Reconsidering Bad Leadership and Bad Theory to Improve Research

Abstract

Given the continued existence, personal success, and organizational damage of ‘bad’ leadership (e.g., with Bankman-Fried, Lay, Welch, Stonecipher), how do we—in academia—do more good than harm in studying and informing others about it? We address that question to improve leadership research more generally. We do so through a dialectic inquiry approach where we join current conversations about teaching and research related to the private benefits and the public costs produced by ‘bad’ leadership. As part of the dialectic, we offer a new model of the phenomenon—where ‘bad’ behaviors can be used to win a contest for leadership where the leader can then act badly or not. Furthermore, we inquire as to the efficacy of how the literature diagnoses ‘bad’ leadership (at either stage), suggests treatments, and attempts to impact practice. We ground our inquiry in two foundational definitions of ‘bad’-ness—one based on psychology’s dark triad, and one based on economics’ self-interest-seeking with guile agent. We critique a relevant, representative set of recent examples to make our points about the significant issues involved in the three elements, across two stages and two definitions. This antithesis-like critique provides the grounds for a set of synthesizing recommendations to improve leadership studies, in light of practical (and often unnecessarily self-imposed) constraints.

1. Introduction

There are two related questions of interest in leadership—how to win the contest to be a leader in an organization, and what the effects are of the winner exercising leadership power in that organization. An important ongoing debate that connects these questions involves the proposition that behaviors labeled as ‘bad’ are more effective in winning the contest to gain leadership, and so should be taught, even though the evidence does not exist supporting the notion that those behaviors can then be ‘turned off’ when the winner-as-leader uses their organizational power (e.g., Cunha et al., 2024). On the hopeful side of the debate, the argument is that teaching about ‘bad’ behaviors (e.g., Machiavellianism) is net beneficial, because it is needed for our students to gain leadership positions over the ‘truly bad’ others (e.g., actual Machiavellians), so that they can then go on to exercise power in ‘good’ ways for the betterment of the firm and society (e.g., Pfeffer, 2021). But, on the cynical side of the debate, there are several reasons to believe that teaching about such ‘bad’ behaviors is net harmful, including that it hurts those who are exploited on the path of the candidate leader trying to win the contest; it makes filtering out truly bad leaders more difficult; it raises awareness in those who are truly bad about the possible ways they may be exposed, allowing them to adjust and not be so easily detected; and it encourages those who learn such behaviors to continue to use them once they gain leadership positions for their continued private benefit (e.g., Clegg et al., 2016).

Our objective is to increase the understanding of ‘bad’ leadership—its forms, its diagnoses, its treatments, and the academy’s roles and responsibilities in teaching others about it (e.g., without in any way endorsing its harmful behaviors). We examine this important debate from a new perspective to direct future efforts in leadership studies away from often static and tautological typologies and towards more fertile ground. Our new perspective considers the study of organizational leadership as a two-stage process and then concentrates on the critical area of ‘bad’ behavior in that process. Specifically, we focus on the cases where, in the first stage of the process, employees who apply ‘bad’ behaviors are more likely to win the contests for leadership positions (e.g., because they manipulate others better, claiming credit for successes while blaming others for failures, often because they are narcissistic or Machiavellian—see Gauglitz et al., 2023, or Lee-Kugler et al., 2025). Arguably, these are the cases where we—as researchers, teachers, consultants, and policy-influencers—can have the most positive impact, specifically by finding ways to reduce the likelihood that ‘bad’ candidates then become ‘bad’ leaders and to mitigate the harms that the ‘bad’ paths to leadership positions can produce (e.g., Fehn & Schütz, 2021).

To move the field towards such positive outcomes, we need to decisively assess how we have tried to make a difference in these areas in the past. Specifically, in this paper, we critically evaluate how past work has considered three relevant challenges: diagnosing the ‘bad’-ness of leadership behaviors and outcomes in the two-stage process; proposing treatments to address that perceived ‘bad’-ness; and having an impact on practice in so doing. From this analysis emerges several paths to advance future leadership research.

The research question that we try to address, then, is the following: how do we do more good than harm in studying and informing others about ’bad’ leadership? We address this question by focusing on two examples of ‘bad’-ness that can be displayed by leaders to their advantage, one from each of leadership’s foundational disciplines—psychology and economics—to illustrate the potential harms and benefits of teaching about, publishing on, and otherwise trying to influence real leadership-related behaviors. We contribute to the leadership literature in several ways: We describe a new model to analyze an important ongoing debate in the field. We analyze the debated issues by grounding them in concepts from foundational disciplines. We provide a structured approach for how ‘bad’ leadership is addressed. We leverage the analysis to direct future work on ‘bad’ leadership using a flowchart-like approach. And we assess the field’s limitations and responsibilities in moving the needed work forward. By so doing, we address gaps in the understanding of the debate over teaching ‘bad’ leadership, over its modeling and management, and over the field’s role in studying and reporting on it.

Our contribution is positioned within the critical leadership studies (CLS) literature, more specifically on the under-researched area of ‘bad’ leadership, and most specifically on the debate over the effects of teaching about ‘bad’ leadership behaviors (e.g., Gelles, 2022; Ghoshal & Moran, 2005; Greenbaum et al., 2015; Kellerman, 2004, 2024; O’Reilly et al., 2014; Sutton, 2017; Willmott, 1993). The research gap we address involves the incompleteness of the current debate, including how the leadership ‘process’ is characterized and how such ‘bad’ leadership can be treated. This gap in the leadership literature includes the causes, forms, effects, and prevention of ‘bad’ leadership (e.g., Pircher Verdorfer et al., 2024; Zhu et al., 2025). Our central argument is that the current debate over ‘bad’ leadership is flawed, and its critical assessment can provide a basis for improving the approaches to understanding and treating ‘bad’ leadership. The aim of the paper is to improve the approaches to studying, teaching, and treating ‘bad’ leadership, thus addressing the gap. We apply a dialectic inquiry approach that includes a critical assessment of the recent debate over whether it should be taught as a useful tool for ‘good’, alternative perspectives (e.g., over what is necessary for removing a ‘bad’ leader), and a synthesis of the two that provides a basis for specific avenues for improvement in addressing ‘bad’ leadership (i.e., its causes, forms, effects, and mitigation). We find that many significant flaws exist in the current literature on ‘bad’ leadership (and its teaching)—regarding its modeling, premises, and suggested treatments. We also find that there are several useful ways forward that exist for the field to consider pursuing. The main implications involve the need for the field to admit its problems and the flaws of past work on this debate and on ‘bad’ leadership more generally, and the need to fix them (e.g., in the areas of diagnosis, treatment, and impact). This also entails breaking down the barriers to such changes in the field. Directions for future studies are detailed and include a step-by-step breakdown assessing the impact, treatments, and diagnoses of ‘bad’ leadership. In the remainder of this paper, we take four steps. First, we provide the relevant background. Second, we introduce a new model along with clear, grounded definitions of what ‘bad’ leadership is. Third, we critique how leadership research has covered the diagnosis and treatment of ‘bad’ leadership, and what impact such research has had, providing alternative theses in a dialectic manner. Fourth, we synthesize our assessment to suggest recommendations for future work to improve the field’s understanding and teaching of ‘bad’ leadership.

