Abstract
We examine the reproduction of executive age profiles across chief executive officer (CEO) succession episodes. Counter to the trend of the general workforce becoming more age-diverse, executives appear to have become more age homogenous and increasingly older at appointment. This is despite increasing frequency of CEO succession episodes, which represent opportunities for demographic change at the top. Combining insights from upper echelons theory, homosocial reproduction theory, and executive succession research, we examine age homophily as an underexamined response to appease some of the disruptiveness surrounding CEO succession episodes. Exploiting CEO succession as a theoretical context for change and an empirical identification strategy, we construct a sample of 391 successions in 297 Standard and Poor’s 500 index companies from 2000 to 2020 and apply a hierarchical linear modeling specification to test hypotheses. We find general support for the notion that age profiles of departing CEOs and incumbent top management teams (TMTs) tend to be reproduced across succession episodes. However, some intriguing patterns emerge when accounting for the origin of the incoming CEO. Notably, some of these general tendencies may be reinforced under outsider CEO successors—counter to the usual expectation for upheaval prompted by outside CEO appointments. We discuss implications for theory and practice.
Introduction
Executive leadership of large publicly listed firms is entrusted to a chief executive officer (CEO) and the top management team (TMT) they interface with (Georgakakis et al., 2022; Hambrick and Mason, 1984; Simsek et al., 2018). A longstanding assumption is that executives’ characteristics should adapt to match broader trends in the environment to avoid inertia (Guthrie et al., 1991; White et al., 1997; Wiersema and Bantel, 1993). However, while large corporations operate in a more age-diverse environment than ever before (United Nations, 2015), the age profile of their executives seems to be following a different trend. In fact, executives tasked with leading an increasingly age-diverse workforce and operating in an age-diverse environment seem to be getting older at appointment (Kelly, 2020; Schnoor, 2020). This is intriguing, given compelling evidence that executive age is linked to a wide range of firm outcomes, often related to inertia (Wang et al., 2016). Age, therefore, re-emerges as a crucial demographic characteristic of interest more generally (De Meulenaere et al., 2016; Zenger and Lawrence, 1989), while raising intriguing new questions about why executives “look the way they look” (Hambrick, 2007).
Characteristics of executives are the baseline input for the upper echelons model. While the core legacy of upper echelons theory (UET) lies in predicting outcomes from input characteristics such as age (Carpenter et al., 2004; Hambrick and Mason, 1984), studies rarely address the distributional patterns of these characteristics in the first place. Consider for instance that the average age of CEOs at appointment increased from 45.9 in 2005 to 56.8 in 2019 (Crist Kolder Associates, 2020). Adding to this intrigue, there is the fact that demographic change at the top is often catalyzed by CEO succession episodes (Georgakakis et al., 2022; Liu et al., 2018b), which have become an increasingly frequent occurrence (Berns and Klarner, 2017). Thus, while there should be ample opportunities to diversify executive age profiles to stave off inertia (White et al., 1997), the question remains: Why does the distribution of executive characteristics, notably age, persistently skew toward homogeneity, despite increasing opportunities for change?
In this study, we examine the reproduction of executive age profiles across CEO succession episodes. We start from the general expectation that while CEO succession episodes are an opportunity for re-composing demographic characteristics of executives (Berns and Klarner, 2017; Ma et al., 2015; Wiersema, 1995), they also represent a disruptive episode for firms (Cvijanović et al., 2023; White et al., 1997). While one could expect firms to diversify the demography of their executives to adapt to the environment (Carpenter et al., 2004; Guthrie et al., 1991), we draw attention to an underexamined response to the disruptions surrounding CEO succession episodes: homosocial reproduction of familiar characteristics (Boone et al., 2004; Georgakakis et al., 2018; Nielsen, 2009). We add directionality to this tendency by focusing on executive age as a characteristic with particular qualities, as studies show that older executives are expected to be more predictable, reliable, and conservative in their resource allocations than their younger counterparts (Child, 1974; Hambrick and Mason, 1984; Wang et al., 2016). Yet, there are also other documented qualities associated with older aged executives, such as commitment to the status quo and risk aversion (Hambrick et al., 1993; Heyden et al., 2017b; Serfling, 2014). This highlights an intriguing puzzle, when one considers that executive age seems to be upward trending (Kelly, 2020; Schnoor, 2020), despite more frequent prompts for change.
Theoretically, we combine insights from UET (Hambrick and Mason, 1984; Wang et al., 2016), homosocial reproduction theory (HRT; Boone et al., 2004; Kanter, 1977), and executive succession research (Berns and Klarner, 2017). Building on these insights, we argue that executive age homophily may be one underexamined response to appease some of the potential disruptions surrounding CEO succession episodes, as it may help expedite social adjustment processes and predictability in executive leadership. We further acknowledge that not all succession episodes are alike (Georgakakis and Ruigrok, 2017; Karaevli and Zajac, 2013; Lauterbach et al., 1999). We thus assign an important boundary condition to our arguments, exploring the possibility that while the age profile of the executive team may be generally upward trending across CEO succession episodes, age homophily could particularly be strengthened under outsider CEO succession. Examining how age distribution in the executive suite evolves across CEO succession episodes offers a timely opportunity to reflect on the conditions that skew distributions on the input side of UET.
We offer several contributions, upon which we elaborate in the discussion. First, we consolidate insights from pressures for demographic diversity (grounded in UET) versus demographic familiarity (grounded in HRT) in executive leadership ranks, to inform theory on the countervailing pressures that may shape “why executive teams look the way they do.” Second, we extend our argument on tendencies to alleviate some of the disruptiveness associated with CEO succession episodes, by theorizing why and how outside CEO appointments could reinforce age homophily in the executive suite. In doing so, we uncover a counterintuitive new explanation for the inconclusive debates on outsider CEO influences on post-succession strategic change. Third, we advance literature on strategic leadership interfaces, consolidating literature that has focused on the same demographic characteristics in either the CEO as a solitary actor, or the TMT as an undifferentiated whole, to advance more ecologically valid interpretations of how the CEO–TMT interface could help identify “net effect” patterns of demographic re-composition over time. Finally, we advance conversations on inertia and adaptive change by adding demographic inertia in upper ranks as an underemphasized complement to this larger literature. In doing so, our approach helps clarify one potential source of persistent mismatches between characteristics of executives and broader demographic trends in the environment.
Conceptual background
UET suggests that an organization is a reflection of their executives (Carpenter et al., 2004; Hambrick and Mason, 1984). The age of corporate leaders has a strong legacy in this tradition, with implications for the fate of the corporations they steer (McClelland et al., 2010; Wang et al., 2016). Age captures the general life experience, generational worldviews, and accrued beliefs of an individual (Hambrick and Mason, 1984), and “CEO age (i.e., the length of time that a CEO has lived) is considered to be an important indicator of CEO experience” (Wang et al., 2016: 780). Scholars commonly theorize the implications of the age of executives in terms of stylized differences between relatively older and younger executives. For instance, older CEOs are expected to be more reliable, efficient, trustworthy, and compliant than their younger counterparts (Adhikari et al., 2015). In turn, younger executives have been shown to be busier in their resource allocations, making larger and more frequent decisions (Child, 1974; Li et al., 2017; Yim, 2013). This pattern resonates with findings showing that organizations led by older executives tend to be more inert, committed to the status quo, and engage in less risky investments (Andreou et al., 2017; Hambrick et al., 1993; Serfling, 2014). Intriguingly, “firms led by younger CEOs at the time of their appointment perform better than others” (see Serfling, 2014; You et al., 2020: Table 2). A compelling line of evidence thus suggests that older and younger executives make different choices. These differences matter.
