Management accounting change as an amplifier of a leadership dispute: an ethnography of convergent and divergent leader – follower relations

Purpose – The purpose of this paper is to explore the role of leadership in management accounting change processes and outcomes. Design/methodology/approach – The paper draws on an ethnographic study in a Southern European company and mobilizes leader – follower relations as a method theory to analyse the observations. Findings – The findings show how a leadership dispute between two top managers can be amplified during the management accounting change process and percolate throughout an organization. The authors identify five contested areas where the role of accounting amplifies the leadership dispute by unfolding its reach to other organizational actors. The leadership dispute can shape and reinforce a fragmented organization, with some organizational members creating convergent leader – follower relations while others divert and fragment withanincreasedturnover.Thisamplificationcanleadtounexpectedoutcomesofthechangeprocessinterms of how and by whom accounting is performed. Research limitations/implications – The authors propose the study of leadership and followership as an important but, to date, largely neglected theme in management accounting research. Originality/value – In contrast to the prior management accounting literature, the paper departs from a leadership-centric and role-based approach and employs a co-constructionist and relational approach to leadership and followership to analyse management accounting change. In addition, it applies and extends Alvesson ’ s (2019a) theory on “ divergent relationalities ” between the presumed leaders and followers. In doing


Introduction
The president of ConstructionCo hired a consultant to facilitate internal communication between the accountants and the construction workers. The scene of the first meeting looked desperate: at the head of the room the president and the consultant, who were chairing the meeting; on one side the accountants who appeared apathetic to be spending their time in an extraordinary meeting; on the other side, the construction workers, who had a mixed interest in it-even though accounting was not their occupation, they realized the importance of the new accounting system for their own work; and, at the back of the room, the vice-president. He discredited the meeting, which he openly revealed by the gestures between him and some of the construction workers. He did not even sit down, but leant on the wall right next to the door.
(Notes from field observations) Leadership has long been recognized as an important factor in the process of management accounting change, such as the introduction of a new accounting system or a role change for the accountants within an organization. It has been argued that leadership facilitates an entire change process, from its initiation to the overcoming of potential barriers to change, to creating momentum for change and to the eventual outcomes (Cobb et al., 1995;Kasurinen, 2002;Munir et al., 2013). Leadership is concerned with what followers are thinking and feeling, and how they can be persuaded and influenced to contribute to a common vision through their roles and tasks (Nicholls, 1987;Weathersby, 1999). In this vein, leadership behaviour is sometimes associated with the notion of sensegiving (Gioia and Chittipeddi, 1991), which has been mobilized by accounting scholars in various conceptual ways, though typically not explicitly discussed as leadership (e.g. Guiliani and Skoog, 2020;Jordan and Messner, 2012;Kraus and Str€ omsten, 2012;Malsch et al., 2012;Meidell and Kaarbøe, 2017;Tillmann and Goddard, 2008). Hence, leadership, as defined in this study, differs from (but can situationally relate to) everyday managerial activities such as using managerial authority, tools and power, because it is concerned with whether leaders win others as followers by conveying values and meanings to them in support of the intended change. Whilst prior management accounting literature has acknowledged the importance of leadership over decades (Brownell, 1983;Hopwood, 1974;Otley and Pierce, 1995), not only are there surprisingly few studies on the theme (see Abernethy et al., 2010;Jansen, 2011), but existing studies by and large share a leadership-centric and role-based assumption. The focus tends to be on the presumed leader in a senior managerial function, their leadership style and/or heroic and cowardly behaviour, and whether and how they ultimately bring about change in the "right" direction.
In contrast, we draw on the alternative assumption that leadership is a more complex and multilayered phenomenon which requires an interpersonal perspective (DeRue and Ashford, 2010;Uhl-Bien et al., 2014). Leadership thus demands an analysis of whether and how the influencing process takes place, considering that even leaders who are seen as inspiring and persuasive might not always do "good" (Alvesson, 2019a). Several strands of the leadership literature therefore highlight the importance of including followership in the analysis (Blom and Alvesson, 2015;Carsten et al., 2010;Kelley, 1992;Oc and Bashshur, 2013). Hence, we argue that central to an understanding of management accounting change processes and outcomes are leader-follower relations, where both the presumed leaders and followers shape social reality (DeRue and Ashford, 2010). This assumption not only considers that leadership depends on followers and following behaviour (Uhl-Bien et al., 2014), but analytically focusses Management accounting and leadership on the meanings that organizational actors attach to their relationship. Drawing on the recent work by Alvesson (2019a), this analysis brings relationality to the foreground and examines whether those meanings, in terms of values, understandings and perceptions, are convergent or divergent between organizational actors.
The aim of this paper is to explore the role of leadership, specifically leader-follower relations, in management accounting change processes and outcomes. Prior management accounting change studies have touched upon related themes. They have looked at power and politics during the change process (Burns, 2000;Markus and Pfeffer, 1983;Scapens and Roberts, 1993); the dividing cultural perceptions, values and norms between or within occupational groups leading to resistance (Bhimani, 2003;Burns and Baldvinsdottir, 2005;Malmi, 1997); or the discrepancies between identity and an aspired future image (Abrahamsson et al., 2011;Goretzki et al., 2013;Taylor and Scapens, 2016). These are all theoretical explanations for human agency (often involving one or several powerful managers in the empirical storyline of studies) in shaping structural change. Other studies have mobilized actor-network theory to describe the situational and experimental means that drive the change process as relational drift, using a flat ontology to analyse the network of human and non-human actors (Andon et al., 2007;Quattrone and Hopper, 2001). However, the theoretical focus here differs from those prior studies because we analytically examine the relationality between (human) organizational actors in the process of influencing changethat is, whether the meanings about their relations converge or diverge.
The choice of our theoretical focus was triggered by a series of intriguing observations at ConstructionCo, a pseudonym for a Southern European construction company, during a threeyear ethnographic study. Moving analytically back and forth between different theoretical perspectives and empirical observations, it was puzzling to us how the adverse relations between the two top managers (see also notes from field observations above) dynamically unfolded topdown and across the organization in an interplay with the emerging management accounting change; how it situationally overruled the formal structure, hierarchy and occupational duties of organizational actors; and ultimately how it redirected the change process and outcome.
Drawing on these empirical insights, we mobilize leader-follower relations (e.g. Alvesson, 2019a;Fairhurst and Uhl-Bien, 2012;Shamir, 2007;Uhl-Bien, 2006;Uhl-Bien et al., 2014) as method theory [1] to inform our observations of how management accounting change unfolds. Building on the recent work on divergent relationalities by Alvesson (2019a, p. 329), we define leadership "as a social influencing process based on the convergence and alignment of meanings in terms of definition and assessment of a leader/follower relationship. Without such an alignment, when there are significant misfits, there is no functioning leadership". Our findings show that a leadership dispute (here labelled leader-leader alignment misfit) amongst top managers can be amplified during the enactment of management accounting change, and distributed across the company, shaping an adverse work atmosphere triggered by the confusion, ambiguity and conflicts surrounding the organization's leadership. We identify five contested areasnecessity, resources, accuracy, ownership and responsibilitywhere management accounting change amplifies the leadership dispute by involving other organizational actors. We analyse how particular characteristics and assessment of leaderfollower relations (Alvesson, 2019a) dynamically unfold and can lead to a divided and fragmented organization, with some organizational actors creating convergent leader-follower relations, while others divert and fragment with an increased turnover. The findings unveil how and why this amplification of the leadership dispute can redirect the outcome of the management accounting change process. The learnings from this case highlight the responsibility of top managers (or any cultural gatekeeper) to establish an inclusive rather than a divisive tone at the top in order to avoid accounting being an arena where disputes are amplified.
This study contributes to the management accounting literature by making the case for leadership studies in accounting research, which, despite its importance, is to date a relatively neglected space. In doing so, we introduce leadership and followership as a relational and co-constructionist process that differs from the traditional leadership-centric, role-based perspective. This approach offers an analytical tool to distinguish leadership from management by separating leader-follower relations from (hierarchical) managersubordinate relations. More specifically, rather than a generalized interpersonal approach, we mobilize leadership and followership divergent relationalities as a new lens through which to view management accounting change (e.g. Abrahamsson et al., 2011;Burns, 2000;Cobb et al., 1995;Jansen, 2011;Kasurinen, 2002;Munir et al., 2013). In doing so, we also contribute to the leadership field by responding to Alvesson's (2019a) call for in-depth ethnographic studies that can capture the meanings that different parties give to their relationship and augment Alvesson's theory by adding leader-leader alignment misfit as a further category to explore convergent and divergent relationalities.
The relational and co-constructionist approach has several important conceptual implications for leadership studies in accounting, including the related sensemaking/ sensegiving literature. Firstly, we illuminate how to conceptually distinguish between traditional role-based and co-constructionist influencing processes and discuss why this distinction is important. Secondly, we show the relevance of understanding the meanings that people give to their relationships in order to examine the quality of the influencing process. Thirdly, we demonstrate that whether those meanings are convergent or divergent affects the functioning of the influencing process (sensegiving process), and ultimately whether no, frictional or functioning leadership and followership occur.
2. Conceptualizing leadership and followership in the management accounting domain 2.1 Leadership-centric approaches The ongoing discourses in management accounting discuss leadership primarily based on a leadership-centric perspective. The notion of leadership has been introduced in normative frameworks (see Figure 1) which seek to explain the factors that enhance or hinder management  (Cobb et al., 1995;Innes and Mitchell, 1990;Kasurinen, 2002). While Innes and Mitchell (1990) originally developed a framework that established the distinction between motivators, catalysts and facilitators during the change process, it was Cobb et al. (1995) who added the dual role of individuals as catalysts and leaders. According to those studies, leadership can overcome barriers to change and individualsfor example, a chief financial officer or divisional financial controllercan take on the role of change agents. The studies also highlight the notion of the momentum for change, which represents the assumption that actors will expect an ongoing change process. In contrast to those studies, Kasurinen (2002) conceptualizes barriers that may hinder, delay or prevent management accounting change. He identifies the absencethe sudden resignationof a strong leader as a barrier to the change process. Munir et al. (2013) recently extended this discourse by studying how individuals, in their roles as leaders and change agents, create, sustain and build momentum for change. Leadership is relatively unproblematic in those studies, however. For example, Munir et al. (2013, p. 210) describe the process of the president of the bank by which he influences others as straightforward: "he helped overcome resistance by top management and first-line managers against the new PMS [performance measurement system] by communicating the need to change management practices throughout the bank. [. . .] Hence, the president was a very influential catalyst for change and was also fundamental in creating the momentum for change".
Reflecting on this prior research, those studies highlight the relevance of leadership to create momentum for change or overcome barriers to change, but they do not theoretically analyse how the relations between organizational actors emerge as an explanation for how the change process unfolds. Leadership is normatively assumed based on role and hierarchical position, and is a relatively unproblematic process that relies on the skills and behaviour of the leader (or manager fulfilling the leadership function) in the pursuit of change. However, the analytical focus does not pay much attention to the followers (or people presumed to be followers) and the relationships between people, which are arguably not always characterized by cohesion but often by confusion, ambiguity and even misalignment.
Another approach is pursued by Jansen (2011), who follows the tradition in accounting research of distinguishing different leadership styles (Brownell, 1983;Hopwood, 1974;Otley and Pierce, 1995). Whilst his study theorizes the relationship between the needs of employees and leadership style in the context of management accounting change, similar to other accounting studies considering leadership (Abernethy et al., 2010;Hartmann et al., 2010), Jansen (2011) draws on an established classification from the leadership literature. This approach advances the management accounting change literature by taking into account the interaction between managers and employees. However, the theorizing of leadership remains leadership-centricthat is, role-based, static and not situational, and intrapersonal rather than interpersonal (Hollander, 1993;Uhl-Bien et al., 2014).