2. Background

What is leadership, and why is it important? Several leadership scholars do not believe that leadership actually has a distinctive definition or theory (e.g., Allio, 2018; Washbush, 2005). Others believe it pertains to the ability to influence a group of people towards the achievement of a vision or a goal. For example, leadership has been defined as ‘a process whereby an individual influences a group of individuals to achieve a common goal’ (Northouse, 2010, p. 3). However, some believe that leadership should not be seen as a fixed role occupied by individuals with superior traits, but as a dynamic relationship shaped by competing interests, structural conditions, and ongoing tensions (Collinson, 2014; Fairhurst & Collinson, 2023). Regardless of the specifics, leadership exists and is important given that it is a training–development industry worth tens of billions of dollars annually (e.g., Yemiscigil et al., 2023), and it significantly affects any group’s goal attainment. As such, it is worth scientific study. Leaders are responsible for getting things done—achieving goals—in the organization. They do so through delegation, monitoring, rewarding and punishing, inspiring, training, and explaining. As such, they are but one type of specialist (of many) in the organization who helps its efficiency and effectiveness of production. They provide an important economic function.

Why does ‘bad’-ness matter in leadership? It matters because ‘bad’ leaders cause harm to their organizations, followers, and others (e.g., Entwistle & Doering, 2024; Li et al., 2024; Nie & Wang, 2025), and because they are difficult to dislodge, given the power, resources, asymmetric information, and other factors afforded to their high positions. Furthermore, ‘bad’ leadership is ubiquitous, pernicious, and significant in its effects (e.g., Fehn & Schütz, 2021; Lee-Kugler et al., 2025; Waples et al., 2025). Some even believe it is growing (Hassan et al., 2023). ‘Bad’ leaders can injure employees physically and mentally (Cialdini et al., 2021; Hancock et al., 2023; Nie & Wang, 2025; Shafqat et al., 2025). Even when it is not mean-spirited in nature—when it is in the form of stretch goals or empowerment from a leader, or when it is embodied in the laissez-faire attitude of a leader—it may still lead to organizational damage caused by the unethical behaviors of that leader’s followers (e.g., when the climate is pressurized—see Perkins et al., 2025; Yam et al., 2022; Zhu et al., 2025).

The main issue here is that ‘bad’ leadership has not been studied at a depth consistent with its importance (e.g., Kellerman, 2024; Zhu et al., 2025). One exception to this observation is the CLS stream. CLS challenges the dominant narratives in leadership research, which often portray leaders as heroic figures who drive organizational success through their individual traits, behaviors, and styles (Alvesson & Spicer, 2012; Collinson, 2011). Traditional leadership studies tend to focus on how leaders influence followers and how leadership effectiveness can be measured; they assume that leadership is inherently a force for good (Collinson, 2014; Fischer & Sitkin, 2023). CLS questions these assumptions by exposing the power dynamics, contradictions, and ideological underpinnings embedded in leadership discourse. One of the core critiques CLS offers is the romanticization of leadership and the overreliance on binary thinking (e.g., leader vs. follower, control vs. resistance, or good vs. bad leadership—see Fairhurst & Collinson, 2023). CLS rejects these binaries, arguing that leadership is fluid, context-dependent, and shaped by ongoing tensions between competing forces (Alvesson & Spicer, 2012; Collinson, 2014).

CLS critiques leadership research in several targeted areas. For example, CLS challenges mainstream leadership research’s failure to interrogate power relations and structural inequalities (Alvesson & Spicer, 2012). Beyond power, CLS also critiques the conceptual and methodological flaws in mainstream leadership research, particularly its reliance on tautological reasoning. Fischer and Sitkin (2023) highlight a tendency in leadership studies to define leadership styles by their positive outcomes rather than as independent variables. Collinson (2014) calls for a critical approach to leadership research that acknowledges the contradictions and unintended consequences of leadership practices. CLS calls for an approach to leadership studies that engages the contradictions, struggles, and dilemmas that feature prominently in practice. Our paper follows in the CLS tradition—challenging the leadership research specifically focused on ‘bad’ leadership and calling for changes in the approaches taken so far, as we outline below.

Our study focuses on ‘bad’ leadership. It joins a growing stream in the leadership literature (e.g., Chaleff, 2015; Ivanov et al., 2021; Kellerman, 2000, 2004, 2021; Lipman-Blumen, 2005; Örtenblad, 2021; Pfeffer, 2015; Pircher Verdorfer et al., 2024) that analyzes various forms of unethical and destructive leadership (see Hassan et al., 2023, for a recent review). However, it is one of the few to conclude that ‘we’, as academics, have a professional responsibility to address the subject in all of its forms to grow the good leaders while mitigating the bad (Kellerman, 2024). This paper is also unusual, as we draw from both the trait- and process-based approaches to leadership (Northouse, 2010).

Of course, leadership’s benefits and harms also affect other research areas, such as corporate social responsibility (CSR) and organization theory (OT). In the CSR realm, ‘good’ leadership is mainly considered as a way to enhance CSR outcomes. For example, it has been found that firms engaged in CSR activities benefit more when leaders are authentic and innovation-oriented (e.g., Banker et al., 2023; Fox et al., 2020; Ndone & Kyriakopoulos, 2024). In the realm of OT, ‘bad’ leadership often involves abuses of power and politics (Carter et al., 2024; Cui & Zhang, 2025). For example, when ‘bad’ leaders rely on coercion to achieve goals, followers engage in unethical behaviors to reach them (e.g., Resick et al., 2023; Tang et al., 2024; Yuki, 2006).

3. Methodology

The aim of the paper is to improve approaches to studying, teaching, and treating ‘bad’ leadership. As with other papers in the CLS stream, this paper applies a new, alternative analysis framework, takes a skeptical position, and raises several pointed issues (e.g., on how ‘we’ can overcome the barriers in the field to affect positive change). The paper uses a dialectic inquiry approach, critiquing the existing positions in the debate, offering counter proposals, and finding a synthesis for recommendations. To understand the current sides of the debate about ‘bad’ leadership and its teaching, we draw on a review of the literature summarized in a recent, highly visible outlet (i.e., AMLE), a review that included examples of well-known ‘bad’ leader types (e.g., ‘dark triad’ leaders). By critiquing the premises of the arguments and examples of that literature, both directly and through offering an alternative model, we provide new insights that improve the understanding of ‘bad’ leadership, address the current gaps, and direct future work1.

