A process (re)turn?

Management Accounting Research 18 (2007) 453–475

A process (re)turn? Path dependencies, institutions and performance management in Swedish central government Sven Modell a,∗ , Kerry Jacobs b , Fredrika Wiesel c a

University of Manchester, United Kingdom The Australian National University, Australia c Stockholm University, School of Business, Sweden b

Abstract Recent central government reforms have signalled a transition away from output-based governance and control to a more citizen-orientated and outcome-focused performance management ethos. Prior research suggests that this may give rise to institutional inconsistencies related to conflicting performance management logics. In this paper, we argue that rather than resulting in growing pre-occupation with outcomes and effectiveness, the emerging reform agenda may somewhat paradoxically reinforce managerial concerns with the operating processes underpinning public service delivery. Drawing on policy studies and management accounting research informed by new institutional sociology, we theorise these developments by invoking the notion of path dependency. We provide an empirical illustration based on field work in Swedish central government and adopt a multi-level approach in examining the evolution of performance management practices. While evidence of strongly constraining path dependencies is found at the overall policy level a case study in the Swedish Tax Agency shows that these may also embody an element of flexibility such that the meaning of inconsistent performance management logics is re-constructed over time. This illustrates how individual government agencies may link outcomes to operating process concerns and manage implementation problems related to conflicting and uncoordinated reform initiatives. © 2007 Elsevier Ltd. All rights reserved. Keywords: Central government; Institutions; Path dependency; Performance management; Process orientation; Public sector reform

∗

Corresponding author. E-mail addresses: [email protected] (S. Modell), [email protected] (K. Jacobs), [email protected] (F. Wiesel).

1044-5005/$ – see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.mar.2006.12.001

454

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

1. Introduction New public management (NPM) reforms in central government over the past 20 years have been characterised by increasing emphasis on devolved performance management, notably epitomised by a transition from heavy reliance on centralised planning to ‘management by objectives’ or ‘managing for results’ as the dominant mode of governance and control. Important legislative and regulatory changes to this end include, but are by no means confined to the Next Steps Initiative in the UK (Carter and Greer, 1993; Hyndman and Eden, 2000; Likierman, 1994), the State Sector Act/Public Finance Act in New Zealand (Jacobs, 1997; McCulloch and Ball, 1992; Pallot, 1998) and the Government Performance and Results Act/National Performance Review in the U.S. (Cavalluzzo and Ittner, 2004; Kravchuk and Schack, 1996; Radin, 1998; Watkins and Arrington, 2007). These reforms paralleled the efforts to clarify the boundaries between policy-making, service delivery and the executive role of central government. An important means to this end was to specify and hold central government agencies accountable for outputs rather than less composite measures of inputs and operating processes to further devolution (Bromwich and Lapsley, 1997; Carter and Greer, 1993; McCulloch and Ball, 1992; Talbot, 2004). In some cases, clear specification of output indicators was also seen as a prerequisite for government to enter into more market-like forms of contracting with public and private agencies (Mascarenhas, 1996; McCulloch and Ball, 1992). Recently, this output-focused mode of performance management has been subject to increasing criticism and seems to be under reconsideration in the very countries which had previously led the way towards such practices (Modell, 2005a; Osborne, 2006). In response to the fragmentation of public service delivery and myopic tendencies calls have been made for a new performance management ethos, pivoting around the outcomes of service delivery as manifested by the more composite and often long-term effects on users, citizens or society as a whole (see e.g., Catas´us and Gr¨onlund, 2005; Neale and Pallot, 2001; Norman and Gregory, 2003). Recent policy changes seem to signify increasing concerns with the effectiveness of public service delivery and the need to visualise the value delivered to the citizenry (McGuire, 2002; Norman and Gregory, 2003; Osborne, 2006). In the present paper, we take a critical view of the propensity for such an outcome- and effectivenessfocused performance management ethos to permeate central government operations. It is argued that rather than resulting in growing pre-occupation with outcomes and effectiveness, the emerging reform agenda may represent a process turn; a return to the managerial concerns with the operating processes underpinning public service delivery. While such tendencies may seem reminiscent of a reactionary return to pre-NPM practices this is overly simplistic. We illustrate that the process focus may take on new meanings through an engagement with wider change processes, especially those related to the influence of the ‘modern’ quality movement in central government. Drawing on recent experiences from Swedish central government, we theorise these developments by invoking the notion of path dependency. Path dependencies are related to historical, political and institutional factors that constrain and mediate the choice and implementation of novel control practices (North, 1990) and have been increasingly used as an analytical lens in policy studies (see e.g., Gains, 2004; Greener, 2005; Hay and Wincott, 1998; Kay, 2005; Pierson, 2000a,b). However, it has been less emphasised in research on management accounting change. While the notion of path dependency was used in Burns and Scapens’s (2000) institutional analysis of management accounting change their references were rather tangential and they did not detail the conceptual foundations and analytical ramifications of

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

455

this concept.1 In the present paper, we explicate how institutional research on performance management in the public sector may be developed by paying more explicit attention to path dependencies. In the following section, we outline how the notion of path dependency may contribute to institutional research on performance management and develop a theoretical framework. We then account for the research design before offering an empirical illustration based on recent developments in Swedish central government. This illustration covers both general developments at the overall policy level and more specific changes mirroring these developments within the Swedish Tax Agency. A discussion of our main findings and implications for further research closes the paper.

2. Theoretical framework Institutional theorists have long grappled with the issue of how institutionalised structures and agency exercised by actors with vested interests influence each other (e.g., Barley and Tolbert, 1997; Covaleski and Dirsmith, 1988; DiMaggio, 1988; Greenwood and Hinings, 1996; Hirsch and Lounsbury, 1997). In contrast to the deterministic predictions concerning the role of institutions in early advances in new institutional sociology (NIS) (DiMaggio and Powell, 1983; Meyer and Rowan, 1977), this research has drawn attention to the intricate interplay between stabilising and dynamising forces in the process of organisational change. Instead of portraying organisational members as over-socialised actors, heavily constrained by rules and norms embedded in legislation, professionalisation and societal expectations, there is now widespread agreement that choices and agency matter. Under certain circumstances actors may influence the process of institutionalisation although they are rarely free to completely evade existing institutional constraints. Until recently, however, NIS has been largely silent about the more specific mechanisms of choice and under what circumstances individual actors, or change agents may influence institutional change (see Beckert, 1999; Dorado, 2005; Hensmans, 2003; Seo and Creed, 2002). Seo and Creed (2002, p. 224) criticised earlier institutional research for failing to satisfactorily answer the question of: ‘. . . when and how do embedded actors individually and collectively come to that conscious choice point—presumably, a point where they recognize the need, the opportunities, and the appropriate courses for collective action for changing existing institutional arrangements?’ Similar criticisms may be levied at research on management accounting change informed by institutional theories. A few recent studies have explored how change agents try to affect shifts in collective consciousness and how this impinges on the process of institutionalisation (e.g., Burns, 2000; Burns and Baldvinsdottir, 2005; Modell, 2005b; Siti-Nabiha and Scapens, 2005). Yet, Modell (2007) observed that little empirical research has addressed the issue of how discrete actions aimed at the transformation of institutionalised structures are mediated by such structures. Hence, even though the tensions and dynamics inherent in the interplay between institutionalised structures and agency have attracted considerable interest over the past decade the theory of choice embedded in this interplay would seem underdeveloped. 1 Burns and Scapens (2000) rather extended their discussion into an analysis of the interplay between rules and routines in the institutionalisation of accounting. As explicated later (see under Section 2 below) this detracts from some basic assumptions associated with the notion of path dependency.