Considering followership and leader-follower relations
Attempting to interpret our empirical insights with those existing studies on leadership in the domain of management accounting change, we find a more complex interplay between people, their perceptions and the meanings of the relations between them than contemporary theorizing in management accounting considers. Drawing on recent leadership literature (Alvesson, 2019a;Blom and Alvesson, 2015;Carsten et al., 2010;DeRue and Ashford, 2010;Oc and Bashshur, 2013), we therefore consider the analytical relevance of followership, which is defined as the "investigation of the nature and impact of followers and following in the leadership process" (Uhl-Bien et al., 2014, p. 89). However, it is important to note that there are several conceptually different strands of research on followership, raising confusion if they are mixed without considering their different underlying assumptions. Uhl-Bien et al. (2014) clarify the roots of this confusion by showing that most studies can be grouped into two conceptually distinct schools: a role-based and a co-constructionist approach, as illustrated in Figure 2 [2].
The role-based approach of followership has emerged (Meindl, 1990;Meindl et al., 1985) to counter the mainstream leadership-centric literature. Researchers using this follower-centric approach reverse the lens (see Shamir, 2007, p. xii) and argue that followers influence the leaders in terms of leaders' perception, behaviour and outcomes (Carsten et al., 2010;Uhl-Bien et al., 2014). However, leadership and followership remain aligned with the hierarchy, where followers are in lower formal ranks or positions than the leader. Hence, despite giving some agency to followers, the follower-centric, role-based approach maintains and reinforces the assumption that leadership is associated with prestige and power and followership with relatively powerless individuals that dutifully carry out the leader's directives (Alvesson, 2019b;Gordon, 2011;Kelley, 1992;Uhl-Bien et al., 2014).
In contrast to the role-based approaches, the co-constructionist approach (Conliffe and Eriksen, 2011;Uhl-Bien, 2006;Uhl-Bien and Ospina, 2012) assumes that there cannot be any leadership without followership and following behaviour. This relational perspective distinguishes leadership from management because leadership is not assumed by formal hierarchy and power. A management position is a relevant indicator and facilitator of leadership, and manager-subordinate relations will often overlap with leader-follower relations to some extent (Carsten et al., 2010). However, not every managerial position is necessarily associated with a leader identity; nor does every subordinate position equal a follower identity (Alvesson et al., 2017;Learmonth and Morrill, 2017). The co-constructionist approach considers that managers sometimes take on following behaviour and subordinates show leading behaviour (Blom and Alvesson, 2015;DeRue and Ashford, 2010). Individuals can be leading in one area and following in another, depending on the focus, situation and context (DeRue and Ashford, 2010).
An important empirical indicator of the existence of leadership is therefore the observation that following behaviours occur (Uhl-Bien and Ospina, 2012). Such behaviours show individuals' willingness to be disproportionately influenced by another in a noncoercive way (Uhl-Bien et al., 2014), by granting a leader identity to another and claiming a follower identity for oneself (DeRue and Ashford, 2010). Hence, the co-constructionist approach changes the narrative of leadership and followership away from viewing the leader as heroic and powerful and the follower as passive and recipient. Instead, leadership and followership need to be analysed more critically as leading and following behaviours in everyday behaviour, where individuals often perform both (see also Jaser, 2020).
The co-constructionist and relational approach is closely related to the literature on sensemaking conceptually (Weick, 2001), especially the concept of sensegiving (Gioia and Chittipeddi, 1991;Maitlis and Lawrence, 2007), which has been mobilized by several accounting scholars (e.g. Guiliani and Skoog, 2020;Jordan and Messner, 2012;Kraus and Str€ omsten, 2012;Malsch et al., 2012;Meidell and Kaarbøe, 2017;Tillmann and Goddard, 2008). Sensegiving and sensemaking in organizations resonate in accounting research because they provide an analytical lens for studying the shaping and influencing of accounting narratives those stories that are created about and around numbers and calculations (Beattie, 2014;Czarniawska, 2017;Llewellyn, 1999). In terms of management accounting change, such narratives concern both the narrative of the purpose of the change itself and the narrative of how the calculations and numbers are (re-) interpreted.
However, taking the critique by Uhl-Bien et al. (2014) about the followership literature seriously, the sensemaking/sensegiving literature is an example of where the assumptions of the role-based and co-constructionist approaches are partially mixed. On the one hand, the sensemaking literature assumes that leaders are top managers, as in the role-based approach (Maitlis, 2005;Maitlis and Christianson, 2014). This assumption is rooted in the original work by Gioia and Chittipeddi (1991), who introduced sensegiving as the process by which top managers influence strategic change. Middle managers, stakeholders or even lower level employees are not conceptualized as leaders but as another category (e.g. Maitlis and Lawrence, 2007), who may also principally engage in sensegiving, yet this tends not to be interpreted as leadership. On the other hand, as in a co-constructionist approach, it is widely established that sensegiving can be applied within and across hierarchical directions: for example, it has been used to examine how middle managers influence top managers (e.g. Meidell and Kaarbøe, 2017). The concept of sensegiving is therefore sometimes interpreted as leadership behaviour and sometimes not. The sensemaking literature is also relatively silent about how and under what conditions sensegiving is received by others (Maitlis and Lawrence, 2007). Hence, we find that the followership literature not only provides useful conceptual implications for the application of leadership studies in accounting in general but also the related sensemaking/sensegiving literature.