To apply our method of critical assessment to the examples from the relevant literature, we now explain the basis and structure of the critique. First, we describe our new model for understanding what leadership is—i.e., in terms of gaining the position and using it. Second, we leverage the model to focus on the interesting cases where ‘bad’ leadership has the largest impact. Third, we explore what ‘bad’ leadership is more fully, choosing two main exemplars grounded in the foundational fields supporting leadership—i.e., one from psychology and one from economics. This all sets the table for the analysis of the recent literature over the focal debate.

Table 1. A taxonomy of good versus bad leadership.

		Leader Behaviors
		Bad	Good
Leader-based Outcomes for Stakeholders	Good	Maverick	Paragon
	Bad	Villain	Naive

depicts our proposed two-stage model representing the basic process of organizational leadership. It highlights three facts: (1) that in order to wield the power of a leader, one needs to win the contest to be a leader; (2) that there are various ways to win that contest; and (3) that the behaviors used to gain power do not necessarily translate into the behaviors used to exercise that power. These are important aspects of the leadership phenomenon, and ones that squarely center on the current debate over how to teach, model, and publish research about ‘bad’ leadership.

To provoke new thinking, we do two things: First, our model dissects the basic leadership process differently from past research, allowing us to focus on the most interesting cells in the matrix—the cells on the left where ‘bad’ behaviors win leadership and its power. (The other two cells are less interesting—either where we can offer accolades for well-behaving emerging leaders to do good things for the firm, or we can provide admonitions for well-behaving emerging leaders to toughen up when using power against outside forces.) It is with these two focal cells that we can do more interesting things, such as help diagnose ‘bad’ behaviors and propose ways to treat it, both on the way to leadership positions and then when leaders gain power. Second, we take a new perspective on the current debate about teaching ‘bad’ leadership. Instead of focusing on how audiences apply our assumed-as-accurate research about ‘bad’ leadership, we focus on the underlying premises involved—premises concerning the accuracy of the research and the interpretation and impact of it on audiences. By critiquing the premises underlying the two focal cells, we can provoke questions that this academic field must address to be better at helping organizations have good leadership.

Scholars approach the subject of ‘bad’ leadership from different angles with different labels (Mackey et al., 2021), including but not limited to abusive supervision (e.g., Tepper, 2000); toxic leadership (Lipman-Blumen, 2005); negative leadership (Schilling, 2009); and destructive leadership (Einarsen et al., 2007; Thoroughgood et al., 2012). Some scholars focus on ‘bad’ traits of the leader, such as those described in a dark personality (i.e., Kaiser et al., 2015). Some use specific ‘bad’ leader behaviors, such as those found in abusive supervision (Tepper, 2000) or destructive leadership (Thoroughgood et al., 2012). Others consider ‘bad’ leadership as a systematic phenomenon that results from a confluence of the leader, the follower, and the context—as captured in the toxic triangle (Padilla et al., 2007) or the bad leadership web (Kellerman, 2005). Nevertheless, the commonality across the multitude of perspectives appears to be that the ‘bad’ leader inflicts harm on other stakeholders by causing ‘bad’ outcomes, often by engaging in unethical behaviors. Broadly speaking, by contrast, there is an agreement that ‘good’ leadership produces good outcomes for stakeholders and ethical behaviors.

We define ‘bad’-ness in leadership in a similar manner to the existing work (see Fischer & Sitkin, 2023, for a review). For the dimensions in our table, ‘bad’/‘good’ is what produces imminent net harms/benefits to focal organizational stakeholders (i.e., followers, peers, shareholders), harms/benefits that could be financial, emotional, reputational, positional, professional, or physical, and that most often come with private benefits for the leader. In this model, we delineate traditionally ‘good’ (i.e., the top right quadrant, where both dimensions rate as ‘good’) and traditionally ‘bad’ (i.e., the lower left quadrant, where both dimensions rate as ‘bad’) leadership. We then add two modalities that are relatively less discussed, but not uncommonly observed in real life (i.e., the top left and the lower right quadrants, where a mix of ‘good’ and ‘bad’ leadership occurs).

To initiate our critical assessment of the relevant leadership literature, one focused on the debate about teaching ‘bad’ leadership, we now turn to exemplifying that ‘bad’-ness from two contrasting perspectives. Each one is famous and grounded in a foundational discipline of leadership studies. One comes from psychology and is inductive, while the other comes from economics and is deductive. From psychology, the recent literature has focused on the dark triad (henceforth denoted as DT) of leadership traits (e.g., Cunha et al., 2024; Lee & Ding, 2024; Paleczek et al., 2018). The triad includes Machiavellianism—as the cynical, planned, strategic callous pursuit of self-serving goals—which is captured in survey measures like ‘The best way to handle people is to tell them what they want to hear’ (Mach-IV); psychopathy—as egocentric, impulsive, deceitful recklessness without remorse—which is captured in survey measures like ‘Pathological lying’ (PCLR); and narcissism—as acting with self-interest and manipulating others without empathy—which is captured in survey measures like ‘I find it easy to manipulate people’ (NPI) (Paulhus & Williams, 2002).

From economics, the recent relevant literature has focused on the decision-making logic of economic man—homo economicus (henceforth denoted as HE)—the benchmark conceptualization of an opportunistic, super-rational agent (Jensen & Meckling, 1976; Williamson, 1975). This decision-maker seeks self-interested outcomes with guile (Williamson, 1993)—expecting all others to do the same. This leader attempts to maximize their own utility, regardless of the effects on others, while assuming every other agent is doing the same. That behavioral assumption simplifies the identification of the decision-maker’s best choices in any decision (e.g., by reducing the optimization calculations to focus on only the expected monetary rewards).

In each example, the leader is ‘bad’ in terms of being amoral if not immoral, and not valuing, if not exploiting, others. These leaders are looking to maximally benefit themselves regardless or more likely at the expense of others. The implicit expectation is that they will do harm to stakeholders on either their path to gain power, or when they use that power, or both.

4. Analysis

With the methodology, model, definitions, and examples understood, we now proceed to assessing how the current leadership studies may harm rather than help organizations deal with the focal problem of ‘bad’ leadership. Specifically, we consider the problem of the ‘bad’ behaviors exhibited by those trying to become, and those who are, their leaders. As leadership researchers, the way ‘we’ can help with such a problem parallels how a doctor treats a sick patient—by diagnosing the problem, formulating its treatment, and then hoping that the recommendation is heard and adopted by the patient. Thus, these are the three areas where we focus our critique. Regarding the diagnosis of the ‘bad’-ness in individuals, we assess whether the suggested real-life measures are fair and accurate. Regarding the suggested treatments, we consider whether these are reasonable and unlikely to backfire. And, regarding the adoption of the field’s findings, we examine how much of a difference to practice our research and teaching about these problems, models, and solutions can and does have.