Age differentials in the executive suite
Studies on executive age date back to the embryonic stages of the upper echelons tradition (Child, 1974; Hart and Mellors, 1970). Several discernable streams have since emerged to document the implications of age differentials in the executive suite. First, the literature on career concerns suggests that executives’ career motivations change across their lifetime, which becomes observable in their patterns of temporal resource allocations (Brickley et al., 1999; Cheng et al., 2024). Gibbons and Murphy (1992), for instance, find that older CEOs will tend to focus on their short-term bottom-line results to “finish on a high note,” conserve their legacy, and boost their reputation, as they position themselves to secure post-career opportunities. This is consistent with Heyden et al. (2017b), who found that as CEOs closer to retirement age tend to display lower investments in long-term innovation.
Second, older and younger executives tend to exhibit different risk-return preferences in their strategic choices (Belenzon et al., 2019; Serfling, 2014). Rising through the competitive corporate ranks to become an executive at a large corporation tends to reflect a career of compounded success and achievements, on the back of which an individual gets promoted to each subsequent rung. As career success accumulates, individuals with proven track records may become more risk-averse to protect their acquired status (Li et al., 2021). Evidence in this stream is consistent with meta-analytic findings from the study of Wang et al. (2016), who show that CEO age has a significantly negative relationship with international diversification, firm risk-taking, and product innovation. This rationale can also be seen in studies that have documented a tendency for older CEOs to reduce business risk and lower operating leverage (Serfling, 2014). In turn, younger CEOs who may still have to establish their reputations, tend to be more aggressive in their acquisitions (Yim, 2013), but also more likely to experience stock price crashes (Andreou et al., 2017).
Finally, the generational imprint literature suggests that the environmental conditions in formative periods (e.g., the start of our careers) have a lasting imprint on our attitudes in the workplace (Tilcsik, 2014). Zenger and Lawrence (1989: 356) noted early on that “[e]mployees of similar age, regardless of their expertise, status, or tenure in an organization, tend to have common non—work-related experiences.” These generationally-imprinted attitudes are observed across the workplace (Twenge et al., 2010). Indeed, studies on executive age cohorts show that the iconic events, traumatic episodes, and technological standards collectively characterizing a formative period may define the attitudes of generations long after (Schuman and Scott, 1989; Van Doorn et al., 2023). For instance, studies have shown that the generational exposure to war and famine have enduring effects on managers’ mindsets and behaviors (Cheng et al., 2021; Malmendier and Nagel, 2011). Although the notion of generational boundaries is still debated (Twenge et al., 2010), the difference between zeitgeists associated with specific age cohorts may still be observable in attitudes toward different technologies and concepts (e.g., Boomers and Gen Z may vary in their initial beliefs about the merits of new technological concepts, such as social media and artificial intelligence).
Together, the above present a non-exhaustive illustration of literatures showing that older and younger executives bring different perspectives and strengths via the behaviors, preferences, and attitudes typically reflected in their age profiles. As a general prescription, then, one could expect that companies would tend to favor executive ranks composed of diverse age profiles to harness the full range of these strengths. Yet as noted, executive age profiles seem to trend counter to this general expectation, as well as contrary to the observed age distribution in the workforce they are entrusted to lead. To the extent that executive composition should adapt to trends in the broader environment to avoid inertia (Guthrie et al., 1991; Thomas and Ramaswamy, 1993; Wiersema and Bantel, 1993), it is useful to understand why the distribution of certain characteristics are reproduced in particular patterns—and perhaps more countercyclical to changes in the larger environment than previously thought.
Theory and hypotheses
Demographic (re)composition of the executive suite
UET would suggest that one cannot fully understand an organization without understanding the characteristics of top executives. It is thus important to understand the conditions under which these characteristics change. CEO succession is perhaps the most visible opportunity for demographic change in the executive suite (Liu et al., 2018b; Wiersema, 1995; Wiersema and Bantel, 1993), and CEO succession episodes have become a more frequent occurrence (Berns and Klarner, 2017). Strategic adaptation perspectives would suggest that this presents firms with more opportunities to actively refresh the executive suite in order to avoid inertia (White et al., 1997). However, these succession episodes also douse the firm in uncertainty (Berns and Klarner, 2017; Cvijanović et al., 2023), as a recalibration of the compositional characteristics of the executive team often ensues, with other changes to follow.
Firms have a natural inclination to reduce uncertainty during episodes of potential change (Thompson, 1967), and CEO succession is a particularly uncertain episode for the firm (Cvijanović et al., 2023). An important aspect of this disruption entails the social re-adjustment processes during which continuity in the executive leadership function may become unsettled—social uncertainty. The required adaptations can be disruptive and costly in the short term (Schepker et al., 2017). One way to cope with—and reduce—some of the social uncertainty surrounding a CEO succession episode, we argue, is to adapt the demographic composition in favor of familiar characteristics with more predictable qualities. This is consistent with a premise from HRT about one means used to facilitate social adjustment in uncertain contexts: homosocial reproduction of familiar demographic profiles (Kanter, 1977). This line of reasoning follows an attraction—selection—attrition logic, where “. . . top executive management teams tend to selectively hire and fire to the effect that their own demographic characteristics are strengthened, thus promoting their social reproduction” (Boone et al., 2004: 635; see also Ertug et al., 2022).
To the extent that demographic re-composition may serve to appease some of the social uncertainty surrounding a CEO succession episode (McDonald et al., 2018), we theorize that pre-succession age profile may provide a familiar demographic template along which social uncertainty is calibrated and reduced. One visible way in which homosocial reproduction manifests in our setting, could be through reproduction of executive age profiles across CEO succession episodes. Thus, although change may be necessary, continuity and predictability in the executive leadership function is also desired (Schepker et al., 2017). Accordingly, firms may favor appointing executives with familiar characteristics, helping to minimize additional disruptions in an episode already rife with uncertainty. While studies have noted external sources of uncertainty such as changes in technology (Boone et al., 2004; Yamak et al., 2014), an internal tendency to reduce social uncertainty presents as a compelling complement as to why certain characteristics and qualities are favored in socially uncertain situations.
Homosocial reproduction of executive characteristics
Executives have been shown to gravitate to and favor similar others (Georgakakis et al., 2018). Similar others are often perceived as a reflection of the self; shared characteristics are often deemed to signal likenesses in deeper-level attitudes, values, and beliefs (Tsui and O’Reilly, 1989). Overall, the intuition would be that “demographically similar individuals accentuate the positive attributes of each other and derive a positive social identity” (Tsui et al., 2002: 901). Social uncertainty then is heightened when this social identity is threatened (e.g., a succession episode when someone with atypical characteristics might be appointed; see also Cutolo and Ferriani, 2024; Wechtler et al., 2022). This could be especially so if a new appointment is in a disproportionately more visible leadership position compared to the rest of the team, such as a new CEO (see e.g., Ritter and Lord, 2007; Van Knippenberg, 2011 for complementary perspectives).
