Identity Leadership in a Crisis: A 5R Framework for Learning from Responses to COVID-19.

The COVID-19 pandemic is the greatest global crisis of our lifetimes, and leadership has been critical to societies' capacity to deal with it. Here effective leadership has brought people together, provided a clear perspective on what is happening and what response is needed, and mobilized the population to act in the most effective ways to bring the pandemic under control. Informed by a model of identity leadership (Haslam, Reicher & Platow, 2020), this review argues that leaders' ability to do these things is grounded in their ability to represent and advance the shared interests of group members and to create and embed a sense of shared social identity among them (a sense of "us-ness"). For leaders, then, this sense of us-ness is the key resource that they need to marshal in order to harness the support and energy of citizens. The review discusses examples of the successes and failures of different leaders during the pandemic and organizes these around five policy priorities related to the 5Rs of identity leadership: readying, reflecting, representing, realizing, and reinforcing. These priorities and associated lessons are relevant not only to the management of COVID-19 but to crisis management and leadership more generally.

Introduction

On March 25 2020, public health officials in New Zealand and Australia issued short text alerts to citizens in their respective countries announcing a range of restrictions designed to protect them from the ravages of COVID‐19 (see Figure 1). The Australian message stated simply “Coronavirus Aus Gov msg: To stop the spread, stay 1.5 m from others, follow rules on social gatherings, wash hands, stay home if sick.” In contrast, the New Zealand message started by stating that “This message is for all New Zealanders. We are depending on you,” before explaining briefly what the restrictions would involve and concluding “Let's all do our bit to unite against COVID‐19. Kia kaha (Maori for “stay strong”; Jetten et al., 2020a).

Fig 1

Alternative approaches to leadership during COVID‐19 (from Jetten et al., 2020a)

The purpose of these two messages was much the same, as was their core content. Nevertheless, they instantiate two very different models of leadership in a crisis. The first, exemplified by the Australian text, embodies a paternalistic model based on the assumed authority and expertise of leaders—leaders who know what needs to be done and who see their role as being simply to impart that knowledge to others as efficiently as possible simply by telling them what to do. In the process, it positions followers as passive recipients of that knowledge, seeing them as having little or no role to play in the leadership process itself. The second model, epitomized by the New Zealand message, instead positions followers as partners in the influence process, regarding them as coproducers of desired outcomes. It recognizes that leadership is as much about creating strong groups as it is about having strong leaders.

In the present analysis, we follow the general consensus among researchers and practitioners in defining leadership as “the process through which one or more people influence other group members in a way that motivates them to contribute to the achievement of group goals” (Haslam et al., 2015, p. 248; see also Rost, 2008; Smith, 1995). Yet while there is general consensus among researchers and practitioners about what leadership is, there is little consensus about what makes it effective. This is exemplified by the understandings that inform the above two messages. As Turner (2005) observed, the difference between these can be seen to hinge on whether one sees leadership as a paternalistic process of “power over” others or as a partnership process of “power through” others. The paternalistic approach (exemplified by the Australian message) has a long history in the annals of psychology, management and political science, and views leadership largely as an exercise in command and control (Adams et al., 2017). It is also informed by what MacGregor (1960) referred to as a Theory X model of human behavior, which treats those who are led as passive vessels whose natural state is one of ignorance, indifference and torpor.

In contrast, the partnership approach (exemplified by the New Zealand message) is emblematic of what we have called “the new psychology of leadership” (Haslam et al., 2020). This sees leadership as a group process that centers on a psychological coupling of leaders and followers as a unit in which all parties see themselves as having shared interests and concerns, and which is effective because (and to the extent that) all members are engaged, energized, and agentic (in ways suggested by a Theory Y model; MacGregor, 1960). This coupling, we argue, proves to be especially important under conditions of crisis—when successful management of the crisis relies on securing widespread group adherence and engagement through a process of shared responsibility.

A core goal of this review is to clarify the practical implications of these two models for leadership during a crisis. In particular, we endeavor to spell out the consequences of the two models for group‐level outcomes and to derive concrete lessons from them. To provide a platform for this analysis—and for the substantive implications that we seek to derive from it—our review starts by fleshing out some of the key theoretical and empirical underpinnings of the new psychology of leadership. This analysis centers on the assertion that it is critical for leaders to build and mobilize a sense of shared group membership with followers (a sense of shared social identity, or “us‐ness”) through a process of identity leadership (Haslam et al., 2020; Steffens et al., 2014). This shared social identity is then a platform for the mutual influence, heedful interrelating, and goal‐focused coordination that are the hallmarks of effective leadership (Ellemers et al., 2004; Haslam, 2001; Turner & Haslam, 2001).

In the sections that follow, we drill down into the specifics of this analysis to make more targeted recommendations supported by theory, empirical research, and the varied experiences of leaders and groups in the first 6 months of the COVID‐19 pandemic. As Figure 2 indicates, these are structured around the 5Rs of identity leadership that have been derived from core aspects of the new psychology of leadership (Haslam et al., 2020, 2017), and that point to practical ways in which leaders can not only build social identity but also advance and embed it. Specifically, we argue that leaders need to priorities processes of reflecting, representing, realizing, and reinforcing shared social identity, as well as readying groups and their members, so that they are in a position to be mobilized when a crisis looms.

Fig 2

Shared social identity as a focus for priorities of identity leadership (the 5Rs) and a platform for engaged followership (adapted from Haslam et al., 2019)

At the same time, our analysis also points to various actions that work against these identity‐mobilizing objectives, and which therefore prove to be problematic for leaders’ capacity to lead groups through a crisis. Like the two text messages in Figure 1, the review therefore juxtaposes dos and don'ts in the process of abstracting key lessons that provide guidance for those who are looking to lead their groups through crises. Alongside this provision of a framework for good leadership that centers on the creation and management of shared social identity, our review also identifies key factors that produce ineffective or toxic leadership. Broadly speaking, these are associated with leadership that divides the leader from their group or, still worse, divides their group against itself. So where effective leadership creates unity of purpose by marshalling a sense of “us‐ness,” ineffective and toxic leadership foments dissent and discord by turning one part of “us” into “them.”

Priority 1: Reflect on Shared Social Identity

Given that part of this response depends on being altruistic and doing the right thing to help other people who we will never meet, having a leader who can articulate how we're all in this together and make a convincing case for why you need to do your part … is very important. (cited in McElroy, 2020)

The above observation by the Canadian epidemiologist David Fisman was cited in an article by Justin McElroy in which he sought to explain why some Canadian provinces—notably British Columbia (BC)—had been more successful than others in controlling the early spread of COVID‐19. His conclusion was that this had a lot to do with the leadership displayed by the provinces’ chief medical officers (CMOs). In particular, he singled out BC's CMO, Bonnie Henry, as someone who had succeeded in connecting with her fellow British Columbians in ways that encouraged them not just to listen to what she had to say, but to embrace the demanding course of action that she was asking them to pursue. When asked by the New York Times journalist Catherine Porter about how she did this, interestingly, Henry had little to say about her own leadership and instead focused primarily on her connection and empathy with the people to whom she had responsibility, commenting “It really is about the recognition that we are all in the same storm” (cited in Porter, 2020).

There are many features of Henry's leadership that are instructive (and that we will return to below), but the most basic is that, throughout the pandemic, she showed a clear appreciation of the fact that leadership was never just about her (see also Sergent & Stajkovic, 2020). The simplicity of this fact points to a realization that ultimately leadership hinges on a person's capacity to mobilize and channel the energies of others in the form of followership—since, as Platow and colleagues observe, “there is no leadership if no‐one follows” (Platow et al., 2015, p. 20; see also Bennis, 1999). This also means that any analysis of leadership that focuses only on the psychology and behavior of leaders as individuals (e.g., as “great men”; Carlyle, 1840) will necessarily come up short. This is all the more obvious in a crisis where it is not the grandstanding of leaders that determines the scale and nature of the disaster that unfolds but rather the behavior of followers (i.e., the general public; Diehl, 2020). For example, during COVID‐19 levels of public adherence with physical distancing measures had the capacity to reduce the spread of the virus by a factor of 10 (Jetten et al., 2020a, p. 4) and this was a critical determinant of communities’ capacity to effectively control the pandemic (Reynolds, 2020).

Focus on Achieving Power Through People Not Power Over Them

Both theoretically and practically, the key question upon which any analysis of leadership during a crisis must hang is therefore “What is it that provides the basis for a leader's influence over followers?” A short answer to this is power. However, as we suggested in the Introduction, there are two very different ways to think about this construct (Turner, 2005; see also Simon & Oakes, 2006). The traditional way is to see power as a force that leaders wield over others by virtue of their capacity to control key resources (e.g., rewards, punishments, information; French & Raven, 1959). A social identity approach to leadership however, argues that the optimal way to exert influence as a leader is by achieving power through those that one leads. Indeed, as Turner (1991) observed, this is true influence because here, where followers take your lead, they do so in the belief that this is the right thing to do and therefore do so willingly and energetically rather than begrudgingly and half‐heartedly. A key point, then, is leaders need to work with the groups they want to lead rather than see them as irrelevant to their leadership or, worse, as impediments to it.