Convergent and divergent leader-follower relations
The role-based and co-constructionist approaches mentioned in the prior section as well as the sensemaking/sensegiving literature share the assumption that leadership is about the influencing of meaning, as emphasized by researchers on symbolic and transformational leadership (e.g. Bass, 1985Bass, , 2008Ladkin, 2010;Smircich and Morgan, 1982). They also share the assumption that there is a high degree of shared meanings between people about their relationship. However, the recent work by Alvesson (2019a) highlights that research neglects that those meanings often diverge between people as explained in the following quote: "There is seldom a consideration that the manager may see the leadership as for example delegating while subordinates view it as laissez-faire, that the former thinks she or he demonstrates engagement while subordinates read this as an unbalanced overreaction, or that the manager believes she demonstrates authenticity while subordinates may have doubts about the sincerity, goodness, or self-awareness of the person claiming authentic leadership." Hence, our analytical frame does not only shed light on situations where people generate largely overlapping understandings about their relationships but assumes that there is often ambiguity, complexity and frictions in their relations and their potential role identities as leaders and followers.
This theoretical perspective is especially useful to examine challenging phases such as during the management accounting change process (e.g. Baldvinsdottir et al., 2009;Granlund, 2001;Malmi, 1997). People might be overwhelmed, have false expectations and wrong judgements about each other's attitudes and behaviours. They may think they agree or disagree, but in fact they might deviate much in their intention, motivation, perception, prejudgements and frustrations (Alvesson, 2019a). We therefore argue that considering convergent and divergent relationalities can explain why and how management accounting change sometimes unfolds rather illogical, unexpected and surprising (Andon et al., 2007;Quattrone and Hopper, 2001).
Drawing on Alvesson (2019a), we establish two criteria to distinguish typical forms of convergent and divergent meanings regarding leadership: (1) whether the meanings converge or diverge about the characteristic of the leader-follower relation (i.e. what actors think is going on) and (2) the assessment given to the leader-follower relation (i.e. whether they perceive the relation positive, irrelevant or negative). Table 1 summarizes the different types of leader-follower relations.
Leadership, the first category in Table 1, represents the conventional and functioning understanding of leadership and occurs when people have high alignment in their Management accounting and leadership understanding of both the character and assessment of their leader-follower relationship. Hence, they converge in their understanding and evaluation of an unequal relationship in terms of the related identities as leader and followers (see also DeRue and Ashford, 2010).
Frictional leadership, the second broad category, represents situations when there is only partial overlap in either the understanding or the evaluation of the relationship. Value divergence, one form of frictional leadership, occurs when people have an overlap in their understanding of what is going on yet disagree in their evaluation of the relationship. For example, both the leader and follower perceive the relationship as authoritative, but they largely disagree on the value of the relationshipthat is, the leader might assess the relationship favourably and the follower unfavourably. Construct divergence, the other form of frictional leadership, occurs when people have different perceptions of the relationship but assess it similarlythat is, the leader assumes inspiring and empowering leadership but the follower feels micromanaged. Nonetheless, despite those different views, in the particular constellation, both experience the relationship favourably.
No leadership, the third category, occurs when there is wide divergence in terms of both the character and assessment of the relationshipwhat Alvesson (2019a) labels multiple breakdown. Such a situation leads to a high degree of ambiguity and confusion about what is going on.
Finally, we add a further category to Alvesson's theory, which we label leadership dispute. This situation occurs when there is an alignment misfit, and two people claim leadership but are unable to grant each other following behaviour (DeRue and Ashford, 2010). In contrast to the previous category where the meanings diverge, leadership dispute is a special situation, because people converge in their experience of both the content and the value of the relationship; however, both claim leadership but do not reciprocally accept those claims, leading to a high degree of conflict and tension (Collinson, 2005(Collinson, , 2014. In sum, we consider divergent relationalities in the study of management accounting change, especially because the bulk of prior literature highlights the importance of human agency in the change process (e.g. Abrahamsson et al., 2011;Burns, 2000;Burns and Scapens, 2000;Granlund, 2001;Makrygiannakis and Jack, 2016;Taylor and Scapens, 2016). Leadership and followership, as theorized here, is dynamic and evolving, enacted through an unstable and relatively uncertain process (Alvesson, 2019a;DeRue and Ashford, 2010) where events emerging during the change process can redirect leader-follower relations. This theoretical lens on divergent relationalities therefore provides a tool to explore the sometimes illogical and ambiguous turns (e.g. Andon et al., 2007;Quattrone and Hopper, 2001) in management accounting change processes.

Ethnographic method
This study is based on an ethnographic method, which is a useful approach for investigations into accounting in action (J€ onsson and Macintosh, 1997). In the anthropological tradition, the purpose of an ethnography is to understand the native's point of view (Geertz, 1973) by immersing oneself in the culture and learning its practices, values and social norms. The challenge for the enquirer is to move from the etic (theoretical, outsider) to the emic (insider) perspective, and to make sense of the empirical insights obtained. This exploration requires prolonged and extensive fieldwork which favours direct participant observation and supplementary interviews in combination with other forms of data collection, archiving and analysis (J€ onsson and Lukka, 2007;Pike, 1954;Van Maanen, 1988). In ethnographies, those experiences are portrayed in thick descriptions (Geertz, 1973) to bring the voices in the field close to the reader, thereby creating cultural knowledge for outsiders.
The strength of ethnography is that the enquirer experiences the unfolding of events and the related narratives of the actors in the field, including their intentions, perceptions and interactions. In doing so, the researcher can include a critical reflection on the narratives of the many "others" who are often silenced in studies using different methods. Llewellyn (1999) highlights that narratives operate at several levels. The researcher explores the different argument-making in the field, and ultimately creates a (research) narrative about the insights gained (see also Bjurklo, 2008;Czarniawska, 1997;Scheytt et al., 2003;Seal and Mattimoe, 2016). Writing up the (research) narrative is part of data collection and analysis, and contains an authentic, though inevitably selective, account of the experiences of the enquirer (see Ahrens and Mollona, 2007). The narrator tells the story in a particular manner to make a point. Narratives are therefore a particular plot of the story that include specific characters and events which make the narrative unique in conveying a particular argument (Llewellyn, 1999), providing the basis for analytical generalizations (Lukka and Kasanen, 1995;Parker and Northcott, 2016).
Ethnographic work raises the ethical responsibility of the researcher, who becomes a trusted companion of the actors in the field. The information collected must be presented in a way that protects individuals in the field (Bryman and Bell, 2007). We have therefore anonymized the datathat is, contextual information has been limited to the minimum necessary for our theoretical argumentation. In addition, time has passed since the field enquiry took place and several actors in our field study have moved on.