Literatures on social identity and relational demography offer insightful clues on some of the mutually reinforcing sub-mechanisms through which social uncertainty can be counteracted (Hogg and Terry, 2000). First, status enhancement serves as a response to the possible dilution of social standing of in-group characteristics through the appointment of dissimilar others (McDonald et al., 2018). As such, incumbents reduce uncertainty by strengthening the status of characteristics associated with their in-groupness (Goldberg et al., 2010). Second, surrounding oneself with demographically similar others facilitates behavioral integration, as similar others display more favorable attitudes toward one another, share better quality information, and behave more consistently with mutual expectations, which may be especially useful in reducing disruptive conflict (Carmeli and Schaubroeck, 2006; Standifer et al., 2013). Finally, anticipatory favoritism may suggest that by promoting those who are similar to us, we can benefit more from their goodwill in the future (Balliet et al., 2014; Westphal and Stern, 2007). Similar others may display higher levels of reciprocity or general goodwill toward one another (McDonald and Westphal, 2011), which may be especially relevant for TMT members who expect to vie for CEO roles themselves someday (McDonald and Westphal, 2011). Accordingly, similar others have an incentive to favor profiles that may benefit them in the future.
Against these reinforcing theoretical mechanisms, we proceed to hypothesize how CEO succession may be a particularly potent context for catalyzing reproduction of age profiles in the executive suite. We exploit the trend of CEO age trending upward at appointment (Crist Kolder Associates, 2020; Kelly, 2020; Schnoor, 2020), providing a timely opportunity to formulate theory that combines insights from UET and HRT. While HRT helps explain homophily under uncertainty more generally, findings of UET on the qualities associated with age, give direction to this change (i.e., executive age profiles skewing toward older age).
Hypotheses on CEO–TMT age
To ground our predictions, we draw on the theoretical framework of Berns and Klarner (2017) who conceptually distinguished three conditions characterizing a succession episode: CEO change, CEO attributes (e.g., age), and CEO origin (e.g., insider or outsider). To further unpack our model, we align with recent studies that have recommended separating, but co-theorizing, the influence of the CEO and the TMT (Georgakakis et al., 2022). As such, all else being equal, 1 we will argue that the pre-succession age profile of the executive suite is expected to be reproduced post-succession in the age of the incoming CEO (H1), as well as the average age of the TMT in the years post-succession (H2). To contextualize the hypotheses further, we introduce the moderating influence of CEO origin in H3, distinguishing between insider and outsider succession.
Post-succession CEO age
Incoming CEOs are expected to add desirable qualities to the executive leadership function of a firm. Yet, CEO turnover is a disruptive episode for the firm in the short term (Schepker et al., 2017). Thus, appointing a CEO is more multilayered than just a simple comparison of the technical track record of available candidates. The literature on person–organization fit (Kristof-Brown et al., 2005), for instance, would suggest that, in addition to the more technical needs of the organization, considerations of intangibles such as values (Agle et al., 1999), leadership capabilities (Kuntz et al., 2019), and cultural fit with the incumbent TMT (Berns and Klarner, 2017) also play a role in transitioning a new CEO into their role.
Age cues shape our impressions, expectations, and behaviors toward others, as well as our self-concept and social identity (Kite et al., 2005). While teams and work-groups at lower levels of the organizational hierarchy can typically afford the time to explore similarities on a deeper level (Harrison et al., 2002), the CEO–TMT interface needs to function from inception. A new CEO has to hit the ground running, and Cvijanović et al. (2023) found that there was swifter adjustment and learning when there was a good fit between the new CEO and the firm. From this interpretation, when a new CEO who is much younger than the outgoing CEO (or much younger than the average TMT age) is appointed, it is more likely that pre-existing exchange processes (between the CEO and the TMT) will be experienced as more disruptive and behaviorally less predictable, as age dissimilarity may give rise to questions around competence level (Andreou et al., 2017), job attitudes (Ng and Feldman, 2010), and role division (Georgakakis et al., 2022). In fact, You et al. (2020: 1236) note that “. . . older executives are more apt to strengthen cohesiveness among organization members, which consequently leads to greater efficiency in strategy execution.” A socially-efficient approach to cope with disruption, then, would be to appoint a new CEO with a familiar age profile to the executive suite.
In the context of our theoretical assertions, it can be anticipated that given greater similarity between the incoming CEO and the incumbent TMT (Byrne et al., 1971), the transition into the executive leadership function will be experienced as less disruptive, requiring fewer behavioral adjustments from others, and allowing for greater expectation of continuity (Harrison et al., 1998), even if the incoming CEO has different backgrounds from those of the outgoing CEO and/or the incumbent TMT (Zenger and Lawrence, 1989). Thus, to the extent that firms may have a tendency to appease some of the social uncertainty around CEO succession episodes, we presuppose a general tendency, all else being equal, to appoint a CEO who is demographically familiar compared to (a) the outgoing CEO and (b) the broader TMT. Taken together, as a baseline we expect,
Hypothesis 1 (H1): The pre-succession age profile of the executive team will be positively related to the age profile of the incoming CEO, such that:
(a) departing CEO age will be positively associated with incoming CEO age
(b) incumbent TMT age will be positively associated with incoming CEO age.
Post-succession TMT age
Firms require some degree of continuity in their executive leadership. CEO succession episodes can introduce potential frictions at the CEO–TMT interface (Christensen-Salem et al., 2022; Liu et al., 2018b). Incoming CEOs often reconfigure the TMT to gain more control over the strategic direction of the organization (Shen and Cannella, 2002; Wangrow et al., 2022). For instance, CEOs may implement reconfigurations of the TMT in order to better reflect their vision of the future strategic direction of the organization, such as when adjusting to environmental requirements (Keck and Tushman, 1993) or entering new markets (Lin and Liu, 2011). Yet, it is difficult to predict how a new team will function, as re-composing an executive team can affect the likelihood of disruptive conflict (Amason, 1996), lower team chemistry (Lin and Rababah, 2014), and slow decision-making (Clark and Maggitti, 2012). The new CEO, thus, faces the challenge of re-composing a team with aligning worldviews to be consistent with their own vision for the future, while minimizing additional disruptions due to miscommunication and conflict. New CEOs may be favorably inclined to prioritize demographic familiarity to increase the likelihood of alignment in expected behaviors, underlying values, and motives with the new TMT. Thus, age homophily between new TMT members and the incumbent TMT could help facilitate behavioral integration and allows the team to strengthen exchange and collaboration patterns (Carmeli et al., 2011, 2012).
Our theoretical assertions around similarity-attraction would then suggest that the incoming CEO would tend to appoint new TMT members of similar age as one way to expedite social adjustment. While deep-level values may take time to become visible, individuals often infer behavioral expectations from observable characteristics (Tsui et al., 2002). Building rapport with the broader TMT can be particularly challenging, especially if the incumbent TMT is demographically quite dissimilar from the incoming CEO (Zenger and Lawrence, 1989). By appointing more age-similar TMT members, CEOs may strengthen alignment with their own worldviews. As a result, the TMT age profile can be expected to gravitate toward greater compatibility with the incoming CEO in the years post-succession (Barron et al., 2011; Kesner and Dalton, 1994). Thus, through age homophily of the TMT post-succession, a CEO can help alignment in expected risk profiles, appetite for change, and generational worldviews. Consistent with the previous hypotheses then, we can expect that the following is the case, post-succession:
Hypothesis 2 (H2): The pre-succession age profile of the executive team will be positively related to the TMT age profile in the years post-CEO succession, such that:
(a) departing CEO age will be positively associated with the age of the TMT in the years post-succession
(b) incumbent TMT age will be positively associated with TMT age in the years post-succession.