4.1. Critiques over the Diagnoses of ‘Bad’ Leadership

One way that the leadership literature can make the challenge of addressing ‘bad’ leadership worse is to apply faulty diagnosis methods. We consider this issue focusing on three problems: (i) the questionable accuracy of observational measures; (ii) the inherent problems with outcome-based measures; and (iii) the not uncommon disregard of alternative explanations for questionable short-term behaviors and outcomes.

The Premise of Accurate Trait-Focused Diagnoses—We analyze the way that DT leadership is diagnosed in order to raise serious questions about whether ‘bad’-ness in behaviors can be accurately measured in real time through observation. What most relevant studies of DT do, however, is to apply only subclinical (i.e., low-severity, often non-observable) measures of some very toxic and rare underlying disorders. Note that clinical psychopathy is very rare, even in prison populations; and, while that is very much not what is being considered in most papers (e.g., Cunha et al., 2024; Malik et al., 2024), it is often implied to casual readers (e.g., business students) who key off on such labels. What is actually being considered in leadership research is the statistically noticeable variance in relatively low-level scores from surveys filled out by individuals about themselves or about co-workers, thus being subject to a myriad of biases and recall issues. Furthermore, the casual accusations of the kind of identified-after-the-harm-is-done ‘bad’ DT-style leadership (e.g., that of Welch, Holmes, Fastow, Fuld, Ghosn, Dunlap, and others), often peppered into the introductions of such papers, actually smacks of an uncomfortable level of unprofessionalism; this is because no expert in psychology would ever offer a diagnosis of such underlying traits without a full clinical assessment.

To be clear, even in the foundational DT paper of Paulhus and Williams (2002), the sole focus was on sub clinical measures. The paper’s assessments of narcissism and psychopathy held positive and significant correlations with ‘good’ traits like extraversion and openness. And it was based on a student-only sample. The authors concluded that the root of their DT members’ social destructiveness was banal, and that the psychopathy measured was within the ‘normal range of personality found in [their] sample’ (p. 561). Furthermore, they held the view ‘that no personality trait is universally adaptive or maladaptive…’ (p. 561). Thus, what is actually being measured and discussed in the DT leadership debate is not what the serious underlying clinical conditions speak to but something much less severe and observable. This raises the question of whether such ‘triggering’ labels have been abused as a cheap means to gain attention (e.g., from editors or the media) rather than to seriously explain ‘bad’ leadership.

Removing the hyperbole of the DT label severely weakens the intuitive connection of ‘DT leaders’ to harmful outcomes. Three serious concerns result: The first involves commonality. If most people display subclinical levels of these traits—some of which are correlated with positive traits—then the idea of filtering out ‘bad apples’ based on them is alarmingly impractical given the strong evidence of such commonality of those traits in the literature (Greenbaum et al., 2015; Paulhus & Williams, 2002; Pfeffer, 2021). The second concern involves a failure to capture the whole person. A leader may have offsetting positive traits, given that both kinds of traits ‘are present and intertwined in even the best of leaders’ (Gamble & Christensen, 2022, p. 590). That would make a myopic focus on filtering by DT traits alone unfair and unwise. The third concern involves functionality. If the leader can control the harmful use of the underlying traits, or if they only are activated under specific situations (e.g., under the ‘toxic triangle’—Padilla et al., 2007), then there may be opportunities to responsibly exploit whatever benefits such behaviors can have without incurring offsetting costs. To sum up, based on the serious concerns raised regarding the use of a DT-based measure to diagnose ‘bad’ leadership, we are doubtful about the validity of the results of such research on bad leadership.

The Problems with Behavior-Focused Models—Diagnosing a ‘bad’ leader after-the-fact—after the damages have been done—is of relatively low value. It may also be an inaccurate attribution. A bad outcome is not always caused solely by the preceding behaviors and actions of a leader. Thus, a major concern about the extant research on both ‘good’ (e.g., transformational, authentic, empowering, ethical, or servant) leadership and ‘bad’ (e.g., abusive or destructive) leadership is the failure to determine causality, specifically the causality of outcomes due to leadership behaviors (Fischer & Sitkin, 2023). One reason for this failure lies in the inherent tautology within many leadership models. That is, many models define their specific leadership style’s goodness or badness in terms of the goodness or badness of the effects on others (e.g., Popper, 1959). ‘Many definitions of authentic leadership also include outcomes’ (Alvesson & Einola, 2019, p. 390). For example, van Knippenberg and Sitkin (2013) identified the confounding of a conceptualization of leadership and its effects as a fatal flaw in the transformational–charismatic leadership research. And, when the cause is defined in terms of the effect, any empirical support of the relationship is questionable.

Beyond the tautology issue, the ‘bad’ leader behaviors measured are often conflated with other factors, which casts a further doubt on the validity of any findings. This is especially noticeable when those allegedly harmed are used to assess the ‘bad’ leadership. For example, in abusive supervision, the assessment of the item ‘my boss blames me to save himself/herself embarrassment’ (Tepper, 2000, p. 190) conflates the leader behavior (i.e., blames me) and the evaluation of the underlying intention (i.e., to save himself/herself embarrassment) (Fischer & Sitkin, 2023). In sum, diagnosing ‘bad’ leadership based on ‘bad’ outcomes is prone to error, given the tautology and conflation concerns, as well as possible measurement issues (e.g., motivated perception) involved.

The Disregard of Alternative Explanations—There are often other viable explanations for ‘bad’ behaviors and ‘bad’ outcomes than simply ‘bad’ leaders. When research fails to test for alternative explanations, the validity of its conclusions is questionable (Antonakis et al., 2010). Consider ‘entrepreneurial orientation’ as an alternative explanation for some often seen as ‘bad’ short-term leader behaviors, such as aggression and narcissism. Entrepreneurs, in particular, are prone to being associated with DT-related behaviors (e.g., Kraus et al., 2020). For example, based on the biographies of entrepreneurs who helped shape the commercial landscape of the United States, Miller (2015) found high levels of DT traits. This is not surprising, given the overconfidence needed to pursue new markets and technologies, and the narcissism it takes to be a creative and generative leader in such emerging, embryonic competitive environments (Humphreys et al., 2011). Yet such ‘entrepreneurially DT’ leaders can be net beneficial to society, for example, by bringing product innovations to market. A separate way that such leaders may provide social benefits is through a different kind of creative destruction (Schumpeter, 1935)—not one based on the introduction of a new technology, product, or organizational form, but on the changing of outdated social morals (Brenkert, 2009). These entrepreneurs may bend, or even break, a given set of current ethical standards for the long-term benefit of society by helping spur people to reconsider the then-current norms as in need of updating (e.g., about women’s rights). This ends-justify-means argument is not new, but for some categories of leaders (e.g., entrepreneurs) it is not given sufficient consideration as an almost necessary form of means.