Data collection process
The ethnographic insights presented in the following section are based on the observations and interviews conducted by one of the authors. She worked part-time in the case company between January 2010 and September 2012, and was introduced to the organizational members as a student, taking on the role of a supporter in the accounting department. This identity as a student (and not an ordinary organizational member) allowed her to have a genuine learning interest in the activities of the organization that went beyond her tasks (see Morales and Lambert, 2013). Her work tasks primarily concerned the execution of ad-hoc tasks in support of organizational members. This student identity is important, as the on-field researcher's access to informants and her ability to move effectively within the field are influenced by how other actors understand her interests and intentions (Grills, 1998).
The on-field researcher had an agreement with the president that, besides being a paid part-time worker, she would conduct a series of semi-structured interviews for an academic study. This study would concern accounting matters, particularly the implementation of the new accounting system in the organization. The focus of this enquiry was kept open in nature: the social insights emerging from the field observations, paired with an interest in the implementation process of the new accounting system, narrowed the research focus over time. Hence, participants whose actions and beliefs were being studied were not aware of the theoretical focus (see Morales and Lambert, 2013).
Working as a supporter and simultaneously conducting a case study on the technical means of a new accounting system granted the possibility to meet all types of organizational member. Being fully embedded in the dynamic web of interactions, the on-field researcher was able to be closely and regularly involved with actors across the organizational hierarchy, including the protagonist top and middle managers, accountants and construction workers.
The researcher kept a diary where she noted details of observations from the field. The facts of the observations were archived by time, place and actors involved. This diary was updated within 24 h of each field day. The notes included details about observed actions, conversations and events, including descriptive details, such as how actors were positioned in the rooms, their gestures or their tone of voice. These notes also contained information gained Management accounting and leadership while shadowing actors during their work, and from casual conversations and small talk held with and amongst actors in hallways and during meals and coffee breaks.
The data were complemented by semi-structured interviews, along with internal and external documents such as memos, reports, plans, slides, tables and e-mails. In total, 85 semistructured interviews were conducted, mainly focussing on organizational roles, the technical operation of accounting, control systems in general and the change initiative. These interviews were pursued with all parties mentioned above (the top and middle managers, accountants and construction workers) as well as an external consultant and four external industry experts, as background knowledge. The interviews' duration varied, lasting between 15 and 120 min. All interviews were recorded and transcribed.

Data analysis
The triangulation of this ethnographic material provided a detailed account of how relationships amongst people emerged over time, thus making it possible to confirm our theoretical argumentation on leader-follower relations using various data sources (Alvesson, 2019a). Two members of the research team held regular discussions to make sense of the empirical data during the fieldwork. The third member of the research team probed these insights ex-post by posing further questions based on the case insights. The combination of two researchers familiar with the insider perspective and one with an outsider perspective proved valuable in making sense of the data and interactively moving back and forth between empirical evidence and its theoretical interpretation (Ahrens and Chapman, 2006;Ahrens and Dent, 1998;Lukka and Modell, 2010). In this process, the authors used graphs to make sense of the data (e.g. Clarke, 2003), including the emerging relations between organizational actors. The research team explored several ways of interpreting the data until the theoretical interpretation emerged (Dent, 1991). The authors kept a research diary during the data analysis process which served as a working document, supporting the authors in keeping track of their progress on theoretical ideas, relevant input and the feedback received.

Empirical findings and analysis
4.1 About the company ConstructionCo (a pseudonym) is a Southern European family business involved in the construction sector. Since its foundation at the end of the 1980s, the company has continually extended its business activities towards becoming a general contractor that is able to serve the entire value chain of the construction business from material procurement to engineering, execution and maintenance. The company has achieved steady growth from its inception, which was boosted from 2007 onwards when the president of ConstructionCo began to promote the development of major real estate projects. Between 2007 and 2012, the company's assets doubled, seeing turnover rates of between V50m and V60m. During the same era, the construction industry suffered due to the macroeconomic financial crisis of 2008 and 2009. This caused a lack of publicly funded construction projects, declining profit margins in the industry and several business failures amongst the company's competitors. Whilst ConstructionCo survived the crisis, the situation raised concerns about the company's strategic positioning in the future.

The emergence of a leadership dispute
In 2010, the company was managed by two brothers, the younger being the president and the older the vice-president. They owned most of the shares, were both managing directors and were legally responsible for the company. Together with their father and other brothers, they founded the company more than 20 years ago. The skills and responsibilities of the two managing directors were complementary and they had built up a long corporate history, managing the growth of the company. In the early stages of our fieldwork, we had the impression that the two top managers maintained relatively stable leader-follower relations, each showing following behaviour in the responsibility area of the other. Their leaderfollower relations were aligned, and both valued the shared leadership arrangement of the company.
However, reflecting on our case insights during the field research period, we observed that the relationship between those two top managers was more ambiguous and conflictual. In an interview, the project portfolio manager expressed confusion about the organization's leadership: "It seems somehow that the company has two heads and they drive their people in two directions." The two top managers agreed on the need to maintain strict cost control, to analyse and manage fixed assets, and to look for possibilities to stabilize and strengthen the company's financial position. However, they differed on how to achieve those strategic priorities.
The president, who was responsible for the administrative and financial side of the business, argued for diversifying the risk of the company by introducing a new business unit that would handle large-scale real estate projects. Real estate suited him, being an expert in financial analysis and skilled in analytical thinking, as the following quote illustrates: The real estate business is our future. Construction is not just build, build, build. . . our brand could be perceived as a global partner if we invest in real estate projects.
Although the president was diplomatic and demonstrated leadership towards those external to the organization, internally, he preferred others to act on his behalf. Accordingly, employees perceived him as distant and not easy to approach. As one of the accountants put it: "He appears phantom-like and he does not stand up for his choices. . . He acts for his convenience only".
Hence, although the president was directly responsible for the administrative staff, he was not displaying leadership behaviour towards employees other than operating through the formal structure. Instead, he concentrated on networking outside the organization, and built relationships with politicians and other actors in the region to facilitate support for public bids. He was interested in building the company's brand and its image to become a more powerful organization on a national level.
In contrast, the vice-president, who was responsible for the operational side of the business, preferred to concentrate the company resources on the traditional core competencies of the construction business. He established a counter-narrative, pointing to their competence in the construction business and argued that real estate might undermine their well-established brand.
We have always managed projects from the beginning to the end. [. . .] Real estate is fantasy, it is just finance and risk.
He claimed leadership of the construction workers, which he demonstrated through his everyday behaviour. Based on our informants and observations, the construction workers accepted that claim and they established a bond as a group within the organization. He created a family feeling for the workers that were his direct responsibility: I have project managers who manage workers well and they all know that I am here in case of problems. I'm always in the construction sites with them. [. . .] As in any family, we have to feel good here. . . The business is about more than making money.
In their meetings, construction workers were very respectful and would not argue with the vice-president nor speak up without being called upon. If one wanted to raise a contentious issue with him, they would address him privately before or after the meeting. He was Management accounting and leadership impulsive, hot-tempered and could shout if things went wrong. However, in spite of being perceived as a micro-manager by the construction workers, if an employee had earned his trust through hard work and loyalty, he would delegate responsibility to them and situationally showed following behaviour. The project portfolio manager explained: I grew up with the vice-president, but it remained clear who the boss was. Over the years, I displayed my will to work hard and my commitment to ConstructionCo. In my view, he seems to recognize it. Some time ago I was not able to do anything without his authorization, but now things have changed.
There was alignment in leader-follower relations between the vice-president and the construction workers in terms of the relatively authoritarian nature of the relationship, and the relationship was valued by both parties.
Based on the views of our informants and our analysis, the leadership dispute about strategic positioning emerged due to the self-interests of the two managers. Real estate provided the president with an opportunity to develop his own operational business, whilst the vice-president preferred to keep all business operations under his control. The two top managers undermined each other's leadership and refused following behaviour, which started to become visible to other employees. For example, in a discussion with the accounting manager, the president expressed his frustration at the vice-president's decision to invest in new personnel and equipment instead of using the resources for the new real estate business unit: Again another vehicle! It is impossible! We do not buy it! He [the vice-president] continues to spend money on workers and vehicles. So we have to work to pay for the operational workers and all their things.
Although the vice-president did not support the new business unit, he was reluctantly willing to compromise with the president. He insisted that the new real estate unit would not make independent business decisions and kept the construction business always as first choice over external contractors (i.e. independent of whether the internal or external offer was more lucrative for the real estate unit).