456

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

The notion of path dependency provides an analytical lens enabling researchers to address this issue without falling victim to some na¨ıve historical determinism (Goldstone, 1998). Rather than conceptualising the constraining role of institutions as a set of relatively fixed, habitualised or taken-for-granted action patterns, it directs our attention to the process of institutionalisation as a series of interdependent choices exercised at specific, critical junctures (Goldstone, 1998; Hall and Taylor, 1996; Hay and Wincott, 1998). Its basic premise is that past choices have an enduring influence on and continue to shape emerging change initiatives by narrowing the range of viable alternatives. Hence, certain trajectories are precluded while others are reinforced (Hay and Wincott, 1998; Kay, 2005; Pierson, 2000a). Path dependencies thus reinforce the evolutionary nature of change (Burns and Scapens, 2000; Greener, 2002). Understanding the process of institutionalisation and change in accordance with the notion of path dependency requires some recognition of what plausible alternatives were available to the ones that gradually come to exert a dominant influence when institutionalised structures become subject to critique or re-consideration (Goldstone, 1998; Greener, 2005; Kay, 2005). Therefore, it is an over-simplification to conclude, as Burns and Scapens (2000, p. 12) do, that institutional change is path-dependent merely because ‘existing routines and institutions will shape the selection and implementation process.’ Analyses of path dependencies should be extended to examine what competing alternatives to existing institutional arrangements were available and why these were rejected at specific points in time. Such analyses complement the general view of institutionalisation as a ‘cumulative process’ (Burns and Scapens, 2000, p. 7) by directing attention to the relatively rare events where more visible agency is exercised and competing interests become manifest and funnelled into deliberate attempts to change institutionalised structures. The self-reinforcing property of path dependencies may perpetuate inefficiencies and unintended outcomes (Hall and Taylor, 1996; Pierson, 2000a,b). The choice processes associated with path-dependent changes are rarely guided by unbounded rationality but may rather ‘lock in’ actors in suboptimal arrangements (Burns and Scapens, 2000; Greener, 2005; Kay, 2005). Despite such inefficiencies path dependencies often exhibit increasing returns as the relative costs of alternative trajectories in terms of potential loss of legitimacy or other resources are even higher. Such constraints increase the attractiveness of existing institutional arrangements and are particularly pertinent where the outcomes of alternative trajectories are contested or uncertain (Pierson, 2000a). The likelihood of path dependencies also hinges on the extent to which alternative trajectories are compatible with existing cultural norms and institutionalised arrangements in a particular social setting (Greener, 2005). Greener (2005) argued that the propensity for existing structures to persist is highest when emergent properties associated with alternatives are complementary to such structures. Under such conditions, emergent properties can be incorporated into existing structures without generating much conflict between actors with vested interests in various alternatives. Conversely, the likelihood of pathdependent change is lowest when alternatives represent incompatible courses of action as this fuels conflict and forces some choice that may imply rejection or radical re-consideration of existing structures. Such insights complement recent advances in NIS tracing the origins of institutional change to incompatible or inconsistent institutional arrangements (Abrahamsson and Gerdin, 2006; Burns and Baldvinsdottir, 2005; Clemens and Cook, 1999; Seo and Creed, 2002). Whereas this research has also emphasised inefficiencies and conflicting interests as ‘triggers’ of change processes, it has not addressed the issue of path dependency explicitly. Recent central government reforms underscore the relevance of using the notion of path dependency to enhance our understanding of performance management. Since the 1980s, such reforms have evolved in an institutional setting historically dominated by bureaucratic values stressing due process and procedures

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

457

and an emphasis on detailed regulation of operating processes as a complement to input-focused budgetary control (Hood, 1991, 1995; Lynn, 2001; Svara, 2001). Despite the international trend towards deregulation, devolution and increasing reliance on output-centred performance management, rules and regulations still exercise an important influence on central government operations (e.g., Graham and Roberts, 2004; Moynihan, 2006; Talbot, 2004). This has resulted in a rather paradoxical situation where inconsistent logics of control operate simultaneously (cf. Hood and Peters, 2004). Still, this has not always generated disruptive tensions and the process of change has in many cases been evolutionary rather than revolutionary (see Pollitt and Talbot, 2004). One plausible explanation for this was offered by Gains (2004), who observed that the implementation of the Next Steps Initiative in the UK was surrounded by a considerable element of negotiation and compromise such that existing structural arrangement had a path-dependent influence on emerging control practices. Inconsistencies such as those outlined above may be amplified by the recent tendency to re-orient performance management towards a heavier emphasis on outcome and effectiveness indicators. However, there is some confusion as to how this might impinge on control in central government. For example, observers of U.S. federal government reforms suggest that process concerns have been compromised in the overall quest for accountability for outcomes (Heinrich, 2002; Radin, 1998). On the other hand, Watkins and Arrington (2007) argue that legislative constraints may reinforce operating process considerations at the expense of the emphasis on outcomes. The former scenario would suggest a tendency towards radical re-consideration of historically dominant control practices while the latter is more in line with the notion of path-dependent change. Conflicting predictions such as those advanced above may need to be nuanced, however, in light of the broader reform movement surrounding growing effectiveness concerns. Recent central government reforms in several countries have been under the dual influence of the striving towards enhanced citizen orientation and the ‘modern’ quality movement, epitomised by such notions as total quality management (TQM) (McGuire, 2002; Modell and Gr¨onlund, 2006). This signifies a return to operating process considerations, but from a more externally focused and customer-orientated perspective than that associated with the ‘bureaucratic’ values historically dominating central government (McGuire, 2002). Thus, Modell (2005a) speculated that potential inconsistencies associated with recent reforms may be reconciled by re-casting citizens as customers and relying on customer satisfaction indicators as proxy outcome measures, closely related to short-term, operating process aspects. This might reinforce the tendencies towards path-dependent change as conflicts related to apparent inconsistencies are reduced (cf. Greener, 2005).

3. Research design The following empirical illustration examines the development of performance management practices across different levels of analysis in an attempt to stimulate multi-level theory building (Klein et al., 1999). Whereas management accounting research has primarily examined path dependencies at the intraorganisational level (Burns and Scapens, 2000), our study responds to recent calls for widening such analyses to examine their interplay with similar phenomena in the broader political and institutional environment (Dillard et al., 2004; Modell, 2003). Hence, our analysis traces path dependencies at the overall policy level in Swedish central government and then explores how these are reflected within a particular government agency.