Now consider a different alternative explanation—this time relating to a ‘bad’ outcome to outsiders but not to insiders. A business case for ‘bad’ behaviors by the leader—with the followers’ approval—is made; here, ‘bad’ leader behaviors are supported by the organization because the outcome is ‘good’ for the firm’s participants even though it is ‘bad’ for society. The narrative depicted by most recent studies fails to capture the business case leveraged by the many so-called DT leaders to gain and hold onto their power (e.g., Cunha et al., 2024; Paleczek et al., 2018; Palmer et al., 2020). Instead of assuming that all such leaders do what they do purely for power or self-interest or to inflict pain on others, as would be the nature of their supposed traits, it seems much more reasonable to assume a more rational alternative. The alternative is that they find followers and at least short-term success because they have built a solid business case that benefits at least some substantial set of stakeholders for the near-future, regardless of the harms it inflicts on others (e.g., through rent-shifting, or due to the negative externalities involved in the underlying business model2). Such models define a toxic triangle and are often easy to rationalize (e.g., with accusations that other organizations are doing it, or that it is not illegal, or that the victims are outsiders)—thus also fulfilling the fraud triangle’s motivation, opportunity, and rationalization requirements (see Cressey, 1953). We do need much more research on the kind of business models that are leverageable by ‘bad’ leaders if any serious attempt is to be made to prevent their harms. Indeed, any diagnosis of ‘bad’ leaders must take alternative explanations of their behaviors and outcomes into account to help improve the state of leadership in the organizations that we seek to advise.

4.2. Critiques over Possible Prescriptions

A second way that the leadership literature can make the challenge of addressing ‘bad’ leadership worse is to prescribe faulty treatments. We consider this issue by focusing on the treatments of DT and of HE. The suggested treatments of DT are based on psychology and speak to observed behaviors and the seemingly practical actions that could be taken to address them. The suggested treatments of HE are based on economics and speak to the theoretical–mathematical modeling of phenomena. For the former, we question whether the treatments warn the users sufficiently of the details and conditions required for success. For the latter, we question whether the treatments warn the users of the strong premises being invoked, and whether any predictions of their use represent benchmarks rather than realistic estimations.

Dangerous Treatments for Addressing DT Leaders—There are two treatments suggested by the literature to address DT leaders. The first involves filtering them out prior to their gaining power. The second involves removing them once they have gained power. In the first case, it has been suggested that our students can be effectively taught to act as psychologists to diagnose their peers, their supervisors, and their organization’s leaders competently to identify the ‘bad’ apples (e.g., Cunha et al., 2024). This, of course, is dangerous. It is dangerous because those students do not actually have the underlying psychological training or the access to the kinds of information required to make a reasonable diagnosis. Furthermore, this treatment is open to many abuses (e.g., by those wishing to become leaders by misdiagnosing their peers and supervisors as ‘bad’). Simply put, it is quite difficult to filter out DT leaders early on, just based on the simple measures taught in leadership courses or found applied in leadership research. To drive that point home, just consider the following: recall that we, in academia, have all too often celebrated those leaders who turned out to be ‘bad’3, as have award-winning investigative journalists, experienced board members, and sophisticated investors—all of whom had much greater incentives to identify the dangers of such leaders to organizational value than a mere student or student employee would.

The second dangerous treatment involves the suggested steps for removing a ‘bad’ leader once they are in power. The quip ‘if you come for the king, you best kill him’ seems strangely disregarded in our literature. The well-intentioned solutions offered by our peers (e.g., in Cunha et al., 2024) raise significant questions over their originality, implementability, and especially their naiveté. Regarding originality, recommending better screening for ‘bad’ apples and greater constraining governing mechanisms for those in power is hardly new when it comes to what are essentially HR practices that organizations should be adhering to anyways (see Cohen & Baruch, 2022, or Schyns et al., 2022). Regarding implementability, many proposed solutions to remove ‘bad’ apples necessitate that there be no ‘bad’ barrel to begin with. But that is exactly what would attract a DT leader. In other words, the conditions for successful implementation may be least likely to be met when they are actually needed (e.g., Li et al., 2024). And, regarding naiveté, there is no shortage in the solutions suggested. To expose, let alone remove, a ‘bad’ leader takes an extraordinarily underappreciated amount of effort. It involves an effort that not only requires the moral courage of specific employees but also the added efforts of outside entities. To be clear, ‘it takes a village’ to confront and defeat an actual entrenched DT leader. It takes an interested press, stakeholders, and often lawyers. It takes access to sensitive information, patience, and unrelenting effort. In other words, successful leader displacements are rare exceptions to the rule; so, teaching that students alone can solve a real-world DT problem with any of the solutions provided is simply bad practice. Real DT leaders will take pleasure in ‘killing any messenger’ who is trying to expose their true traits, and will do so in many ways and through many parties, including through coopted HR departments and complaints systems. So, it is extremely dangerous and irresponsible to under-prepare students for the likely reaction to opposing any real DT leader. In sum, the kind of treatments that ‘we’ as leadership scholars can suggest are often too optimistic and too simple to be applied; as such, we need to be very careful in warning about the conditions in which they may work and, especially, in which they may fail.

The Dangers of Basing Policies on Benchmark Models—Now consider a different type of treatment of ‘bad’ leadership. This is not about filtering out or removing leaders but instead is about dealing with them being ‘bad’. Treatments involve crafting policies based on assuming a type of ‘bad’ leader, here an HE leader. Specifically, we consider policies that are based on assuming that all employees are HE. While this may not seem like a reasonable assumption, it only takes a few employees ‘being HE’ to generate a self-fulfilling prophecy where mutual distrust leads to others adopting the same approach.

The treatments of HE embody the majority of neo-classical economic modeling. And that modeling is advantageous in producing clear results. The mathematical simplicity of capturing a universal decision-making rationale provides clear, provable, explainable answers to a wide variety of important interactions—between principals and agents, between contracting parties (inside and outside a hierarchy), and between policymakers and firms (e.g., Gil & Ruzzier, 2018; Williamson, 1975). The cost of such clarity and optimizability of models, however, can be significant. It is most significant when the predictions and prescriptions from these unrealistic models are not recognized as benchmarking an extreme scenario but, instead, are implemented ‘as is’ in a non-extreme context. When that occurs, it can negatively affect contractual terms, negotiations, property rights, investments in specific assets, choices of governance forms, information sharing, trust, and communications. In fact, it mostly guarantees that only suboptimal solutions and further distrust and defensiveness will result. That bred distrust will cost affected parties, often heavily, in terms of what ‘could have been’ under a trusting, longer-term, more-than-transactional relationship. For example, it will rule out the kinds of relationships that involve mutually beneficial innovation along the supply chain or between alliance partners, the learning and benchmarking of better practices, and greater flexibility (e.g., Xie et al., 2022).