COVID‐19 has provided us with many examples of the former leadership model. Most obviously perhaps, it was exemplified by the actions of US President Trump when he declared that he (rather than state governors) would decide whether state borders were reopened (after they had been closed to contain the spread of COVID‐19) because “When somebody is president of the United States, the authority is total” (Wilkie, 2020). While to some this may seem compelling as a model for leadership (notably because it projects an image of leader strength), as Trump's own experience attests, it has a range of limitations (for a critique, see A. Brown, 2014). The most basic is that it alienates people and casts followers as opponents rather than allies. As a result, even if leaders are able to control other people's behavior (e.g., by administering rewards or punishment), they are unlikely to generate much enthusiasm for their initiatives and will often encounter fierce resistance. Evidence of this was seen in South Africa when, as part of their policy for managing COVID‐19, the government unilaterally imposed a ban on smoking (Imray, 2020) and likewise in France when the government unilaterally mandated the wearing of masks in public indoor spaces (Patel, 2020). In both cases, the failure to take stock of public opinion and, more importantly, to get the public “onside,” led to widespread public protest which eventually led governments to back down.

Moreover, because one needs to police adherence much more when it is based on external rewards or punishment (rather than internal beliefs), this approach is itself energy‐sapping and resource consuming (Tyler & Blader, 2003). One also needs to recruit people willing to do this policing and, the moment they stop doing this, people's willingness to engage in desired forms of behavior rapidly diminishes (Tyler, 2006, 2012). The bottom line, then, is that with reference to the definition we offered above, the paternalistic imposition of power is the antithesis of leadership (Haslam et al., 2020; Turner, 2005). Indeed, this is something that Trump himself discovered in the face of widespread backlash to his assertion of authority—backlash that forced him to make the face‐saving concession that he would “authorize each of the states to adjust plans as they see fit” (Fallows, 2020).

Related to this point, when things go wrong it is unhelpful for leaders to respond by blaming group members for their (supposed) shortcomings. Not only is this because (as we will see below) that blame is often unwarranted, but also because it undermines the sense that leaders and citizens are on the same side, and invites backlash. This was evidenced, for example, when the Singaporean trade minister, Chan Chun Sing chided Singaporeans’ depletion of the country's supply of surgical masks as “xia suay” or “disgraceful”—describing them as “idiots” and adding “we embarrass ourselves, … we disgrace ourselves” (World News, 2020). It was also seen in Australia when the Prime Minister Scott Morrison denounced those who had stockpiled food by telling them to “Stop hoarding. I can't be more blunt about it. Stop it. It is not sensible, it is not helpful and it has been one of the most disappointing things I have seen in Australian behaviour in response to this crisis” (Martin, 2020). Elsewhere, the singling out of particular communities for blame (e.g., students, young people, party‐goers, protestors) generally proved counterproductive because as well as often being unjustified (Reicher, 2020), it also positioned those communities as enemies rather than (potential) supporters, in ways that made their nonadherence to desired norms more (not less) likely.

Critically, though, the idea that leaders need to work in partnership with those they lead does not mean that they do not need to take “tough” decisions in a crisis. They clearly do. However, it is critical to recognize that toughness is not a virtue in itself. So, when displayed, it always needs to be understood, communicated, and experienced as something that is for the good of the group not the aggrandizement of the leader or the belittling of others.

Lesson 1

Leaders will be in a better position to manage a crisis effectively if they focus on achieving power through followers.

Leaders who seek to manage a crisis by relying on the power that they have over followers will tend to be less effective—because they are likely to alienate them.

Focus on Recognizing Groups as the Solution Not the Problem

A key lesson to be drawn from the foregoing observations is that leadership centers on the mobilization of group‐based power and that this is achieved when leaders and followers internalize a given group into their sense of self—as part of their shared social identity (e.g., as “us Americans,” “us doctors”). This implies, then, that groups and group psychology are the cornerstone of effective leadership.

There is abundant evidence that this was the case in the COVID‐19 pandemic. It was perhaps seen most clearly in the spontaneous emergence of community self‐help groups that formed to address and resolve local problems surrounding access to resources and services. This occurred, inter alia, in neighborhood groups that developed systems to check on and support the vulnerable, in online groups that were established to share vital information, and in community mutual‐aid groups that developed to tackle everything from childcare to homelessness (Jetten et al., 2020b; Monbiot, 2020; Purdy, 2020; Walker, 2020). Indeed, in the United Kingdom alone there were over 4,000 such groups involving over three million people and here, as Walker remarked, “the selfless solidarity and sacrifice in the face of a life‐threatening national emergency in 2020 showed the very best we are capable of being” (2020, p. 53). As the journalist Jill Herron concluded from a survey of the various ways that New Zealanders in Otago responded to the challenges of COVID‐19, “it takes a village to beat a virus” (Herron, 2020).

It needs to be recognized, though, that such observations go against the wisdom of much received psychology in which groups and their psychology are seen not as solutions to a crisis but as a source of problems. The origins of such thinking can be traced back to the influential writings of Le Bon at the end of the 19th century (e.g., Le Bon, 1895), which equated the group with a loss of morality, decency, and rationality (Bergman, 2020; Reicher, 1987; Neville & Reicher, 2020). However, in the intervening century antigroup sentiment has become commonplace in psychology and indeed is foundational to much of the received wisdom surrounding effective leadership in a crisis.

In particular, suspicion of the group is at the core of what has hitherto been one of most influential models of crisis mismanagement: the groupthink model developed by Irving Janis (1972, 1983). Janis defined groupthink as the process whereby “members of any small cohesive group tend to maintain esprit de corps by unconsciously developing a number of shared illusions and related norms that interfere with critical thinking and reality testing” (1983, p. 35, emphasis added). The core symptoms of groupthink fall into three classes: (a) overestimations of the power and morality of an ingroup, (b) closed‐mindedness, and (c) pressures for uniformity. According to Janis, a group that has fallen prey to the syndrome tends to believe it is more moral, powerful, and invulnerable than it really is, and is also adept at dismissing criticism, in part because it prizes consensus and puts pressure on those who threaten it.

The idea that leaders readily succumb to groupthink in ways that compromise their judgement and efficacy has been used to explain poor leadership in the context of a range of disasters that have occurred in the last 60 or so years. These include the management by the Centre for Disease Control (CDC) of an outbreak of swine flu in New Jersey in 1976 (Neustadt & Fineberg, 1978), and the performance of the Johnson‐led United Kingdom government in responding to COVID‐19 (Coker, 2020; Heneghan & Mahtani, 2020). Yet while Janis's analysis has perennial appeal and is routinely invoked as a source of advice for leaders in a crisis, the groupthink model suffers from a range of fundamental problems (see Haslam, 2001; Haslam et al., 2006). The most basic is that even though many groups create a context for poor leadership decisions, the core claim that groupthink is a universal hallmark of poor leadership and poor decision making is not supported by available empirical evidence (Fuller & Aldag, 1998; Packer & Ungson, 2017). For example, in a systematic review of extant research, Peterson and colleagues (1998) found that while organizational groups that displayed symptoms of groupthink tended to make bolder decisions than those in which these symptoms were absent, there was no evidence that these decisions were, on average, any worse. Counter to the groupthink model, there is also evidence that group cohesiveness is more often an asset than a liability (Esser, 1998). Moreover, when leadership that is held up as exemplary (notably Kennedy's handling of the Cuban missile crisis; Janis, 1982) is subjected to empirical scrutiny, it is found to contain as much evidence of groupthink as leadership which is held up to be deficient (Kramer, 1998).

What is true though, is that when a group fails to achieve its goals, leaders often blame this on groupthink (and group psychology more generally) rather than on their own failings—and in this way the group becomes a convenient scapegoat (Haslam, 2001). Yet rather than blaming groupthink for problematic decisions (e.g., as Heneghan & Mahtani, 2020, do in the case of the Johnson government's response to COVID‐19), it is often more parsimonious to see these simply as evidence of poor leadership (a point we return to below).

Moreover, turning the logic of Janis's groupthink model on its head, there is evidence that it is the capacity for leaders to bring the group and its members to the center of their thinking that enables them to manage crises successfully. Rather than focusing on the elimination of groups as meaningful psychological entities (the core implicit recommendation of Janis's model), it is therefore important to see groups and their energies as a necessary part of the solution in dealing with and recovering from any crisis. Certainly, these energies need to be channeled in ways that encourage critical reflection on group activities rather than slavish acquiescence. But, importantly, research by Postmes and colleagues (2001) shows that critical thinking is something that flows from the norms that leaders establish as part of a group's operating principles and practices. Moreover, there is evidence that, providing they are guided by appropriate (e.g., critical) norms and that groups are not defined too narrowly (see Lesson 6 below), cohesive groups with a strong sense of shared social identity typically produce outcomes that are superior to those produced by groups that are uncohesive (Fuller & Aldag, 1998; Peterson et al., 1998)—at least in the sense of being more likely to meet the goals that the group sets itself (Haslam et al., 2006). Fear of groupthink, then, should never blind leaders to the need to think group.

Lesson 2

Leaders who treat groups as the solution to a crisis are likely to be more effective in harnessing the power of those groups.

Leaders who treat groups as if they are a—or the—problem in a crisis are likely to be ineffective.

Focus on Unlocking People's Capacity for Strength

A key point, then, is that in a crisis—especially one like a pandemic—leaders need to make groups as strong as possible. Nevertheless, it is generally the case that when disaster strikes a community or society, one of the first things to surface are stories about human weakness. In part, this is a reflection of cultural narratives which are built around the trope that when their sense of normality is threatened, people go to pieces (Monbiot, 2020). In disaster movies, for example, the threat of a hurricane or a flood or an alien life force is typically the cue for scenes of mass panic in which people flee for the exit with little thought for the consequences of their actions and the well‐being of their neighbor. Similarly, newspaper headlines routinely focus on the consequences of such behavior: blocked roads, queues for petrol, and empty supermarket shelves.