The new accounting system as centre stage of the leadership dispute
The introduction of the real estate sector to the business mix did not only require financial resources for new competencies and infrastructure: it also demanded an improved accounting system. Its design took centre stage of the leadership dispute because the impact of the real estate unit was calculated, objectified and made visible in the new system. The existing accounting system had been developed in the early days of the company and consisted of a mix between direct and full costing based on arbitrary overhead allocations. This system worked as long as the business was relatively consistent with construction projects. Moreover, the vice-president shaped a narrative in which numbers played a secondary role, since he had a gut feeling about what was right for the company: I know if a construction site goes ahead or not. I have no need to see a lot of numbers.
However, the characteristics and consumption of overhead costs in the real estate sector were different. The allocation of overhead costs and the related full costing and transfer pricing system had to be improved. The president highlighted the importance of an accurate accounting system: . . .we are growing and need to have the numbers under control, especially overheads and financial flows. Without precise control, the future cannot be predicted.
The new system should separate the finances of the different business units while addressing pressing requirements from banks, private equity funds, and large investors and customers.
At this stage, however, the vice-president did not realize the importance of the design of the accounting system in terms of how it would affect the perceived costing and profitability of the traditional construction units.

The role of the change agent
To develop the new real estate business unit, the president hired the extraordinary financial manager (EFM) in 2009. This manager had an MBA degree from a prestigious university, which was important for the president, as it seemed to add to the image of the company. The newly recruited manager had previously worked in a multinational company specializing in real estate and, in the eyes of the president, brought the necessary experience to manage financial flows and systems in the real estate sector. The EFM reported directly to the president and was responsible for the development of real estate, as well as the testing and implementation of the new accounting system. The president reflected on his choice to hire the EFM as change agent: He has an excellent study background and he worked at [a multinational manufacturing company] as a real estate planner. He knows the real estate business and he has also good contacts in that world.
In an interview, the EFM highlighted his previous work experience, and that he felt the company had a long way to go to manage real estate projects. I used to create different scenarios and simulate operations for a major understanding of project cash flows and the return of investment. Here, no one manages projects in this way and there are no systems to support these activities.
The president and the EFM established high alignment in the nature and evaluation of their leader-follower relationship. Based on our empirical material, they worked well together at the outset, had similar priorities and values, and assessed their work relationship favourably. At company events, they would present the company results in well aligned speeches that demonstrated their collaboration and acceptance of each other's roles.
Since the president delegated his operational work to the EFM, the latter formally also became the manager of the accountants. In January 2010, the EFM expressed how he felt about working with them: I am a newcomer; I do not know exactly what they do. But now they have the opportunity to use a new system and increase their interaction with other people [in the organization].
Although he was in regular contact with the accountants, the EFM noticed that they seemed relatively reserved towards him: When I go to their open space office they are silent. When I ask them something, they look at me. . . But that does not matter [. . .], you will see that they will follow me.
Formally being their manager, the EFM believed that he could change the accountants' work culture by increasing their involvement with construction workers. However, the accountants' perceptions were different. They did not understand how the role of the EFM related to them because the president had not communicated it to them. The EFM was located in the newly formed real estate branch (which was 40 km away) and was mostly in contact with them via emails. If he visited their office, he typically asked them for information with regard to real estate or the design of the new accounting system, yet he did not communicate the purpose of the requested information. He seemed to act based on his formal authority. An accountant reflected: When he comes here, he is a people-pleaser. He asks for some data and afterwards goes back to [the real estate office].

Management accounting and leadership
More problematic was that he disrespected the existing team hierarchy amongst the accountants. Instead of liaising with the accounting manager, the team head of the accountants, he spoke directly with anyone he chose to; this irritated the accounting manager, who felt disrespected. The accounting manager was accepted by the accountants, though, who soon called the EFM, contemptuously, the "university guy". Hence, while the EFM thought of himself as leader of the accountants, they did not perceive him as such. Leadership by the EFM, as defined in this study, was at most frictional but mostly absent. The relationship of the EFM with the accounting manager and the accountants was characterized by confusion, ambiguity and perceived irrelevance on both sides, what Alvesson (2019a) labels a situation of divergent relationalities. Figure 3 provides an illustration of the leaderfollower relations at the outset of the management accounting change process.
In the following sections, we explain how leader-follower relations and the management accounting change unfolded, and specifically how accounting amplified the leadership dispute between the two top managers by involving others. Table 2 provides a summary of our analysis. Each of the following five sections addresses a specific dispute and contains a figure providing a schematic summary of our amplification analysis.

Necessity dispute
Early in 2010, the EFM had drafted a new accounting system in an Excel application that was useful for managing real estate projects but which had not yet been adjusted to manage the construction business. The EFM contacted the accountants and construction workers to obtain data to understand how the current system worked and what needed to be improved. It was important to involve both groups as there were two independent IT systems in useone for accounting information and one for project-based operational information. The current accounting system was maintained and updated by the controller, who was absent from headquarters at the time of the system design. The president decided to relocate the controller, who was the most skilful management accountant, to the construction site of a significant and financially risky real estate project. The controller reflected on how quickly the president reallocated his work tasks: It was strange . . ., the president told me: "You have to work day and night in order to implement a system for the allocation of indirect costs." The next day he told me: "We have a new project, we have no time, you have to go to the construction site and check all the costs related to that project." The president preferred to have the controller at the real estate site due to the high financial risk this project involved for the entire company. However, he used his formal power for the reallocation but did not explain what tasks he expected from the Controller beyond the work at the construction site. During the process of setting up the new accounting system, the accountants provided the information required by the EFM. However, due to the absence of  The president blames the EFM and the controller for misdesign of the new accounting system, in front of employees and in the absence of the vice-president; he does not reflect on his own mistakes and responsibility; increased turnover of accountants in the subsequent period leads to a fragmented organization accounting and leadership the controller, they were not able to make any suggestions regarding the design of the new accounting system.
To design a standard costing system that based on more accurate overhead allocations, the EFM also depended on information from construction workers who were able to administer the project-based IT system. However, due to the leadership dispute between the president and vice-president and the loyalty of the construction workers to the vicepresident, the EFM faced difficulties. The construction workers refused to provide the EFM with information, arguing they had already done so to the controller earlier and did not see any purpose in sharing it again. The fact that the EFM was working directly on behalf of the president did not make any difference. Construction workers delayed filling in their data. The project planner highlighted that they had the support of the vice-president in this decision.
I talked to the vice-president and he suggested I continue my work. We have other things to do and we don't have time for this.
Hence, the vice-president actively supported undermining the new accounting system and approved the construction workers' resistance. The following quote from the project manager further shows the emerging tensions in the company, particularly how the construction workers perceived the real estate staff of the EFM: The people who work in the real estate business arrive at the office at 9.30 a.m. At 11 a.m., they take a coffee break, including some brioches. They go to the office well dressed, making a show that they will do something [important and busy work]. They "ask us" (idiomatic expression in local language) for some data and we are here working hard and without paid overtime! When the president heard about the barriers in the system development in spring 2011, he called several construction workers over the phone to find out why they were not co-operating. They answered that they already had systems in place to manage their projects and were not interested in developing a new one. They also confessed that they did not know nor understand what the purpose of the new system was. The EFM had omitted to inform them about the need for the system and they seemed to be clueless as to what needed to be done and why. Hence, we label this first amplifier "necessity dispute" as there was a struggle over the purpose and necessity of the new accounting system amongst organizational members (see Figure 4).

