Research offers conflicting answers to whether boards favor women for high-risk leadership roles, revealing a gap between experimental findings and real-world CEO appointments.
A senior McKinsey & Co. partner questioned on LinkedIn whether women are handed the hardest chief executive jobs, and the firm distanced itself from the remarks, according to Bloomberg and Fast Company. The discussion has revived a long-standing question in corporate leadership: Are women more likely to be appointed when a company is already in trouble?
Researchers have examined the issue from different angles, and their findings do not point to a single conclusion. Some experiments suggest that decision-makers favor women when an organization is struggling. Studies of actual CEO appointments, however, have produced mixed results. The distinction matters because finding women at the helm of struggling companies does not, by itself, establish that boards selected them because of their gender or that they faced greater risks than comparable male appointees.
What the glass cliff claims asserts
The phenomenon is known as the glass cliff, a term coined by psychologists Michelle Ryan and Alexander Haslam to describe a tendency for women to be appointed to leadership positions more often when organizations are facing difficult or precarious circumstances.
The argument goes beyond the observation that women sometimes take charge of struggling companies. Men also become CEOs of businesses in distress. To establish a glass cliff, researchers must show that women are more likely than comparable men to receive leadership roles under riskier conditions.
That raises a difficult question: At what point in the appointment process does the disadvantage emerge?
Where risk can enter a CEO appointment
A board acts after a trigger, such as a retirement, a dismissal or a decline in results. It then decides what the company needs, assembles candidates, selects one and later judges the outcome.
Risk could enter at any of these steps. The board may define the company's problems in a particular way, narrow the candidate pool, favor one candidate over another or appoint someone to a role with limited room to succeed.
Experiments typically isolate the selection step by asking participants to choose between candidates, while studies of real appointments capture the whole sequence, including who was in the pool and who accepted the job. That sets the order of the evidence: first whether a preference exists at the selection step, then whether real appointments show it.
Experiments show a preference for women in crises
At the selection step, the evidence is consistent. In three experiments by Haslam and Ryan, each with fewer than 100 participants, a woman was more likely to be chosen over an equally qualified man when an organization’s performance was declining.
Another experiment found the preference appeared when poor results were blamed on past leadership, and that a woman’s potential to signal change, not her quality as a leader, explained it. A 2020 meta-analysis agreed on selection. If real boards chose this way, appointments should show it.
Real appointments show no consistent pattern
Research examining actual leadership transitions has produced mixed findings.
Alison Cook and Christy Glass, analyzing 15 years of Fortune 500 transitions, found that white women and people of color were more often than white men promoted to chief executive at poorly performing companies.
A study of 233 German and British companies found that firms did not decline more before appointing women than before appointing men.
A Villanova University analysis of more than 10,000 U.S. CEOs from 1998 to 2022 found no clear evidence that women were appointed in riskier circumstances. These conflicting results suggest the studies are not measuring risk the same way.
The findings challenge the idea that women routinely inherit more troubled companies. They also show why the glass cliff debate cannot be settled by looking only at the number of women leading businesses in difficulty. Researchers must establish whether those appointments involved greater risks than comparable appointments of men.
Why studies of the same question diverge
One analysis found that firms close to default were more likely to appoint women as top executives, while the German and British study tracked accounting and stock market performance, so a company can look precarious by one measure and stable by another. Roles differ too: the default-risk pattern appeared among top executives but not board directors.
Decisionmakers matter too: Cook and Glass separately found that diversity among those making the choice, rather than firm performance, raised the likelihood of a woman’s promotion. Each study captures a different slice of the process, so none settles the question alone.
How boards can test their own process
Because published studies cannot show whether a given board’s choices are tilted, the testable part is the selection step, where experiments found the preference. A board can ask whether it chose for demonstrated fit with the turnaround or for what the appointment would signal, the factor that explained the preference in the experiments. It can record how it defined the company’s risk before the search began, since definitions appear to shape findings. The evidence is thin: women made up about 5% of the Villanova sample, and the experiments were small.
The evidence does not support the claim that women are appointed only when risks are hardest to manage. Even the studies that find a tilt describe a higher likelihood, not exclusivity, and the largest sample found no such pattern. Experiments and a few field studies show that a crisis can change whom decision-makers favor. The open question is where that preference operates, and a board can begin to answer it by examining its own succession decisions.
