Behind Smoke and Mirrors: A Political Approach to Decoupling

693268 OSS0010.1177/0170840617693268Organization StudiesKern et al. research-article2017 Article Behind Smoke and Mirrors: A Political Approach to Decoupling Organization Studies 2018, Vol. 39(4) 543–564 © The Author(s) 2017 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav https://doi.org/10.1177/0170840617693268 DOI: 10.1177/0170840617693268 www.egosnet.org/os Anja Kern Baden-Würrtemberg Cooperative State University Mosbach, Germany Aziza Laguecir IESEG School of Management, France Bernard Leca ESSEC Business School, France Abstract In this paper, we examine neglected dimensions of decoupling - i.e., its power and political aspects. We draw on an empirical study of the reaction of two hospital subunits and an external agency (the Regional Health Agency) to a policy implementation, to contribute to the recent renewed interest in decoupling. We first reconsider the distinction between internal and external actors by investigating how they interact in their responses to the new policy implementation. While observing different forms of decoupling, we show how power and politics allow us to understand how these forms are articulated and related. Furthermore, we highlight that contexts characterized by institutional complexity are particularly propitious for decoupling. Finally, we outline that how actors use logics to justify their claims might differ significantly from how they enact those logics. More broadly, this paper contributes by bringing back power and politics into the analysis of institutional processes. Keywords agency, decoupling, healthcare management, institutional theory, power dependence theory Introduction Until recently, decoupling seemed a well-studied topic, with decades of research allowing a good understanding of how organizations manage to ceremonially adopt formal rules while keeping their internal practices untouched (Bromley & Powell, 2012, for a critical review; Meyer & Rowan, Corresponding author: Bernard Leca, Accounting Department, ESSEC Business School, 3, avenue Bernard Hirsch, CS 50105, 95021 CergyPontoise Cedex, France. Email: [email protected] 544 Organization Studies 39(4) 1977). Nevertheless, some authors are revisiting the notion, pointing to important aspects of decoupling remaining unaddressed. First, decoupling has been mainly considered at the organizational level, and little is known regarding the decoupling of subunits and how such decoupling can vary within the same organization due to the complex dynamics underlying such processes (Moll & Hoque, 2011). Second, while existing research focuses on symbolic policy adoption, other different forms of decoupling remain unexplored. Bromley and Powell (2012) emphasize the need to distinguish symbolic adoption from symbolic implementation and to further explore it. Third, the authors highlight that institutional complexity favours decoupling because, in such situations, it is impossible to establish whether a policy achieves the exact intended goals (Binder, 2007; Bromley and Powell, 2012). This creates a gap in our knowledge on intra-organizational decoupling as related to institutional complexity. We fill this gap by illuminating why subunits differ in their adoption and implementation of policies in situations of institutional complexity. To do so, we conducted an in-depth inductive field study analysing the reactions of different organizational subunits to a new policy – i.e., the casemix – in a public hospital in France. Our findings illuminate the role played by power and politics in the different responses. The relative power of each subunit played a key role in their response to the policy. Following an abductive approach, we further theorize and discuss this finding by pointing to the relevance of Power Dependence Theory (Emerson, 1962) for a better understanding of decoupling and to the need to pay more attention to power and politics when examining such organizational phenomena. For the sake of clarity, we organize our paper as follows. In the next section, we review recent research on decoupling, pointing to the need to bring in power and politics, and introduce power dependence theory (PDT). Subsequently, we present the research setting and method. Then, drawing on PDT, we analyse the responses of three groups of actors to a new policy, examine how they decoupled their practices, and identify policy-related tools. Finally, we discuss the results and detail the implications for future research. Revisiting Decoupling Recent research suggests that decoupling is a fundamental aspect of organizational life, and its importance may have been underestimated, particularly within institutional analysis (Binder, 2007; Hallett, 2010). Here, we elaborate further on the three dimensions of decoupling that have recently been revisited. The first re-evaluation relates to the level of analysis. A traditional assumption of decoupling studies is that they construe the organization as a whole that either couples or decouples its activities from policy and institutional pressures (Meyer & Rowan, 1977). Decoupling is mainly examined at the organizational level, assuming that organizations are homogenous. For instance, Westphal and Zajac (2001) show that the greater the CEO’s power over the board, the greater the extent to which companies decouple their financial investments from formally adopted repurchase programmes, despite pressure from investors to adopt them. This approach pays scant attention to intra-organizational diversity and consequently neglects the decoupling phenomenon taking place at the level of subunits. Recent research has started to focus on how different subunits decouple or not, within organizations (e.g., Binder, 2007; Pache & Santos, 2013; Sandholtz, 2012; Tilcsik, 2010). The second re-evaluation was introduced by Bromley and Powell (2012), who suggested that research should be extended to other forms of decoupling beyond the symbolic adoption or violation Kern et al. 545 of policies and the related tools. They label this traditional form of decoupling ‘policy-practice decoupling’ (PPD), and identify another form ‘Means-ends decoupling’ (MED) (Bromley, Hwang, & Powell, 2012; Bromley & Powell, 2012; Wijen, 2014) which refers to situations where policy is actually adopted but implementation is symbolic. In such cases, compliance with what Wijen (2014) terms ‘the letter’, is achieved, yet the goals (‘the spirit’) for which those rules were defined are not achieved, therefore the implementation does not have the intended consequences. In such cases, the contribution of practices, aligned along formal structures to promote the policy’s primary goals, may be tenuous or opaque at best (Bromley et al., 2012; Bromley & Powell, 2012). One example of such MED is provided by the study by Wijen (2014) of fair trade practices, which have inadvertently aggravated exploitation by cocoa bosses in strict opposition to the goals for which those practices were initially started. The third re-evaluation comes from the increasing interest in research on ‘institutional complexity’, where actors confront incompatible prescriptions from multiple logics (Greenwood, Raynard, Kodeih, Micelotta, & Lounsbury, 2011). In such complex settings, where multiple logics intersect, intra-organizational decoupling is more likely to occur (Binder, 2007). Corporate social responsibility (CSR) exemplifies such situations. When engaged into CSR, organizations must address multiple and potentially contradictory issues (e.g., social issues, environmental issues, relations with multiple external stakeholders) using multiple and potentially contradictory logics. As Wijen (2014, pp. 307–308) puts it, ‘The sheer number and multifaceted nature of these issues render a full understanding challenging for sustainability experts (Atkinson, Dietz, & Neumayer, 2007), leave alone for firms whose business is to do business.’ Those situations favour PPD as well as MED because this complexity is likely to favour opacity in the relations between means and ends (Wijen, 2014). The Missing Political Dimension of Decoupling A neglected dimension in the study of decoupling is the importance of power and politics. Recent research suggests that power and politics might be of importance for explaining decoupling (Tilcsik, 2010) but does not further elaborate. The limited interest in power and politics is unsurprising because current institutional research is notoriously considered to have a ‘rather sanitary view of the world’ (Munir, 2015, p. 91) that lacks consideration of power (e.g., Clegg, 2010; Willmott, 2015). While the ‘Old’ institutionalism used to consider power as central (e.g., Gouldner, 1954; Selznick, 1949), the ‘New’ institutionalism (which emerged in the 1970/80s) focused on institutions as cognitive schemes, myths and ceremonies and has paid less attention to power, interests and conflicts (Dimaggio & Powell, 1991). Along with other scholars (e.g., Lawrence, 2008; Munir, 2015), we argue that power should be reintroduced as a chief concern to understand institutional processes. To do so, specifically regarding decoupling, we choose to draw upon PDT (Emerson, 1962). PDT does consider power as an attribute of isolated individuals but also as a property of the social relation examined. In this view, A will have power over B if B aspires to goals or gratifications whose achievement will be facilitated by appropriate actions from A (Emerson, 1962). Thus, power is a function of two dimensions: mutual dependence and power imbalance (Casciaro & Piskorski, 2005; Emerson, 1962). Mutual dependence is the existence of bilateral dependencies between A and B regardless of whether they are balanced or imbalanced. Power imbalance accounts for the difference in power between A and B – i.e., the difference between the two actors’ dependencies. Mutual dependence accounts for the given power relations that actors need to manage. The power imbalance is the more political dimension because this balance can be managed by actors. They can balance operations to reduce, or increase, asymmetry of power that might exist. Emerson 546 Organization Studies 39(4) (1962) insists that mutual dependence and power imbalance must both be considered. At stake here is agency, more than resources, thus being distinguished from Resource Dependence Theory (RDT) (Pfeffer & Salancik, 1978). Building on this approach, we examine subunits’ responses to a new policy in a complex setting. Our intention is to contribute to a better understanding of why organizational subunits differ in their reactions to a new policy, focusing on power and political relations. Further, we examine the articulation between different forms of decoupling and their relation in a situation where actors engage under institutional complexity. We accomplish this through a detailed analysis of negotiations and power plays around the introduction of a formal policy, namely the casemix, and the responses of the two subunits we observed during a field study of a French hospital. Research Site The field study was conducted from 2001 to 2004. Facing major cuts in public funding, at the time of our study, the government decided to change from an expense-based to a performance-based hospital funding system. Following the logic of the initial US DRG (Diagnostic Related Group) system, the French government developed its own system, classifying patients into homogenous groups, DRGs,1 based on medical and economic criteria (Kimberly, De Pouvourville, & D’Aunno, 2008). A national agency compiled the DRG data in the casemix, including two types of data: for each DRG the regional tariff (average costs across all, or a sample of, hospitals in the region) and average costs at a particular hospital. The casemix measures a hospital’s output over a defined period by comparing hospital costs with the regional tariff for each DRG. The tariff is the income that the hospital receives for each DRG. When the hospital has a DRG cost lower than the regional tariff, it is making a profit; when the DRG is higher than the tariff, it is losing money. The Regional Health Authority (RHA) used the casemix to negotiate its global budgets, strategic plans, and investment with each hospital. The goal of casemix introduction was to make healthcare services more efficient in hospitals, notably by increasing activity levels with the same capacity. This change included new practices for budget allocation. Indeed, budget negotiations with the RHA2 would now be based on each subunit’s casemix performance, rather than on overall hospital performance (Michelot & Rodrigues, 2008). In 2001, Hospital-F had approximately 35,000 admissions per year and a total of 1089 beds. It introduced the casemix at the global hospital level in 1994; since then, casemix performance has influenced only total hospital budget increases or decreases. We studied this new policy implementation at the clinical level by focusing on three interdependent groups of actors: the RHA and the cardiology and surgery subunits. Methodology This study emerged from a larger research project on the DRG implementation in France and at Hospital-F. This research was an in-depth abductive case study (Timmermans & Tavory, 2012) covering a three-and-a-half year period, during which one of the field researchers conducted participant observation (Jorgensen, 1989; Spradley, 1990). The field researcher committed to provide the hospital with feedback during the study, and the hospital committed to support the study by granting access to the information required. This material informed a broader analysis of the casemix and provided a framework for operational practices. During this research, we became aware that the casemix implementation at the clinical practices level was more complex and difficult than expected, leading us to further investigate the different types of reactions to the casemix implementation. Kern et al. 547 Data collection Data collection primarily relied on direct observations and interviews. During the five-month preparation and immersion phase, the field researcher spent one day a month in the hospital. During the 18-month direct observation period, she spent one to two days per week there, and then she spent an additional one day a month over a further 18-month period. She was based in the medical control subunit and participated in the daily practices of employees. She also attended relevant meetings between the medical control and clinical subunits addressing casemix performance and budgets. She attended the cardiology and surgery subunits’ project and staff meetings, and those of other hospital departments. She also observed actors during lunch and coffee breaks and in chats in the office, listened to communications and watched daily operations. To gain a basic understanding of the clinical processes, she scheduled particular days to follow clinicians and nurses. Once familiar with the clinical staff, she was also invited to other meetings, such as medical staff meetings and meetings to discuss future clinical department strategies, and she was given access to all related documents. Further, she was invited to meetings of the hospital committee that was preparing and overseeing RHA budget negotiations. This gave her access to higher level discussions between hospital management, physicians, union representatives, directors and other RHA representatives. She also took extensive minutes, including exchanges in formal observations with archival data analysis. She took notes during work and more detailed field notes immediately afterwards, which she ordered chronologically in a research diary for each field visit, each visit resulting in 3–5 pages of notes for a total of approximately 750 pages for the entire study. This prolonged engagement allowed her to become aware of issues relevant to the actors, thus enriching our understanding of responses to the casemix system at the clinical practice level. We combined direct observations with archival data analysis: mainly documentation from the hospital accounting systems, reports and presentations discussed at all attended meetings, specifically casemix performance and clinical unit budget data. The researcher also conducted formal semi-structured interviews with key actors including national casemix experts, RHA representatives, the CEO and CFO, the medical controller, department heads, physicians, and nurses.3 The formal interviews lasted an average of 60 minutes each, resulting in 26 fully transcribed formal interviews. Analysis For the sake of clarity, we distinguish here between three different stages. However, these stages of analysis often overlapped, as we moved back and forth iteratively between data and theory. First, we organized the collected data to analyse responses to casemix at the clinical level. We used interviews and archival documents to elaborate on and confirm issues arising from our observations. Because our initial approach was inductive, the main questions emerged from this organization of data. Second, we reorganized the data around emerging issues of significance related to the dynamics of casemix adoption and implementation. We constructed within-case narratives and followed with case comparisons. Third, using PDT concepts, we returned to the data and focused our analysis on identifying what motivated subunits’ responses to casemix, and we compared their responses, also iterating between the data and the decoupling and power dependence literature. We gradually rearranged the narratives chronologically and by analysis level, emphasizing similarities and differences between the two subunits as the casemix was introduced (Eisenhardt, 1989). Taking the RHA into account, we also analysed decoupling between casemix and clinical practices within the subunits. We organized all relevant data according to this perspective and produced an abbreviated narrative, which is presented in the next section. 548 Organization Studies 39(4) As a final step, because we performed our analysis after completing the data collection, we confirmed our interpretations with key actors (e.g., RHA representatives, physicians), discussing our final results with them so that we could include their feedback and reflections in our analysis. We used the variety of cases studied as an opportunity to contrast different situations and elaborate theoretical relationships (Eisenhardt, 1989; Yin, 2009). Below, we address the different responses of the two hospital subunits and the RHA to the new funding policy and examine how they addressed power imbalances and mutual dependencies. Findings In the next section, we describe the new policy and responses from two subunits (surgeons and cardiologists). To investigate this, we mobilized PDT. Describing surgeons’ and cardiologists’ responses to their environment and comparing these responses enabled us to examine the motivations triggered by the application of a new formal policy and the power imbalances and mutual dependencies between the hospital subunits and the RHA. The new policy At the time the study began (2002), a new policy was implemented. Initially, the policy funded each hospital based on its global casemix performance. However, the policy shifted to become based on a more detailed casemix performance level, funding each subunit based on the casemix performance of their specific homogenous patient group (DRG) (Michelot & Rodrigues, 2008). At the same time, Hospital-F management faced declining economic performance and, to address this, decided to proactively implement casemix at clinical levels earlier than the RHA required. The spirit of the new policy was detailed in an official letter in which