Economists have developed alternative views to the HE model to assess its inherent drawbacks. For example, the homo reciprocans model suggests that humans are often motivated by fairness and reciprocity beyond pure self-interest. Such humans often engage in cooperative behaviors (Bowles & Gintis, 2002). These economic theories suggest that, although humans possess egocentric, even dark, personality traits, it is hardly practical for a leader to act purely on the HE assumptions in a well-functioning social environment, especially outside the short-term. In sum, treatments as policy, based on dealing with ‘bad’ leaders ‘as they are’ as captured in their behaviors predicted by the HE model, are not only wasteful, but can also build distrust and, ironically, further endorse such ‘bad’ leadership.

4.3. Critiques over the Impact of Leadership Research

In this third area of critique, we move beyond diagnosing the patient (i.e., as a ‘bad’ leader) and the suggestion of treatments (i.e., to mitigate their harms). We now focus on assessing whether those prescriptions are followed and have the intended impact. To do so, we consider both the impact of our teaching and our research on leadership.

The Impact of Teaching about ‘Bad’ Leadership—To begin with, there simply is no evidence to support the premise that business school teaching has had a significant real-world impact on business, other than through the introduction of a handful of finance-focused techniques, like the Black–Scholes option valuation method. Every piece of so-called evidence of impact in past studies is questionable. For example, and specifically related to the impact of teaching about ‘bad’ leadership, the highly cited 2001 Aspen study espousing significant findings has been contradicted by the 2003 study (Aspen Institute, 2003). Even the refreshing change in attitudes towards ‘bad’ leaders noted in the 2003 study was not due to our teaching but instead due to the spate of 2002 corporate scandals that heightened respondents’ understandings that bad behavior could be punished. That lack of impact is further supported by the finding that the survey question about the possible causes of corporate misconduct, suggesting that it is attributable to ‘the priorities communicated during the MBA program’, was ranked lowest by respondents of nine possible factors. Other studies suggesting significant influence on our students often do not control for self-selection, contextual factors, contributory personal characteristics, and previous ethical education, or consider behaviors outside of university or lab conditions, even while acknowledging the importance of such factors (e.g., Gioia, 2003). Even the highly cited McCabe and Trevino (1995, p. 213) paper raises doubts over our teaching impact in this area, since the cynicism of our students often blunts the points we try to make about ethical issues, like ‘bad’-ness in leaders. Other papers often cited for evidence of teaching impact, such as that by Williams et al. (2000), have their share of questionable empirical choices, like having firm size appear on both sides of the regression, sampling from only the largest firms, and interpreting only the negative results. If there was hard evidence of ‘our’ instructional impact—especially for promoting or suggesting the benefits of ‘bad’ behaviors, then top-tier schools would have been sued already by the victims of their graduates.

Constraints on Our Teaching Impact—We are constrained in what we teach about leadership in both its specificity and its realism. What we teach about leadership in the classroom is necessarily at a general level. Thus, it is not immediately implementable in the real world because contingencies and details matter in the translation of simple ideas to hands-on practice (Bennis & O’Toole, 2005). Furthermore, what we teach is almost exclusively based on picking up linear patterns from backwards-looking empirical studies of the real world. So, it is almost always the case that what is working already—what is being done already about leadership—is what we are teaching and prescribing. In other words, it is more often the case that we, as teachers, are being taught and influenced by business rather than the other way around.

Despite these constraints, several scholars in the ‘bad’ leadership stream believe that our lessons can have an impact on future student-as-leader behaviors (e.g., Cunha et al., 2024; Pfeffer, 2021). We question those beliefs. We doubt that such behaviors can be taught, let alone be taught to be applied strategically, in a business school classroom. It is no secret that people can be taught to lie and manipulate others (either directly or indirectly through imitation and experience) and to do so in specific settings. (Intelligence agencies do this regularly.) Whether ‘we’ can do so effectively—and specifically for the DT traits—is unknown, but somewhat doubtful, given much of this would require tacit training rather than simply providing a set of operating procedures. (Instructing someone to lie convincingly under pressure is much different than them being able to do so.) The point is that, if we cannot teach students how to be DT leaders effectively, then our possible negative impact on such behaviors should perhaps be ignored.

A further concern that is raised in the debate about the teaching of DT traits involves the misuse of those lessons. For example, teaching actual DT leaders how ‘we’ might act to filter them out or otherwise frustrate them may just give them a better means to defeat ‘us’ attempting such acts. This possibility—where there is an escalation of countermeasures by parties with competing interests—is challenging to model completely, and so leaves unanswered the question of the ‘level’ of knowledge we need to teach to be effective in our intentions. Is it when we provide the solutions to mitigating the actions of DT leaders who are informed of the first set of solutions, or the next set, or some other set? Where we end up is likely to muddy the waters for all relevant parties. In sum, our impact on ‘bad’ leadership through teaching is likely constrained, if not minimal or mitigated by countermeasures in reality.

The Impact of Research on ‘Bad’ Leadership—Ghoshal (2005) labels a theory as ‘bad’ when it assumes the worst about human behavior. For example, Williamson’s (1975) assumption that people always act guilefully in their own self-interests makes for ‘bad’ theorizing. But this is an incredibly ‘bad’ (i.e., inappropriate) label for assessing theory (e.g., Cunha et al., 2024), given that objectively bad theorizing has nothing to do with someone’s ethical judgment of a choice of premises and most everything to do with the quality of an explanation for a focal phenomenon (see Arend et al., 2015). That label also implies an inaccurate cause of real phenomena. ‘Bad’ leadership behaviors are unlikely to be caused by those leaders being exposed to our teaching or our research articles. In this case, blaming the poor treatment of workers or the abuses by CEOs on academic work espousing industry analysis frameworks and principal–agent models (as visible targets of Ghoshal) is simply unreasonable4. There is no evidence that our published research has had significant impact on business behavior; in fact, quite the opposite is true, given most research is empirical. There is no evidence that we even provide a consistent message promoting bad behaviors or pessimistic assumptions. There is evidence that we often do just the opposite, given the 40-plus years of teaching about ethical behavior, Theory Y, compassionate leadership, and CSR5. When taught correctly, students will know that the benchmark characterized by HE is an abstracted, one-dimensional, extreme model of a decision-maker. (It is a ‘hack’ to make the analysis of already simplified games and optimizations mathematically tractable.) Furthermore, consider the questionable weight given to the highly cited quote from the AMLE Ghoshal piece regarding ‘our’ research’s influence on business. First, note that it is actually from the very last paragraph in Keynes’ 1936 book. Keynes speculates, without any evidence, that what drives practical men, at a time between two world wars, is the ideas of economists and political philosophers rather than their own vested interests. (Note that the self-interested Nazism that soon followed did not abide by any economic philosophy of the time—see http://www.nber.org/chapters/c9476, accessed on 2 February 2025) How Keynes’ pure speculation applies to events 80-plus years on (i.e., after major wars, major economic cycles, and upsurges of contrasting economic, political, and philosophical theories) is a mystery to us, as it should have been to Ghoshal and to those who quoted him on that quote (e.g., Cunha et al., 2024). Keynes provides no real evidence of ‘our’ influence at all.