Resource dispute
The president was disappointed in the work of the EFM who failed in testing and implementing the new accounting system. The president asked for a fresh start to the pilot project test. While he was not usually involved, in this situation he used his formal power and Necessity dispute amplification analysis employees performed the design and testing of the new accounting system. To achieve a good result, the president asked for the co-operation of employees and suggested testing the system based on one of their current infrastructure projects. The vice-president agreed to this and decided to test in one of the more complex infrastructure projects, which construction workers believed was accurately measured. The project portfolio manager explained: . . . this is the most complex project we are managing and the most resource-consuming. But it is also the project we are managing in a more accurate way.
The construction workers were responsible for providing information about the quantities of resources used, such as labour hours per project and materials. The cost drivers were provided by the accountants. For example, overhead rates included the cost of workers per hour, depending on their qualifications, as well as the cost of materials based on simulations that calculate the transformation of raw materials into final products. The EFM scheduled several meetings with accountants and construction workers to determine overhead allocations, transfer prices and the project margin based on the draft accounting system.
Calculating the project margins with the new system, the project, which had looked profitable with the old system, was now suddenly showing a loss! As information about the new system spread, the construction workers began to realize the impact of redesigning the new accounting system. It appeared the overhead rates were much higher now that ConstructionCo had entered the real estate sector, introducing higher additional overhead costs for marketing, administrative and sales personnel, infrastructure and vehicles. The project planner highlighted: When I saw the loss I jumped from my chair. We worked accurately on the budget, calculating the amount of resources needed. . . and now, we are aligned with forecasts, but those estimates give a loss in the end! We have to better understand the situation! The construction workers' realization of the potential effects of the accounting system on resource allocation and performance measurement was a turning point in the management accounting change process. Construction workers, as loyal followers of the vice-president and initially not interested in accounting matters, now started to be concerned about the accounting system design. This resource dispute created a new momentum in the management accounting change process because construction workers now refrained from resistance and wanted to be involved in the process (see Figure 5).

Accuracy dispute
The EFM and the president had stressed to accountants the importance of updating the overhead rates and transfer prices in the current accounting system. However, the Management accounting and leadership accountants had ignored this task due to unclear allocation of responsibilities. The controller, who had usually done those updates, was absent and was overwhelmed with work at the real estate site. Once the project manager of the pilot project realized the overhead rates were based on outdated accounting calculations, he informed the project planner. She had provided accurate estimates for the project by adopting the transfer prices and the overhead costs the controller had communicated to her, but she showed leading behaviour in this situation: as she had access to the software used by the accountants, she downloaded bookkeeping data and detailed the costs allocated to the cost centres. She analysed the data, but found that the overhead costs and transfer prices did not match the ones given to her by the controller. She realized that the outdatedcompared to the updatedoverhead cost were 25% instead of the 15% used by the pilot project. According to her calculation, the transfer prices for internal materials seemed 5% higher than they should be, which had a significant effect on the total project cost. The project planner highlighted: I worked hard for three days to check all the values. With the project manager, we retraced all the estimates we had made before starting the project. If we consider these costs, the project loses money! . . . I have done my work accurately, but someone else probably not. I'm angry! She and the other construction workers informed the vice-president. He shouted his head off and was angry with the president and with all the accountants, as they seemed an incompetent team. He argued that the president had to pay for his "mania of grandeur". The dispute between the two top managers became amplified and even more visible to the construction workers. In discussion with the vice-president during overtime in the evening, the project portfolio manager and the supply manager decided to monitor the accuracy of the new accounting system more closely. The project portfolio manager explained: It is unacceptable that we work accurately and those who should know how to do the accounting disappear. . . They give us wrong calculations! I am worried about them managing the situation. . . if they also manage real estate projects in this way. . .
The vice-president showed following behaviour, approved the worries of the project portfolio manager and the supply manager, and during their next closed meeting, which occurred every Saturday, informed all construction workers about their plan to monitor the accounting system more closely. Being loyal followers of the vice-president, several construction workers volunteered to do overtime and learn how overhead costs and the transfer pricing system worked. We label this questioning of the calculations between organizational members the "accuracy dispute": it remained ongoing, as construction workers continued to discuss the accounting system in their upcoming meetings (see Figure 6).

Ownership dispute
The accountants' lack of ownership of the current and new accounting system jeopardized the president's real estate initiative. A main problem was the lack of involvement of the accountants and construction workers. The president explained that he had moved some of the accountants closer to the construction workers to encourage interaction: They have the opportunity to interact with other members and see other things. But they prefer to stay in their open space office with their things and without any distraction or news. Probably I will have to fire some of them. The accounting system and our prospects require smart people.
This quote shows how he perceived his employees and how he intended to use his hierarchical power if they failed to deliver on his expectations. Based on our ethnographic data, we understood that the EFM had to execute the president's wish: He [the president] decided that some accountants had to move their office in order to make them aware of the importance [of communicating with the construction workers] to their work, otherwise they would have been fired. [. . .] and I have to express that message to them and control their behaviour.
Since neither the president nor the EFM maintained convergent leader-follower relations with the accountants, these relatively forceful and threatening instructions based on the formal hierarchy worsened the atmosphere while not increasing the interaction between accountants and construction workers.
In May 2011, having realized that the accountants were still not involved with the construction workers, the president hired a consultant to mediate communication between the groups (see excerpt from field notes in the introduction). The consultant was introduced by the president to facilitate communication. However, the accountants were not prepared for the meeting and appeared bored and apathetic. This was surprising to the construction workers, who expected the accountants to be active in this meeting. The project planner reflected: Accountants appeared not interested in the meeting. The consultant talked about the role of the new accounting system and the importance to have up-to-date data and an integrated system. I heard a couple of accountants saying that the controller is in charge of those activities.
The controller did not participate in the meeting: he had a meeting for the headquarters project he was directly managing. Although the topic related to the new accounting system, the accountants did not ask or say anything, with the exception of the accounting manager. The vice-president and many construction workers attended the event. The consultant explicitly addressed the vice-president. However, the latter did not respond. Later during the meeting, the vice-president's phone rang with a loud ring tone. He did not apologize: instead, he took the call and started a phone conversation inside the room. Then he asked the project portfolio manager to follow him outside. A short while later, they came back into the room and extended their discussion to include other construction workers.
The president and the consultant were irritated by the behaviour of the vice-president: the consultant twice had to interrupt his speech, yet the construction workers around the vicepresident did not care and continued the discussion triggered by the phone call. Instead of supporting the president and the consultant, the vice-president displayed a confrontational attitude towards the president's initiative by directly involving the construction workers in a task that was supposedly more important than the meeting. Through his behaviour, the vicepresident signalled to employees the divide of the two top managers. Instead of taking their top manager positions seriously, their personal dispute, their unwillingness to compromise, affected the entire organization.

Management accounting and leadership
Since the accountants did not take ownership of the accounting system, the construction workers insisted on integrating the new cost allocations, transfer prices and estimations into their own IT system rather than into the new system suggested by the EFM. The project portfolio manager explained that they thought the new system was not useful for their work: The system [the EFM used] is not useful! With that system, we do not accurately monitor construction and demolition projects. We should use our system and, if necessary, we should integrate it with the missing phases. This way we will be able to control everything better.
Ultimately, the construction workers took control of accounting matters. The new accounting system was implemented, although it was integrated in the operational system controlled by the construction workers. We label this "ownership dispute" to signify the struggle over who took ownership of the calculations and system administration (see Figure 7).