458

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

Empirical data were drawn from a larger research programme examining the development of performance management practices in Swedish central government. This research programme is based on intensive networking and exchange of experiences with central government officials and a large number of government agencies. Hence, our insights into broader developments at the overall policy level emanate from continuous dialogue with a broad range of civil servants involved in performance management issues as well as prior research, official investigations, policy documents and four formal interviews. These interviews targeted the key actors involved in the development of novel performance management practices, such as the Prime Minister’s Office, the National Financial Management Authority, the National Council for Quality and Development and the Swedish Agency for Public Management. The three latter organisations have important advisory functions in the development of performance management practices in central government. Their respective roles and relative influence on evolving performance management practices are explicated in the empirical analysis. To provide more specific insights into how the emerging reform agenda may impinge on intraorganisational change processes we offer a case study in the Swedish Tax Agency. This case was chosen for illustrative purposes as recent changes in its performance management practices have clearly been influenced by the Government’s overriding concerns with enhanced citizen orientation. The case study also shows how tensions between seemingly inconsistent quality control and ‘managing for results’ practices emerging at the policy level may be managed within government agencies. Case study data were collected between 2004 and 2005 and primarily consist of ten semi-structured interviews and archival sources such as annual reports and internal planning documents and performance reports. Documents describing the results of internally initiated projects aimed at developing performance management practices were also examined. In addition, important insights into the development of novel quality management practices in the organisation were gleaned from a prior, longitudinal field study (Quist, 2003). Interviews were conducted with managerial staff extensively involved in performance management issues at headquarters and in five of the ten regional divisions of the organisation. The interviews lasted between 1 and 2 h and were in some cases followed up via additional personal meetings. In addition to agency staff, we interviewed the civil servant in charge of monitoring the performance of the Swedish Tax Agency within the Ministry of Finance to solicit information on more specific governance issues. The interviews were only partially tape-recorded to avoid bias related to politically sensitive topics and we observe some caution in using longer interview quotes. However, we relied extensively on respondent validation by returning extensive summaries of each interview for interviewee feedback and clarifications. In the case of the Swedish Tax Agency, our findings were also validated through continuous debriefing with and feedback on a preliminary case study narrative from two key informants (the Chief Planning Officer and another member of headquarters staff).

4. Developments at the policy level within Swedish central government 4.1. The implementation of ‘managing for results’ By international comparisons, Swedish central government is relatively unique as it has a long history of devolved responsibility for operating matters handled by agencies enjoying a considerable degree of autonomy. The possibilities of direct political control of agencies have been limited as ministerial

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

459

intervention in operating matters is prohibited by law. Hence, formal parliamentary and governmental control of agencies has largely been confined to the legislative process, annual appropriations and the appointment of Director-Generals (Molander et al., 1999; Pierre, 2004). As explicated below, however, Swedish central government has not been immune to the global reform tendencies emerging since the 1980s. The official introduction of a more pronounced ‘managing for results’ ethos is generally dated to 1988, when Parliament passed a Bill outlining new management principles in central government (Swedish Parliament, 1987/1988). The reform was based on the notion that the responsibilities of politicians and civil servants both can and should be separated (Sundstr¨om, 2003). The reform was couched in a rhetoric emphasizing the need to reduce the allegedly pervasive element of detailed political control in order to remove constraints on managerial freedom (Jacobsson and Sundstr¨om, 2001; Lindstr¨om, 1997). This may seem paradoxical as Swedish government agencies are renowned for being highly autonomous. However, the introduction of ‘managing for results’ was accompanied by a strongly felt need to deregulate central government operations. The principles underpinning this results-based performance management ethos gradually evolved in the 1990s. In particular, two types of official documents have gained increasing prominence; the annual appropriation directives from the Government and the annual reports of government agencies (Sundstr¨om, 2003). Evaluations of the reform during the 1990s suggest that the main emphasis in the annual appropriation directives has been on objectives pertaining to outputs and efficiency rather than outcomes and effectiveness. However, the introduction of ‘managing for results’ did not reduce the reporting requirements of agencies. Concerns were thus raised that the new control system merely implied a new means of regulating agency activities (Sundstr¨om, 2003). Nevertheless, there is evidence of extensive use of the information generated through the ‘managing for results’ system for internal control purposes by government agencies (National Financial Management Authority, 1999a). The Government has also persisted in defending the position of ‘managing for results’ as an overriding governance mechanism (Holmblad Brunsson, 2002) and it is questioned by few actors within central government (Modell, 2006). Throughout this gradual institutionalisation of ‘managing for results’ implementation has been supported by key actors, such as the National Audit Office and the National Financial Management Authority, prescribing ways of improving this control model (Modell, 2006; Sundstr¨om, 2003).2 4.2. Rise of the TQM and citizen orientation Parallel to the institutionalisation of ‘managing for results,’ alternative perspectives on performance management have been introduced and developed by different actors. Two such actors are the Swedish Institute for Quality Assurance (SIQ) and the National Council for Quality and Development. The SIQ is a non-governmental organisation dedicated to promoting ‘modern’ quality management practices within the private and public sectors. In 1992, the organisation launched its model; the SIQ Model for Performance Excellence. The model was developed as a tool for internal evaluation and improvement and was heavily inspired by various TQM practice. Important sources of inspiration were the Malcolm Baldridge Award in the U.S. and the European Foundation for Quality Management (EFQM). 2

The National Financial Management Authority was formed as an advisory agency in 1998 as the consultative branch of the National Audit Office was separated from state auditing.

460

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

The customer- and process-orientated SIQ model for performance excellence has diffused to a growing number of central government agencies since the mid-1990s, including the one under closer examination below (Modell, 2006; Quist, 2003; Wiesel, 2006). As explicated later, the SIQ has also influenced actors with an important advisory role on policy matters in the sector, such as the National Council for Quality and Development. While the ‘modern’ quality movement started to penetrate government agencies without much involvement by the Government, efforts to enhance their emphasis on the beneficiary-perceived value and quality of services were also initiated at the political level. In 1998, Parliament passed a new Bill, outlining principles for development of central government operations from a more pronounced citizen perspective (Swedish Parliament, 1997/1998). Although few direct references to the emerging, process-orientated quality management practices were made in the Bill, traces of the ethos of the ‘modern’ quality movement can be found. For example, the Bill prescribed a new reform trajectory as follows: ‘The Government intends to emphasise the quality of government operations to a greater extent than earlier. High quality in government services as well as in government work processes and competence requires continuous development of operations.’ (Swedish Parliament, 1997/1998, p. 13) Neither the new Bill nor the extensive official investigations preceding it (see Ministry of Finance, 1997) saw the combination of a more citizen-orientated quality focus with established ‘managing for results’ practices as particularly problematic. Rather, their complementary nature was emphasised as the following quote illustrates: ‘As far as quality work is concerned, it is a matter of utilising the possibilities that systematic quality work and systematic comparisons (benchmarking) with other actors generate. Quality work constitutes a complement to the process of managing for results and is a basis for the performance dialogue with heads of agencies.’ (Swedish Parliament, 1997/1998, p. 16) However, the Bill placed little emphasis on re-orienting ‘managing for results’ practices towards a heavier emphasis on outcomes. On the contrary, efficiency and financial probity were seen as the key concerns of performance management. This was arguably due to the dominating influence of the Ministry of Finance in the final formulation of the Bill (Sundstr¨om, 2003). Although the importance of compiling outcome indicators was recognised, the compilation of such indicators proved problematic. Such difficulties have prompted several government agencies to replace or complement quantifiable indicators with broader, verbal descriptions of how operations impinge on various stakeholders and society in general (Modell, 2006; Sundstr¨om, 2003). Nevertheless, there are some indications that the growing emphasis on citizen value has reinforced concerns with outcomes. Surveys conducted by the National Financial Management Authority (1999b, 2000) suggest that the number of outcome-related objectives in appropriation directives increased considerably in the late 1990s, although outcome indicators are primarily used for external reporting rather than internal control by government agencies. Despite these changes in performance reporting the development of quality controls has largely progressed in isolation from ‘managing for results’ practices. A new advisory agency, the National Council for Quality and Development, was formed in 1999 as a result of the aforementioned Government Bill (i.e., Swedish Parliament, 1997/1998). This organisation was established with the objective of stimulating and supporting long-term quality development in government agencies and was motivated as follows by the Government:

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

461

‘The Government needs a special organ to influence and develop the administration more actively in the desired direction in strategically important areas such as competence development and quality development.’ (Swedish Parliament, 1997/1998, p. 22) Consistent with its view of quality management as compatible with existing control practices the Government instructed the new agency to ensure that: ‘. . . quality issues receive broadly based attention and that work on such issues is integrated as a natural part of agency operations.’ (Swedish Parliament, 1997/1998, p. 22) However, the National Council for Quality and Development was to operate as an independent advisory agency with particular focus on quality issues. For example, the Government did not require the organisation to coordinate its activities with other advisory agencies with a key role in developing the ‘managing for results’ system, such as the National Financial Management Authority. Subsequently, the Council has developed into a key spokesman for the adoption of TQM practices inspired by the SIQ model for performance excellence and similar process-orientated models being developed by government agencies. 4.3. Emerging inconsistencies and tensions Despite the official government view of ‘modern’ quality management practices as complementary to ‘managing for results,’ their parallel evolution has resulted in some inconsistencies exacerbating implementation of the former. As explicated below, recent attempts to reconcile the logics of control embedded in these techniques have amplified rather than reduced such tensions. According to the National Council for Quality and Development, the key to enhance quality and nurturing a customer focus lies in the development of process-orientated management practices. The Council saw this as an important means of continuous improvement and ensuring customer satisfaction. The Council has conducted several studies of the implementation of process-orientated management practices in Swedish government agencies (National Council for Quality and Development, 2003a,b, 2004). However, considerable difficulties in combining such practices with the ethos of ‘managing for results’ were observed, as process-orientated management practices partly violate the vertical and hierarchal logic of the former model. In particular, it has proved problematic to extend process-orientated management practices within government agencies to coordinating activities involving more than one agency. For example, in a broadly based review of such practices the Council concluded that: ‘Managing for results has largely focused on maximising the performance of the operations of each individual agency without linking this clearly to effects from a holistic perspective.’ (National Council for Quality and Development, 2004, p. 77) In our interviews and continuous contacts with Council representatives a picture of relatively mixed experiences with process orientation within individual government agencies also emerged. For example, a Council interviewee suggested that: ‘. . . the introduction of TQM and quality models doesn’t have a very good track record in central government.’ He continued by ascribing such shortcomings to the lack of political incentives to adopt TQM-inspired work practices and the limited capacity of senior management in government agencies to follow up the

462

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

implementation of such practices. When queried about these developments at a later occasion he reaffirmed the difficulties in reconciling TQM-inspired practices with the ethos of ‘managing for results’ in government agencies. Although there was an initial flurry of experimenting with process-orientated quality management within a number of government agencies some of the enthusiasm seems to have waned. Also, few government agencies have adopted ‘pure’ forms of process orientation, but have generally modified these to suit operating realities (National Council for Quality and Development, 2003a). Process-orientated management practices have also had limited influence on other efforts to develop citizen-focused performance management models in Swedish central government, such as those associated with the notion of the ‘24/7 Agency.’ In 2003, the Minister for Local Government and Financial Markets stated that the ‘24/7 Agency’ should symbolise the movement from an old bureaucratic central government to a modern central government whose main focus is the interests of the citizenry. The Government subsequently commissioned the Swedish Agency for Public Management to devise a ‘generic’ model for evaluating and developing the ‘24/7 Agency’ concept. The results of this assignment were reported in 2005 (Swedish Agency for Public Management, 2005). In developing the new model, the Swedish Agency for Public Management consulted representatives from several other government agencies, including the National Council for Quality and Development. Several well-established quality management models were also evaluated to provide sources of inspiration. The notion of customer orientation constitutes a key stepping stone of the model subsequently developed. However, the agency deliberately chose not to adopt a more process-orientated management approach. According to our interviewees, the reason for this was that the agencies being consulted were concerned that the model would imply yet another layer of controls and be inconsistent with the ‘managing for results’ system. Such concerns were particularly related to the problems of reconciling the existing system of control with inter-agency coordination. For example, a representative of the Swedish Agency for Public Management argued that: ‘. . . the process-spanning perspective conflicts with the Swedish model of governance. Swedish government agencies are autonomous units and there is a lack of overarching process owners. The agencies don’t let anybody onto their turf just like that.’ To accommodate such concerns, the model takes existing, politically established objectives as its point of departure and describes value creation in terms of three corresponding ‘value areas,’ namely service, openness and efficiency, which are then divided into a number of underlying sub-areas and performance indicators focusing on individual agencies (see Table 1). It may be argued that the indicators in Table 1 are closely linked to operating-level processes, such as the administration of customer contacts, operating routines, and fulfillment of service guarantees, rather than objectively verifiable outcomes. However, the National Council for Quality and Development disproved of the model as it did not build on a more pronounced process-orientated approach, similar to that underpinning its own models. In our interview, a representative of the Swedish Agency for Public Management claimed that the National Council for Quality and Development: ‘. . . criticised the model for building on an inside-out perspective and that we had an introvert view.’ One Council representative also confirmed this critical posture in informal discussions. However, such objections seem to have had little influence on subsequent political decisions. The Government has recently endorsed the model developed by the Swedish Agency for Public Management as a key

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

463

Table 1 The model for measuring value creation in government agencies proposed by the Swedish Agency for Public Management (2005, p. 37) Value area

Sub-areas

Indicators

Customer satisfaction

Accessibility Service guarantees

Total customer satisfaction Customer satisfaction regarding: perceived service; perceived openness; perceived efficiency Number of customer contacts per contact channel Proportion of service guarantees or service objectives fulfilled

Transparency Rule of law Participation

Proportion of electronically traceable service deliveries Proportion of appeals; proportion of appeals resulting in redress Existence of routines for systematic administration of viewpoints

Process efficiency

Lead times, three major service types; proportion of effective processing time of total lead time Proportion of automatic information exchange with other actors Total productivity or total unit costs per service type; workforce productivity or personnel cost per service type; operating cost reductions from investments

Service

Openness

Efficiency Information efficiency Cost efficiency

mechanism for evaluating the steps taken towards enhanced citizen orientation (Ministry of Finance, 2006). 4.4. Summary of path dependencies at the policy level In summary, there is evidence of path dependencies reinforcing inconsistencies between recent policy initiatives in Swedish central government. Despite the official Government view of the enhanced emphasis on citizen orientation and ‘modern’ quality management as compatible with ‘managing for results,’ implementation of these practices has not been tightly coordinated. In addition, the advisory agency created to support the new reform agenda (i.e., the National Council for Quality and Development) has developed vested interests in a management model that is not wholly consistent with the ethos of ‘managing for results.’ This may be described as unintended effects of the reforms introduced in the late 1990s. These unintended effects have since been cemented by path-dependent developments. The difficulties in establishing process-orientated management practices more firmly within Swedish central government may be traced to the constraints embedded in the dominant mode of governance pivoting around the vertically orientated ‘managing for results’ system. The limited progress in implementing process-orientated management practices has also weakened the influence of advocates of such practices, such as the National Council for Quality and Development. This is notably manifested in the recent choice of the Swedish Agency for Public Management not to nurture a more process-orientated approach while integrating other properties of the ‘modern’ quality movement, such as the emphasis on customer orientation, into its performance management model. This choice was also conditioned by widespread concerns about the potential perils of adopting an alternative performance management model among government agencies. Taken together, these developments suggest that uncertainties surrounding alternatives to the ‘managing for results’ system have reinforced increasing returns and the tendency towards path-dependent change (cf. Pierson, 2000a).