the RHA director in charge of conducting budget negotiations with Hospital-F set conditions for the upcoming budget negotiations (framing letter of 28 May 2002). The budget negotiations would be based on an analysis of past activity, Hospital F’s current economic situation and clinical outcomes, and regional needs. More specifically, the negotiations would establish the financial means the hospital would receive from the RHA, specifically for infrastructure investments. The framing letter detailed the new policy and its implications. It outlined Hospital F’s situation, detailed its priorities,4 and identified projects to be discussed and the methodology to be applied during the negotiations. In this framing, the RHA explained that for the next budget round, casemix use would become obligatory, as the negotiations would be conducted at the more detailed clinical subunit level. The letter also emphasized the financial difficulties Hospital-F had recently experienced. By 2002, Hospital-F had accumulated a deficit of several million euros and had experienced rising costs accompanied by decreasing clinical activity (fewer patients). In November 2002, the RHA director opened the first general meeting to negotiate the budget for the next five-year period with a serious tone: These negotiations are going to be a difficult exercise. There is no flexibility. There is no money left, neither in the hospital nor the region … We have to change hospital management. It has to move from hospital level to service lines [i.e., DRGs, clinical levels], that is to say, to clinical activities. This is going to be inscribed in the budget contract we will be making. Big hospitals, such as Hospital F, have to adopt this ‘management by service lines’ approach. The hospital must decrease its costs while at least maintaining the present service level if it wants to stay competitive. In the past two years, costs have risen above the regional average while activity has decreased. The hospital is now in a bad financial situation, and the only way out is to manage its costs more rigorously. (RHA director) Kern et al. 549 As one of the largest general hospitals in the region, Hospital-F is situated in an agglomeration of more than 1 million inhabitants characterized by a rising population marked by poverty. Hospital-F has a quasi-monopoly for hospitalization and, more specifically, for surgeries: it covers 50% of the region’s total patient stays, 75% of its public sector hospital stays, and more than 50% of the region’s total surgeries. Therefore, as a key player in the region’s health sector, Hospital-F has received particular attention from its RHA. The hospital’s financial situation declined over a very short time period (within a year). The hospital’s executive managers benchmarked the hospital’s performance against the regional average in the months before negotiations began. As their performance had fallen below the regional average, they knew this would become an issue in the negotiations. In particular, rising costs for medical staff (doctors, nurses, etc.) (+18.6%) and medical logistics (+120%), both above the national average, caused this declining performance. The detailed economic performance of different clinical activities was unknown to both clinicians and Hospital-F managers. During the interviews, it became clear that many clinicians were unaware of which DRGs their patients belonged to and whether the DRG was below or above tariff (i.e., if it was saving or losing money). As illustrated by one clinician’s statement: ‘I do not know what the DRGs of my patients are. No idea.’ Further, at the beginning of our study, clinicians, and staff more widely, had no access to casemix data. According to the CFO, this was because the former medical controller5 was not in favour of using casemix for internal management and so never issued any such data. The CFO explained that he had gained access to this data only recently with the appointment of a new medical controller. Below, we describe how the surgery and cardiology subunits experienced the anticipated adoption of casemix at the clinical subunit level, followed by a description of the results of the negotiations with the RHA. The surgeons Rather than ceremonially complying and decoupling their practices, the surgeons refused to implement casemix. Our findings elucidate the motivations behind this refusal to comply, focusing on power dependence. Surgery has always been the centre of attention for both hospital management and the RHA; it is this particular hospital’s “raison d’être”. Without surgery, Hospital-F would be doomed to close because all other subunits were subordinate to surgery. The surgeons considered their clinical practice and expertise as being “too complex to be evaluated by casemix or management. An evaluation of our practice can really be done only by peers” (Surgeon 4). Consistently maintaining this perspective in the past, surgeons were reluctant to see any management intrusion in their practice. They clearly perceived themselves as experts, highly trained to deal with a sophisticated practice, and other subunits at Hospital-F shared their perception, along with RHA staff, as the following quotes show: Surgeons require a long training time; it is a difficult and very technical field. In order to operate, surgeons need to have experience and be highly skilled. (Intensive care physician) Surgery is key for the hospital’s existence. If we want to keep our status as a major hospital in this area, we must keep and extend surgery. (Management employee) Surgery is an important medical discipline; heavy and complex surgery is carried out mainly in public hospitals. So this is an important task. (RHA representative) Apparently, the surgeons are valued more than we are. I guess surgery is more technical and requires heavier infrastructure and equipment than we do. So it is just more impressive than prescribing medicine. (Cardiologist 3) 550 Organization Studies 39(4) The casemix reflected surgery’s central and complex nature. The DRG systems favoured technical activity, such as surgery, over non-technical medicine. Tariffs for surgical procedures were higher than those for medical procedures because they involved more resource consumption and therefore higher costs. For instance, while the tariff6 for chronic heart failure acts (including other pathologies) was EUR 4337.47 in 2004, the tariff for installing a permanent pacemaker for heart failure was EUR 9598, and valve replacement surgery with cardiopulmonary bypass was EUR 18257. This means that when surgical procedures were performed, hospitals received more money than when they were not. The casemix and DRG systems favoured surgical acts by incentivizing surgery, even though patient benefits might have been limited, as this would ensure more funding. This dynamic placed surgery at the very core of the DRG system. Formerly funded at the hospital level, Hospital-F surgeons did not have to be concerned with funding or management issues. However, although the surgeons did not voluntarily use casemix beforehand, the framing letter stated quite clearly that they would have to use it in their next RHA budget negotiation. In the framing letter, the RHA pointed out that surgery activity had dropped in contrast to other medical services. Further, while several surgical speciality activities had stagnated or declined, overall expenses had increased by 7.7% and, in some surgical subunits, by as much as 15% (internal document). Casemix introduction at the clinical level for the surgery subunit. The clinical level casemix introduction began in the summer of 2002, after the framing letter was circulated internally and the proposed goals were formally presented to the clinical staff. In line with other clinical specialities, the surgery subunit established a working group with the goal of implementing casemix at the clinical level and thereby integrating it into surgery activities. The working group, which included surgeons, the head of surgery, nurses, RHA representatives, hospital managers, the medical controller, the CFO, and a management accountant, met over the next few months. During these meetings, casemix figures were presented, with the surgeons contesting them from the start. One surgeon explained: These figures are very crude and cannot take into account the complexity of clinical practice. It would be dangerous to base decisions solely on this data. We have to see the whole picture. (Surgeon 1) Another noted: We all know that casemix figures are not very accurate. Also they do not include the main clinical indicators, such as quality of care and outcomes. We should not use them as a basis for decision-making. (Surgeon 2) In the subsequent surgery working group discussions, the surgeons deliberately did not draw on casemix figures. The head of surgery was the most salient opponent to casemix figures. He pointed out that medical aspects were more relevant to discussing surgery activities and to justifying the need for investments and budgeting: This investment [i.e., surgery equipment] is necessary, as the equipment in our operating rooms has to be updated. It does not correspond to state-of-the-art technology … If there is no surgeon at Hospital-F, then this also calls into question the existence of other areas such as the anaesthetists, medical resuscitation and, in the end, the hospital as such. (Head of surgery) Using this argument, the head of surgery is clearly warning other specialities and professionals in the hospital that questioning the existence of surgical activity at Hospital-F would, by extension, affect others. This view was also spread within the hospital, as the following excerpt shows: Kern et al. 551 It does not matter what the casemix says; one thing is certain, surgery has a central place in the hospital. If there is no