Early on in the COVID‐19 pandemic, there was widespread evidence of these same narratives playing out. In particular, reports of “panic buying” focused on shoppers’ unreasonable hoarding of such things as frozen food, bottled water, rice, face masks, and toilet paper (Lufkin, 2020). Moreover, as we saw above, some leaders were quick to blame the public for what they saw as the weakness and selfishness that such behavior exposed. Yet while it is undoubtedly the case that some people did engage in excessive stockpiling (particularly in the early days of the pandemic), it is less clear that this was simply a manifestation of “idiocy” or lack of concern for others (Reicher et al., 2020b). Indeed, in many cases the behavior in question reflected people's desire to do something—and to be seen to be doing something—to protect their families (Taylor, 2019). Importantly too, evidence suggests that in the vast majority of crises, problems are much more likely to be caused by people's underreaction than by their overreaction (Drury et al., 2009; Quarantelli, 2001). Furthermore, evidence gathered from those who have been caught up in a wide range of disasters suggests that in such situations, people's first instinct is in fact to help others rather than to abandon them (Drury & Tukin Guven, 2020).

Rather, then, than exposing people's inherent irrationality, selfishness, and weakness, evidence suggests that disasters tend to reveal the opposite—namely that people are (or at least can be) rational, civic‐minded, and strong (Reicher et al., 2020a). Nevertheless, leaders’ decisions are often informed more by the “frailty” model of human nature (akin to MacGregor's Theory X) than by the “fortitude” model (Theory Y). Moreover, this can sometimes reflect their own lack of trust in, and identification with, citizens (Haslam, 2001; Steffens et al., 2018).

Again, this was very apparent in some leaders’ preferred strategies for containing the spread of the virus in the early stages of the COVID‐19 pandemic. In particular, while the World Health Organization (WHO) recommended that countries pursue a strategy of aggressive containment, several countries were reluctant to go down this path for fear that people would lack the mental fortitude to comply. This belief was most clearly in evidence in the United Kingdom and the Netherlands, where policy was informed by suggestions that a lockdown was impractical as people would quickly succumb to “behavioral fatigue” (Boseley, 2020; Dujardin, 2020; Hahn et al., 2020; Sodha, 2020; for discussion, see Mols, 2020). Indeed, this strategy was consistent with the assumptions underlying behavioral insight psychology (e.g., after Thaler & Sunstein, 2009), which proposes that rather than seeking to change people's behavior through influence and persuasion, it is often easier and more effective to change their behavior through a series of small steers (or “nudges”) which covertly redirect their behavior. In particular, such strategies are designed to take advantage of the fact that people are understood to be “cognitive misers” whose hard‐wired cognitive biases incline them to process social information in a way that prioritizes fault‐ridden efficiency over resource‐intensive accuracy (Fiske & Taylor, 1984).

Although the analysis that informed such policy is extremely influential, it has a range of shortcomings—many of which were exposed by unfolding events during the pandemic (Gigerenzer, 2018; Mols et al., 2015; Oakes et al., 1994). The most obvious is that rather than succumbing to behavioral fatigue, people's capacity to endure lockdown proved to be far greater than many policy makers (and modelers) had anticipated. Indeed, especially where they had been exposed to leadership that made a strong case for the importance of acting collectively (e.g., in New Zealand and Singapore rather than Brazil and the United States), people proved remarkably willing to adhere to policies which required them to endure extreme privation (Jetten et al., 2020a; Wilson, 2020). Moreover, rather than exposing an inability to act responsibly towards collective goals, it was clear that people were deeply motivated to act in the interests of others (especially the vulnerable) rather than themselves (Purtill, 2020). Indeed, this fact led 681 British social scientists to publicly decry the lack of evidence for the notion of behavioral fatigue, and for the idea to be quickly disowned by those to whom it had been attributed (Mills, 2020).

In line with Mols and colleagues’ (2015, 2020) argument that when one is seeking to produce radical forms of meaningful behavior change it is far better to pursue a strategy of identity‐based persuasion than to rely on nudges, it is also clear that adherence with guidelines and regulations during the pandemic (e.g., relating to physical distancing and the wearing of face masks) was predicated on people's identification with leaders and authorities, and an associated sense that this was “the right thing to do” (e.g., as argued by Haslam & Reicher, 2017; Turner, 1991). Accordingly, when people failed to go along with lockdown instructions, this generally reflected the fact not that they were weak, but either (a) that they did not identify with the leaders who were issuing those instructions or (b) that they identified with leaders who were discouraging adherence. This pattern was again seen clearly in the United States—where adherence with the edicts of state governors was far higher in areas where residents shared those governors’ party affiliation (Grossman et al., 2020; Rothgerber et al., 2020) and in states where governors (often women) were keen to engage with residents’ concerns by displaying compassion (Sergent & Stajkovic, 2020). They were also higher in states where social distancing and face‐mask wearing mapped closely onto people's identification either with health officials (who endorsed these actions) or with the president (who did not; Painter & Qui, 2020). Moreover, as Oakes and colleagues (1994) have argued, it is apparent that the forces that structured such behavior were social and political rather than purely cognitive in origin (Heath, 2020).

Taken as a whole, then, the COVID‐19 pandemic has done much to explode the myth that people are inherently weak, lazy, and egotistical and has instead revealed citizens’ capacity for strength, endurance, and altruism. Critically, though, and in line with the logic developed in the previous section, this capacity has also been shown to be contingent on leadership which treats citizens not as sheep who are in need of leaders’ paternalism, but as lions whose strength is unlocked through partnership with leaders that centers on a shared social identity which motivates them to contribute to collective goals which they have internalized as self‐defining (Steffens et al., 2018). This, for example, was the strategy successfully pursued in Canada by Bonnie Henry when she remarked that “If you tell people what they need to do and why, and give them the means to do it, most people will do what you need” (cited in Porter, 2020). And it was also evident in the approach taken by the German Chancellor Angela Merkel when she announced in an address to the nation that “Since World War II, there has been no greater challenge to our country that depends so much on us acting together in solidarity” (Davidson, 2020; Posner, 2020). As Davidson (2020) observed, in this “What gave [Merkel's] address its force was her tone, which was direct, honest, and searingly empathic. She laid bare not just the test we all face but also the solace that leadership can provide.”

Lesson 3

Leaders will be more likely to encourage adherence with their policies if they recognize that people can be resilient but need to understand what they are being asked to do.

Leaders will be less effective if they see perceive people to be weak and rely on coercion, nudging, or punishment to drive adherence to their policies.

Priority 2: Represent “Us” and Our Goals

We are all in this together. The virus threatens everyone. Human rights uplift everyone. By respecting human rights in this time of crisis, we will build more effective and inclusive solutions for the emergency of today and the recovery for tomorrow. (Guterres, 2020)

As we flagged in the Introduction, research inspired by the new psychology of leadership suggests that influence over followers is something that leaders achieve through leadership that serves to create and harness a sense of shared social identity (a sense of “us‐ness”) within the group as a whole (Haslam et al., 2020; Steffens et al., 2014). As a large body of research confirms, the key reason for this is that the capacity for social influence is grounded in perceptions of shared social identity (Turner, 1991; Turner & Oakes, 1986). This means that people tend to be more open to the influence of others if (and to the extent that) they perceive them to be representative of a shared ingroup membership (i.e., to be “one of us” rather than “one of them”; Ellemers et al., 2004; Hogg, 2001; Steffens et al., 2020; Turner & Haslam, 2001).

Social identity refers to individuals’ sense of internalized group membership. It is a sense of self associated with an awareness that one belongs to a particular social group (e.g., a nation, an organization, a community), and that this group membership is important and meaningful (Tajfel, 1972). It is this, then, that allows people to refer to the self in the first‐person plural—as “we” and “us” (e.g., “us Australians,” “us health care workers”). Indeed, a key insight of social identity theory (Tajfel & Turner, 1979) is that such references reflect an understanding of self that is qualitatively distinct from that associated with references framed in the first person singular (“I” and “me”), but one that is every bit as real and authentic (Turner, 1985). Moreover, rather than reflecting a loss of self (a self which psychological theory typically locates more or less exclusively in the person as an individual), social identity reflects a gain in self that is achieved by seeing oneself as a group member who is thereby psychologically connected to other members of that group (one's ingroup) and therefore open to their influence (Baray et al., 2009; Haslam, Jetten et al., 2018; Reicher, 1987).

Put simply, the reason why social identity is a basis for influence is that when we define ourselves as group members, other ingroup members are no longer external to our self but rather become part of our self. Consequently, they become integral to our selfhood and hence relevant to (and a point of reference for) our sense of self‐interest (Platow et al., 2003). During the course of COVID‐19, this was a point that the New York's Governor, Andrew Cuomo, captured when he harangued those who were protesting against the lockdown that he had imposed by reminding them that “It's not just about you. You have a responsibility to me. Get your head around the ‘we concept’” (Slattery, 2020; cited in Jetten et al., 2020a, p. 10).