As with our teaching, our research is also constrained in many ways that make it unlikely to have a significant effect on practice (other than in very specific instances, like in finance). The logic of the double-hermeneutic—i.e., that the fellow humans we study can adjust behavior to what we publish, in a co-evolving system—assumes that our initial impact, if any, will not be long-lasting. That is a significant constraint on the longer-term impact of any research. Further, we study and report on only parts of a complex, interconnected system, doing so with a substantial lag to its observation due to lengthy publishing processes, while we compete with more proximate reporting observers (e.g., the business press, consultants, professional networks) in so doing. So, it is no surprise that there is little evidence that we alone, and in real time, are altering practice in any significant manner. Even if we could, with the equivalent of telling people to ‘eat healthier diets and exercise more’, there is no guarantee that we would ever actually alter others’ behaviors. Thus, compared to the natural sciences, and partially due to such constraints, it should be no surprise that ‘the achievements of the social and behavioral sciences and their impact on society may not appear particularly impressive and their relevance for practice can appear doubtful…’ (Ferraro et al., 2005, p. 8).

5. Findings and Discussion

We now summarize our main findings and discuss their implications.

5.1. Main Findings

Applying a dialectic inquiry approach that leveraged a structured, critical assessment of representative opinions over an important debate in the leadership literature has proven fruitful. It has revealed that the diagnosis, treatment, and impact claims about ‘bad’ leadership are all questionable. Regarding the diagnosing of ‘bad’ leaders, the validity of the kinds of psychological measures used is disputable, as is the application of outcome-based measures. Furthermore, alternative explanations for behaviors and outcomes are too often ignored in the process. Regarding treatments for ‘bad’ leaders, these are also dubious, given that filtering requires the use of questionable diagnosis tools, and that removal requires resources well beyond any one employee’s control. Furthermore, conducting policy defensively—based on the assumption that all actors are ‘bad’, including leaders—is likely both wasteful and self-fulfilling. And, despite questionable studies and historical quotes to the contrary, our impact on ‘bad’ leadership through teaching and research is constrained and minor in reality. Ironically, this is likely a ‘good’ thing given the distrust in the value of the field’s body of knowledge espoused by several prominent scholars (e.g., Kellerman, 2024; Washbush, 2005). We interpret these results as evidence of a flawed research area. Specifically, leadership, as an academic field, is less than helpful in resolving the debate over teaching ‘bad’ leadership. Currently, the field involves questionable premises and a poor underlying model that fails to consider both steps in the leadership journey (i.e., of gaining and using that powerful role). As such, the field requires some effortful reconstitution to reach a valid conclusion on the debate.

5.2. Implications

The results from our dialectic inquiry have both theoretical and practical implications. On the theoretical side, we added a new model to the field to help peers envision the leadership journey (see Table 1) and analyze what ‘bad’-ness is involved. We also provided arguments to question the premises of current theoretical sides of the focal debate about teaching ‘bad’ leadership. We concluded that the current state of the field’s relevant definitions and premises are weak. That weakness contaminates its measures, constructs, and theories (Allio, 2018; Washbush, 2005; Yoos, 1984). The field needs to strengthen its fundamentals. One way to do so would be to take CLS seriously, and explore dysfunction in leadership (Collinson & Tourish, 2015) from a plurality of perspectives (Learmonth & Morrell, 2017).

On the practical side, there are two suggestions we can make for management. First, take any advice from our journals about ‘bad’ leadership with skepticism, given that the underlying diagnosis tools are questionable and the consequent treatments likely flawed, if not also naive. Second, pay more attention to the ‘bad’ leadership, ‘bad’ followership, and ‘bad’ contexts at your organization. Use postmortems, 360-degree evaluations, and additional training to improve the understanding of leader behaviors. Although costly, doing so can be paid back in terms of the ‘bad’ leadership-based harms avoided. For example, building ‘good’ norms can mitigate the damaging behaviors of ‘bad’ leaders (Bidner & Francois, 2013). Addressing negative employee behaviors early on reduces the chances of subsequent destructive leadership (Li et al., 2024). And growing the wisdom-enhancing competencies of nascent leaders through such reviewing and training can increase the chances of lasting ‘good’ leadership behaviors (Kerns, 2020).

The main practical implications of the paper, however, are directed at academics, and specifically those in the leadership field. The practical suggestions are focused on improving the modeling, understanding, and education about ‘bad’ leadership to do less harm and provide more benefits and insights from multiple perspectives. We advocate not prematurely shutting down conversations about the complex phenomenon of ‘bad’ leadership. The specific directions and expected impediments to implementing these suggestions are detailed below:

5.3. Consequent Directions for Improvement

Synthesizing the analysis thus far, we propose that improving the current state of the field, especially with respect to ‘bad’ leadership studies, requires following a logical flow of action, charting in reverse the critical assessment we followed so far. Start with the impact of the field. Then move to building treatments for ‘bad’ leadership. And, finally, consider robust ‘bad’-ness identification methods.

Regarding impact, addressing our critique-based concerns should start by figuring out whether we make a difference or not to leaders, leadership, organizations, policymakers, and society. If we do not, then we should look to other fields that do make a difference to such audiences to determine if we can do what they do. But, if we find evidence that we do make an impact, then we need to check whether the lessons we provide and the prescriptions we base on theory and empirical testing are absorbed and used as intended (i.e., in the direction expected—for ‘good’). If not, then we should consider how to alter our message and add safeguards to obtain the direction desired. But, if so, then we should determine what we can do to make the message more effective—to build on it and expand it in ways helpful to the field and its stakeholders.

Taking control of this impact, however, means taking greater responsibility and asking tough questions, like the following: Should the schools (e.g., Ivy), programs (e.g., MBA), and pedagogies (e.g., experiential) that have more power to make an impact on tomorrow’s leadership be made more accountable for the negative impacts on society caused by their graduates, and what would that look like? Do we need to create pedagogical standards for teaching leadership, such as teaching students to challenge the binary notion of ‘good’ versus ‘bad’ leadership, to abandon the heroic view of leadership (Collinson & Tourish, 2015), and to facilitate an integrative and nuanced understanding of leadership that is located in the web of cultural, power, and ideological tensions? Can we assume more control of relevant research, in terms of our responsibility for taking difficult actions, such as questioning, critiquing, and then explicitly setting aside bad theories and leaders and professional structures and pedagogical approaches, all in order to move society to healthier and less harmful outcomes?