surgery, this hospital is at risk; its existence is questioned. (Physician from the intensive care unit) In a surgery working group meeting, surgical service line market figures were also discussed. These showed that, for several surgical specialities, private hospitals, which performed 46% of all surgical procedures in the catchment area, including complex interventions (long stay), provided strong competition. Despite these figures and the declining economic performance of surgical activities, the surgical subunit existence at Hospital-F had never been openly questioned during working group meetings. This contrasted with the declaration during a previous general budget meeting, when the RHA representative made the following statement: We are going to discuss all fields of activity, and we should discuss without taboo. For example, in the presence of a strong private sector and a poor performing public service, we should ask if this activity should be continued in the future. (RHA representative) This was a general remark, and the RHA representative made no explicit mention of surgical activity. Further, the surgery infrastructure, particularly the operating rooms and wards, were outdated and did not meet state-of-the-art standards, so if the surgical subunit was to be maintained, it would need modernization and investment. During all of the budget negotiations, the surgeons continued to refuse to use casemix, claiming that they followed different logic and that they had no intention of compromising. In one meeting, a discussion started over which methodology should be employed for budget negotiations at the service-line level, and the surgeons stated firmly that: Our only objective is the patient. All our thinking is and must be focused on the patient. Casemix does not allow to evaluate our activity from a clinical point of view. The quality of the clinical activities is not included in the casemix. So, from my perspective, it is absolutely necessary to provide us with tools that allow proper analysis of the clinical dimensions and that are more refined than casemix. (Surgeon 2) If a patient arrives in the middle of the night, we treat him or her, so all of this is not properly covered by in casemix. (Surgeon 3) Although the surgeons were aware that their costs were above DRG tariffs, they opted not to bother to analyse casemix figures more closely to better understand the reasons for their casemix performance. Casemix figures were not distributed or discussed at internal surgical subunit meetings, such as staff meetings, and the surgeons did not use them to follow up on clinical activities: Our staff meetings have to focus on discussing individual patients. It is on the basis of these discussions that we improve our practice and outcomes. Everything else is a distraction and does not help the patients. (Surgeon 3) They stated that the patient was their primary interest and that their practice and decisions revolved around improving patient care, regardless of the casemix. The cardiologists In contrast to the surgeons, the cardiologists claimed that they would proactively use the casemix within their clinical activities. Indeed, the cardiology subunit volunteered to use casemix for 552 Organization Studies 39(4) internal management before budget negotiations with the RHA began. Nonetheless, closer investigation reveals that coupling the casemix with the cardiologists’ clinical activities was not that simple, as the cardiologists adopted casemix in pursuit of their own agenda and to minimize impact on their activities. Indeed, in examining our data, we found that the cardiologists did not embrace the casemix logic, but instead drew upon it to achieve their own goals. Before casemix was introduced, the cardiology subunit performed quite well in cardiology exams, the cardiology intensive care unit, and minor interventions such as catheters and stents, registering an increase in activities and a relative cost reduction at the beginning of our study. The overall hospital management had a positive view of the subunit, praising their efficiency: Cardiology is one of our key services. It is not only crucial in terms of patient numbers and budget, but it is also efficient. (Hospital general management) Nevertheless, management never managed to gather the resources to fund a chronic heart failure unit project that would have provided complete patient handling (without performing any surgery). The cardiology department considered this the most important project. Chronic heart failure will become one of the most important diseases in the world in terms of patient number and costs. We are trying to establish a new way of dealing with this disease, taking into account the patient, his family and context. The way we treat these patients will contribute to decreased costs at a system level, while increasing the quality of life for these patients. (Cardiologist 1) Cardiology was also among the RHA’s priorities. By adopting casemix, cardiologists found an opportunity to obtain the necessary resources to create this unit. The Head of cardiology, who was in favour of casemix, assumed that the casemix measures would bring greater fairness to the budget allocation process: Casemix will enable internal budget allocation. Rather than being the outcome of a political game, the budget will be linked to actual activities … it is also very clear that this is a very incomplete tool. We, clinicians, use much more refined data at the patient level to analyse our practice. It is this clinical data, which is most important in the development of this. (Head of cardiology) He anticipated that casemix would place his subunit in a more favourable position than before, as illustrated by the following: I think that we are a profit-making department, and therefore I want to have a share of the profit to invest in my department. At the moment, the entire profit is used to make up for the losses of other clinical services. Casemix could stop this practice of cross-subsidy … If we get back some of the profit we make, we can improve clinical care for patients. For example, there are new and better quality stents on the market, causing less inflammation for patients. However, those stents are not currently covered by the casemix tariff. If we can get some of those benefits back, we can use them to buy more of those new, expensive stents, which are of better quality. (Head of cardiology) A cardiology working group, including cardiologists, the head of cardiology, nurses, RHA representatives, hospital managers, the medical controller, the CFO, and a management accountant met regularly to prepare for its implementation. In the first casemix meeting, the medical controller proposed editing the casemix report to give the team an overview of the financial performance per DRG. This report contained all cardiology DRGs and their costs compared to the tariff (as set by the French government on the basis of average regional performance). 553 Kern et al. Table 1. Evolution of chronic heart failure related DRG between 2000 and 2004. 2000 2004 DRG Total patient number Patient number in % Total costs in Euros Total costs in % Difference in % with tariff 184. Chronic heart failure 05M09V chronic heart failure without other pathology (Former DRG184) 05M09W chronic heart failure with other pathologies (former DRG184) 302 270 16% 25% 1 194 947.86 797 260 21% 10% 14% –2% 120 11% 493 956.40 6% 5% At the next meeting, the participants analysed the casemix report. Their attention was drawn to DRG 184 (chronic heart failure), which was important in terms of patient numbers, costs, and income, as well as being 14% above the tariff (Table 1). This was critical for the cardiologists because the chronic heart failure unit that they wanted to create was intended to specialize in this DRG. The fact that this DRG operated at a loss posed a potential problem for the funding of the chronic heart failure unit because RHA approval of the investment depended on the DRG’s economic performance. The head of cardiology tried to understand the reasons for this overrun: We need more detailed analysis of what is going on with DRG 184. But first and foremost, this has to be discussed with the [senior clinical physician in charge of heart failure], who is in charge of chronic heart failure in our department. (Head of cardiology) As the cardiologists investigated further, they were astonished to find out that a significant proportion of patients in DRG 184 were actually treated in other departments. Further, costs for patients treated outside the cardiology department were indeed higher. These patients had multiple pathologies. The head of cardiology agreed with the cardiologists that DRG 184 should comprise only patients treated inside the cardiology subunit, and those who were suffering from chronic heart failure but treated outside the subunit should be excluded from this DRG because they had multiple pathologies. The medical controller agreed that these patients could be classified according to another one of their pathologies. The medical controller and the cardiologist in charge of the heart failure unit project presented the initial casemix use results by cardiologist at the clinical level to other clinical department heads to show them the potential benefits in terms of both economic performance and clinical quality. Specifically, these results showed that the average cost per patient for DRG 184 had gone down over a six-month period. The cardiology subunit, and particularly DRG 184, was presented as the most striking example of how casemix could be used to pursue the goals set by the underlying policy. Nevertheless what was not discussed during the presentations was how those spectacular results had been achieved – by diverting some multiple pathology patients to another DRG to improve their own DRG