Focus on Building Shared Identity

For leaders, then, this sense of social identity is the key resource that they need to build and tap into in order to mobilize the support and energies of others. Accordingly, during COVID‐19 the message that “we are all in this together” has been critical to leaders’ appeals to the public (Wilson, 2020). This indeed was the title of the report on the pandemic authored by the Secretary‐General of the United Nations, António Guterres, from which we quoted above. This report was framed around a sense of common humanity, and it was on this basis that it sought to motivate citizens and their leaders to champion the cause of human rights in ways that would protect them not only from the virus but also from some of its toxic social impacts (e.g., in the form of increased inequality, division, and prejudice; Crimston & Selvanathan, 2020; Hall et al., 2020; Huo, 2020; Jetten, 2020b; Van Bavel, 2020).

By the same token, leaders’ appeals to shared social identity also animate citizens’ adherence to their specific directives. As we saw in discussing Priority 1, this was a key feature of Bonnie Henry's efforts to enjoin her fellow Canadians to comply with her office's demanding health directives. Likewise, it was for this reason that the President of the American College of Surgeons, Tim Eberlein, appealed to his fellow surgeons to maintain their efforts on the surgical frontline by assuring them that

We have never been more proud to be a member of the American College of Surgeons. FACS—Fellowship in the American College of Surgeons means that we are all in this together. We are here for you. (Eberlien, 2020)

Indeed, returning to the text messages that we discussed in the Introduction (see Figure 1), we can see it was the injection of social identity messaging (“Let's all do our bit to unite against COVID‐19”) that differentiated the New Zealand communiqué from the Australian one. Research in the social identity tradition leads us to expect that this would have encouraged the target audience not only to process the message more deeply but also to take on board its content more fully (Bentley et al., 2017; Greenaway et al., 2015; Mackie et al., 1990; McGarty et al., 1994).

Lesson 4

Leaders secure followership by building, and drawing on, a sense of shared social identity (a sense of “us‐ness”).

Leaders will undermine followership if they are perceived to represent the interests of an outgroup, a subgroup, or themselves.

Focus on Treating People Respectfully

We argued above that social identity is the basis for effective leadership primarily because it is this that creates the psychological bonds with followers that motivate them to engage in the acts of followership that are necessary for the group as a whole to succeed. Yet as important as the rhetoric of “us” is, social identity is not just about what leaders say, it is also about what they do. A basic point here, then, is that leaders need to treat followers as ingroup members. Indeed, any action by leaders that serves to create a social identity faultline either (a) between them and followers or (b) between different sets of ingroup members is likely to prove detrimental to their leadership.

Empirical confirmation of this point is provided by a program of experimental research conducted in New Zealand by Platow and colleagues (1997, 2000) that explored how group members’ support for a leader varied as a function of that leaders’ treatment of ingroup members. More specifically, in a paradigm eerily prescient of realities associated with COVID‐19, group members were told about the decisions that a (male) leader had made in allocating access to life‐saving medical technology to members of different groups. The key finding was that support for the leader was enhanced when he gave ingroup members (New Zealanders) more access than outgroup members (non‐New Zealanders), but diminished if he gave one ingroup member more access than another ingroup member (i.e., if he made arbitrary distinctions between ingroup members that suggested that they were not “all in the same boat”). Importantly too, this support was also translated into followership (see also Haslam & Platow, 2001). For when group members were given the opportunity to write a letter in support of the leader's actions, their willingness to do so varied as a function of the way that leader had treated their ingroup. Put simply, it was only when followers saw the leader both as standing for (and standing up for) their ingroup and as having treated ingroup members fairly (i.e., equally) that those group members were prepared to “go into bat’ for him (see also Meyer et al., 2015).

There is plenty of evidence that these same dynamics play out in the context of crises. For example, it has been argued that the Presidency of George W. Bush suffered a significant blow when photographs emerged of him surveying the devastation caused by Hurricane Katrina from the window of Air Force One, rather than being on the ground alongside the victims and support workers (Walsh, 2015). As Bush himself noted later: “That photo of me hovering over the damage suggested I was detached from the suffering on the ground. That was not how I felt. But once that impression was formed, I couldn't change it.” (Bush, 2010; cited in Han, 2018, p. 352). Conversely, during World War II, the standing of the British royal family was considerably enhanced by images of King George VI and Queen Elizabeth walking through the bombed‐out houses of East Londoners whose communities had been destroyed in the blitz. Indeed, after Buckingham Palace was also bombed, Elizabeth famously observed “I am glad we've been bombed. It makes me feel that I can look the East End in the face again” (cited in Russell, 2014, p. 44).

The importance of leaders being able “to look followers in the face” is something that has also been very much in evidence during the COVID‐19 pandemic. In particular, it has been closely linked to the changing fortunes of the British government that we alluded to earlier. Ironically, for the UK's Prime Minister Johnson, the high‐point of his leadership came when he emerged from hospital after he had contracted and been successfully treated for, the coronavirus. For, like the Royal Family in the blitz, this created a connection not only between him and other victims, but also with the health workers who treated him (a connection that Johnson himself reinforced by naming his newborn child after his doctors; Ng, 2020). At the time Johnson also earned praise for his humility, and for his explicit acknowledgement that “there is such a thing as society” (a direct repudiation of the assertion by his predecessor, Margaret Thatcher, that there was not; Williams, 2020).

Yet the political capital that Johnson gained from his hospitalization dissolved rapidly soon afterwards when his chief advisor, Dominic Cummings, was found to have violated the lockdown that the government had put in place in an attempt to curb the spread of the virus (Fancourt et al., 2020). Interestingly, the key problem here was not the violation itself, but the management of the violation. Elsewhere, the standing of leaders under similar circumstances had not been harmed when members of their group committed similar violations—for example, when Scotland's Chief Medical Officer, Catherine Calderwood, flouted her own Department's advice to reduce unnecessary travel (Carrell, 2020), or when New Zealand's Health Minister, David Clark, broke his own government's lockdown by going mountain biking (McKay, 2020). The critical difference, however, was that Nicola Sturgeon (Scotland's First Minister) and Jacinda Ardern (New Zealand's Prime Minister) moved quickly to remove those who had transgressed from office. In contrast, Johnson refused to give Cummings his marching orders (or to criticize him at all), but rather defended him as having “acted responsibly, legally and with integrity” (Diver & Cameron‐Chileshe, 2020). The reasons for this have been debated, but the impact on Johnson's leadership was clear. Not only did his approval rating drop by 20 points in 4 days (Carpani, 2020), but, in the face of widespread anger, adherence to lockdown rules also slumped dramatically (Fancourt et al., 2020). As a British junior doctor commented at the time:

For the last three months [I have] told relatives they can't visit patients in hospital. While patients lay dying of this virus, I was forced to narrate their deaths over the phone to people who kept to the rules.…. We all sacrificed something, we all gave up something of ourselves. Except for him [Cummings]; he has nothing but contempt for us all. (Blank, 2020).

The implications of Cummings’ actions for management of the pandemic were also clarified by an epidemiologist on one of the UK government's main scientific advisory boards who observed: “One of the things that's always stood out is that for … targeted measures to work, we need public adherence to quarantine to be very high. But I fear it's now going to be far more difficult to achieve this” (Kucharski, cited in Colson, 2020). This certainly proved to be the case. Indeed, by the end of the week in which Johnson defended Cummings, mobility levels among the British general public were close to prelockdown levels (Calcea & Rea, 2020). As one of us (Reicher) explained in an analysis of the Cummings saga:

If you give the impression there's one rule for them and one rule for us, you fatally undermine that sense of “we're all in this together” and you undermine adherence to the forms of behavior which have got us through this crisis. … Because of these actions, because of undermining trust in the government, because of undermining adherence to the rules that we all need to follow, people are going to die. (cited in Colson, 2020)

Lesson 5

Leaders will be more effective if they treat group members, respectfully, fairly and as equal partners.

Leaders will undermine followership, and hence be less effective, if they create faultlines within the groups they lead.

Focus on Defining Ingroups Inclusively

The foregoing discussion has focused on the importance of leaders working to create a sense of shared social identity within the communities they lead in order to mobilize those communities to work towards shared goals. As we have noted, they are generally helped in this regard by the fact that the emergence of a common enemy (in the case of COVID‐19, the coronavirus) creates a sense of common fate that motivates a sense of shared identity and a desire for unity of both understanding and action (Drury, 2012; Levy, 2020). Indeed, early on in the pandemic, this unity of shared identity and purpose was widely called for and widely in evidence (e.g., N. Brown, 2020; Shimooka, 2020; Sibal, 2020).

However, as the pandemic wore on, cracks in the coalitions that had initially been formed to fight the virus started to appear (e.g., Wollny, 2020; Yuan, 2020). In large part, these cracks followed the contours of preexisting social divisions, and, moreover, they were generally more marked the deeper those divisions had been (Jetten, 2020a). That is to say, preexisting material and psychological realities place a significant constraint on the capacity for leaders to promote—and for followers to buy into—the rhetoric of a united “us” (Borkowska & Laurence, 2020; Oakes et al., 1994; Reicher & Stott, 2020b; a point we return to in discussing Priority 5 below). This was perhaps most apparent in the United States—where preexisting tensions between Republicans and Democrats meant that Americans approached the crisis from very different vantage points . As a prominent example of this, Democrats suggested that the Trump administration was treating the virus as a hoax (Greogorian, 2020; Reider, 2020) while Republicans suggested that the Democratic leadership were politicizing the pandemic (Frank, 2020).