Regarding treatments, assuming we have proven that we can have an impact in the leadership realm, then we want to try to formulate reasonable and robust prescriptions for addressing ‘bad’ leadership. Such prescriptions should be explicit in the minimum conditions required for success and include contingencies. If research points to the conclusion that no ‘first-best’ solution is possible, then we need more efforts in identifying the tolerable and feasible ‘second-best’ remedies (e.g., attacking the other two parts of the toxic triangle) and the tradeoffs those incur. We then need to synthesize any conflicting treatments currently prescribed by finding the contingencies that underlie those conflicts. We need to be explicit regarding at which stage a treatment is more likely to work, and the risks and costs involved in its implementation, along with any legal and ethical constraints. If there are no generalizable solutions, then we also need to be explicit on that point and perhaps focus on case studies to make students and practitioners more aware of the limitations on what we can responsibly prescribe at that level.

Finally, regarding diagnosis, assuming that we can both make an impact and formulate responsible treatments to address ‘bad’ leadership, then we want to try to find responsible, fair, and objective ways to identify ‘bad’ leaders, preferably before they can do their largest harms. Rather than relying on clinical psychology or waiting for the harm to begin to be done, perhaps more and deeper post-mortems on legitimate organizational failures by leaders need to be conducted. Those results would not only help us to better determine their exact causes (e.g., DT traits) but also to better understand why so many stakeholders (e.g., trained investigative journalists and experienced investors) were fooled for so long. If it turns out, however, that we can only verify a bad leader after the fact, then we need to admit that and then focus much more on big investigations and punishments to dissuade more leaders from doing bad. Furthermore, we need to do more research to identify where in the organization (e.g., level; function) the verification of a ‘bad’ leader is best accomplished and whether any proposed filtering there could run afoul of legal or ethical issues (e.g., such as allegations of discrimination given that, for instance, DT is more common in males).

5.4. Barriers to Change

There should be no debate that ‘we’ need to do something to improve leadership studies. We need to pursue research and teaching that especially reduces ‘bad’ leadership. But, in reality, there does not appear to be any viable way forward that guarantees the prevention of ‘bad’ leadership, let alone that makes such leaders accountable for the harm they do. While some may have believed that the field could make a positive difference (e.g., Gioia, 2003), that seems to have been contradicted by the ‘no leader stopped or held accountable for’ the 2008/9 financial crisis or Boeing’s 737Max-related deaths by negligence. We have a responsibility to change that.

We bear a serious responsibility to ourselves and to others to do good things in our teaching, research, and service to improve leadership quality (e.g., Kellerman, 2024). But we have done almost nothing to make it so. We have not built a global professional organization to oversee the carrying out of that responsibility. The academies, committees, and societies we do have are all ineffective at enforcing whatever feeble ethics codes or pedagogical standards they promote. We have balked at any openly coordinated attempts to find consensus over fundamental definitions and constructs that could lead more efficiently to effective treatments for phenomena like ‘bad’ leadership. Instead, we worry it could harm our academic freedoms (more than the current, opaque system controlled by a handful of editors does), as well as the incomes of the peers who exploit that lack of consensus of ideas in a very lucrative leadership ‘industry’ (see Kellerman, 2024). Most of our doctoral programs do not require any special training in academic or pedagogical ethics. And our tenure and promotion systems do not try to assess the social benefits and harms that our research and teaching have caused. As such, we have not even tried to be responsible. It has even been suggested that ‘the academy’, and especially that focused on business, economics, and politics (i.e., major fields related to leadership), embodies a robust form of unaccountability itself (see Wedel, 2014). Indeed, it seems like we have very little legitimacy upon which to pontificate about practical solutions to potential organizational harms, including those for addressing ‘bad’ people in leadership positions. Such preaching may even appear distasteful, given the not-unusual phenomenon of doctoral student abuses and rankings frauds seen in our schools (Cohen & Baruch, 2022). To put it bluntly, then, there are many barriers to carrying out the suggestions to improve the leadership field. That said, there exist ways to overcome these barriers. For example, we could do more to highlight academic integrity (Fukukawa et al., 2013); involve more stakeholders; develop adoptable strategies (Laursen & De Welde, 2019); and use emotion (Salk, 2012).

6. Limitations and Conclusions

In this paper, we offered a critical assessment of the premises of an ongoing, important debate in the leadership field as part of a dialectic inquiry. We made our points through modeling, logic, the use of grounded, representative examples, and a synthesis of ideas. That noted, the conclusions are open to several limitations, the main one involving generalization. Given that this paper did not itself review the entire literature over the debate, let alone over ‘bad’ leadership or over leadership more broadly (choosing instead to draw on the conclusions of recent reviews and a summary of the debate in a top journal), the comprehensiveness and specification of its critiques are somewhat restricted. That said, we believe that we have covered the main areas of concern found in the relevant literature and have been representative in the models and examples we have used.

Returning to our research question—concerning how we do more good than harm in studying and informing others about ‘bad’ leadership—our analysis has provided explicit directions for how to improve the field’s performance in pursuing that quest. Our directions are based on an improved understanding of the current harms being done that emerged from a structured analysis. We introduced a new model for assessing leadership, focusing on ‘bad’ leader behavior, to address the current debate about what it is, how it can be used, and what we should do about it. We then conducted a focused critical analysis of the leadership area’s work within that debate. To do so, we considered three primary aspects of that work—the diagnosis of ‘bad’ leaders, the treatments for addressing ‘bad’ leadership, and the impact we have had on practice through our teaching and publications in this area of ‘bad’ leadership. We focused on two foundational definitions of ‘bad’-ness (i.e., DT and HE leadership). Leveraging those three aspects and two definitions allowed us to identify the significant concerns in the debate, and then in the leadership field more generally. This led us to a synthesis of ideas that formed our suggested directions for moving the debate and the field forward towards less ‘bad’ and more ‘good’ leadership outcomes.

Our dialectic inquiry has contributed to the literature by addressing several gaps in the current understanding of ‘bad’ leadership—regarding its causes (e.g., diagnoses), its forms (e.g., DT and HE), its effects (e.g., its impacts and ours in teaching about it), and its possible treatments (e.g., involving prevention and removal). Our new model (that provides an improved understanding of the current debate over teaching ‘bad’ leadership behaviors as a means for ‘good’ outcomes) and our new flowchart-based directions for future work also provide valuable contributions to this particular stream in the leadership field.

In closing, we note that the world will continue to face many challenges over improving the state of leadership. This includes challenges in the diagnosis and treatment of ‘bad’ leadership, in affecting the contests to win leadership positions, and in how such powers are used. As academics, we also face challenges in improving our impact on leadership practice, and in our own structural impediments to making such improvements. However, it is well past the time for us to begin the process of becoming the ‘good’ leaders and good scholars who take on those challenges. We need to start soon (and to be more wary of the proclivities of our leaders to the kinds of behaviors that may stop us).