Skip to main content

Discovering the Glass Cliff: Insights into addressing subtle gender discrimination in the workplace

The glass cliff refers to the phenomenon whereby women (and members of other minority groups, such as those based on race or disability) are over-represented in leadership positions that are risky and precarious – think UK Prime Minister Teresa May and Brexit. The metaphor evokes a woman who has reached the heights of senior leadership, but nonetheless finds herself teetering on the edge. These leadership positions can occur at the highest levels of in large multi-national companies, in local libraries and primary schools, in politics, and even in sport. Indeed, there are examples of glass cliffs in every realm in which there are leadership positions. The glass cliff is a robust phenomenon that has been demonstrated empirically through over a decade’s worth of archival and experimental research, including work in the UK, the US, The Netherlands, Australia, and Germany.

Michelle Ryan and Alex Haslam (now at the University of Queensland) coined the phrase in 2003 in response to a newspaper article from the front page of the business section of The Times. The article reported statistics that suggested that FTSE 100 companies that had the most women on their boards of directors tended to perform the least well in terms of their average annual share price. On this basis, the article suggested that women were “wreaking havoc’ within organisations and concluded that “corporate Britain would be better off without women on board”.

Such a conclusion is clearly a significant setback to those who argue for the promotion of gender equality at the top of organizations. However, as social scientists our first reaction was to question the strong conclusion made in the article because the evidence that it reported was entirely correlational. It is possible that while having more women on boards of directors and poor share price performance may be associated, the causal claim that women cause poor company could be unfounded. Indeed, we theorised that a very different causal relationship could be at play, such that women may be more likely to secure leadership positions only when companies are doing poorly.

To examine this alternative hypothesis we decided to conduct in-depth archival research to fully examine the data that were reported in The Times. We examined the share-price performance of all nineteen companies that had appointed women to their boards of directors in the previous year, as well the performance of a matched sample of companies that had appointed men. The results clearly demonstrated that while the appointment of women did coincide with poor company performance, their appointment tended to be preceded by a period of consistent poor share-price performance, rather than performance declining after women were appointed. Importantly, there was no evidence of a similar pattern for men.

These results demonstrated that while women were clearing breaking through the glass ceiling, it was also clear that the types of positions they were taking on were very different from that of their male counterparts. More specifically, we interpreted the data as suggesting that women were being appointed to positions of leadership that could be seen as relatively precarious as a consequence of the poor track record of company performance. We argued that such circumstances made it likely that women’s leadership would be placed under a close scrutiny and that they risked shouldering the blame for problems set in train long before their appointment, as clearly exemplified by the in original Times article.

Subsequent archival analysis has demonstrated that glass cliff appointments are not restricted to corporate board positions in the UK’s largest companies. For example, archival analysis of US Fortune 500 firms revealed that those companies that had experienced a scandal in recent years were more likely to have female executives on their boards. Similarly, Cook and Glass (2014a) examined Fortune 500 companies from 1996 to 2000 and found that weakly performing firms were more likely to promote women to CEO positions than they were to promote White men.

Within the political sphere, archival analysis of the 2005 UK general election also demonstrated the precariousness of women’s leadership positions. Here, within the Conservative party, women won significantly fewer votes than their male counterparts. However, this discrepancy in performance was completely explained by the fact that female candidates contested seats that were significantly more difficult to win because the candidate they were competing against had won a higher proportion of votes at the previous election.

The glass cliff has also been shown to extend beyond gender to other minority groups, such as those based on race and ethnicity. Having demonstrated that women may face a political glass cliff, we investigated whether Black and ethnic minority (BME) groups, who are also significantly under-represented in politics, may face a comparable disadvantage. In an analysis of the past three UK General Elections (from 2001 to 2010), Kulich, Ryan, and Haslam (2014) demonstrated that, again within the Conservative Party, individuals from BME backgrounds also confronted glass cliffs.  Again this meant that their lack of success could be fully explained by the fact that, compared to White candidates, the candidates they confronted had previously secured a higher proportion of the vote.

Similar findings have been observed within both the business and the sporting realm. In an analysis of Fortune 500 companies over a 15-year window Cook and Glass (2014a) demonstrated that, compared to White men, members of occupational minorities (i.e., Black men and Black and White women) are more likely to be promoted to CEO positions in firms that are experiencing a decline in performance. Cook and Glass (2013) also uncovered a glass cliff for ethnic minorities in a sporting context after examining the promotions of US College men’s basketball coaches over a 30-year period. Compared to White coaches, they found that ethnic minority coaches were more likely to be appointed as head coaches to teams with a history of losses in the previous year.

One key question that arises from the initial literature on the glass cliff is whether women are preferentially selected by others to be leaders in times of crisis. To answer this question we conducted a number of experimental studies to investigate people’s hiring decisions under conditions of varying company performance.