Responsibility dispute
The president had become increasingly disappointed with the work of the accountants, and particularly the EFM. He also realized that the controller had lost control over the accounting system. In an informal meeting with our on-site field researcher, the consultant explained: [The president] is still furious. He noticed that the controller does not have the company under control. He [the controller] has become totally absorbed with the headquarters project and does not work for the overall company. And he did not say anything about this to the president.
At the beginning of a meeting with the consultant in 2011, the consultant asked the audience (consisting of construction workers and accountants) if there were some daily activities which were useful to discuss. One of the construction workers raised an issue related to the incorrect overhead allocations that were identified during the testing of the accounting system. Everybody knew that the accountants were being addressed on this point.
The consultant followed this up, asking for more details, but then the president intervened. The president addressed the EFM, saying it must be easy for him to answer that question and explain the design of the new accounting system to everybody. The EFM did not anticipate this situation and gave an answer that did not refer to the system but to organizational control in general. The president pressed further and invited him to the blackboard for an illustration of his thoughts. Again, the president reinforced that he would expect an answer. The EFM lost the thread of his speech. He then started to discuss how real estate projects related to the accounting system. The president interrupted again, saying that graduates from his university have a "good reputation", yet the EFM seemed to be unable to convey that. The two managers had drifted apart and broken their leader-follower alignment. The president then turned to the controller and did the same with him.
It was silent and tense in the room. The accountants and construction workers were stunned. The meeting finished soon after. The president had scrutinized his direct reports in front of employees, showing how their leader-follower relations diverged. The EFM, the controller and some of the accountants left immediately. The treasury manager was shocked by the president's behaviour. If someone did not suit the president, he simply put them under the "guillotine". Others agreed. They guessed that the president would have not done that if the vice-president had been at the meeting. The treasury manager further commented: Today, I think the president revealed his disappointment with the EFM. During the last months, I have noticed [the EFM's] incompetence with the management of the accounting system and that he does not mix with us. [. . .] He is a real estate man and he is full of himself.
In the remainder of our case observation, until the end of 2012, the image of the accountants in the organization suffered as a result of these developments. One of the project managers confessed that the accountants "lost face during these months". The accounting department developed an increased turnover rate and the EFM left the company soon after our field research ended. The leader-follower relations between the president and the EFM, the president and the controller, and the president and the accountants dissolved. Ultimately, the atmosphere triggered by the leadership dispute amongst the two top managers had affected all employees in the organization. Drawing on this last part of our empirical insights, we add the label "responsibility dispute" to signify the situation when organizational members blame each other for mistakes during the management accounting change process (see Figure 8).
5. Discussion and conclusion 5.1 Reflection on the findings The aim of this paper was to explore the role of leadership in management accounting change processes and outcomes. In contrast to prior literature, our study does not support the assumption that leading and following behaviours echo the hierarchical structure. Instead, leading and following behaviours occurred in various directions and originated from diverse people, and were sometimes hindered by diverging perceptions between people about their relationship (Alvesson, 2019a). These leader-follower relations and their dynamics affected the unfolding of the management accounting change process and outcomes.
Whilst we highlight that the occurrence of leadership and followership is independent of the hierarchical structure, our findings also stress the importance of a functioning leadership at the top of the organization (e.g. Bhimani, 2003;J€ arvenp€ a€ a, 2007;Pfister, 2009). The development and testing of the new accounting system shaped a central stage where the leadership dispute between the top managers was amplified throughout the organization. By amplification, we mean that the accounting system triggered power struggles and fears of being disadvantaged amongst organizational actors, thereby reinforcing existing and shaping new tensions between organizational groups. As our analysis at ConstructionCo shows, disputes emerged surrounding the need for the accounting system, fears about the resource allocation, concerns about the accuracy of the calculations, contests about system ownership and blame for the responsibility for the breakdowns in setting up the system. These disputes involved typical resistance behaviour (e.g. Burns, 2000;Granlund, 2001;Malmi, 1997) such as intentional delays in the process, non-co-operation in meetings and reciprocal blame which serve the self-interest of individuals or specific groups at the expense of overall constructive organizational progress. Hence, our case opened up several classic management accounting change areas of dispute, conflict and struggle (Burns, 2000;Burns and Baldvinsdottir, 2005;Granlund, 2001;Kasurinen, 2002;Malmi, 1997).
However, most striking, and different from the prior literature, was the inability of the top managers to reconcile their strategic dispute, which ultimately became amplified during the design and use of the new accounting system. Hence, we assert not only that management accounting change depends on leadership, and how leaders instil and inspire values and meanings in followers in order to create momentum for change (Cobb et al., 1995;Kasurinen, 2002;Munir et al., 2013), but also, and importantly, that management accounting change can amplify a leadership dispute if top managers pull the organization in different directions and involve other actors in their dispute. Importantly, as we discuss below, how this conflicting tone at the top percolates top-down in the organization is influenced by the leader-follower relations between the top managers and other organizational actors, as well as the leader-follower relations amongst other actors themselves.
On the one side, the president established a convergent leader-follower relationship with the EFMthat is, they established alignment in the meaning and evaluation of the relationship between them, and we observed leading and following behaviours in both actors (Uhl-Bien et al., 2014). Interestingly, the president and the EFM perceived themselves as leaders of the accountants and saw their relations with them at the outset of the fieldwork as relatively unproblematic and favourable. However, the views of the accounting manager and other accountants of their relationship with the president and the EFM diverged significantly (Alvesson, 2019a). The accountants perceived the president as distant, someone who would only act "for his convenience", as an accountant said, and the EFM as an outsider who did not respect their team hierarchy and would "not mix" with them, as said by the treasury manager. Those insights show that there was no following behaviour from the accountants towards those managers, and their relationship was characterized by confusion and ambiguity (Alvesson, 2019a).
The president primarily used his managerial power to allocate plans and monitor the work of accountants. He initially allocated the design and use of the new accounting system to the controller, but then changed his mind and allocated him to the real estate site to manage a large and important project. The president did not explain to the controller why it was important to keep an eye on the accounting system despite the heavy workload at the real estate site. However, the president expected the controller to make sense of both tasks and return to him with issues (for example, if he had a lack of time to manage both). Similarly, the EFM was not able to progress with the new accounting system as the president had hoped, and was not able to develop a leader-follower relationship with either the accounting manager or the accountants. Ultimately, over the course of the fieldwork, neither the EFM nor the controller showed the leadership behaviour expected by the president. These developments culminated in the responsibility dispute where the president put them "under the guillotine". However, the president was not self-reflective in terms of assessing his own behaviour in failing to inspire and position the EFM and the Controller in the organization. Overall, there was a lack of leading behaviour by the President, the EFM and the Controller.
In contrast, the vice-president and the construction workers established what Alvesson (2019a) calls a relatively strong convergence in their leader-follower relations in terms of both the nature and assessment of their relationship. Although the vice-president was hottempered and authoritative, the construction workers knew this manager would care for them and support their interest. However, we also saw that the project portfolio manager, the project planner and the supply manager took the initiative and showed leading behaviour in investigating the accounting system. In those situations, the vice-president exhibited following behaviour and granted them leadership. Hence, our theoretical perspective differs from the traditional managerial leadership perspective, where the leadership/followership role of the individual is assumed (Blom and Alvesson, 2015;Uhl-Bien et al., 2014). Instead, we show that subordinates can take on leadership.
Importantly, however, our study also shows how the vice-president used his leaderfollower relationship to divide the organization. He created a counter-narrative about the need for the new accounting system and, for example, approved construction workers' delay of the system development. He also undermined the leading behaviour of the external consultant by disrupting the meeting and boycotting subsequent meetings, even though the president aimed to bridge the business units in the organization. From a counterfactual perspective, the vice-president could have reinforced strategic unity by demanding that his workers collaborate and by supporting the consultant. Instead, he used his leader-follower relations for his self-interest, divided the organization and hampered constructive organizational progress.
The uneven distribution of convergent and divergent leader-follower relations in the organization affected how the management accounting change process unfolded. On the one hand, the construction workers realized that accounting would affect the resource allocation and performance evaluation, which triggered their involvement and created a sudden momentum for change (e.g. Cobb et al., 1995;Munir et al., 2013). They became sensitized and would challenge the accountants' work, thereby improving the quality of accounting. However, on the other hand, the leadership dispute undermined the work of the accountants. They were not inspired and motivated to contribute to the initiative, partially because they were lacking leadership. There was a lack of leadership from the president, the EFM, the controller and the accounting manager, but also the aforementioned leadership from the vice-president, which excluded the accountants and undermined a collaboration between accountants and construction workers. In summary, although the implementation of the new accounting system was, somewhat surprisingly, accomplished, the way it was designed and used, and by whom it was administered and controlled, changed over the course of our fieldwork as a consequence of the leadership dispute. The evolving relationships between organizational actors explain how and why the change intended by the president was accompanied by several unintended consequences (e.g. Burns and Scapens, 2000).