464

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

Hence, contrary to Greener’s (2005) predictions the prevalence of inconsistent performance management practices has not prompted any radical re-consideration of the dominant mode of governance in Swedish central government. This is partly due to the limited mobilisation of powerful interests around alternatives to the dominant form of governance. However, the following case study in the Swedish Tax Agency offers a complementary explanation pivoting around the flexibility of established performance management practices and the re-construction of process-orientated management (cf. Gains, 2004; Modell, 2005a).

5. The case of the Swedish Tax Agency 5.1. Agency background The Swedish Tax Agency was formed in January 2004 following the merger of ten regional tax agencies with the National Tax Board. The latter organisation previously filled an important advisory function in the area of tax policy but exercised little direct authority over the regional tax agencies. Although the regional organisation of tax services has largely remained intact a number of organisational changes followed the merger. This has entailed some centralisation as the former National Tax Board now has a clearer headquarters function. A separate organisational unit has also been created in the area of business taxation to improve services provided to large corporations on a national basis. However, most tax services are still administered regionally and are divided into five operating areas: personal and business taxation, administration of the national population register, property taxation, crime prevention and administration of estate inventory proceedings. Internal performance management practices largely pivot around the annual budgetary process and the ‘managing for results’ model embedded therein. Short-term performance monitoring is based on monthly and triannual reporting from the regions and cover issues such as financial performance, productivity, sick-leave statistics, service levels and activity levels across the five operating areas. However, the annual planning and reporting cycle is separated from long-term planning covering a period of 3–5 years. The Swedish Tax Agency is directly accountable to the Ministry of Finance. Of central concern is the effectiveness of taxation, which is defined as the ‘gap’ between theoretically available taxes and the taxes being levied. However, this notion of effectiveness is difficult to operationalise and translate into meaningful and reliable performance indicators for a variety of reasons.3 First, estimating theoretically available taxes is extremely difficult as this fluctuates with economic cycles. Second, the effectiveness of taxation is not only dependent on controllable factors linked to the administration of tax services but also hinges on legislative changes determining the limits of taxation. Although the Swedish Tax Agency has some advisory role in the latter area, it is ultimately vested in the political decision-making domain. Hence, according to the government official in charge of monitoring the performance of the Swedish Tax Agency, this is not only a matter of evaluating agency operations, but also of assessing how policy changes might affect the system of taxation as a whole. For reasons explicated above, a more meaningful indicator of the overall performance of the Swedish Tax Agency and the outcome of its activities is the effectiveness of tax collection, defined as the ‘gap’ 3

In 2003, the ‘taxation error’ (i.e. the gap between theoretically available taxes and the taxes being levied) was estimated at 8.5% of theoretically available taxes.

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

465

between the taxes levied and those actually paid. While this ‘gap’ is typically smaller than that signifying the effectiveness of taxation it is still associated with an element of uncertainty as the taxes actually paid are partly dependent on legal proceedings and may only be known after sometime.4 The combined outcome of tax levying and tax collection is of particular interest in external performance monitoring. While recognising the limited possibilities of influencing such outcomes in the short term, the government official in charge of performance monitoring explained that: ‘I am extremely interested in what is not included and what is not collected through the system of taxation. The same goes for the political level.’ Within the agency, a relatively detailed regulatory framework prescribing which control activities should be undertaken by the regional divisions to maximise tax collection forms the cornerstone of day-to-day operations. Even though the divisions have some discretion to devise additional rules and regulations the efforts to centralise the coordination of these by headquarters have been stepped up in recent years. Notwithstanding this traditional, internal emphasis on detailed regulation of operating processes, the Government grants the agency considerable discretion and has limited its reporting requirements. Interviewees claimed that the agency enjoyed considerable freedom in formulating more specific objectives and targets and that the number of targets had been reduced over time. For example, a member of headquarters staff stated that: ‘By and large, we set our own short-term targets. . . . It gets better and better but earlier it was rather difficult to absorb and translate the [annual] appropriation directives.’ Our interviews with regional managers provided similar evidence, or as two of them proclaimed: ‘A couple of years ago, you didn’t see the wood for all the trees. Now, goals are fewer and better motivated.’ ‘Headquarters control has become much better. We have fewer goals and tighter control.’ In contrast to many other government agencies, the objectives in the annual appropriation directives of the Swedish Tax Agency are sparse and usually not quantified. Members of headquarters staff described this as something relatively ‘unique’ within Swedish central government and argued that this reflected the Government’s trust in the agency.5 As explicated below, however, a tendency towards more detailed control of operating processes is noticeable as a result of changing operating priorities and the emerging reform agenda. 5.2. The introduction of process-orientated management practices The former National Tax Board started to implement process-orientated quality management practices, heavily inspired by the SIQ model for performance excellence, in the late 1990s. As evidenced in a deeper, longitudinal study covering the period up until 2002 (Quist, 2003), this initiative was strongly endorsed 4

Between 2002 and 2005, the gap between the taxes levied and those actually paid decreased steadily from 0.53 to 0.35%. This ‘uniqueness’ is underscored by more broadly based studies conducted by the National Financial Management Authority (1999b, 2003, 2004) indicating that the number of objectives and the element of detailed control of operating aspects in the governance of agencies are generally increasing. 5

466

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

by the former Director-General and resulted in relatively straightforward adoption of the model as a tool for process orientation and quality enhancement. However, this study also reveals that the initial experimenting with process-orientated quality management encountered considerable implementation problems due to the lack of adjustments of the model to established work and control practices. Similar findings were reported from one of the regional tax agencies following the example of the National Tax Board. Within the National Tax Board, considerable adjustments of the SIQ-inspired model followed the appointment of a new Director-General in 1999, who was less pre-occupied with implementing the model in its original form (Quist, 2003). Still, the subsequent development of the model was riddled with conflicts over the meaning of quality management and process orientation. Implementation of the model seems to have stopped short of detailed descriptions of operating processes and evoked little change in established control practices (Quist, 2003). Process-orientated management practices were not genuinely integrated with the ‘managing for results’ model forming the backbone of performance management and the implementation process now seems to have lost much of its momentum. In our interviews, concerns with process orientation in its original form featured relatively sparsely and few large-scale projects to further develop such practices are under way. Also, whereas the work with the SIQ-inspired model was previously lauded in the annual reports of the National Tax Board, external reporting on these activities has subsequently been de-emphasised. 5.3. Conformance to the new reform agenda Parallel to its unfolding work with process orientation the National Tax Board took the initiative to translate the emerging governmental reform agenda aimed at enhanced citizen orientation to their own operating realities. This resulted in a broadly based mapping of the demands of citizens and the business community, strongly emphasising such performance aspects as accessibility and service-minded treatment (National Tax Board, 2002). Increasing emphasis was also placed on citizens’ trust in the system of taxation and their norms and attitudes towards paying taxes. Perception-based information concerning service aspects as well as trust, norms and attitudes is collected annually via large-scale surveys (targeting individual citizens and businesses every second year). The results of these surveys now feature prominently in the annual reports of the Swedish Tax Agency. This is complemented with more direct contacts through ‘stakeholder councils’ and focus groups at the regional as well as national levels. The mapping of tax-payers’ demands claimed to adhere to a ‘customer-orientated perspective’ (National Tax Board, 2002, p. 11). These efforts to compile information about tax-payers on a more systematic basis were not tightly coordinated with the ongoing, but gradually de-emphasised work with process-orientated quality management. Nevertheless, they have left an important legacy. A notable development is that the ethos of tax administration has partly shifted towards factors underpinning the willingness to pay taxes rather than outcomes such as those reflected by the effectiveness of taxation and tax collection. A member of headquarters staff explained that: ‘In the early 1990s it [i.e. internal performance evaluation] was mainly a matter of measuring the effectiveness of taxation. But since the mid-1990s we have drifted away from this and now we hardly talk about it all.’ Although the outcomes of taxation are still of interest increasing attention is paid to the role of citizen and business participation in the administration of tax services. Regional managers nearly invariably