New CEO origin
Succession episodes are uncertain in part because new CEOs are expected to introduce visible changes to organizational strategy (Lauterbach et al., 1999; Schepker et al., 2017). The extent of these expected changes commonly differs between those CEOs promoted from within or those hired from outside organizational boundaries (Berns and Klarner, 2017; Georgakakis and Ruigrok, 2017; Schepker et al., 2018). Outside succession is often expected to be a precursor to other major changes as CEOs are often hired specifically to plug key leadership gaps and/or introduce a new vision (Kavadis et al., 2020; Ma et al., 2015). As such, outsider succession adds the prospect of further disruption to an already uncertain episode, given that it is less common, particularly when organizations lack experience navigating associated challenges.
When a CEO is hired from outside, they may be less attuned to existing team informal norms and more prone to making adjustments via new TMT appointments to expedite alignment. This adds to the uncertainty surrounding a succession episode as the outsider may take longer to understand existing social patterns, as well as the (in)formal roles played by different TMT members in relation to pre-existing strategy. Given that outsider CEOs are often appointed with a view to reposition strategy, the increasing divestment risk for prior activities puts pressure on the existing exchange relationships in the broader TMT. Uncertainty is further heightened due to inherent tensions between the newcomer and the incumbent TMT given that the internal talent pool was deemed unfit to fill the vacant CEO position (a threat to the competence of the shared social identity). To effect change, the outsider CEO then needs to balance harnessing the strengths of the incumbent TMT with making key appointments that will expedite alignment between the new CEO and the evolving TMT post-succession (Messersmith et al., 2014).
Outsider CEOs may need more time to adjust to and understand firm-specific assumptions as well as how the firm is positioned relative to the extant set of environmental opportunities (Agrawal et al., 2006). Thus, introducing an outsider to the executive ranks may particularly instill friction in social exchanges between the CEO and the TMT (Bauer et al., 2007; Chen et al., 2025). Industry observers have echoed this point, with Heidrick & Struggles (2024: 2) noting that “58% of the highest-priority hires—new executives hired from the outside—fail to gain fit within their new positions within 18 months.” Given that older CEOs are expected to be more reliable, efficient, and compliant than their younger counterparts (Adhikari et al., 2015), age similarity with the TMT can thus reduce the likelihood of behavioral unpredictability, facilitating outsider integration into the executive ranks. For the incoming CEO, then, familiarity in age may facilitate common-basis-of-general-life-experience (Fasbender et al., 2020), which can facilitate newcomer socialization (Ellis et al., 2015). Thus, although traditionally studies would suggest that outsider CEOs represent an even greater expectation for change, we posit that,
Hypothesis 3 (H3): Outsider CEO succession will strengthen the reproduction of age profiles in the executive suite across succession episodes, such that:
(a) the predicted post-succession CEO age will be higher when the incoming CEO is an outsider
(b) the predicted TMT age in the years after succession will be higher when the incoming CEO is an outsider.
Data and methods
Sample and data collection
We drew on the Standard and Poor’s 500 (S&P500) index as our sample frame, taking publicly listed firms indexed as at 2020, then extending the window for identifying succession cases in these firms back through 2000. We collected archival data from Compustat’s Execucomp database from 2000 to 2020, via institutional access to the Wharton Research Data Services. Execucomp identifies and archives information on the key officers of a company, which allowed us to identify the CEO and remaining members of an executive team of a company on an annual basis.
Given that CEO turnover is the key event in our empirical identification strategy, our chosen timeline takes into account that CEOs of publicly listed corporations commonly stay in their focal company role between 5 and 10 years, but rarely more than 20 years (Favaro et al., 2010; Karlsson et al., 2019). This rationale enables a reasonable chance of at least one CEO succession event taking place for each company in the sampling frame, while also factoring in companies that have been on the index long enough to identify lagged patterns. We then restricted our sample to the companies which had experienced (a) at least one CEO succession over the 20-year period, and (b) successions with a minimum of 3-year longevity (i.e., the new CEO kept their functions for at least 3 years), as the discernable impact of a CEO may take at least 2–3 years to become visible (Bandiera et al., 2020; Heyden et al., 2017a). This restriction also excludes the particular case of serial CEO succession indicating companies going through crises otherwise unaccounted for. To complement the dataset, firm-level accounting data (e.g., firm size, age, performance) were collected from Compustat Fundamental Annual and board-level data (e.g., board gender ratio, board age diversity) were collected from the BoardEx database. Non-systematic missing data (e.g., related to the tenure) were manually recorded from publicly available records, such as the EDGAR database of the Security and Exchange Commission of the USA.
A dataset was then structured at the succession level, forming a sample of 391 successions in 297 companies with corresponding lags from 2000 to 2020 for S&P500 index companies in the United States. Together, this comprehensive dataset allowed us both to collect information from a substantial portion of the US companies with various sizes and industries and to capture the essential information related to the CEO and associated TMT for each succession.
Measures
Dependent variables
We measured age in the demographic tradition and identified the year of birth indicated in the archival sources consulted. We then captured age of all executives listed in Execucomp as a continuous measure of yearly increments in two ways. For H1, where we are interested in the pre-succession age profile of incoming CEO age, we recorded the age of the incoming CEO at the year of succession. Second, for H2, where we are interested in the post-succession reproduction of age of the incoming CEO, we took the average age of the post-succession TMT executives 3 years after the change, as changes prompted by a new CEO may take some time to manifest (Bandiera et al., 2020; Heyden et al., 2017a).
Predictors
Consistent with the above, we measured the outgoing age of the CEO (H1a) at succession and the average TMT age (without the CEO) at succession for H1b. For H2a, we take the incoming CEO age and for H2b TMT age at time of succession (without the CEO).
Moderator
For our moderator (H3a–b), consistent with established approaches in CEO succession studies (Georgakakis and Ruigrok, 2017; Shen and Cannella, 2002), we measured CEO outsider successor as a binary variable, taking the value of 1 if the new CEO had not been an executive in the same company prior to being appointed CEO and 0 otherwise.