Relation to the prior literature
Our findings contribute to the management accounting literature by introducing and proposing the study of leadership and followership as an important but, to date, largely neglected theme. In contrast to the traditional leader-centric and role-based approaches (Cobb et al., 1995;Hopwood, 1974;Kasurinen, 2002), we view leaders and followers as co-constructed and (analytically) independent of the formal structure, and suggest that for any leadership to occur, it requires following behaviour. Moreover, rather than a general relational approach, we introduce and apply recent work on divergent relationalities (Alvesson, 2019a) to theorize Management accounting and leadership and show how management accounting change can amplify a leadership dispute between top managers throughout the organization. In doing so, we also contribute to the leadership field by extending Alvesson's (2019a) theory and adding leader-leader alignment misfit as a further category to explore convergent and divergent relationalities.
Our contribution has not only important conceptual implications for leadership studies in accounting but also the related sensemaking/sensegiving literature (e.g. Guiliani and Skoog, 2020;Jordan and Messner, 2012;Kraus and Str€ omsten, 2012;Malsch et al., 2012;Meidell and Kaarbøe, 2017;Tillmann and Goddard, 2008). Specifically, we highlight the relevance of understanding the meanings that people give to their relationships to examine the quality of the influencing (sensegiving) process. Whether those meanings are convergent or divergent affects the functioning of the influencing process, and ultimately whether and how leadership and followership occur. Hence, the co-constructionist approach applied here provides an analytical tool to consider the potential ambiguity, confusion and conflict in the relationships between people when studying the influencing processes amongst them.
More broadly, the findings of our study entail several insights for the management accounting change literature (e.g. Burns and Scapens, 2000;Ezzamel and Burns, 2005;Cobb et al., 1995;Granlund, 2001;Jansen, 2011;Kasurinen, 2002;Lukka, 2007;Malmi, 1997;Munir et al., 2013;Scapens and Roberts, 1993). The case of ConstructionCo has provided a platform to explore how leadership facilitates the overcoming of organizational barriers (Cobb et al., 1995;Kasurinen, 2002;Munir et al., 2013). However, it also, and differently from prior literature, demonstrates how leadership creates barriers, ambiguities and conflict in the face of change due to the divergent relationalities (Alvesson, 2019a) and alignment misfits among people. Therefore, similarly to Kasurinen (2002), this case study analyses the barriers to and factors hindering accounting change, and not only the factors that promote it. However, in contrast to Kasurinen (2002) and others who draw on a static categorization of different sources of barriers to change, we take a dynamic perspective, illuminating how the emerging negotiations and re-negotiations of meanings amongst organizational actors about their relationship change their positioning during the management accounting change process.
Prior literature has emphasized the important role of change agents in the management accounting change process (Cobb et al., 1995;Kasurinen, 2002;Munir et al., 2013). We show that when a change agent (here, the EFM) is not introduced in the formal hierarchy nor a good fit as a leader for specific subordinates, the change agent is ineffective. As Blom and Alvesson (2015, p. 273) put it: "You may accept and comply with the manager's formal mandate, but when it comes to management of meaning (values, ideas, beliefs, understandings) subordinates can more or less choose if they take a follower position or not".
We also provide insights into the role of the momentum for change (Cobb et al., 1995;Munir et al., 2013). While the construction workers were initially what Kasurinen (2002) labels "frustraters" (i.e. resisting change), they later became major drivers in the process when the vice-president began to realize the importance of the new accounting system. Interestingly, the emerging visibility of new accounting rationales played a major role in creating the momentum for change (Hopwood, 1987). Increased visibilityfor example, in the implementation of an ABC system (Malmi, 1997) or production cost system (Scapens and Roberts, 1993)shapes accountabilities, which might lead to certain organizational actors resisting them. However, our findings contrast with this prior literature (Burns and Scapens, 2000;Granlund, 2001;Malmi, 1997;Scapens and Roberts, 1993) because we find that the emerging visibility of new accounting rationales creates a momentum for change. Organizational actors do not resist due to tighter accountability (Malmi, 1997;Scapens and Roberts, 1993) or being excluded in the system design (Granlund, 2001). Instead, the resisting vice-president and his loyal followers started to realize how subjectively constructed accounting actually was (Hines, 1988) and how its redesign might negatively affect their business unit performance and resource allocation, and as a consequence became active.
The co-constructionist and relational approach to leadership provides an alternative perspective to existing method theories in the management accounting change literature. Whilst the space here is limited, we will briefly discuss four such aspects. Firstly, the theorizing of this study further unpacks prior literature on power and politics (e.g. Burns, 2000;Markus and Pfeffer, 1983;Scapens and Roberts, 1993) by focussing on the meanings of relations, and showing how those relations, in interaction with management accounting change, (re-)shape power structures within the organization. For example, Burns (2000) used Hardy's (1996) framework to distinguish power over resources, decisions, meanings and systems. Through the perspective of leader-follower relations, we explain nuances to power over meanings by showing how convergent and divergent leader-follower relations enhance or reduce influence in the organization, creating momentum or barriers during management accounting change.
Secondly, a related stream of literature has drawn attention to the cultural orientation of specific organizational members and groups of actors (e.g. Bhimani, 2003;Dent, 1991;Taylor and Scapens, 2016). This literature has also highlighted that cultural interventions (Bhimani, 2003;J€ arvenp€ a€ a, 2007) might be necessary to find acceptance for management accounting change. Our theorizing enhances an analysis of such interventions, for example, whether and how leaders reinforce (amplify) the divide between organizational units, or bridge them.
Thirdly, a stream of research on management accounting change has focussed on identity work, either at the micro level of individual managers (Goretzki et al., 2013) or the organizational level (Abrahamsson et al., 2011). This literature has focussed on the discrepancy between the current identity and a future image as a driving force of the change process. Our study entails two top managers that developed a different narrative for the future image of the company. We show how these conflicting future images affect the change process, how individuals at the micro level create and shape leader and follower identities, and how this ultimately affects the change process.
Finally, whilst applying a different ontology, the divergent relationalities lens mobilized here aims to consider the complexity and dynamics in the change process, which has been emphasized by researchers using actor-network theory (Andon et al., 2007;Quattrone and Hopper, 2001).

Implications for practice
The learnings from this study highlight the responsibility of top managers (or any cultural gatekeeper) to establish an inclusive rather than a divisive tone at the top to avoid accounting being an area where disputes might be amplified. Leadership and followership are not attached to formal rank and prestige, but leading and following behaviours are equally important and need to be situationally assessed. Leadership from top managers can inspire and encourage other employees to lead during management accounting change, but it can also create a toxic work environment if top managers put their self-interest ahead of the larger collective they are responsible for.

Future research opportunities
The accounting literature has focussed much on management but has largely neglected research on leadership. We have presented an approach to distinguish analytically between leader-follower relations and manager-subordinate relations, and introduced followers and following as an important part of the leadership process. Based on the recent work on divergent relationalities by Alvesson (2019a), we have further distinguished between leadership, frictional leadership, no leadership and a leadership dispute. This theoretical lens offers opportunities that can be applied in various accounting and control domains, possibly using similar, different or a combination of methods for the inquiry. For example, future research could explore in more depth how family firm dynamics affect the leader-follower Management accounting and leadership relations in management accounting change (Leotta et al., 2017), or apply the lens of leadership and followership in a different domain such as social and environmental accounting research. Our co-constructionist approach also offers possibilities to take a more critical perspective. Accounting is often a tool to create ranks and hierarchies based on which the labels "leaders" and "leadership" are distributed. But what kind of followership and following behaviour are behind this? Looking at the effects of calculative practices on shaping the terms "leaders" and "leadership" opens many research opportunities in future accounting and control research. Notes 1. According to Lukka andVinnari (2014, p. 1309), a method theory (or theoretical lens) is a "meta-level conceptual system for studying the substantive issue(s)" of the domain in focus.