performance, thereby focusing on less costly patients and increasing their internal subunit activity level (bed occupancy rate). This omission had consequences at the hospital level. While the cardiologists improved their performance, the expenses were actually just transferred to other services. Accordingly, the nomenclature of the DRGs changed in 2004. To set aside multiple pathologies cases, DRG 184 was split into two DRGs: 05M09V and 05M09W (Table 1). The first category kept chronic heart failure activity including only cases without other pathologies, while the second 554 Organization Studies 39(4) included chronic heart failure with other pathologies. As shown in Table 1, in 2004, the first DRG (without other pathologies) was below the tariff established by casemix (by 2%), thereby making some profit. However, the second was above tariff by 5%. According to casemix logic, these changes in nomenclature also changed the related tariff. As the tariff is an average of the costs for the entire region, the tariff for the first DRG (without pathologies) went down, yet the performance of Hospital-F remained above this tariff. Further, the tariff of the second (with other pathologies) increased. Thus, although the related clinical practices and costs in Hospital-F did not improve, the performance as benchmarked against the tariff improved because the tariff rose. As a result, while the cardiology subunit’s performance improved, overall hospital performance remained the same because the performance of the subunits that subsumed the cardiology expenses had declined. This contradicted the casemix goal, which was to improve overall hospital performance. The cardiologists depended on the RHA to obtain the additional funding they needed to develop a chronic heart failure unit. The power imbalance was not in their favour, as they had virtually no bargaining power in their relationship with the RHA. Consequently, they implemented casemix as a means to pursue their own ends – to obtain a chronic heart failure unit – rather than aiming to achieve the RHA’s primary goal. By shifting patients to other subunits, they managed to improve the performance of their unit, with no impact on the hospital’s global performance. The result of cardiologists perverting the goal of the casemix was an intentional MED. To summarize, although casemix was supposed to be implemented by both cardiologists and surgeons at the clinical level, the antecedents and effects of their responses have little in common. Within the surgery subunit, surgeons were aware that their central role in the hospital and the region gave them power over the RHA. To maintain activity in the hospital, the RHA needed them and therefore could not take any decision that might upset them. Therefore, the surgeons actively resisted casemix in an attempt to preserve their professional autonomy and power, despite the fact that casemix could work in their favour to some extent. Conversely, the cardiology subunit adopted casemix as a funding necessity and implemented it as such, with minimal disruption to their existing clinical practices. The Regional Health Authority The RHA was initially in charge of enforcing the new casemix policy at the clinical levels and funding hospitals based on their DRG performance details. Prior to that time, the government mandated that the RHA provide all hospital-related medically necessary services for the population – with easy access – within their geographic boundaries (i.e., the region).8 Although health funding came from the government, the RHA was responsible for budgets and resource allocation. Therefore, the RHA’s mission was to provide hospital-related healthcare services and to enforce the new policy. According to the initial framing letter and the first casemix figures on the cardiology DRGs, the RHA considered the cardiology department to be a good performer. Therefore, the cardiologists expected support in terms of project funding. Indeed, the hospital general management presented their casemix experience to the RHA during the negotiations (which started in June 2002 and lasted until May 2004, longer than expected) as evidence for successful casemix use at the clinical level; the cardiologists themselves were introduced in the budget committee meeting in December 2002 as the ‘best students’ in terms of casemix: We have had our first successful experience in Hospital-F of how casemix is used internally by the clinical departments to manage their activity. The cardiology department was the first to do this. Our experience shows how inefficiencies can be revealed, efficiency increased and at the same time the quality of processes improved. [The senior cardiologist] will now present the results of this experience. (CFO) Kern et al. 555 During the RHA negotiations, DRG 184 costs decreased. Further, overall cardiology costs went down by 21.6% as well (Contract of objectives and means, p. 36). Nevertheless, the chronic heart failure project disappeared from the ‘priority axis’ for funding, and the cardiology department received no means to acquire it (Contract of objectives and means, p. 46 and p. 65), much to their surprise: Of course this was a surprise to us. We are the only department using casemix internally. We improved our efficiency. Chronic heart failure moved from being a loss-maker to a profit-making activity, and then the project gets deleted from the list of projects to be funded. In some ways, it just shows that casemix changes nothing with regards to how activities are funded. (Cardiologist 4) The cardiologist expressed his frustration that although rules had been established before the budget negotiations started, they appeared to have changed during the process A few months into the RHA budget negotiations, the project to modernize the existing surgical infrastructure became part of the ‘priority axis of development’, even though it had not initially been included (framing letter) and despite the fact that the surgeons did not implement casemix at their clinical level. This decision was justified using the surgeons’ previous arguments that their current infrastructure was too old and needed to be renewed, and that if Hospital-F continued to offer surgery, then investment in the surgical infrastructure was inevitable. Also, the surgeons promised to introduce some changes in their clinical practices in the future. This decision was contradictory to the casemix logic, in which funding was not supposed to be delivered on the basis of the politics or promises but on actual implementation, efforts made, and results achieved. Yet in this specific case, the RHA decided to provide funding based on the surgeons’ promises, which, if kept, would be consistent with the casemix logic, as the following dialogue shows: Surgery will become a competitive unit to offer high-class service for patients in the region. The most important measures are that we introduce an outpatient section for certain surgical procedures and modernize the operating theatre. With a higher number of patients, we will be able to become more cost-efficient. RHA representative: I am pleased that we have come to such a good agreement that safeguards the operations of this hospital and the surgical service in the future. (Field notes taken during the COM Meeting 02.07.2003) Surgeon: The contrast in how the RHA evaluates surgery versus cardiology is striking when comparing the discussions between the RHA representatives and the members of those two units. When talking to the surgeons, the RHA proved ready to accept promises, pay attention to the idiosyncrasies of this activity, and eventually provide financial support, as the following discussion shows: Surgeon: RHA representative: Physician from intensive care unit: I think it is difficult to judge costs and activity in this way. I mean, you really need to look at the complexity of the cases. Sure, we are aware that there are medical specialities that are more costly than others, and we take that into account. The intesive care unit is a key technical platform in the hospital. Of course we are more costly than others, so what ? 556 RHA representative: Surgeon: RHA representative: Organization Studies 39(4) Yes, we know that the hospital needs these key functions, but nevertheless there is pressure to make some effort to cut costs. We agree to introduce outpatient surgery for some procedures; this can make us more competitive, also in terms of costs. But we really must renovate our operating theatres. If we want to be competitive, we need state-of-the-art surgery. This is what we also think. The hospital needs a high-quality surgery and intensive care unit. This is key for the survival and development of this hospital. (Field notes taken during the COM Meeting 02.07.2003) In sharp contrast, when talking to the cardiologists, the RHA acknowledged that the unit had been exemplary in implementing casemix, but refused to make investment decisions based on this as was initially planned. The following dialogue shows this: Cardiologist: RHA representative: Cardiologist: RHA representative: We are the most efficient service in the hospital. We are not only the only service that has introduced key figures and PMSI management at service-line level, we also used these tools to improve our service. So we are a very efficient unit compared to other cardiology units. This is very good that you are using the PMSI; other units in the hospital should take this as an example. What we are aiming for is creating a new unit for chronic heart failure patients, the first unit of this kind in France. So our objective is to obtain the financial means to create this unit. Let’s analyse the figures, and in times of cost pressure, everybody has to make efforts. Let’s discuss it further in the next meeting. But it is important that you are focused on saving in order