustian-Underdahl, S. C., Walker, L. S., & Woehr, D. J. 2014. Gender and perceptions of leadership effectiveness: A meta-analysis of contextual moderators. The Journal of Applied Psychology; Advance online publication.

Pepsi. 2016. Diversity & engagement. Retrieved from http://www.pepsico.com

Quigley, T. J., & Hambrick, D. C. 2012. When the former CEO stays on as board chair: Effects on successor discretion, strategic change, and performance. Stra- tegic Management Journal, 33: 834–859.

Ragin, C. C. 2000. Fuzzy-set social science. Chicago, IL: University of Chicago Press.

Ragin, C. C. 2006. Set relations in social research: Evalu- ating their consistency and coverage. Political Anal- ysis, 14: 291–310.

Ragin,C.C.2008.Redesigningsocialinquiry:Fuzzysetsand beyond. Chicago, IL: University of Chicago Press.

Reskin, B. F., & Padavic, I. 1988. Supervisors as gate- keepers: Male supervisors’ response to women’s in- tegration in plant jobs. Social Problems, 35: 536–550.

Rihoux, B., & Ragin, C. C. 2009. Configurational compar- ative methods: Qualitative comparative analysis (QCA) and related techniques. Atlanta, GA: Sage.

2018 403Dwivedi, Joshi, and Misangyi

Rosette, A. S., & Tost, L. P. 2010. Agentic women and communal leadership: How role prescriptions confer advantage to top women leaders. The Journal of Ap- plied Psychology, 95: 221–235.

Ryan, M. K., & Haslam, S. A. 2005. The glass cliff: Evidence that women are over-represented in precarious lead- ership positions. British Journal of Management, 16: 81–90.

Ryan, M. K., & Haslam, S. A. 2007. The glass cliff: Exploring the dynamics surrounding the appointment of women to precarious leadership positions. Academy of Management Review, 32: 549–572.

Ryan,M.K.,Haslam,S.A.,Hersby,M.D.,&Bongiorno,R.2011. Think crisis–think female: The glass cliff and contextual variation in the think manager–think male stereotype. The Journal of Applied Psychology, 96: 470–484.

Schein, V. E. 1973. The relationship between sex role ste- reotypes and requisite management characteristics. The Journal of Applied Psychology, 57: 95–100.

Schein, V. E. 1977. Individual power and political behav- iors in organizations: An inadequately explored re- ality. Academy of Management Review, 2: 64–72.

Schein, V. E. 2007. Women in management: Reflections and projections. Women in Management Review, 22: 6–18.

Schrock, D., & Schwalbe, M. 2009. Men, masculinity, and manhoodacts. AnnualReviewofSociology,35:277–295.

Shen, W., & Cannella, A. A. 2002. Revisiting the perfor- mance consequences of CEO succession: The impacts of successor type, postsuccession senior executive turnover, and departing CEO tenure. Academy of Management Journal, 45: 717–733.

Smith, C. 2015. Retired PepsiCo CEO: Leadership “starts with understanding yourself”. Retrieved from http://www.charlotteobserver.com/news/business/ article50299410.html.

Strategy &. 2013. The 2013 Chief Executive Study. PWC. Retrieved from https://www.strategyand.pwc.com/ media/file/Strategyand_The-2013-Chief-Executive- Study.pdf.

TheConferenceBoard.2014.DepartingCEOTenure.Retrieved from https://www.conference-board.org/retrievefile.cfm? filename5tcb_cw-0561.pdf&type5subsite.

Vandello, J. A., & Bosson, J. K. 2013. Hard won and easily lost: A review and synthesis of theory and research on precarious manhood. Psychology of Men & Mascu- linity, 14: 101.

Wade-Benzoni, K. A. 2002. A golden rule over time: Reciprocity in intergenerational allocation decisions. Academy of Management Journal, 45: 1011–1028.

Wade-Benzoni, K. A., Hernandez, M., Medvec, V., & Messick, D. 2008. In fairness to future generations: The roleofegocentrism,uncertainty,power,andstewardship in judgments of intergenerational allocations. Journal of Experimental Social Psychology, 44: 233–245.

WellPoint. 2007. Report. Retrieved from https://www.sec.gov/ Archives/edgar/data/1156039/000119312507038978/ dex991.htm

Zajac, E. J., & Westphal, J. D. 1996. Who shall succeed? How CEO/board preferences and power affect the choice of new CEOs. Academy of Management Journal, 39: 64–90.

Zajac, E. J., & Westphal, J. D. 2004. The social construction of market value: Institutionalization and learning perspectives on stock market reactions. American Sociological Review, 69: 433–457.

Zhang, Y., & Qu, H. 2016. The impact of CEO succession with gender change on firm performance and succes- sor early departure: Evidence from China’s publicly listed companies in 1997–2010. Academy of Man- agement Journal, 59: 1845–1868.

Zhang, Y., & Rajagopalan, N. 2004. When the known devil is better than an unknown god: An empirical study of the antecedents and consequences of relay CEO successions. Academy of Management Journal, 47: 483–500.

Priyanka Dwivedi ([email protected]) is Assistant Professor of Management at the Mays Business School, Texas A&M University. She received her PhD from the Pennsylvania State University. Her research interests in- clude gender and diversity in the upper echelons, CEO successions, and socio-cognitive and affective mecha- nisms that influence strategic decision-making.

Aparna Joshi ([email protected]) is the Arnold Family Professor of Management at the Smeal College of Business, Penn State University. She received her PhD from the SMLR, Rutgers University. Her research interests include gender inequality, diversity, inter-generational dynamics, and expertise recognition in teams.

Vilmos F. Misangyi ([email protected]) is Professor of Stra- tegic Management, BNY Mellon Fellow, and Chair of the M&O Department in the Smeal College of Business, Penn State University. He received his PhD from the University of Florida. His research focuses on how top executives and organizations influence, and are influenced by, their sur- rounding environments and constituents.

404 AprilAcademy of Management Journal

Copyright of Academy of Management Journal is the property of Academy of Management and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.