Control variables
We account for several alternative explanations that have featured as important enablers and constraints of inertia and adaptation, to demonstrate how and where our study contributes above and beyond existing insights, notably at firm-, board-, executive-, and industry-level. At the firm level, consistent with other research on CEO succession (Georgakakis and Ruigrok, 2017; Shen and Cannella, 2002; Zhang and Rajagopalan, 2010), we control for the overall financial stability of the firm using pre-succession firm performance, measured as the 3-year average return on assets (ROA) at the succession date. This metric reflects the baseline operational success of the firm that might influence board decisions regarding continuity or change in leadership more generally. In addition, we included the squared term of prior performance to capture the non-linear effects that extreme levels of past performance have on strategic choices, thus addressing more dynamic aspects of firm behavior under varying performance scenarios. We then controlled for firm size as “perhaps the most powerful explanatory organizational covariate in strategic analysis” (Josefy et al., 2015: 715) and a well-known correlate of inertia more generally (Hannan and Freeman, 1984), taking the logarithm of the number of employees. Larger firms, with their more established processes and greater internal candidate pools, may thus differ in the nature of demographic changes made around succession episodes. Further, as firms tend to reflect their founding generational imprints (Boeker, 1989), we took firm age as the number of years between the founding year and the current year (Un, 2016), with founding year approximated by the first year available in Compustat for each company. This control is important as older firms might have more entrenched cultures and norms that include age-related succession preferences, influencing the reproduction of similar age profiles in leadership transitions.
From the perspective of the board and governance (Johnson et al., 2013), we integrated key indicators that could influence dispositions toward change more generally, which could be reflected in similar or different age profiles in CEO and TMT appointments, such as board size, measuring the number of members of the board (Goodstein et al., 1994), board gender ratio, measuring the proportion of women on the board (Baghdadi et al., 2023), and board age diversity, measured by the standard deviation of the age of the board members (Ali et al., 2014). To capture the degree of independence of the board, we included CEO duality measured by a dummy variable standing for 1 if the CEO is also the chair of the board (Krause et al., 2014), and board independence, calculated as the proportion of non-executive board members (Boone et al., 2007). In addition, to account for the potential influence of share-ownership dynamics on CEO succession and subsequent strategic decisions, we have incorporated measures of CEO shareholdings and board ownership concentration (Banerjee and Homroy, 2018). High CEO shareholdings might encourage decisions that favor stability, including maintaining age homogeneity, while significant board ownership by external entities like institutional investors might push for changes that include age diversity. CEO shareholdings were measured as the proportion of the shares of the company held by the outgoing CEO at the time of succession. Board ownership concentration was operationalized as the proportion of shares held by the largest shareholders, which often includes institutional investors and founder-directors. We also captured CEO–TMT age dissimilarity (Georgakakis et al., 2022) by calculating CEO–TMT age distance using the mean Euclidean distance formula (e.g., Westphal and Zajac, 1995):
Here, Xi represents the age of the CEO i, Xj represents the age of each non-CEO executive j, and n is the number of TMT members.
At the executive-level, we further accounted for TMT size to consider the number of TMT members that play a key role in the selection of outsider CEOs, as they represent the pool of internal candidates for the CEO position (Withers et al., 2023). By accounting for TMT size in the panel structure, we further captured time-varying changes in the TMT that cannot be observed otherwise, such as attrition (Boone et al., 2004). TMT gender ratio was accounted for by incorporating the proportion of TMT women relative to its total size (Mah et al., 2023). In our context and in line with other research, we also controlled for CEO tenure, measured as the number of years the CEO spent in their function (Miller, 1991), and CEO gender, as gender diversity remains a category around which we can also expect demographic discontinuity, which we captured here based on a binary classification of sex assigned at birth measured as a dummy variable indicating 1 if the CEO is a woman and 0 otherwise.
We can anticipate that environmental shocks, where sudden and unpredictable shifts in the environment of a firm compel drastic internal adaptations, may trigger more radical shifts in the age profiles, either as a reactionary safeguard against unforeseen circumstances or as an aggressive strategy to capitalize on momentous successes. Therefore, to contextualize the firm within its broader industry dynamics, which can influence shifts in leadership requirements and the consequent age profile of the CEO, we controlled for firm environmental instability. Following Shen and Cannella (2002), we measured industry instability as the variance in the four-firm sales concentration ratio of the industry over the first 3 years following CEO succession. To account for different operating business environments, we further include industry fixed effects based on Standard Industrial Classification (SIC) codes reported in Compustat, while also adding year dummies to capture unobservable historical background heterogeneity.
Empirical strategy and estimation considerations
To test our hypotheses and taking the non-independence of observations in our sample into account, we applied hierarchical-linear-modeling (HLM) estimations and we grand-mean centered our data beforehand (Algina and Swaminathan, 2011). This factored in our expectation that over our observation period multiple successions could occur for the same organization at different points, so it was appropriate to nest our observations at the organization-level. In addition, CEO succession episodes are not randomly distributed, but often linked to performance issues, creating potential endogeneity challenges from sample-selection biases. To address this issue, we used the Heckman two-stage correction for selection bias (Certo et al., 2016; Heckman, 1979). In the first step model, we estimated the likelihood of CEO succession over the period taking into account fundamental firm characteristics (age, size, industry, and performance), key board and governance features (board gender ratio, board age diversity, CEO duality, board independence, and board size), and CEO characteristics (age, tenure). From the Heckman’s first-step model, we used the inverse Mills ratio of CEO succession and integrated this ratio into the main models to account for potential endogeneity in our subsequent multivariate analyses (see Table A in the Supplemental Materials).
Analyses and key findings
Univariate and bivariate results
Descriptive statistics and correlations are presented in Table 1, with some noteworthy observations. First, our data show that the median age of CEO successors has increased substantially over the last two decades (see Figure in the Supplemental Materials). In addition, for a quarter of the sample CEO age was, in 2019, 58 years or older, getting closer and closer to the conventional retirement age of 65 years (see Table B in the Supplemental Materials). Table 2 presents the descriptive statistics on the age differences between outgoing and incoming CEOs and how these patterns differ between internal and external hires. Some interesting empirical scenarios reveal the following insights. First, the age comparison between departing and incoming CEOs reveals that in about 26% of cases the successor is more than 10 years younger than their predecessor. Then, with an average CEO tenure in our sample of 8.77 years, and a tendency to see successors either roughly the same age as their predecessors at the time of succession or the same age as when the outgoing CEO started their role (about 67%), the CEO’s age tends to increase overall. This would suggest that the observed upward trend in average age of CEOs is particularly driven by incoming CEOs being older than outgoing CEOs, using the age of the latter when they were appointed as the basis of comparison. In addition, 8.18% of successors are older than their predecessors in our sample, reinforcing the trend. Additionally, outsider CEOs appear likely to be significantly older than their predecessors in comparison to their insider counterparts (Chi-square test; p = 0.02). Furthermore, we compared some descriptives between the S&P500 index top 10, top 25, and the rest of the S&P500 index sample. These findings suggest that the most prestigious companies of the S&P500 index tend to have slightly older CEOs than the rest. Yet, the difference is not significant, and overall, the trends in the age and age delta underscore a consistent age-related pattern across the S&P500 index. Against these detailed additional insights into the phenomenon, we proceed to infer multivariate empirical patterns based on our hypotheses.
Descriptive statistics and correlations.
CEO: chief executive officer; CEOO: chief executive officer outsider; TMT: top management team.
CEO age succession key characteristics and illustrative scenarios.
CEO: chief executive officer.
Multivariate analyses
We applied our HLM regression analyses in a stepwise fashion for our two dependent variables, that is, (1) incoming CEO age and (2) the age of the TMT 3 years after CEO succession (see Table 3). The first models contained the following control variables (Model 1 and 3, respectively): the inverse Mills ratio, firm size, firm age, firm, performance, board gender ratio, board age diversity, CEO duality, board independence, board size, TMT size, CEO tenure, CEO gender, industry, and year effects. In the next models (Models 2 and 4), the main effects were entered to the model.