In the face of just such divisions, leaders who want to encourage a united response to a crisis need to engage in a particularly vigorous form of identity leadership if they are to create a shared sense of “us” that eclipses the previously dominant sense of “us” and “them” (Dovidio et al., 2020). If leaders are unable or unmotivated to do this, then the crisis will tend to exacerbate previous divisions in ways that can lead to sharply polarized responses to both the crisis and their leadership (Crimston & Selvanathan, 2020; Jetten et al., in press). Again, this scenario has played out dramatically in the United States, where the enthusiasm with which supporters of Donald Trump embraced his leadership and policies, was matched only by the derision with which they were met by his opponents (e.g., as extensively documented by the Pew Research Center, 2020).

The key theoretical point here, then, is that citizens’ behavior follows the contours of those shared social identities that prevail in a given society and which leaders have a critical hand in creating (Haslam et al., 2020; Reicher et al., 2005; Reicher & Stott, 2020b). By extension, the practical point is that where those identities are ultimately defined at a subordinate rather than a superordinate level of inclusiveness (i.e., one in which a unified “us” self‐categorization is eclipsed by opposing “us–them” self‐categorizations) a group's capacity to mount a coordinated response to a crisis will be compromised.

Moreover, this applies at multiple levels of inclusiveness (in line with the basic tents of self‐categorization theory; Turner et al., 1987, 1994). So whereas the inability of Americans to unite around shared national identity adversely affected the United States’ ability to fight the virus (in ways that showed up in rates of infection and death; Yamen & Wenham, 2020), so too the inability of different nations to unite around shared international identity meant that the global response to the pandemic was suboptimal (Guterres, 2020). And this too was ultimately a failure of leadership. As Mark Dybul the codirector of Center for Global Health Practice and Impact at Georgetown University observed “It's very difficult for the international system to respond when we're still all in our corners, and we're in our corners because no one's leading” (cited in Igou, 2020).

Lesson 6

Leaders will be more likely to encourage broad support for their policies if they advance a broad and inclusive definition of their ingroup.

Leaders will secure less uniform support if they advance a narrow and exclusive definition of the ingroup.

Priority 3: Realize Shared Identity in Plans and Policy

This is a storm that's affecting the world. But we are not in the same boats, so we can't make assumptions about other people. I am going to give you everything we know so you can do your best to keep afloat. (Bonnie Henry cited in Porter, 2020)

The forgoing lessons underscore the importance of fairness for leadership. Yet while fairness will often involve treating all group members alike, there are times when it does not. More particularly, in their dealings with group members, leaders’ sense of fairness needs to be built around notions of equity rather than equality (Wenzel et al., 2002). This means that if the needs and circumstances of group members are very similar then they need to be treated equally, but if they are very different then they need to be treated differentially (i.e., unequally). The significance of this point is underlined by extensive programs of research in social psychology which show that people's respect for the law and their adherence to it is underpinned by their sense that principles of procedural and distributive justice have been upheld (e.g., Tyler, 2006, 2012; Tyler & Blader, 2003).

Focus on Appreciating People's Differing Needs and Circumstances

The importance of being sensitive to group members’ differing circumstances has also been brought home in the context of the COVID‐19 pandemic—where it quickly became clear that people's risk of contracting the virus and their capacity to follow health guidelines were unequally distributed. In particular, members of disadvantaged and vulnerable groups (e.g., the homeless, the elderly, the poor) were much more likely both to be in, and to find it hard to get out of, harm's way. Not only, then, were people who were old and poor much more likely to fall ill, but they were also more likely to need to expose themselves to danger. Estimates from Britain, for example, showed that the poorest people were six times more likely than the most affluent to have to go out to work during the pandemic and three times more likely to find it hard to self‐isolate (Atchison et al., 2020; Bibby et al., 2020; Smith, 2020). This also meant that when tight restrictions on movement were imposed, these same people were much more likely to violate them. As a resident of the Parisian suburb of Clichy‐sous‐Bois (a suburb with a high proportion of residents of North African descent) observed:

People are trying to respect the lockdown, but what do you do if you're a family of five or more in a small apartment on the 15th floor? How do you keep children in? How do you feed them when the markets where you buy cheap fruit and vegetables have closed and you can't afford supermarkets? How can families whose children normally eat in school canteens now make three meals a day?” (cited in Jetten et al., 2020a, p. 7)

To be effective, leadership therefore needs to be sensitive to these differential realities (Templeton et al., 2020). It was for this reason that the German government introduced a policy of paying workers on low incomes 70% of their pay for the first 2 weeks that they needed to self‐isolate, while in Australia people were eligible for a A$1,500 “Pandemic Leave Disaster Payment,” and a similar policy in China paid the food, rent, and essential bills of those who needed to be quarantined. Similar programs—generally centered on the provision of paid sick leave (OECD, 2020a)—operated in many other countries. And, in all these cases the policy in question was identified as a key contributor to the high levels of adherence necessary to restrict spread of the virus, with policies generally being more effective the more generous they were and the more quickly they were implemented (see OECD, 2020b, for details).

However, aside from simply distributing financial support, leaders’ sensitivity to group members also needs to attend to their broader circumstances and needs. In Pakistan, for example, where around two‐fifths of households live in poverty and hence do not have the means to access funds that are transferred electronically, the government moved quickly to implement a program of unconditional cash transfers that allowed those who were in need to collect money in person from a designated bank (Jafri, 2020). The program was accessed by over 80 million people, and the World Economic Forum concluded that it was a very effective way of providing humanitarian assistance (Nishtar, 2020). In Taiwan, as well as rolling out a program of financial support for those who needed to quarantine, the government also implemented a system in which village leaders delivered them “goodie bags” containing food, books, and movie services (Yang, 2020). Every day a local official also called them to thank them for doing their part to contain the virus (Farr & Gao, 2020).

The corollary of this, though, is that if leaders are insensitive to group members’ needs and circumstances, this will generally undermine support for them and their policies. This is important because all too often it is the disadvantaged, marginalized, and vulnerable members of society who are in the blind spots of public policy. Thus one reason why infection and death rates were ultimately lower in Pakistan than in India (despite India spending 50% more per capita on health) was that “in a textbook case of what not to do” the Indian government provided relatively little support to those people who were particularly in need, and, when the government introduced a lockdown, people were given just four hours’ notice to prepare (Chandraseker & Ghosh, 2020). Likewise, in France, a punitive system which led to over 350,000 people being fined for lockdown violations led to widespread backlash—including rioting in suburbs like Clichy‐sous‐Bois that we alluded to above (FR24 News, 2020; Reicher & Stott, 2020a). Primarily this was because, as we have seen, many of those violations resulted from necessity, not choice. As one of the rioters observed:

Perhaps for the French elite we are second‐class citizens, but our DNA is French and we are protesting to defend our rights—mainly to be able to live. That's all. (cited in Borges, 2020)

As another put it:

These are not riots just for the sake of rioting. What's happening in these districts is a cry for help—to the government, to the state, to the president. It might not be the best way to do it, but it's the only way for many here to be heard. (cited in Borges, 2020)

For leaders, then, it is critical not just to speak the language of “being in this together,” but to put in place policies and structures that allow people to have the lived experience of equity. In a pandemic, this requires them to recognize that the crisis affects different groups of people very differently. This point is captured eloquently by the penultimate lines of a poem entitled “We are not all in the same boat” that was shared widely on social media during the COVID‐19 pandemic:

We are not in the same boat. We are going through a time when our perceptions and needs are completely different. …

It is important to see beyond what is seen at first glance. Not just looking, actually seeing. (Cuthbert, 2020)

Lesson 7

Leaders will be more effective if they implement policies that are sensitive to the differing circumstances of different group members.

Leaders who ignore disadvantaged and marginalized members of their groups will fail to secure the broad support that they need for their policies to be effective.

Focus on Being Empathic Rather than Punitive

A general point here, then, is that policies need to be informed by empathy for the plight of others. This is no less true when people fail to adhere to relevant guidelines or, more generally, to “do the right thing”—and where (as we saw earlier) leaders are often tempted to dish out blame, criticism, and punishment (Reicher, 2020).

On the ground, there is evidence that empathy was a critical ingredient of effective health policy and policing during the pandemic. For example, it was central to messaging around the virus by the New Zealand government (McGuire et al., 2020) and to the “4E” guidance provided to all UK local police forces (College of Policing, 2020; Stott & Radburn, 2020). The former focused on efforts to build trust and a sense of common purpose between authorities and the public (Carter et al., 2020). Similarly, the latter recommended that police officers’ first strategy should be to “Engage” with the public and then to “Explain” what people are required to do and to “Encourage” them to do so. The fourth strategy—to “Enforce”—was then flagged only as a last resort, to be pursued only if approaches based on respectful dialogue had been thoroughly explored and exhausted. Clearly, there are moments when such enforcement will be necessary in a crisis; however, evidence suggests that it will more effective if it is seen to be an expression of community consensus rather than an independent assertion of authority for its own sake (Stott & Rabburn, 2020).

Where it was adopted in the United Kingdom, this “policing by consent” approach (and an associated reluctance to use special powers that were granted by law, as per Lesson 1 above) was widely credited with reducing incidents of public disorder and law breaking, and with helping to consolidate a belief that police and health agencies were working with and for the communities they served, rather than against them (Carter et al., 2015, 2020; Reicher & Stott, 2020a; Stott, 2020; Wood et al., 2020). Critically too, it also contributed to the fact that compliance—and desire to comply—with relevant laws and guidelines generally stayed very high (at least until Lesson 5 was not heeded; Fancourt et al., 2020). This accords with Carter and colleagues’ observation that “compliance will only ensue when authorities respect and trust the public … [and] that only when trust is mutual [does] it becomes possible to formulate, internalize and unite around shared norms concerning health protective behaviours” (Carter et al., 2020, p. 89).