to fund the new unit. This could be a possibility in such hard times. (Field notes taken during the COM Meeting 02.07.2003) Eventually, the RHA suggested that the cardiologists should self-fund the unit, which was against the initial casemix rules set by the RHA. In March 2003, during a meeting (COM Meeting 20.2.2003), the specialized unit for chronic heart failure was still among the ‘star’ projects, whereas in the July meeting, this project had been deleted from the list of projects to be funded. This was a surprise to the cardiologists as there had been no official letter informing them and no direct discussion with them. Despite decoupling its funding practice from the policy, the RHA nevertheless refers to the policy even though this policy is not respected by the surgeons. Just as the surgeons knew that they could resist the RHA, the RHA realized its own dependence on the surgeons; the RHA’s mission was to provide hospital healthcare at the regional level, and surgery was the core of this mission and the hospital’s raison d’être. The power imbalance was in favour of the ‘regulated’ and not the ‘regulator’. While referring to the policy when funding surgery might have been intended to keep up appearances, it did not fool the cardiologists. As the RHA’s decision to fund the surgical infrastructure was inconsistent with the casemix data, the cardiologists were astonished and frustrated to learn that the surgical infrastructure was to be updated while their own project remained unfunded, as noted here: They don’t take DRG results into account in allocating the means … This questions the management in this hospital. Everything we’ve done was for us, so that we know better where we stand in terms of Kern et al. 557 performance, but in terms of the budget negotiation, it was not useful … The outcome would have been the same if we had not done anything. This is complete madness. (Cardiologist 3) While this RHA decision was not consistent with the new policy of detailed casemix performance funding, it was consistent with the need to maintain a hospital in the region, with surgery being the very core of that hospital. However, the decision to fund the surgery project instead of the cardiology project clearly conflicted with the respective performances of the two departments, as reflected in the casemix. The budget and investment were not, therefore, based on casemix criteria but on other factors. Discussion In this paper, we have sought to better understand why subunits differ in their adoption and implementation of policies and the related decoupling practices in a context of institutional complexity. We did so through a close examination of different forms of decoupling around the adoption and implementation of a new policy (i.e., the casemix) in a hospital. We focused on the importance of power and politics in explaining the reaction of different organizational actors. The findings provide helpful insights to elaborating a political approach to decoupling. They also provide insights on the decoupling of subunits, means-end decoupling and the role of institutional complexity in decoupling. Furthermore, the study has potential implications for institutional research, particularly regarding the role of power in institutional processes and the ways in which actors engage with institutional logics. We highlight the implications for those research streams, discuss our research limits and offer suggestions for future research. A political approach to decoupling Our study underlines the centrality of power and dependence relations between organizational actors, for a better understanding of decoupling. This finding led us to further theorize the role of power using PDT (Emerson, 1962). Our findings emphasize three behaviour types regarding decoupling; when analysed through the lens of PDT, these highlight three types of power dependence relations. Surgeons decided to openly reject casemix, although it might have proved beneficial for them. While this system was designed to favour surgery, adopting it would have meant accepting external supervision by managers. Refusing it was a way for the surgeons to affirm their power and their capacity to resist external norms. The surgeons resisted it to maintain their autonomy and their favourable power balance toward the RHA, even if surgeons and the RHA were mutually dependent. The surgeons knew that the RHA had no choice but to fund them to keep the hospital operating. In the local context, it was crucial for the RHA to maintain this hospital, which would not be possible if the surgeons decided to leave and engage in private practice. The surgeons were well aware that there would be no consequences, as the RHA could not exert any constraints. The cardiologists were in a very different situation regarding mutual dependence and the power imbalance. They depended on the RHA to obtain funding for their chronic heart failure project. The RHA depended less on them. Consequently, the cardiologists suffered from a power imbalance and had little room to bargain with the RHA or to potentially resist casemix. They used decoupling as a balancing operation to reduce the existing asymmetry of power in their disfavour. They symbolically implemented casemix, yet engaged in MED. The cardiologists did not follow the ‘spirit’ of the policy (Wijen, 2014), i.e., its ends, which was to reduce 558 Organization Studies 39(4) Table 2. Power dependence and forms of decoupling. Surgeons Cardiologists RHA Mutual dependence Power imbalance A depends on B Differences in actors’ dependencies •• Depend on RHA •• RHA depends on surgeons •• Depend on RHA •• Power imbalance in their favour versus RHA •• Power imbalance in their disfavour versus RHA •• Depends on surgeons •• Surgeons and cardiologists depend on RHA •• Power imbalance in their favour versus cardiologists •• Power imbalance in their disfavour versus surgeons Balancing operations •• Reject casemix to increase this power imbalance •• Try to draw upon casemix to obtain this funding. •• Symbolic implementation to avoid scrutiny and reduce power imbalance •• Acknowledge power imbalance regarding surgeons when not considering casemix for funding Form of decoupling No decoupling. Openly reject casemix Means-Ends decoupling Policy-Practice decoupling overall costs at the hospital level. They limited their compliance to the ‘letter’ of the policy (Wijen, 2014), i.e., its means, by implementing casemix and monitoring its results. They pursued their own ends: obtaining funding for the chronic heart failure unit. Good results were achieved by diverting some patients and the related costs to other services to improve their DRG results without dramatically changing their work practices. Thus, at the overall hospital level, these results did not contribute to reducing the overall costs. The RHA depended on the surgeons’ activities to maintain the hospital, but not on the cardiologists whose activities could be externalized without threatening the hospital existence. Consequently, the RHA had no choice but to eventually decouple its practice from its policy, renouncing hospital funding based on detailed casemix performance at the clinical level. Instead, they decided to support the surgeons and not to fund the new heart failure unit. Overall, these findings suggest that decoupling can be understood as being related to dependence and power relations among actors. Decisions to decouple or not (and the form of decoupling) depend on the impact of the decoupling on power imbalances and on the actors’ own interests. The RHA could implement PPD because of its position. Indeed, as the authority, no other actor could sanction the RHA for such a decision. Meanwhile, the cardiologists needed to hide their decoupling, as this could have led to immediate sanctions if discovered. Overall, this study suggests that power relations are an essential component of the reactions of a subunit regarding a new policy, be they decisions to decouple or to simply ignore the policy. Incidentally, we also contribute by bringing Emerson’s PDT (Emerson, 1962) back into organizational research. In line with Casciaro and Piskorski (2005), who used PDT to reformulate the resource dependence model, we suggest that PDT can contribute to improving our understanding of the different conditions under which decoupling occurs by accounting for the importance of power relations. Kern et al. 559 Contributions to research on multiple forms of decoupling This research expands the decoupling literature along three dimensions. Firstly, regarding the level of analysis, recent research highlights two main needs: to go beyond the environmental (Meyer & Rowan, 1977) and organizational levels (Westphal & Zajac, 2001) and to examine different reactions within the organizations (e.g., Pache & Santos, 2010; Sandholtz, 2012). While our research enriches this area by examining decoupling at the intra-organizational level, it also suggests that decoupling should be studied across levels, revising our assumptions regarding the relations between extra-organizational actors and subunits. Existing research assumes that extra-organizational actors – such as regulators (Ang & Cummings, 1997), standard-setting organizations (Boiral, 2007; Sandholtz, 2012), financial analysts (Rao & Sivakumar, 1999), etc. – exert power and influence on organizations, presuming a top-down process, where organizations either comply or decouple. The approach we take here, suggests that a central element of decoupling is the power relations between actors, rather than the level at which they are situated. Our study shows that sometimes the power imbalance between the organizational subunits and extra-organizational actors is in the subunits’ favour. It also shows that when the generally assumed power relation between a regulatory agency and practitioners is