HLM estimations for incoming CEO age (H1) and TMT age post-succession (H2).
CEO: chief executive officer; CEOO: chief executive officer outsider; HLM: hierarchical-linear-modeling; TMT: top management team.
N = 391 successions—297 companies. b: HLM estimates.
Before interpreting our main results, we highlight some notable insights from our control models (Table 3, models 1 and 3) that help highlight how we complement existing explanations. 2 At the firm-level, unsurprisingly, we see firm age as positively related to both incoming CEO and post-succession TMT (Josefy et al., 2015). We also see that past firm underperformance is linked to the appointment of younger CEOs, which is consistent with a broader literature on appointments of dissimilar others to precarious positions (Adams et al., 2009). In line with our theoretical reasons for inclusion of a range of governance-level controls, we find some noteworthy patterns. For instance, while more than half of the governance-level indicators show significant effects on the incoming CEO age, notably board age diversity (b = −0.48; p < 0.01), CEO duality (b = 1.15; p < 0.05), CEO shareholdings (b = 0.54; p < 0.07), and board size (b = 0.30; p < 0.001); only board gender ratio has a significant effect on post-succession TMT age (b = −5.13; p = 0.03). Overall, these firm- and governance-level controls exhibit patterns that are consistent with established research and provide a robust base from which to interpret the value-added of the focal hypotheses of our study.
Hypothesis 1 addressed the association between the pre-succession CEO and TMT age on CEO age after succession. As a baseline sense for the notion that homophily mechanisms are particularly activated in situations of uncertainty (Boone et al., 2004), and viewing CEO succession episodes as ripe with uncertainty (Cvijanović et al., 2023), as a background assertion for our core assumptions, we might expect that if executive age increases across CEO succession contexts, then the sub-sample of firms that have experienced multiple succession episodes in our observation window should exhibit a generally higher average CEO age. We conducted a sub-sample t-test analysis for the sub-sample of firms in our sample that had only one succession episode (n = 199) compared to those that had two or more succession episodes (n = 192), as summarized in Table C in the Supplemental Materials. Our notable findings show that for the sample of firms with only one succession episode, the average outgoing CEO age was 60.87 years and for the incoming CEO this was 52.98. Interestingly, for the sub-sample of firms with two or more successions the average outgoing CEO age is 59.38, which is significantly different from the sub-sample with one succession (∆1.49; p < 0.01). In turn, this sub-sample has a higher average age for the incoming CEO at 55.48 years (∆2.50; p < 0.01). Based on this baseline finding that corroborates our core assumption, we can proceed to interpret our hypothesis as per Table 3 (Model 2): departing CEO age was positively and significantly associated with the age of the incoming CEO (b = 0.15; p = 0.05, Model 2), providing indicative support for H1a. In turn, incumbent TMT age was found to be positively and significantly associated with CEO successor age (b = 0.34; p < 0.01, Model 2), providing strong statistical support for H1b. Hypothesis 1 is supported, showing that the strongest effect is that of pre-succession TMT age (H1b), suggesting that the incoming CEO age tends to reflect the age profile collectively held by the larger part of the group.
Hypothesis 2 addressed how the pre-succession TMT age profile was reproduced in the years post-succession (3 years after). Similarly, as we did for H1, we looked at t-tests (Table C in the Supplemental Materials). Interestingly, there is a statistically significant difference between the average TMT age of the sub-sample with one succession episode (51.08 years) and the TMT age of the sub-sample with multiple succession episodes (52.20; ∆1.12; p < 0.01). However, for the TMT average age at 3 years post-CEO succession, there seems to be no direct statistically significant difference between the sub-sample of firms experiencing one succession as compared to those experiencing two or more succession episodes. Thus, as CEO succession episodes have varying baseline effects on the age of the CEO and the TMT (Georgakakis et al., 2022), this (a) ratifies our rationale for disaggregating the CEO from the rest of the TMT and supports the need to use CEOs (and their characteristics) to help understand patterns of homosocial reproduction, while (b) also encouraging additional examination of boundary conditions through our contingency hypotheses discussed later in H3. Against these baseline observations, we proceed to estimate models for H2. As summarized in Table 3 (Model 4), departing CEO age is significantly associated with TMT age 3 years after CEO succession (b = 0.12; p = 0.01), providing support for H2a. For H2b, we see that the incumbent TMT age has also a strong, positive, and statistically significant association with TMT age 3 years after succession (b = 0.45; p < 0.01; Model 4). Our second hypothesis is thus supported.
For Hypothesis 3, we introduced an important boundary condition and expected that the aforementioned effects would be strengthened when the incoming CEO was an outsider. The interaction terms used to test H3 were introduced in Models 1–4 in Table 4 and the baseline control models are reported in Models 1 and 3 in Table 3. For H3a we document marginal statistical significance in support of the notion that the link between departing CEO age and incoming CEO age will be stronger when the incoming CEO is an outsider (b = 0.23; p = 0.09), as shown in Model 1, providing marginal support for H3a. In turn, the interaction effect regarding incumbent TMT age and CEO outsider succession in Model 2 is statistically significant (b = −0.67; p = 0.05), yet intriguingly negative, contrary to our expectations, suggesting that companies may still seek some complementarity (rather than similarity) between outsider CEOs and incumbent TMTs under this condition—highlighting the complexity of the relations under examination and need for further research on this front. For H3b (Models 3–4), we find the link between departing CEO age and post-succession TMT age (3 years later) to be positive and statistically significant when the incoming CEO is an outsider, providing supporting evidence for the conjecture (b = 0.19; p = 0.02). Finally, Model 4 shows that the influence of incumbent TMT age at time of succession on post-succession TMT age (+3 years) is not strengthened by outsider CEO succession (b = −0.02; p = 0.91), and so is not supportive of this hypothesis.
HLM results moderating effect of CEO origin (H3).
CEO: chief executive officer; CEOO: chief executive officer outsider; HLM: hierarchical-linear-modeling; TMT: top management team.
N = 391 successions—297 companies. b: HLM estimates.
The statistically significant parameters are depicted visually in Figures 1–3, where we show the ranges where the effect appears to be. Plotting the ranges is useful as the occurrence of outsider CEOs (8% of cases) is less common than that of insiders, which is consistent with previous studies, and highlights the potential range sensitivity of the effects observed in inferring H3a–b. We discuss the implications of our findings and theorizing next.
Interaction plot for departing CEO age, new CEO origin, and post-succession CEO age.
Interaction plot for incumbent TMT age, new CEO origin, and incoming CEO age.
Interaction plot for departing CEO age, new CEO origin, and post-succession TMT age.