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

467

associated the growing emphasis on citizen orientation with the shift from a control-focused approach to tax collection to one emphasising the facilitation of citizen participation. For example, one regional manager explained that: ‘Operations have traditionally concentrated on control [of tax-payers]. This is still a natural part but in recent years the focus has shifted towards helping citizens to do the right thing from the start.’ Strategies for changing citizen attitudes to critical issues such as tax evasion and fraud are also being developed. This changing operating ethos is strongly reflected in various policy statements issued by the agency. For example, in the annual report of 2004, a cornerstone of the vision of the Swedish Tax Agency was described as contributing to ‘a society where everyone wishes to pay one’s way’ (Swedish Tax Agency, 2005, p. 6). The changes in the ethos of tax administration are mirrored by the model recently developed to structure performance information for internal control purposes (see Fig. 1). The model outlined in Fig. 1 features important vestiges of the ideas underpinning ‘modern’ quality management practices, not least the emphasis on pre-emptive error avoidance as a means of maximising the effectiveness of taxation and tax collection. The new performance information model was developed with the specific aim of tracing outcomes, but also reinforces the current re-orientation away from more ‘objective’ outcome indicators such as the effectiveness of taxation and tax collection. The difficulties in measuring the effectiveness of taxation accurately are acknowledged and considerable emphasis is placed on nurturing norms and attitudes conducive to voluntary participation in the process of taxation. The primary source of information in this respect is the aforementioned annual surveys of tax-payers. Efforts are also made to link this information to short-term operating aspects such as service levels and provision of information (Swedish Tax Agency, 2004).

Fig. 1. Proposed links between performance aspects underpinning the effectiveness of taxation and tax collection. Source: adapted from the Swedish Tax Agency (2004, p. 18).

468

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

5.4. A process (re)turn While the model outlined in Fig. 1 was not fully operational at the time of our study it reinforces the emphasis on operating processes linked to serving tax-payers. Similar tendencies are discernible in recent developments in the agency’s existing performance management practices. Short-term monitoring of service levels is primarily geared towards indicators such as telephone response times and lead times across various operating areas. There were also suggestions that the emphasis on such performance aspects had increased since the formation of the new organisation in 2004. One reason for this is that the Government has increasingly emphasised the elimination of regional differences in service levels to further citizen orientation and sees the new organisation as a vehicle to this end. For example, a member of headquarters staff argued that a notable change since the formation of the agency was that: ‘. . . the Government especially controls lead times and shortening of these. They also focus a lot on eliminating regional differences. It concerns things like lead times, control activities and service to tax-payers.’ Similarly, when queried about key quality indicators the government official in charge of monitoring the performance of the Swedish Tax Agency expounded that the Government had: ‘. . . recently emphasised the importance of uniformity and equal treatment by the [Swedish Tax] Agency. Earlier, there have been unjustified regional differences in this respect. This has led to enhanced focus on these aspects from the Government.’ These pressures have filtered through to internal control practices. The number of quantifiable targets representing service aspects is increasing in annual reports as well as internal planning documents. This tendency is especially obvious with respect to targets for lead times across various types of services. Monthly performance reports from the regions take the shape of league tables based on a ‘traffic light system’6 indicating the extent to which the regions achieve these targets and whether performance improvements are noticeable. However, headquarters staff emphasised that the primary purpose of regional comparisons is not to stimulate competition but rather to visualise regional differences in order to successively even these out in line with the Government’s intentions. Similarly, a regional manager argued that a notable shift had occurred in control practices, saying that: ‘There is more top-down control today. We have a better overview and make sure everyone is working on what is important for the country as a whole. Earlier, the managers of each [regional tax] agency could decide what to do pretty much on their own. . . . Our present controls enable uniform administration across the country.’ Another regional manager described a similar trajectory: ‘Headquarters do not only control what to do, as specified by targets, they have also become better at issuing guidelines prescribing how we should work at the regional level. This has to do with uniformity.’ 6

This type of system indicates the level of performance and goal achievement based on colour schemes and has gained increasing popularity among Swedish government agencies.

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

469

She went on by defending the development towards more centralised control of operating details arguing that: ‘. . . it concerns such important areas that tighter control is needed to achieve uniformity and placing citizens centre stage.’ Whereas short-term monitoring of service levels is heavily geared towards operating process concerns rather than perception-based information such as that generated by the surveys of tax-payers, the latter is of some importance for long-term planning. Headquarters staff argued that long-term planning activities had become more externally focused as a result of a clearer incorporation of the mapping of tax-payers’ attitudes and needs via surveys in recent years. Increasing efforts are also made to tailor services to the more specific needs of various categories of tax-payers. One example of this is the aforementioned specialisation of tax administration for large corporations. Other examples include special efforts to enhance the comprehensibility of tax information to immigrants and to develop various electronic services to enhance the organisation’s accessibility to tax-payers. While reliance on various electronic solutions constitutes a key element in the ongoing efforts of the Swedish Tax Agency to enhance the level of service to tax-payers, this has also reinforced the tendency towards more detailed control of internal, operating aspects. Investments in IT have increased dramatically as a result of these efforts and now account for about one-sixth of the agency’s annual appropriations. A member of headquarters staff explained that this had resulted in increasingly detailed reporting requirements from the Government despite the freedom typically granted to the agency in this respect. Examples of such reporting requirements are efficiency gains from specific IT investments, the impact of such investments on tax-payers, management of IT development and the steps taken to provide incentives to tax-payers to utilise electronic solutions. Despite this growing attention to operating-level details in the wake of enhanced citizen orientation, few interviewees at any level experienced any pronounced tensions between this development and the existing ‘managing for results’ system. Similar to headquarters staff, regional managers rather emphasised that planning activities have been increasingly influenced by the initiatives aimed at enhanced citizen orientation such that performance targets now emanate from a tax-payer perspective to a greater extent. One regional manager specified this as follows: ‘The target of increasing the number of persons submitting their tax returns electronically leads to higher efficiency within [the Swedish Tax Agency] as well as facilitating citizens’ tax returns. Other targets [linked to the citizen perspective] are that we need to facilitate things for citizens with an immigrant background and shorten lead times.’ Even though the SIQ-inspired model for process-orientated management has been de-emphasised, several interviewees suggested that it had also influenced current planning and goal-setting activities. Two regional managers explicated this as follows: ‘I think our work on developing quality and operations according to the SIQ model has contributed strongly to incorporate the citizen perspective as something natural. Even though this may be hard – you know, we’re so special – you have to ask yourself the question ‘who do we perform certain tasks for and how do we find out what they want us to do?’ ‘We are very happy with the citizen perspective and see it as a natural part of operations. The perspective is one of the 13 basic values in the SIQ model and it is all about how we should approach citizens.’

470

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

However, the notion of process orientation has assumed a very different meaning and does not conform to the traditional TQM ethos. In contrast to the prior work with process-orientated quality management, which prescribed very detailed disaggregation of operating processes (Quist, 2003), such processes are now largely equated with the five overriding, operating areas within tax administration. A member of headquarters staff described this as taking a more ‘balanced’ view of process-orientated quality management practices and striving to integrate these into traditional planning and evaluation procedures. He continued by summing up the development over the past few years as follows: ‘You easily drive off the road if you just take certain models off the shelf and these clash [with existing controls]. We now keep a rather low profile as far as the SIQ model is concerned but it led to the appointment of process coordinators and process groups. We have subsequently adopted a broader approach rather than only focusing on the SIQ model and now things have cooled down. . . . Our present Director-General also takes a pretty down-to-earth approach to these kinds of methods while the former Director-General was more focused on the original model.’