inverted, it can lead the agency to decouple from the rule that it was supposed to implement. Here, the RHA is an extra-organizational funding agency that has no real power over surgeons and that cannot force them to comply with the policy; furthermore, it is well aware that the surgeons are openly refusing to comply. Ultimately, the surgeons drove the RHA to deviate from the policy it was meant to implement because the RHA depended on them. While existing research rather assumes that subunits depend on those funding organizations based in their environment (e.g., Binder, 2007; Meyer & Rowan, 1977), our study shows that funding organizations can be dependent on the subunits as well. Our results underline that the RHA, an extra-organizational funding agency, decoupled its own policy from its practices because of surgeons’ relative power. Secondly, regarding emerging research on MED, our study suggests that this form of decoupling can be used strategically by actors with limited power in attempting to achieve their goals while providing evidence that they comply with the policy. The MED by the cardiologists exemplifies such a strategy. As the cardiologists had limited power with the RHA and were interested in gaining its financial support to develop a chronic heart failure unit, they implemented casemix and showed good results. However, they achieved those results by redefining the parameters of their activities and excluding costly cases with multiple pathologies from their subunit DRGs, leaving other subunits to deal with those cases and subsume the associated costs. While recent research on MED has focused mainly on unintentional MED (e.g., Bromley et al., 2012; Wijen, 2014), our study documents intentional MED, suggesting that it can be a way for actors with limited power to decouple in a less spectacular way than PPD. Our findings contradict previous results (Covaleski, Dirsmith, & Michelman, 1993) on the inability of weaker hospital subunits, in our case cardiologists, to circumvent the DRG system and casemix surveillance practices to some extent to achieve their own ends. These findings also complement previous literature by highlighting that the structuring impact of policy (Carruthers, 1995) is not as intended, because the cardiologists’ core clinical practices actually remained unchanged. Thirdly, this research highlights how institutional complexity generates opportunities for decoupling strategies. Research has documented how institutional complexity is likely to provoke decoupling (Bromley et al., 2012) due to an inability to trace the consequences of policy in such complex settings. Our study extends this research by observing that institutional complexity can be used by actors to strategically decouple. Particularly, it underlines the strategic use of MED in a context of institutional complexity. Sauder and Espeland (2009, p. 77) label as ‘gaming’ this type of strategic 560 Organization Studies 39(4) decoupling ‘not motivated by concerns over legitimacy [but] to protect one’s … from the penalties of a poor ranking and, reassuringly, to do something in the face of great uncertainty’. Our study further suggests that in such complex settings, in order to survive and preserve some autonomy, actors tend to take a manipulative stance toward those very tools that they are supposed to comply with. This creates situations where these rationalized managerial tools eventually undercut the accountability and effectiveness they purport to target. Implications for institutional research More generally, we contribute to institutional research by further investigating the role of power in institutional processes and how actors engage with institutional logics. We contribute by bringing power back into institutional theory. Power relations were initially an important part of institutionalism (e.g., Gouldner, 1954; Selznick, 1949), but received less attention within new institutionalism, as authors were focusing more on collective isomorphism and less on agency. While authors have since worked to bring power back into institutional theory (e.g., Lawrence, Suddaby, & Leca, 2009), critiques still highlight that this integration remains marginal (Munir, 2015). The case study suggests that power relations are central in decisions to reject new norms and standards or to decouple from them. The case study emphasizes the complexity of power relations and reminds us that power is not always held by the hierarchy or external regulatory bodies. It shows that such power relations must be considered without prior assumptions regarding which actors are presumably powerful. Indeed, examining cases of decoupling might prove very efficient for unveiling power dependence relations among actors within and between organizations. Critics of institutional research also insist on the lack of works treating institutionalized norms as ‘media of domination’ (Willmott, 2015, p. 105). Ironically, while researchers might miss this dimension, our study suggests that actors are aware of it, and they organize resistance and decoupling as a strategic means to avoid submitting to such domination. Another interesting aspect emerging from this study regards how actors engage with institutional logics. The findings suggest a difference between how actors use institutional logics to justify their position and how they enact those logics. On the one hand, logics can be used publicly to justify action. For instance, the surgeons used the professional logic of broader medical care to justify their refusal to implement casemix. They insisted that only patient interests should prevail and that any other logic was irrelevant. Used in such a rhetorical manner, logics can be a powerful justification because it pre-exists in the meaning system negotiation where it is eventually used to bargain with. On the other hand, logics are enacted by actors. In the case of surgeons, we found no evidence, other than rhetorical, that patient interest was the surgeons’ main concern. Surgeons justified their reaction by putting up front the component of their professional logic related to patient interest, yet they appear to have been motivated by the component related to the importance of their status in terms of the RHA and of other physicians. Incidentally, this analysis also contributes to refining the examination of evolving logics within healthcare. When referring to institutional logics, research on healthcare management often emphasizes radical change due to the increasing prevalence of financial and bureaucratic logics prevailing over medical logics (e.g., Reay & Hinings, 2005; Reay & Hinings, 2009). We provide an example in which powerful actors – i.e., surgeons – resist this change and eventually overcome it. While research has come to consider such change as inevitable, we suggest that this might not always be the case, as extra-organizational actors (in charge of implementing this change) might not always be the more powerful. Eventually, this suggests a less deterministic institutional process, the outcome of which might be more uncertain than generally considered. Kern et al. 561 Limitations and Further Research Our study has several limitations. The first one is the uniqueness of our case study. Other case studies examining other power relations and institutional contexts (e.g., contexts where surgeons have no access to private practice) would certainly provide alternative results. Despite the limited generality of our empirical data, their richness allows an in-depth examination of the different mechanisms present, particularly the power relations, thus confronting discourses with actions. Further, we theoretically analyse decoupling while focusing on power relations. Alternative approaches, such as focusing on how actors inhabit institutions (Binder, 2007; Sandholtz, 2012), might also have been interesting. Focusing on power does not contradict those approaches but rather complements them by insisting that this lost dimension as a crucial aspect to understand decoupling. Indeed, these studies pay little attention to power, often considering it as secondary or manageable through negotiations (e.g., Binder, 2007; Hallett, 2010). Another limitation relates to the specific framework used in analysing power dependence. Resource Dependence Theory (RDT) (Pfeffer & Salancik, 1978) is more popular in organizational research, particularly for examining institutional processes (Oliver, 1991). Nonetheless, we choose to use PDT because it focuses on power relations between actors and on their attempts to influence other actors, rather than accessing symbolic and material resources as RDT does. Moreover, Casciaro and Piskorski (2005) recently suggested that in order to ensure the potential of RDT, it should be reformulated. To do so, they suggested introducing a distinction between mutual dependence and power imbalance (e.g., the two dimensions of PDT as introduced by Emerson (1962)). Incidentally, our paper contributes to further explorations of these two dimensions. Finally, the paper raises several directions for future research. Firstly, it suggests revisiting those behaviours and processes often analysed through institutional theory, such as decoupling, through a power-focused lens. Following other authors, we suggest that power is an essential dimension of institutional processes (Lawrence, 2008), although it is arguably a lost one (Munir, 2015). We contribute to bring back this aspect as central in institutional analysis. When we advocate for the potential of PDT in the analysis of institutional mechanisms, we more broadly advocate an interest in using multiple lenses to bring the multiple faces of power into institutional analysis. Secondly, this work insists that what actors say about logics might differ significantly from how they enact them. Distinctions between how actors justify