In these studies, participants are typically asked to read a scenario involving an organization that is either performing well or performing badly. They are then given details of a vacant leadership position within the organization and asked either to rank or evaluate the suitability of male and female candidates for the position. For example, in an initial experimental study, we asked participants to read a fictitious scenario about a leadership position in a company that had experienced either improving or declining performance in recent years. Participants were then presented with brief CVs of three candidates, a male and a female candidate who were equally qualified for the position and matched on all relevant dimensions (with CVs counterbalanced between participants) and one less-qualified male candidate who had been included to increase the realism of the scenario. The study revealed that when the company had been performing well, the two well-qualified candidates were ranked equally overall. However, when the company was in decline, participants showed an overwhelming preference for the female candidate.

A subsequent study with business leaders explored perceptions of men’s and women’s suitability for a leadership position in a company that was performing well or performing poorly. Here participants evaluated a single candidate, either a man or a woman (again equally qualified for the position). When the company was doing well, participants considered the leadership ability and suitability of the candidates to be equal. However, when company performance was declining, participants evaluated the female candidate more positively than the male candidate.

One factor that has consistently been shown to underlie gender discrimination in the workplace more broadly are gender stereotypes, that is, gendered beliefs about (a) women and men and (b) leadership. Regarding the former, women tend to be perceived as communal (e.g., being warm, good-natured, understanding, and caring), while men tend to be perceived as agentic (e.g., being competent, independent, competitive, and confident). Importantly, stereotypically masculine traits are remarkably similar to our stereotypical beliefs about good leadership, while stereotypically feminine traits are not. This phenomenon is known as the “think manager-think male” association.

However, research suggests that the association between masculinity and effective leadership is attenuated in times of crisis, opening up the possibility that’s women’s perceived communality — an attribute that is in demand when it comes to dealing with crisis — may contribute to the glass cliff.  Indeed, Ryan, Haslam, Hersby, and Bongiorno (2011, Study 2) found that prescriptive stereotypes about the qualities that managers of unsuccessful companies should have, were much more strongly associated with stereotypically female attributes than with stereotypically male attributes. Clearly, this "think crisis – think-female" association could help explain why women tend to be preferred for leadership positions when organizations are in difficulty.

Experimental research provides some support for these ideas. In particular, Bruckmüller and Branscombe (2010, Study 2) provided participants with a scenario in which they described a leadership position in either a successful company or a company in crisis. Participants rated a male and a female candidate in terms of stereotypically masculine characteristics associated with the think manager–think male stereotype (e.g., independence, decisiveness) and stereotypically feminine characteristics associated with the think crisis–think female stereotype (e.g., communication skills, ability to encourage others). In the context of a successful company, participants were more likely to choose the male candidate, and, at least in part, this was because he was described as more stereotypically masculine. On the other hand, in the context of a company in crisis, participants were more likely to choose the female candidate, and this was because, at least in part, she was described as more stereotypically feminine.

Along related lines, Gartzia and colleagues (2012) asked participants to evaluate male and female candidates for a leadership position within a company facing a crisis, with each candidate described as either stereotypically masculine or stereotypically feminine. In line with other experimental evidence, results revealed a clear preference for female leaders. However, over and above this effect for candidate gender, there was also an overall preference for candidates who were stereotypically feminine.

There is evidence that suggests that the glass cliff phenomenon may also occur as organizations strategically move away from standard leadership practices in times of crisis — appointing a non-prototypical leader as a signal to stakeholders that they are undertaking change. Consistent with this idea, experimental research has shown that manipulating an organization’s history of leadership can impact on the glass cliff. Specifically, Bruckmüller and Branscombe (2010, Study 1) demonstrated that the tendency to prefer to appoint women to leadership positions in a time of crisis was contingent on an organization having a history of male leadership. Where the company had a history of female leadership, there was no preference for female leaders.

Brown, Diekman, and Schneider (2011) found further evidence for this pattern. In their first study, they manipulated the salience of threat (e.g., by priming this through word-completion tasks) and found that this led participants to prefer change over stability. In a second study implicit associations were also observed to link women with change and men with stability. Furthermore, while control participants favored a man over a woman for a leadership position, exposure to a threat manipulation led either to the attenuation of this preference (Study 3) or to its reversal (Study 4). Together, these studies suggest that, to the extent that organizational crises are threatening to decision-makers, the experience of threat may promote a desire for change and hence a preference for female leaders.

Up until this point we have focused on the processes that may influence decision makers — either those involved in leader selection or those involved in formulating organisational strategy. However, it has also been suggested that the glass cliff could be due, at least in part, to the preferences and decisions of women themselves. For example, Patricia Peter, Head of Corporate Governance at the British Institute of Directors has argued, "I know of women who don't want to sit on a board that isn't a challenge, and who feel that if they go to a company that's doing quite well, they might not be noticed”. At the same time it is also the case that women may be more willing than men to accept leadership positions in times of crisis because they have relatively few opportunities to reach the top; after all, beggars can’t be choosers.