6. Concluding discussion Our empirical findings illustrate how recent central government reforms prescribing enhanced citizen orientation and a focus on long-term outcomes may somewhat paradoxically reinforce short-term managerial concerns with operating processes within an individual government agency. Hence, there is some basis for arguing that the emerging reform agenda may represent a turn to process models of control. While such models have typically been associated with the ‘bureaucratic’ paradigm embedded in pre-NPM practices (Hood, 1991, 1995), the ‘new’ process concerns testify to a more externally focused reform agenda. However, we also find remnants of ‘old’ public administration practices, such as increasingly centralised performance monitoring and detailed reporting requirements to ensure equitable service to tax-payers. All this occurred without seriously challenging the ‘managing for results system’ in the Swedish Tax Agency. On the other hand, this system has proved largely incompatible with TQM-inspired practices at the policy level, despite political intentions to the contrary. As explicated below, the notion of path dependency provides a means of theorising why such intricate and partly unintended trajectories unfold at different levels of analysis. At the policy level of Swedish central government, there is clear evidence of ‘managing for results’ entailing increasing returns, which preclude alternative performance management models from seriously challenging its institutionalised position (Pierson, 2000a). Few actors have been compelled to subscribe wholeheartedly to ‘modern’ quality management techniques. Doing so would appear risky and potentially costly as a result of their mixed track record and uncertain future in Swedish central government (cf. Pierson, 2000a). Instead, some selective adoption of properties deemed consistent with the ‘managing for results’ model, such as the emphasis on customer satisfaction, is discernible in the recent attempt by the Swedish Agency for Public Management to reconcile this model with the citizen-focused reform agenda. This has reinforced the dominant position of ‘managing for results’ at the expense of process-orientated management practices and underlines the path-dependent and evolutionary nature of change. Similarly, in developing its performance management practices the Swedish Tax Agency has been more inclined to follow the emerging, government-initiated reform agenda than pursuing ‘pure,’ processorientated quality management practices. Instead of fully subscribing to a performance management

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

471

model which has been contested at the policy level and riddled with internal implementation problems, the agency has integrated perception-based measures of tax-payers’ views and closely related service indicators into the vertically orientated ‘managing for results’ system. Although the agency’s earlier experimenting with process orientation has left some legacy, the basic tenets of such practices have been de-emphasised in the process of adjustment. These findings confirm the highly constraining effects of increasing returns associated institutionalised control practices (Pierson, 2000a). Such constraints may outweigh the change dynamics associated with institutional inconsistencies (cf. Seo and Creed, 2002). However, our multi-level analysis nuances this observation by drawing attention to the flexible and socially constructed dimensions of institutional inconsistencies and increasing returns. While it has proved difficult to reconcile TQM-inspired practices with the hierarchical logic embedded in ‘managing for results’ at the policy level, the re-construction of such practices in the Swedish Tax Agency suggests that these barriers can partly be overcome. The notions of process orientation and quality management have here changed their meanings in the course of implementation. Similarly, the meaning of outcomes has been re-constructed to entail greater recognition of subjective properties, such as citizens’ perceptions of tax services, linked to short-term operating concerns. This supports the prediction that the ongoing re-interpretation of current reform initiatives by individual ‘grass-roots’ organisations may resolve seemingly ‘irreconcilable’ performance management logics (Modell, 2005a). Prior research on path dependencies has not fully recognised such interpretive dynamics. For example, in contrast to Greener’s (2005) conceptualisation we recognise that the ‘degree of consistency’ between alternative performance management practices is not ‘given’ but open to re-construction over time (cf. Clemens and Cook, 1999; Seo and Creed, 2002). Emerging performance management models may subsequently be decomposed such that the relative advantages of selectively adopting certain elements of these are enhanced. This might, in turn, alter increasing returns functions in favour of emerging practices. These insights have important implications for institutional research on management accounting change and public sector reforms. In responding to recent calls to bridge the gap between the policy and intraorganisational levels of analysis (Dillard et al., 2004; Modell, 2003) and by drawing more explicitly on the conceptual development of the notion of path dependency in policy studies, we explicate how institutional inconsistencies are (partly) resolved over time such that conflicts and pressures for change are muted. This goes beyond conceptualisations of management accounting change mainly ascribing the resilience of institutionalised controls to embedded organisational routines (e.g., Burns and Scapens, 2000) and specifies the conditions under which institutional inconsistencies may constitute a source of institutional change (cf. Abrahamsson and Gerdin, 2006; Burns and Baldvinsdottir, 2005; Modell, 2005b; Seo and Creed, 2002). At a more general level, this illustrates the merits of combining institutional analyses with insights from policy studies (Hall and Taylor, 1996). Such an integrative perspective has been curiously absent in research on public sector performance management due to the lack of communication between the public administration and accounting literatures (van Helden et al., 2006). However, more research inspired by such an approach is required to examine how intra-organisational changes inform policy development to fully grasp the ongoing, dialectic and multi-level nature of institutional change (see also Modell, 2007; Pierre, 2004). Further research into whether reform trajectories similar to those observed in the present study are emerging in other parts of the public sector would also seem warranted. Recent research in the UK National Health Service shows that the implementation of process-orientated management techniques

472

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

has encountered similar difficulties and has not led to more radical transformation of control practices (McNulty and Ferlie, 2004). Further investigations into these issues will reveal whether the emphasis on enhanced citizen orientation and adoption of TQM-inspired techniques reflect a ‘new’ reform paradigm or whether prior NPM reforms will continue to influence emerging control practices. Our study demonstrates that the emerging reform agenda may signify a ‘process turn’ although this does not necessarily follow the trajectory prescribed by TQM advocates. The notion of path dependency constitutes a useful analytical lens for examining such issues further within an institutional framework. Acknowledgements Previous versions of this paper were presented at the 29th annual congress of the European Accounting Association, Dublin, and the 4th International Conference on Accounting, Auditing and Management in Public Sector Reforms, Siena. We thank commentators at these presentations and two anonymous reviewers for insightful and helpful comments. The research assistance of Matthias Blom and Nina Haglind is also greatly appreciated. This paper was partly completed while the first author was a visiting professor at the University of Technology Sydney. The research was partly funded by the Swedish Research Council and Stockholm University. References Abrahamsson, G., Gerdin, J., 2006. Exploiting institutional contradictions. The role of management accounting in continuous improvement implementation. Qual. Res. Account. Manage. 3, 126–144. Barley, S.R., Tolbert, P.S., 1997. Institutionalization and structuration: studying the links between action and institution. Organ. Stud. 18, 93–117. Beckert, J., 1999. Agency, entrepreneurs, and institutional change. The role of strategic choice and institutionalized practice in organizations. Organ. Stud. 20, 777–799. Bromwich, M., Lapsley, I., 1997. Decentralisation and management accounting in central government: recycling old ideas? Finan. Account. Manage. 13, 181–201. Burns, J., 2000. The dynamics of accounting change: inter-play between new practices, routines, institutions, power and politics. Account. Audit. Account. J. 13, 566–596. Burns, J., Baldvinsdottir, G., 2005. An institutional perspective of accountants’ new roles—the interplay of contradictions and praxis. Eur. Account. Rev. 14, 725–757. Burns, J., Scapens, R.W., 2000. Conceptualising management accounting change: an institutional framework. Manage. Account. Res. 11, 3–25. Carter, N., Greer, P., 1993. Evaluating agencies: next steps and performance indicators. Public Admin. 71, 407–416. Catas´us, B., Gr¨onlund, A., 2005. More peace for less money: measurement and accountability in the Swedish Armed Forces. Finan. Account. Manage. 21, 467–484. Cavalluzzo, K.S., Ittner, C.G., 2004. Implementing performance measurement innovations: evidence from government. Account. Organ. Soc. 29, 243–267. Clemens, E.S., Cook, J.M., 1999. Politics and institutionalism: explaining durability and change. Annu. Rev. Sociol. 25, 441– 466. Covaleski, M.A., Dirsmith, M.W., 1988. An institutional perspective on the rise, fall and social transformation of a university budget category. Admin. Sci. Quart. 33, 562–587. Dillard, J.F., et al., 2004. The making and remaking of organization context: duality and the institutionalization process. Account. Audit. Account. J. 17, 506–542. DiMaggio, P.J., 1988. Interest and agency in institutional theory. In: Zucker, L.G. (Ed.), Institutional Patterns and Organizations: Culture and Environment. Ballinger, Cambridge, MA.