Lesson 8

Leaders will be more effective if their policies are seen to be informed by empathy with others and their plight.

Leaders whose strategy centers on punishing transgressors will generally fail to inspire the majority of the population to engage in necessary acts of citizenship.

Priority 4: Reinforce Shared Identity through Ongoing Action

The unprecedented COVID‐19 pandemic is a powerful reminder of our interconnectedness and vulnerabilities. The virus respects no borders. Combatting this pandemic calls for a transparent, robust, coordinated, large‐scale and science‐based global response in the spirit of solidarity. We are strongly committed to presenting a united front against this common threat. … The G20 is committed to do whatever it takes to overcome the pandemic. (G20 Leaders statement, cited in , OECD, 2020c)

Leadership in a pandemic is not only about securing the support of people on the ground, it is also about giving support to those people. Indeed, as we noted in discussing Priority 3, in a crisis one of the key functions of leaders is to allocate resources to those who have been most affected by it and who are most in need of assistance. It is apparent too that in the present pandemic, these needs are many and varied—and include such things as access to medical testing, protective equipment, childcare, and income support, as well as medical treatment itself. Importantly too, these needs are just manifest at the start of a crisis but often endure—and in many cases increase—as the crisis progresses.

Focus on Providing Ongoing Support to Those Who Most Need it

In this context, it might seem rather obvious that leaders should provide support where it is needed and that those who are given help should be those whose needs are the greatest at any given point in time. Certainly, this is the sentiment that has been repeatedly expressed by leaders of global charities and aid agencies throughout the pandemic (e.g., Louayza, 2020). However, things are seldom this straightforward. A key reason for this is that just as people's support for leaders follows the contours of shared social identity, so too does the support those leaders give people in return. Indeed, a corollary of observation we made above that leaders have typically seen the pandemic more through a national lens than through an international one is that their concern has typically been much more for the health and welfare of their own citizens than for the citizens of other countries (Runde et al., 2020). In this, their behavior aligns with that seen in previous disasters where leaders in government and business have been found to allocate a far greater proportion of their budgets to the management of crises that occur within rather than outside national borders (Muller & Whiteman, 2009).

One might argue here that the tendency for leaders to favor fellow nationals is natural and reflects a general expectation that their charity “begins at home”—and indeed that this is the appropriate focus for identity leadership. However, this argument neglects the fact that, as we noted in discussing Priority 2, what counts as “home”—and hence where charity begins and ends—varies as a function of the inclusiveness and content of the social identity that informs leaders’ identity leadership. Home, for example, can be one's actual home, one's neighborhood, one's town, one's region, or one's country—or indeed the entire planet (Levine & Thompson, 2004). Equally, who precisely leaders deem worthy of their charity typically depends upon who they see as belonging to the home in question (Haslam et al., 2012; Reicher & Haslam, 2009). Furthermore, even when we see others as part of our community, the place that the most vulnerable have in our collective consciousness varies so that leaders and groups differ in how deserving they see different (particularly the most disadvantaged) members to be. A case in point was the fierce debate that raged in the United Kingdom around whether, and for how long, children should be provided with free school meals (a policy that had been introduced early on in the pandemic but which was subsequently reversed). For some this was a necessary act of compassion, for others it was gratuitous “virtue signalling” (Hinsliff, 2020).

A key point here is that leaders’ category definitions play a key role in structuring the provision of supports (Levine et al., 2005; Reicher et al., 2006). This point is supported by experimental research which has shown that people offer much more help to others when they are encouraged to define themselves in terms of a social identity that includes those others (Levine et al., 2005). It is also confirmed by archival research which has shown how Bulgaria's leaders in World War II protected members of their country's Jewish population against deportation to Nazi extermination camps through sustained efforts to define Jews, not as an outgroup, but rather as an essential part of the national ingroup (Reicher et al., 2006). Again, these dynamics have been very much in evidence throughout the pandemic. Indeed, from a leadership perspective, one of the most striking features of early responses to COVID‐19 was the unprecedented scale of the support that governments provided their citizens in the face of those citizens’ unprecedented needs. For example, Western European countries allocated nearly US$4 trillion to a wide range of relief schemes—an amount nearly 30 times larger (in today's monetary value) than that devoted to the Marshall Plan (Cassim et al., 2020). This rapid policy shift was perhaps most remarkable in countries like the United States, the United Kingdom, the Netherlands, Germany, and Australia where conservative governments that had previously eschewed heavy social spending programs, rapidly embraced them with great enthusiasm. As Rawnsley (2020) observed in relation to the Johnson administration's volte face: “Policy ideas that were decried as madness yesterday are being redefined as the only sane response to this emergency. Positions once held to be immutable are being tossed into a great bonfire of discarded orthodoxies.”

For many leaders, then, the enlarged sense of shared identity created by the pandemic led them to pursue greatly enlarged social support programs and these in turn served generally to bolster support for their leadership. Epitomized by the G20 statement that we quoted at the start of this section, this led a number of commentators to observe at the time that it appeared as if “we are all socialists now” (e.g., Abernathy, 2020; Salutin, 2020). Yet while this was in many cases a widespread and sustained response, its reach was neither uniform nor constant and it admitted exceptions. In India, for example, the ultranationalist Modi government persisted in treating Muslim migrant workers—and the poor more generally—as “outsiders” and exposed them to great harm by imposing a range of restrictions on their movement (e.g., requiring them to walk large distances on foot rather than use public transport; Ramasubramanyam, 2020). In Singapore too, while Prime Minister Lee Hsien Loong was generally praised for his government's response to the pandemic and his assertion that “we feel that we are all in this together, and we do not leave anyone behind” (Parker, 2020), it is apparent that migrant workers were largely excluded from this collective self‐definition. And because those migrant workers were provided with no alternative to crowded living conditions through which COVID‐19 could run rampant, they became its primary victims—accounting for nearly 90% of infections and deaths in the country by May 2020 (Koh, 2020; Ratcliffe, 2020; Yea, 2020).

There were also signs that, as the pandemic endured, leaders in other countries similarly fell back on less inclusive definitions of “us” that meant their charity became more circumscribed. In Australia, for example (as in many other countries), the government was criticized for failing to recognize the particular needs of people with disabilities (Henriques‐Gomes, 2020; see also Frost, 2020; Ryan, 2020). And in Britain, the government's “Eat Out to Help Out” policy initiative designed to support the restaurant industry (and costing far more than the free school meals policy that it rejected; Hinsliff, 2020) favored those who could afford to eat out while doing nothing for people in poverty who were struggling to eat at all (Patrick, 2020). Hence far from being contexts in which sensitivity to the special needs of diverse groups prevails, pandemics and other crises can quickly become policy landscapes in which it is leaders’ prejudices that flourish (Cohn, 2007, 2018). And while this may win leaders support in some quarters, in line with the logic of Lesson 6, it will ultimately tend to reduce their capacity to manage the crisis effectively.

Lesson 9

Leaders maintain support through the ongoing distribution of resources to those who most need them.

Leaders will become less effective over time if they withdraw support from ingroup members or if the support they give undermines a sense of shared social identity.

Focus on Achieving Outcomes That People Most Value

The foregoing discussion leads into broader questions about what it is that allows leaders to succeed in the long term—so that they emerge from a crisis with their reputations and legacies enhanced rather than diminished. This has been a key question within the leadership and crisis management literatures more generally. And here the most influential ideas have focused on the importance of having leaders who are able to reassure and inspire followers by dint of their charismatic personality (Bligh et al., 2004). This, indeed, is one of the cornerstones of traditional understandings of leadership and, in particular, of the “great man” approach which extols the distinctive virtues of leaders as Übermensch (or supermen; Carlyle, 1840; Nietzsche, 1885). The special power of such leaders—which is seen to be especially valuable in a crisis—was first discussed by Weber (1946,) in an influential analysis of Bismarck's ability to achieve Germany unification in the wake of the Austro‐Prussian war and the period of intense regional conflict that preceded it. It has remained a central plank of analyses of leadership in times of crisis ever since, and, in particular, is routinely invoked by commentators and historians to explain the capacity for leaders such as Churchill, Kennedy, Mandela, and Obama to chart a path through crisis and disaster to security and stability (Bligh & Kohles, 2009; Hunt et al., 1999; van Vugt, 2013).

The basis of such charisma is typically seen to lie in the personal character and qualities of the individual leader. Indeed, this sense is associated with the Greek meaning of charisma (χάρισμα) as a “special gift” for prophecy and influence that particular leaders possess (Den Hartog & Verburg, 1997; Marturano & Arsenault, 2008). And in line with this analysis, there is plenty of evidence that during the COVID‐19 crisis people have looked to charismatic leaders to help them make sense of events and give them a sense of purpose and direction (Crayne & Medeiros, 2020).

However, while charisma is undoubtedly an important feature of the unfolding leadership landscape in a crisis, there are at least three reasons to doubt that it is an inherent quality of leaders themselves. The first is evidence that there is often considerable disagreement about precisely how charismatic a particular leader is and considerable variation in the degree to which people are influenced and mobilized by a leader who is ostensibly charismatic. For example, Crayne and Medeiros (2020) observe that while many Canadians responded enthusiastically to Prime Minister Justin Trudeau's charismatic vision for managing the pandemic, he also had many critics who perceived him to be inconsistent and inauthentic and who were left cold by the “path towards a better, more equal society” that he laid out (p. 5). Similarly, while the Premier of the Australian State of Victoria, Daniel Andrews, was fêted by Labor party supporters as a hero for steering his state through a second wave of infection (by imposing a strict lockdown), he continued to be vilified by many conservatives as “Dictator Dan” (Hall, 2020).