In line with this suggestion, research by Ashby and colleagues (2007) found that when participants were asked to choose between candidates for a risky leadership position (leading a legal case that was very likely to fail) their perceptions of opportunity were dependent on the gender of the candidate they chose. Here, a high-risk position was seen as providing a much better opportunity for the female candidate than her male counterpart. Moreover, those participants who chose the male candidate in a time of crisis (a minority of participants), recognized the increased risk associated with the crisis, but those who chose the female candidate did not.

Yet despite the very particular types of ‘opportunity’ that challenging leadership positions offer women, empirical research provides very little support that’s women’s own preferences and choices can explain the existence of glass cliffs. Indeed, experimental work by Rink, Ryan, and Stoker (2012) found that women generally regard risky leadership positions as less attractive than do men and are less likely to say that they would accept such positions. However, it remains possible that these perceptions of desirability are tempered by availability — so that cognitive dissonance leads risky leadership positions to become more attractive once women discover that they are the main option that is open to them.

Since it was first identified, the glass cliff has attracted considerable interest outside academia. In the first reports on the phenomenon it was common for media coverage to illustrate key research findings by elaborating on the career histories of women who had faced glass cliffs. Early examples were Carly Fiorina, appointed CEO of Hewlett Packard as the Tech bubble burst and Margaret Thatcher who faced a series of crises in her rise through UK politics.

This focus has become more pronounced in recent years — buoyed by the large number of cases thrown up by the Global Financial Crisis.  For example, at the onset of the GFC and not long after her appointment as Chief Financial Officer (the first women to ever sit on their executive committee). Erin Callan became Lehman Brothers' company spokesperson, but she resigned soon afterwards and the company declared bankruptcy three months later. A few years later Jill Abramson became the first female chief editor of the New York Times in the face of declining advertising revenues and uncertainty in relation to digital technologies. Like many of her male counterparts, Abramson introduced substantial staff redundancies in a drive to improve profits, but met extremely strong criticism and was fired. More recently, Mary Barra was appointed CEO of General Motors (the first woman to lead a global automobile company) only weeks before it announced the recall of 1.6 million cars due to electrical faults allegedly linked to 13 deaths. Other widely discussed examples include Marissa Mayer, appointed CEO of a declining Yahoo; Julia Pierson, short-lived Director of the US secret service; and Anne Mulcahy, appointed Xerox CEO when the company was close to bankruptcy and as their accountancy practices were under investigation by the Securities and Exchange Commission.

Political Glass Cliffs are also in evidence. One obvious example is that of UK Prime Minister Teresa May, who came into power immediately following the UKs referendum to Brexit the EU. Teresa May’s appointment coincided with a flurry of male politicians standing down from (or away from) significant leadership roles – such as David Cameron, Nigel Farage, and Boris Johnson. Australian Prime Minister Julia Gillard, Canadian Prime Minister Kim Campbell and Icelandic Prime Minister Johanna Sigurdardottir could also be said to have been appointed under glass cliff conditions.

With sample sizes of one, these various anecdotal examples clearly lack the empirical rigor of academic research. Moreover, their engagement with the nuances of the academic literature is often very limited and their capacity to resolve points of controversy is negligible (despite the fact that they are routinely invoked for this purpose).  Nevertheless, they have served to capture the imagination of the media and to bring the glass cliff — and broader issues of gender and leadership — to the attention of the public. Importantly too, in this context the metaphor of the glass cliff provides commentators with an alternative narrative of women and leadership that draws attention to the types of positions women occupy rather than simply blaming them for poor outcomes that are set in train long before their appointment.

1. Ryan & Haslam, 2005
2. Brady, Isaacs, Reeves, Burroway, & Reynolds, 2011
3. Cook and Glass, 2014a
4. Ryan, Haslam, & Kulich, 2010
5. Kulich, Ryan, and Haslam, 2014
6. Cook and Glass, 2014a
7. Ashby, Ryan, & Haslam, 2007
8. Haslam & Ryan, 2008
9. Haslam & Ryan, 2008, Study 1
10. Haslam & Ryan, 2008, Study 3
11. Koenig, Eagly, Mitchell & Ristikari, 2011; Fiske, Cuddy, Glick, & Xu, 2002
12. Schein, 1973
13. Pillai, 1996
14. Ryan, Haslam, Hersby, and Bongiorno 2011, Study 2
15. Bruckmüller and Branscombe 2010, (Study 2)
16. Gartzia and colleagues, 2012
17. Kaplan & Minton, 1994; Lee & James, 2004
18. Bruckmüller and Branscombe 2010, Study 1
19. Brown, Diekman, and Schneider, 2011
20. Woods, 2004, (p. 1)
21. Ashby and colleagues, 2007
22. Rink, Ryan, and Stoker, 2012