their actions and what actually motivates them, or what they actually do, has been underlined in other works (Vaisey, 2009), but it has not yet been explored in works on institutional logics. We argue that this distinction can help to further articulate the notion of logics. A potential important implication for future research is that if researchers want to account for logics, they should not only rely on discourse and language but also need to confront discourse with practice and examine potential contradictions. Acknowledgements The authors contributed equally to this research. We would like to thank Royston Greenwood for having been instrumental in the emergence of this paper. We are indebted Guido Möllering for his fantastic editing which allowed us to make our ideas clearer without renouncing them and to the reviewers for their help and comments. Earlier drafts of this research have been presented at the University of Amsterdam, the Ecole des Mines, and Laval University. We are indebted toward all the participants to these events for the generosity of their comments in particular toward Patricia Bromley, Yves Gendron, Mark Dirsmith, Marion Brivot, Bertrand Malsch, Aurélien Acquier, Eva Boxenbaum, Eero Vaara, Brendan O’Dwyer, Niamh O’Sullivan and Zamzulaila Zakaria. The responsibility for any error in the text in solely the authors’ own. 562 Organization Studies 39(4) Funding This work has been supported by a grant from HaCIRIC, an EPSRC sponsored research Centre at Imperial College London, a Marie Curie Reintegration grant (Grant agreement PERG06-GA- 2009-2565620), and a grant from the French Accounting Association (AFC). Notes 1. 2. 3. 4. 5. 6. 7. Originally termed Homogenous Patient Groups in French (Michelot & Rodrigues, 2008). The Regional Health Authority is a local authority for healthcare that is responsible for allocating resources to hospitals. Details available upon request to the authors. The hospital and the RHA agreed on the hospital’s future strategy in the Project of the Hospital (originally called in French ‘projet d’etablissement’). This document follows article L6143.2 of the public health code (code de la santé publique) and order no. 96-346 of 24 April 1996. It was signed on 27 December 2001. Medical controllers in France are physicians who have chosen to specialize in medical controlling. Patient data privacy is ensured by granting only the medical controller access to DRG data. Hospital managers must ask the medical controller’s permission for access. Tariff for 2004; the tariff is the amount established by the casemix as a regional average that constitutes the budgeted price given by the RHA to the hospital. After the study, RHA responsibilities grew, as they had to provide all medically necessary services (of all kinds and not limited to hospitals). References Ang, S., & Cummings, L. L. (1997). Strategic response to institutional influences on information systems outsourcing. Organization Science, 8, 235–256. Atkinson, G., Dietz, S., & Neumayer, E. (2007). Introduction: Handbook of sustainable development. In G. Atkinson, S. Dietz, & E. Neumayer (Eds.), Handbook of sustainable development (pp. 1–27). Northampton, UK: Edward Elgar Publishing House. Binder, A. (2007). For love and money: Organizations’ creative responses to multiple environmental logics. Theory and Society, 36, 547–571. Boiral, O. (2007). Corporate greening through ISO 14001: A rational myth? Organization Science, 18, 127–146. Bromley, P., Hwang, H., & Powell, W. W. (2012). Decoupling revisited: Common pressures, divergent strategies in the U.S. nonprofit sector. M@N@Gement, 15, 468–501. Bromley, P., & Powell, W. W. (2012). From smoke and mirrors to walking the talk: Decoupling in the contemporary world. The Academy of Management Annals, 6, 483–530. Carruthers, B. G. (1995). Accounting, ambiguity, and the new institutionalism. Accounting, Organizations and Society, 20, 313–328. Casciaro, T., & Piskorski, M. J. (2005). Power imbalance, mutual dependence, and constraint absorption: A closer look at resource dependence theory. Administrative Science Quarterly, 50, 167–199. Clegg, S. (2010). The state, power, and agency: missing in action in institutional theory? Journal of Management Inquiry, 19, 4–13. Covaleski, M. A., Dirsmith, M. W., & Michelman, J. E. (1993). An institutional theory perspective on the DRG framework, case-mix accounting systems and health-care organizations. Accounting, Organizations and Society, 18, 65–80. Dimaggio, P. J., & Powell, W. W. (1991). Introduction. In W. W. Powell & P. J. Dimaggio (Eds.), The new institutionalism in organization analysis. Chicago, IL: University of Chicago Press. Eisenhardt, K. M. (1989). Building theories from case study research. Academy of Management Review, 14, 532–550. Emerson, R. M. (1962). Power-dependence relations. American Sociological Review, 27, 31–40. Gouldner, A. W. (1954). Patterns of industrial bureaucracy. Glencoe, IL: Free Press. Kern et al. 563 Greenwood, R., Raynard, M., Kodeih, F., Micelotta, E. R., & Lounsbury, M. (2011). Institutional complexity and organizational responses. The Academy of Management Annals, 5, 317–371. Hallett, T. (2010). The myth incarnate: Recoupling processes, turmoil, and inhabited institutions in an urban elementary School. American Sociological Review, 75, 52–74. Jorgensen, D. L. (1989). Participant observation. Newbury Park, CA: SAGE Publications. Kimberly, J. R., De Pouvourville, G., & D’Aunno, T. A. (2008). The globalization of managerial innovation in health care. Cambridge, MA: Cambridge University Press. Lawrence, T. (2008). Power, institutions and organizations. In R. Greenwood, C. Oliver, K. Sahlin, & R. Suddaby (Eds.), SAGE handbook of organizational institutionalism (pp. 170–197). London, UK: SAGE Publications. Lawrence, T., Suddaby, R., & Leca, B. (2009). Institutional work: Actors and agency in institutional studies of organizations. Cambridge, UK: Cambridge University Press. Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340–363. Michelot, X., & Rodrigues, J. M. (2008). DRGs in France. In J. R. Kimberly, G. de Pouvourville, & T. D’Aunno (Eds.), The globalization of managerial innovation in health care (pp. 116–143). Cambridge, MA: Cambridge University Press. Moll, J., & Hoque, Z. (2011). Budgeting for legitimacy: The case of an Australian university. Accounting, Organizations and Society, 36, 86–101. Munir, K. A. (2015). A loss of power in institutional theory. Journal of Management Inquiry, 24, 90–92. Oliver, C. (1991). Strategic responses to institutional processes. Academy of Management Review, 16, 145–179. Pache, A., & Santos, F. (2010). When worlds collide: The internal dynamics of organizational responses to conflicting institutional demands. Academy of Management Review, 35, 455–476. Pache, A.-C., & Santos, F. (2013). Inside the hybrid organization: Selective coupling as a response to competing institutional logics. Academy of Management Journal, 56, 972–1001. Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations: A resource dependence perspective. New York, NY: Harper & Row. Rao, H., & Sivakumar, K. (1999). Institutional sources of boundary-spanning structures: The establishment of investor relations departments in the fortune 500 industrials. Organization Science, 10, 27–42. Reay, T., & Hinings, C. R. (2005). The recomposition of an organizational field: Health care in Alberta. Organization Studies, 26, 351–384. Reay, T., & Hinings, C. R. (2009). Managing the rivalry of competing institutional logics. Organization Studies, 30, 629–652. Sandholtz, K. W. (2012). Making standards stick: A theory of coupled vs. decoupled compliance. Organization Studies, 33, 655–679. Sauder, M., & Espeland, W. N. (2009). The discipline of rankings: Tight coupling and organizational change. American Sociological Review, 74, 63–82. Selznick, P. (1949). TVA and the grass roots. Berkeley, CA: University of California Press. Spradley, J. P. (1990). Participant observation. New York, NY: Holt, Rinehart &Winston. Tilcsik, A. (2010). From ritual to reality: Demography, ideology, and decoupling in a post-communist government agency. Academy of Management Journal, 53, 1474–1498. Timmermans, S., & Tavory, I. (2012). Theory construction in qualitative research: From grounded theory to abductive analysis. Sociological Theory, 30, 167–186. Vaisey, S. (2009). Motivation and justification: A dual-process model of culture in action. American Journal of Sociology, 114, 1675–1715. Westphal, J. D., & Zajac, E. J. (2001). Decoupling policy from practice: The case of stock repurchase programs. Administrative Science Quarterly, 46, 202–228. Wijen, F. (2014). Means versus ends in opaque institutional fields: Trading off compliance and achievement in sustainability standard adoption. Academy of Management Review, 39, 302–323. Willmott, H. (2015). Why institutional theory cannot be critical. Journal of Management Inquiry, 24, 105–111. Yin, R. K. (2009). Case study research. design and methods, 4th ed. Thousand Oaks, CA: SAGE Publications. 564 Organization Studies 39(4) Author biographies Anja Kern worked several years as a management accountant in the healthcare sector. She is currently Professor at Baden Wurttemberg Cooperative State University of Mosbach (Germany). Her research focuses on management accounting and cost systems in the healthcare sector. Aziza Laguecir is Associate Professor in Management Accounting and Control at IESEG School of Management (France). Her research adopts sociological approaches of management accounting, and explores the interplay between management accounting and organizational practices. Bernard Leca is Professor in Management Accounting and Control at ESSEC Business School (France). His main research focuses on institutional theory and the way organizations or individuals can initiate and implement institutional change.