Relatedly, second, leaders’ charisma itself is often found to fluctuate over the course of a crisis. For example, as we noted above, while George W. Bush's charisma decreased in the wake of Hurricane Katrina, 4 years previously it had increased dramatically following the 9/11 attacks on the World Trade Centre (Bligh et al., 2004). Likewise, as the death toll from COVID‐19 started to rise in the United Kingdom, supporters of Boris Johnson started to question his charisma, or at least to doubt its usefulness (Kirkup, 2020). More interestingly, there is also evidence that the relationship between crisis and perceived charisma varies over time. Specifically, when a crisis first becomes salient, leaders are typically seen to be more charismatic than they were before the crisis (or when the crisis was not salient; Halverson et al., 2004; Merolla et al., 2007; Stacey & Pickard, 2020; Williams et al., 2012; Willner, 1984). However, if the crisis endures or, worse, escalates then leaders’ charisma typically declines (Pillai & Meindl, 1998). Indeed, if followers come to perceive that leaders are in some way responsible for—and unable to resolve—the crisis then their charisma typically plummets. In the case of the COVID‐19 pandemic, this pattern has perhaps been most evident in the waning appeal of authoritarian and populist leaders who sought to downplay the significance of the SARS‐CoV‐2 virus (e.g., by referring to it as a hoax, a neurosis, or just a bad cold) and failed to mount an orchestrated societal response (e.g., in countries like Belarus, Turkmenistan, and the United States; Light, 2020; Kramer, 2020).

This speaks to a third problem with the characterological view of charisma—namely that it is ultimately an attribution or inference made by followers (Conger et al., 2000; Steffens et al., 2017). Indeed, this was a point first made by Weber when he argued “what is alone important is how the individual is regarded by those subjected to charismatic authority, by his [or her] ‘followers’ or ‘disciples’” (Weber, 1947, p. 359). This suggests that rather than being a gift that leaders possess, charisma is a gift that followers bestow on them (Platow et al., 2006). More particularly, a social identity analysis suggests that the bestowing of charisma on leaders varies as function of those leaders’ perceived capacity to represent, advance, and protect interests associated with a social identity that they share with followers—that is, to be seen to be “one of us” who is “doing it for us” (Haslam et al., 2001; Haslam & Platow, 2001; Platow et al., 2006; Steffens et al., 2017). This, then, explains why leaders’ charisma—and their capacity to influence group members (i.e., to do leadership) fluctuates with the fortunes of the group. While at the start of a crisis (at a point where “us” has suddenly become salient; Drury, 2012) leaders may benefit from an initial period of grace in which charisma flows from a stronger sense of “us” that leaders are seen to be representative of, over time followers increasingly look to leaders to deliver outcomes that serve to “advance us.”

As the examples above attest, this fluctuation has been very much in evidence over the course of the COVID‐19 pandemic where enthusiasm for national leaders like New Zealand's Jacinda Ardern, Denmark's Mette Frederiksen, Germany's Angela Merkel, Scotland's Nicola Sturgeon, and Taiwan's Tsai Ing‐Wen has been closely linked to their capacity to keep infection and death rates down (Bell, 2020; Leaders League, 2020; Tu, 2020). So, while initially leaders like Trump and Johnson initially benefitted from people's appetite for a collective response to the virus, their charisma and influence generally wore off as the infection and death toll mounted (Russell, 2020; ).

Interestingly, research suggests that this is likely to be the case even if leaders have no hand in unfolding events (Meindl et al., 1985; Pillai & Meindl, 1998). However, in the case of COVID‐19 it was clearly exacerbated to the extent that the ongoing crisis was linked to actions that leaders had taken (or failed to take; e.g., as discussed in previous lessons). At the same time, though, social identity processes also have a role to play in this linkage in so far as they structure followers’ apprehension of both (a) the quality of leaders’ actions and (b) the outcomes that matter in a crisis. Accordingly, while the charisma of Trump and Johnson generally declined as cases of COVID‐19 grew in the United States and United Kingdom, there was still a core of followers for whom it remained undimmed (Graham et al., 2020). Nonetheless, the core point that emerges from the forgoing considerations is that charisma is less an input into the leadership process than an output (Haslam et al., 2020; Steffens et al., 2014). In a crisis it is therefore better for leaders to see charisma as a reward for successful leadership of their group and for delivering outcomes that the group really values (in a pandemic, staying alive) than as a resource to make their leadership successful.

Lesson 10

Leaders will be more effective and seen as more charismatic if they are associated with the achievement of outcomes that are highly valued by those they lead.

Leaders who fail to deliver valued outcomes (or who deliver unwanted outcomes) will become increasingly ineffective over time.

Priority 5: Ready the Group for Mobilization

The readiness is all (Hamlet, Act 5, Scene 2; Shakespeare, 1603/1968, p. 215)

The previous four priorities have focused largely on things that leaders can do to manage a crisis once it is underway. However, there are also important things that they can do to prepare their group for a crisis, so that if a storm strikes they are in a good position to weather it. Indeed, in many ways, this should be seen as a leader's first priority (as it is in our 5R leadership development program; Haslam et al., 2017), but the importance of these things is easier to appreciate now that we have explained why social identity and the identity leadership that develops and sustains it is so important in a crisis.

Focus on Preparing Groups Materially and Psychologically for a Crisis

Most obviously, this preparation takes a material form. This was clearly evidenced in the COVID‐19 pandemic, where countries that had crisis management plans in place were able to respond much more quickly and more effectively to the initial threat that the coronavirus poses. Indeed, this was identified as a key factor of the success of countries like Germany, Iceland, South Korea, and Taiwan in keeping the virus under control (Hsleh & Child, 2020; Farr & Gao, 2020; McLaughlin, 2020; Wang, 2020). Around the world, the cost of not preparing for a pandemic—either by failing to invest resources or by strategic disinvestment—has also been readily apparent. In particular, cuts to public health spending were a forerunner of chaotic health responses that contributed to the rapid spread of infection in a number of countries—notably the USA (Bilmes, 2020; ; Sellers et al., 2020), United Kingdom (Scally, 2020), Brazil (Nunes et al., 2020), and Spain (Hedgecoe, 2020).

Preparation for a crisis, however, is not just material it is also psychological. Indeed, if it is the case that leaders need to cultivate a sense of shared social identity in order to encourage adherence and followership during a crisis (as suggested by Priorities 1, 2, and 3), then it follows that their task should be easier if that sense of shared identity also predates the crisis (Oakes et al., 1994). In line with this proposition, it is generally the case that communities tend to respond more adaptively to a crisis, and to recover from it more quickly, to the extent that they go into that crisis with high levels of social identity capital (i.e., social capital that derives from, and helps to build, a sense of shared social identity; Aldrich, 2012, 2017; Haslam et al., 2018; Helliwell et al., 2014; Jetten et al., in press; Ntontis et al., 2019; Pitas & Ehmer, 2020; Reininger et al., 2013; Williams & Drury, 2009).

Consistent with this proposition, there is plenty of evidence that social capital has been a key resource during the COVID‐19 pandemic. For example, at a community level, a one‐standard‐deviation decrease in the social support subindex of social capital in the United States was associated with a 24% increase in the number of confirmed COVID‐19 cases (P. Brown, 2020). This same pattern was apparent at the country level too, with European countries that had high social capital accumulating between 12% and 32% fewer cases than those with low social capital (Bartscher et al., 2020).

As a corollary, though, it is apparent that four of the countries that initially struggled most to manage the virus—Belgium, the United Kingdom, Brazil, and the United States (see Table 2)—had all been characterized by bitter social division in the period leading up to its arrival. This not only made it harder for leaders to mobilize a sense of common purpose in fighting the virus but also made them less inclined to do so (Gopnik, 2020; Ortega & Orsini, 2020). For example, when commentators urged President Trump to strike a more conciliatory and empathic tone in his press briefings, an advisor noted that “it's not his first go‐to emotion” (Holland, 2020). Likewise, in Belgium (a country that had been unable to form a government for over a year prior to the pandemic), Withrow (2020) observes that “instead of galvanizing Belgium's divided political parties to work together, the pandemic has put further pressure on an already fragile polity.”

Table 2

COVID‐19 cases and deaths in selected countries as at September 1, 2020, ordered by the number of deaths per million people

Country	Population	Number of cases	Number of deaths	Cases per million	Deaths per million
Belgium	11.42	84,948	9,894	7,439	866.4
United Kingdom	66.49	334,471	41,499	5,030	624.1
Brazil	209.47	3,846,153	120,462	18,361	575.1
United States	327.17	5,899,504	183,069	18,032	559.6
Netherlands	17.23	70,071	6,215	4,067	360.7
Canada	37.06	127,613	9,113	3,443	245.9
Russia	144.48	995,319	17,176	6,889	118.9
Germany	82.93	242,381	9,298	2,923	112.1
Denmark	5.80	16,700	624	2,879	107.6
India	1,352.62	3,621,245	67,469	2,677	49.9
Pakistan	212.22	295,849	6,294	1,394	29.7
Iceland	0.36	2,105	10	5,847	27.8
Australia	24.99	25,670	611	1,027	24.4
Singapore	5.64	57,771	27	10,243	4.8
South Korea	51.64	19,947	235	386	4.6
New Zealand	4.89	1,387	22	284	4.5
China	1,392.73	90,383	4,729	65	3.4
Taiwan	23.78	488	7	21	0.3

Source: World Health Organization (2020).