Discussion
Why does the distribution of executive characteristics, such as age, evolve in particular and set directions, despite increasing opportunities for change? To the extent that (a) age of the executive team matters for firm outcomes (Hambrick and Mason, 1984), (b) one of the first visible changes that new CEOs bring about is in the re-composition of the TMT (Liu et al., 2018b), (c) frequency of CEO succession has increased to cope with increasing pace of change (Berns and Klarner, 2017), and (d) outsider CEOs may be particularly expected to bring about change (Karaevli and Zajac, 2013), one could expect executives to diversify their age profiles. Yet, CEOs are becoming older at appointment (Crist Kolder Associates, 2020), with implications for the compositional attributes of the TMTs they lead (Georgakakis et al., 2022). Motivated by these observations, we have combined insights from UET, HRT, and executive succession research to examine the reproduction of executive age profiles across CEO succession episodes. Our findings based on hierarchical linear modeling specified to dataset of 391 successions in 297 S&P500 index companies between 2000 and 2020 provide some intriguing results, in that some aspects of age profiles in the executive suite reproduced across the CEO–TMT interface. The patterns we have documented with our study allow us to offer several theoretical contributions and have some notable implications.
Theoretical contributions and implications
Reproducing familiarity versus diversifying characteristics in executive ranks
Firms face tensions between diversifying and homogenizing the characteristics of their executives. UET studies often emphasize the outcomes of diversity in demographic characteristics (e.g., for information processing; see Carpenter et al., 2004), which have long been linked to outcomes such as strategic change (Wiersema and Bantel, 1992), innovation (Bantel and Jackson, 1989), and performance (Buyl et al., 2011). Yet, demographic diversity still seems to elude the upper echelons of many organizations. Some studies attribute this to challenges associated with more diverse teams, such as conflict (Pelled et al., 1999), which can counteract the purported benefits. In turn, a handful of studies drawing on HRT have more deliberately sought to understand demographic homophily (e.g., Boone et al., 2004; Georgakakis et al., 2018; Nielsen, 2009). A particular insight from HRT is that homophily may be a response to help offset uncertainty (Kanter, 1977). We have shown that combining UET and HRT can be insightful for understanding the tensions described above, as while HRT helps explain the tendency for homophily to appease some degree of uncertainty, UET explains the direction of this homogenization based on the qualities associated with a particular characteristic (in our case, age). Our approach adds to the small, but insightful, literature on the antecedents of executive composition (Boone et al., 2004; Nielsen, 2009), helping us to understand why executive teams “look the way they look” (Hambrick, 2007).
We have couched our theorizing around the notion of social uncertainty surrounding a CEO succession episode—the anticipated disruption from potentially introducing unfamiliar or uncommon characteristics to an executive team. By understanding why certain attributes are favored over others in socially uncertain episodes, we can start to understand patterned changes in the reproduction of certain demographic profiles. As Boone et al. (2004: 634) note: “to understand the origins and consequences of specific differences in the demographic distributions of organizations, it is essential to study their metabolisms—that is, the entry and exit of (characteristics of) individuals into and out of these social aggregates.” Specifically, we have engaged with the plausibility that one aspect to navigate CEO succession episodes could be by favoring characteristics associated with predictability, maturity, and compliance; captured here by age. This represents an important new insight, as while studies have long documented that younger CEOs may be expected to further shake things up (Andreou et al., 2017; Serfling, 2014), the tendency to reduce some of the social uncertainty surrounding CEO succession episodes would be toward increasing the age of members appointed to the team. These insights help clarify that under conditions of social uncertainty, instead of becoming more diverse, age profiles skew to the right.
Together, we lay a foundation for understanding how tensions between demographic familiarity and demographic diversity come to shape distributional patterns of demographic characteristics in the upper echelons of organizations over time. This is important, as while scholars have long examined how characteristics of senior executives become reflected in the organizational outcomes of the firms they steer, we lay the groundwork for pinpointing key sources of why, and under what conditions, executive characteristics (e.g., age profiles) become skewedly distributed on the input side. By drawing theoretical emphasis to the input side of the upper echelons model, we raise a call for balancing out a vast body of literature focusing on the outcomes (Wang et al., 2016), as the input side distribution of demographic characteristics may have implications for the predictive validity of models seeking to infer outcomes from persistently skewed characteristics.
CEO succession episodes, new CEO origin, and demographic re-composition
CEO succession is considered an opportunity for change, but is also an uncertain episode for firms (Cvijanović et al., 2023; Schepker et al., 2017). An incoming CEO is expected to bring a new set of insights for the future of the organization (Graffin et al., 2013; Kavadis et al., 2020). Traditionally, CEOs promoted from within are expected to advocate and refine continuity (Schepker et al., 2018), while those recruited from outside are expected to bring novel perspectives to drive change (Georgakakis and Ruigrok, 2017; Karaevli and Zajac, 2013; Lauterbach et al., 1999). Yet, expected changes commonly fail to materialize; with literature remaining inconclusive on the effects of successor CEO origin on expected outcomes such as strategic change (Zhang and Rajagopalan, 2010). With our study, we add a new explanation to these mixed findings: while CEO succession is often expected to be a spark for change, CEO succession episodes may actually catalyze demographic homophily around certain attributes.
Interestingly, our findings corroborate the expectation that certain age homophily tendencies may be strengthened under outsider CEO successors. This adds an important—and counterintuitive—explanation for the mixed results on post-succession strategic change. A re-composition of the executive team’s demographic profile to support a new strategy is often one of the first visible changes by a new CEO; and a precursor to broader changes. Yet, while studies have documented that older CEOs may be expected to be more efficient, trustworthy, and compliant than their younger counterparts (Adhikari et al., 2015; Malm et al., 2021), there are also other documented challenges with older CEOs, such as commitment to the status quo and risk aversion (Hambrick et al., 1993; Heyden et al., 2017b; Serfling, 2014), that may stifle change. This demographic response had remained largely silent in findings from previous comprehensive studies seeking to inform this debate (Schepker et al., 2017) and our theorizing allows us to extend, for instance, Ma et al. (2015) who reviewed practices of incoming CEOs to minimize disruptions around CEO succession, with limited accounting for the demographic dimension. Together, through the tendency for age homophily to appease some of the disruptiveness surrounding a succession episode, we uncover a new explanation for why outsider CEO succession may fail to realize the oft-purported expectation for change.
Strategic leadership interfaces
We also inform recent developments in strategic leadership interfaces, “the purposive contact points where the separate worlds of actors intersect and are central to facilitating the transfer of influence, information, and resources” (Simsek et al., 2018: 281). A CEO and the TMT they interface with exert disproportionate influence over the fate of an organization (Georgakakis et al., 2022; Liu et al., 2018b; Simsek et al., 2018). While some studies have either focused on the CEO or the TMT, we align with an emerging call for differentiating, but co-theorizing, their effects (Georgakakis et al., 2022). For instance, while some studies have focused on CEO age (Belenzon et al., 2019; Serfling, 2014; Wang et al., 2016; Yim, 2013), others have focused on the collective TMT age (Richard and Shelor, 2002; Richard et al., 2019). Our approach in this study aligns us with what Georgakakis et al. (2022) term the social interactionism perspective on the CEO–TMT interface. This CEO–TMT approach allows us to approximate “net effects” of complex CEO–TMT social interactions on one another over time. The crossover effects we document at the CEO–TMT interface (e.g., incumbent TMT characteristics on CEO characteristics post-succession; outgoing CEO characteristics on TMT characteristics post-succession) point to the complex nature of interfacing social processes in the executive suite, responding to calls for more ecologically valid interpretations of how executive composition co-evolves between the CEO and their TMT (Georgakakis et al., 2022). From this vantage point, we have been able to illustrate an important way in which the interface approach to strategic leadership research can help us pinpoint who drives the upward tendency in executive age at appointment.