Note: Table only includes data for countries discussed in this review. Data are imperfect insofar as different countries have different reporting procedures and criteria. For example, Belgium has very inclusive criteria for recording deaths as COVID‐related (Shields, 2020), but Russia has very conservative criteria (Burn‐Murdoch & Foy, 2020).

On the other hand, for counties in which there was less marked division prior to the pandemic, the task of uniting citizens around a common strategy for fighting the virus generally proved easier. Contrasting Germany with the United States, the editor of Der Tagesspiegel, Anna Sauerbrey, thus noted that “the polarization is not that great here,” and that this had allowed leaders from different parties to drop their partisan rhetoric and “cooperate and create a common set of rules”. This unity was a particular hallmark of responses in Scandinavian countries where leaders were able to draw on a strong preexisting sense of “we‐ness” to secure high levels of approval for, and adherence to, their pandemic‐related policies (Rawat & Wu, 2020). As Danish Prime Minister Mette Frederiksen (2020) explained, in the context of making a 50‐m pledge to support global research into the virus:

We stand united in fighting the virus…. The virus is new but our response echoes our experience. No country, no company and no organization can win this battle alone. We have to team up and share our solutions.

In a marked shift from Hamlet's time, far from there being “something rotten in the state of Denmark,” Frederiksen was thus able to lead the way in drawing on social identity capital to support an effective response to the crisis. Not only did this mean that her approval ratings surged by 40 points (Statistica, 2020), but, more importantly, it also put Denmark in a far better position to tackle the pandemic. Indeed, its infection rate was around half that of countries like the United States and the United Kingdom (with 108 deaths per million residents vs. 565 in the United Kingdom and 611 in the United States by September 1 2020; see Table 2), but it was one of the first countries to reopen after lockdown and one of relatively few countries to bring about an end to a first wave of infection without experiencing a second wave soon afterwards (Cuthbertson, 2020; Milne, 2020).

Lesson 11

Leaders will be more effective if they have done the groundwork to prepare their group materially and psychologically for a crisis.

Leaders whose leadership has been built on the fomentation of social division will encounter particular difficulties when seeking to mobilize communities in a crisis.

Focus on Developing Identity Leadership Rather than Acquiring a Leader Identity

Writing at the time, many commentators argued that the success of leaders like Fredericksen and Merkel (as well as Ardern, Ing‐Wen, and Jakobsdóttir) could be attributed partially (if not wholly) to the fact that they were women (Bell, 2020; Garikipati & Kambhampati, 2020; Persaud, 2020; Purkayastha et al., 2020; Sergent & Stajkovic, 2020). There is certainly some evidence that female leaders did a better job at keeping COVID‐19 cases and fatalities lower than male leaders. In particular, in the most rigorous study to date, Sergent and Stajkovic (2020) noted that this was true for female state governors in the United States. In line with Lesson 8, they also reported qualitative findings which suggest that this was partly because women governors were more empathic.

As Sargent and Stajkovic (2020) observe, this evidence also chimes with other research which suggests that women tend generally to be more concerned than men with the communal needs of their group (as opposed to their own personal agency; Eagly & Steffen, 1984), and that this in turn makes them appear more suitable—at least from a social identity perspective—for the task of managing a crisis (Eagly et al., 1995; Ryan & Haslam, 2007; Ryan et al., 2011).

Rather, though, than reflecting anything fundamental or immutable about gender, we would suggest that this evidence speaks to the fact that women tend generally to be more attuned than men to the importance of doing identity leadership and hence to the logic that informs the present analysis—and also more practiced in this. Or, to turn this around, they tend to be less in thrall to toxic “masculine” models of leadership that priorities having a leader identity (e.g., as decisive, strong, and agentic; Guillén et al., 2015) over the need to do identity leadership (A. Brown, 2014; Steffens & Haslam, 2020). There are likely a range of reasons for this, but in part this may reflect the fact that women are more likely than men to have a history of “glass cliff” leadership roles which require them both to attend to the needs of others (Eagly & Karau et al., 1995) and to extricate groups and organizations from calamity (Ryan et al., 1997).

Lesson 12

Leaders will be more effective if they are attuned, thorough practice, to the needs of the group they lead.

Leaders who are in thrall to “masculine” models of leadership that place an emphasis on being seen as a strong leader are prone to failure.

Conclusion

This review has sought to shed light on the processes that enable leaders to lead groups effectively through a crisis. More specifically, we have drawn on principles associated with “the new psychology of leadership” (Haslam et al., 2020) to understand the nature of effective leadership during the first 6 months of the COVID‐19 pandemic and to identify key priorities for leaders in a crisis as well as key lessons associated with these priorities. As spelled out above and also summarized in Table 1, these lessons revolve around a model of identity leadership—represented schematically in Figure 2—in which leaders secure buy‐in for their policies by promoting interests associated with a sense of social identity (“us‐ness”) that they share with those they lead. More specifically, this model suggests that one of leaders’ core tasks in a crisis is to recognize the importance of shared identity (Priority 1: Reflecting) and then to build and sustain this through their actions (Priorities 2 and 3: Representing and Realizing). Over time it then becomes increasingly important for leaders to provide support to the group and its members and to deliver outcomes that matter for them (Priority 4: Reinforcing), but this task will be easier if the crisis itself makes shared identity salient (as it typically will, at least initially) and if they prepare the group materially and psychologically for a crisis (Priority 5: Readying).

Table 1

Twelve lessons of identity leadership associated with effective crisis management

Leaders will be more effective if they focus on trying to
1. Achieve power through people not over them
2. Recognize groups as the solution not the problem
3. Unlock people's capacity for strength
4. Build shared identity
5. Treat people respectfully
6. Define ingroups inclusively
7. Appreciate people's differing needs and circumstances
8. Be empathic rather than punitive
9. Provide ongoing support to those who need it
10. Achieve outcomes that people most value
11. Prepare groups materially and psychologically for a crisis
12. Develop identity leadership rather than leader identity

As we have seen, the COVID‐19 pandemic has provided a wealth of evidence which speaks to the importance of identity leadership for crisis management and which supports the practical lessons for crisis management that we have derived from this theoretical framework (see also Jetten et al., 2020a; Van Bavel et al., 2020). Nevertheless, our conclusions need to be qualified by two very significant caveats. The first of these relates to fact that the pandemic is still ongoing. Thus, while our theoretical analysis is supported by several decades of research (e.g., as reviewed in Ellemers et al., 2004; Haslam et al., 2020; Hogg, 2001), tests of the specific lessons that we draw from this in the context of the pandemic are necessarily limited. In particular, much of the research we have discussed awaits peer review and, more generally, our claims remain to be subjected to rigorous quantitative and qualitative analysis. It also remains to be seen how these claims will fare in the fullness of time—as the pandemic peaks, passes, and recedes into memory. Nevertheless, despite this necessary provisionality, even if there are changes in the specific conditions upon which the principles that we have outlined operate (as there inevitably will be), the strength of prior empirical support for these principles (e.g., as confirmed meta‐analytically by Barreto & Hogg, 2017; Steffens et al., 2020) gives us some confidence that the lessons we have outlined will have enduring relevance not just for the management of COVID‐19, but also for the management of other future crises that require large‐scale social coordination and cooperation.

Related to this point, though, a second caveat pertains to the criteria against which we have gauged effective leadership. Throughout this review, we have taken the view that effective management of the particular crisis on which we were focusing (the COVID‐19 pandemic) could be assessed with reference to leaders’ success in keeping infection and death rates as low as possible. By this measure, as the data in Table 2 suggest, leadership in the Belgium, the United Kingdom, Brazil, and the United States can be judged to have been less effective than leadership in New Zealand, South Korea, Australia, Denmark, and Germany. Nevertheless, it is clearly the case (a) that things may change significantly as the pandemic unfolds and (b) that it is also possible to use other metrics to gauge leadership effectiveness (e.g., social and economic indicators). Indeed, social identity principles (and the logic of Priorities 3 and 4) suggest that if groups start to perform poorly against these metrics, their leaders will seek to bolster their self‐esteem (and their leadership credentials) by arguing for the importance of alternative measures on which they perform better (Tajfel & Turner, 1979). Thus Donald Trump argued that his success in handling the pandemic was confirmed by the high number of performed tests and the resulting low proportion which returned positive (Thorp, 2020) and by comparing United States statistics with those from a restricted pool of other countries (Mackey, 2020). Likewise, Boris Johnson argued similarly that his government's prowess was apparent from the large number of tests that had been conducted in the United Kingdom (Woodcock, 2020).

Again, the final verdict on these matters will only be delivered in the fullness of time. For now, all we can say is that these are the most commonly used metrics in this domain, and that they are ones that are most widely used by health organizations and agencies (e.g., the CDC, the WHO). We also imagine that if leaders were doing a better job of restricting infection and death, these are the measures they would use too (as Donald Trump did when criticizing Barack Obama for his handling of the 2019 Ebola outbreak in the United States). Moreover, in a normative sense, we would argue that the most fundamental thing that citizens look to their leaders to do in a crisis is to keep them and their fellow citizens alive. Ultimately, then, it is because effective identity leadership allows leaders to do this that it proves so important. Likewise, the most basic reason for wanting to avoid poor identity leadership is that it is a killer.