Strategic adaptation and inertia
Our study offers insights into the tensions between strategic adaptation and inertia more broadly. Specifically, the patterns we have theorized and may also underpin demographic impediments to adaptation. Age at the top is a visible attribute along which to add to such a conversation, given longstanding findings on the qualities associated with the age of our leaders; and the fact that the imprint of executives on the fate of their organizations has been increasing in strength (Quigley and Hambrick, 2015). Theory would suggest that companies could benefit from an age-diverse executive leadership to adapt to a more age-diverse environment and avoid inertia (Guthrie et al., 1991; Pegels et al., 2000; Thomas and Ramaswamy, 1993; Wiersema and Bantel, 1993). Yet, we have shown that executive age seems to tick upward across CEO succession episodes. Our approach helps clarify why mismatches between characteristics in the upper ranks of organizations and broader demographic trends in the environment may persist, despite opportunities for change (e.g., CEO succession).
We have shown that demographic characteristics such as age may be stubborn to change, becoming skewed in the direction of familiarity over time. While the emphasis our study relates to literature on demographic aspects of inertia and impediments to change (Boone et al., 2004; Georgakakis et al., 2018), we complement a broader literature on non-demographic aspects of inertia (Hannan and Freeman, 1984; Perini et al., 2024), that has traditionally emphasized factors such as structural (Hannan and Freeman, 1984), strategic (Perini et al., 2024), and socio-cultural drivers of inertia (Armenta et al., 2023). Accordingly, our theoretical model complements existing conversations on inertia and adaptive change by adding demographic inertia in upper ranks as an underemphasized complement to this literature.
Managerial and policy contributions
While younger and older workers share in the fate of the company, decisions are driven by an increasingly older group of executives. Accordingly, our study has real-life policy implications, as CEO succession is clearly on the rise: In 2015, 2500 largest companies of the world had a 16.6% CEO turnover rate, the highest rate in the preceding 16 years (Berns and Klarner, 2017), while CEOs are evidenced as becoming older at appointment (Schnoor, 2020). Over time, we caution that demographic characteristics may come to stick with job identity stereotypes. This interpretation allows us to contribute managerially to the conversation on “glass slippers” (Ashcraft, 2013), discussing the stickiness of age stereotypes for executive positions. Age is an important part of the demographic expectations associated with certain corporate roles (Posthuma and Campion, 2009; Wechtler et al., 2022). The literature on job identity stereotypes informs us that, jobs, for better or worse, but usually worse, have enduring labor identities. Reproduction of age profiles may create implicit associations that underpin a “gerontocratic ceiling” and form an age-based glass slipper for executive leadership roles (Truxillo et al., 2015). If the age stereotypes of executive leadership jobs are being reinforced, companies may benefit from thinking strategically about challenging demographic mold for executive roles and the implicit association of “how executives should look.”
Future research and limitations
Our study is prone to some limitations that offer key opportunities for future inquiries. First, we invite cross-national and cross-cultural interpretations and extensions. Future research could benefit from broader considerations of international corporate governance perspectives, for instance by acknowledging the rich contexts in which corporate governance practices are embedded around the world (e.g., ownership structures that may include sovereign wealth funds, private institutional investors, and other state interests), which interrelate with local context, such as cultural dynamics in state-owned enterprises in China (Chen et al., 2022), ownership practices in Korean chaebols (Liu et al., 2018a), and investment emphases of large pension funds in Belgium (Defau and De Moor, 2018). Future research could also focus on the role of culture as a factor influencing homophily in the upper ranks of corporations (Fliaster and Marr, 2000; Kaplan and Minton, 1994).
Next, we share a limitation with some previous studies (e.g., Boone et al., 2004), in that the full pool of candidates cannot be observed (Withers et al., 2023). Experimental complements with hypothetical pools of candidates and scenarios varying in social and technical dimensions of uncertainty could help provide more nuance to the mechanisms we have theorized to be at play behind the pattern of results observed. Furthermore, future investigations of reproduction tendencies across the CEO–TMT interface could include a broader range of compositional attributes, such as gender and ethnicity that are well-known potential contributors to conflict, while also exploring categories gaining theoretical attention such as caste (see Bapuji et al., 2024).
Finally, while the focus on S&P500 index companies enriches our understanding of CEO succession dynamics in high-profile influential firms, it also presents certain limitations. The generalizability of our findings may be restricted due to the unique characteristics of these firms, which differ significantly from smaller, private, or family-owned businesses, or firms operating in emerging markets. The strategic imperatives and stakeholder expectations shaping leadership decisions in S&P500 index companies may not reflect those in less visible firms (e.g., activist shareholders), potentially limiting the applicability of our conclusions beyond the context of major corporate entities. We also encourage future studies to examine external sources of homosocial reproduction, such as potential peer-imitation effects (Neuwirth and Frederick, 2004). While we have empirically examined whether there were differences between the top-performing firms and found no clear evidence either way, we believe a study dedicated to addressing this question with a tailored research design (e.g., matched pairs designs) could further be illustrative and examine whether there are peer-imitation effects at play. Overall, our study lays the foundation for an exciting research program on the tensions between reproduction and diversification of demographic profiles in the executive suite, while inviting extensions through additional boundary conditions in different settings.
Conclusion
As age composition of the general workforce changes, age-related tensions are re-emerging as central in scholarly inquiries of organizational demography. While older workers remain in the workforce even longer, younger workers leverage new technologies to tackle challenges that would have previously taken decades to master, progressing quicker up the corporate ladder. Against evidence that the general workforce is becoming more age-diverse, we have shown that pre-succession age profiles in the executive suite have an enduring imprint on the age profiles of successive executive teams, despite opportunities for change. By extension we add to a debate on succession planning and career progression, as age reproduction may be an underexamined barrier to development, promotion, and retention of younger managerial talent. Together, the trend of age reproduction in the executive suite may create barriers to ensuring future executive age profiles mirror the specific complexities and opportunities of a broader age-dynamic environment. Addressing these tensions is not as simplistic as replacing older executives, but rather ensuring the benefits of an age-diverse workforce are also leveraged in elite corporate leadership ranks, thus complementing the wisdom of age with the vitality of youth.
Supplemental Material
sj-pdf-1-hum-10.1177_00187267251327955 – Supplemental material for Wrinkle of change? The reproduction of executive age profiles across CEO succession episodes
Supplemental material, sj-pdf-1-hum-10.1177_00187267251327955 for Wrinkle of change? The reproduction of executive age profiles across CEO succession episodes by Mariano LM Heyden, Heidi M Wechtler and Sebastiaan van Doorn in Human Relations
Footnotes
Acknowledgements
We thank Dimitrios Georgakakis, Kohyar Kiazad, and Marko Reimer for developmental input on earlier drafts, as well as Dirk Boehe and seminar participants at the Africa Business School, Alejandro Escribá-Esteve and seminar participants at the University of Valencia, Lane Matthews, and seminar participants at Queen’s University Belfast, and Dan Wang and participants at Monash’s Department of Management Research Writing Retreat. We would further like to thank our action editor Chidiebere Ogbonnaya and three anonymous reviewers for their constructive guidance throughout.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