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

473

DiMaggio, P.J., Powell, W.W., 1983. The iron cage revisited: institutional isomorphism in organizational fields. Am. Sociol. Rev. 48, 147–160. Dorado, S., 2005. Institutional entrepreneurship, partaking and convening. Organ. Stud. 26, 385–414. Gains, F., 2004. Adapting the agency concept, variations within ‘next steps’. In: Pollitt, C., Talbot, C. (Eds.), Unbundled Government. A Critical Analysis of the Global Trend to Agencies, Quangos and Contractualisation. Routledge, London. Goldstone, J.A., 1998. Initial conditions, general laws, path dependence, and explanations in historical sociology. Am. J. Sociol. 104, 829–845. Graham, A., Roberts, A., 2004. The agency concept in North America: failure, adaptation and incremental change. In: Pollitt, C., Talbot, C. (Eds.), Unbundled Government. A Critical Analysis of the Global Trend to Agencies, Quangos and Contractualisation. Routledge, London. Greener, I., 2002. Theorising path-dependency: how history come to matter in organisations? Manage. Decision 4, 614–619. Greener, I., 2005. The potential of path dependence in political studies. Politics 25, 62–72. Greenwood, R., Hinings, C.R., 1996. Understanding radical organizational change: bringing together the old and new institutionalism. Acad. Manage. Rev. 21, 1022–1054. Hall, P.A., Taylor, R.C.R., 1996. Political science and the three new institutionalisms. Politic. Stud. 44, 936–957. Hay, C., Wincott, D., 1998. Structure, agency and historical institutionalism. Politic. Stud. 46, 951–957. Heinrich, C.J., 2002. Outcome-based performance management in the public sector: implications for government accountability and effectiveness. Public Admin. Rev. 62, 712–725. Hensmans, M., 2003. Social movement organizations: a metaphor for strategic actors in institutional fields. Organ. Stud. 24, 355–381. Hirsch, P.M., Lounsbury, M., 1997. Ending the family quarrel. Toward a reconciliation of “old” and “new” institutionalisms. Am. Behav. Sci. 40, 406–418. Holmblad Brunsson, K., 2002. Management or politics – or both? How management by objectives may be managed – a Swedish example. Finan. Account. Manage. 18, 189–209. Hood, C., 1991. A public management for all seasons? Public Admin. 69, 3–19. Hood, C., 1995. The “new public management” in the 1980s: variations on a theme. Account. Organ. Soc. 20, 93–109. Hood, C., Peters, G., 2004. The middle ageing of New Public Management: into an age of paradox? J. Public Admin. Res. Theory 14, 267–282. Hyndman, N., Eden, R., 2000. A study of the coordination of mission, objectives and targets in U.K. executive agencies. Manage. Account. Res. 11, 175–191. Jacobs, K., 1997. The decentralization debate and accounting controls in the New Zealand public sector. Finan. Account. Manage. 13, 331–343. Jacobsson, B., Sundstr¨om, G., 2001. Resultat utan l¨arande? erfarenheter fr˚an tre decennier av resultatstyrning (Results without learning? Experiences from three decades of managing for results) score. Stockholm University, Stockholm. Kay, A., 2005. A critique of the use of path dependency in policy studies. Public Admin. 83, 553–571. Klein, K.J., et al., 1999. Multilevel theory building: benefits, barriers, and new developments. Acad. Manage. Rev. 24, 243–248. Kravchuk, R.S., Schack, R.W., 1996. Designing effective performance measurement systems under the Government Performance and Results Act of 1993. Public Admin. Rev. 56, 348–358. Likierman, A., 1994. Management accounting in UK central government—some research issues. Finan. Account. Manage. 10, 93–115. Lindstr¨om, S., 1997. Att kunna h˚alla de andra krafterna st˚angen (To resist the other forces). In The Ministry of Finance. Det sv˚ara samspelet. Resultatstyrningens framv¨axt och problematik. (The difficult interplay. The development and problems of managing for results), Stockholm. Lynn Jr., L.E., 2001. The myth of the bureaucratic paradigm: what traditional public administration really stood for. Public Admin. Rev. 61, 144–160. Mascarenhas, R.C., 1996. Searching for efficiency in the public sector: interim evaluation of performance budgeting in New Zealand. Public Budget. Finan., 13–27. McCulloch, B.W., Ball, I., 1992. Accounting in the context of public sector management reform. Finan. Account. Manage. 8, 7–12. McGuire, L., 2002. Service charters—global convergence or national divergence? A comparison of initiatives in Australia, the United Kingdom and the United States. Public Manage. Rev. 4, 493–524.

474

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

McNulty, T., Ferlie, E., 2004. Process transformation: limitations to radical organizational change within public service organizations. Organ. Stud. 25, 1389–1412. Meyer, J.W., Rowan, B., 1977. Institutionalized organizations: formal structure as myth and ceremony. Am. J. Sociol. 83, 340–363. Ministry of Finance, 1997. I medborgarnas tj¨anst (In the service of citizens). Ministry of Finance, Stockholm. Ministry of Finance, 2006. B¨attre service f¨or varje skattekrona—Strategi f¨orfortsatt utveckling av elektronisk f¨orvaltning (More value for money—a strategy for continued development of electronic government). Ministry of Finance, Stockholm. Modell, S., 2003. Goals versus institutions: the development of performance measurement in the Swedish university sector. Manage. Account. Res. 14, 333–359. Modell, S., 2005a. Performance management in the public sector: past experiences, current practices and future challenges. Aust. Account. Rev. 15 (3), 56–66. Modell, S., 2005b. Students as consumers? An institutional field-level analysis of the construction of performance measurement practices. Account. Audit. Account. J. 18, 537–563. Modell, S., 2006. Den statliga myndighetssektorn som organisatoriskt f¨alt (Central government as an organisational field). In: Modell, S., Gr¨onlund, A. (Eds.), Effektivitet och styrning i statliga myndigheter (Performance Management in Central Government). Studentlitteratur, Lund. Modell, S., 2007. Managing accounting change. In: Hopper, T., et al. (Eds.), Issues in Management Accounting, 3rd ed. Pearsons, London. Modell, S., Gr¨onlund, A. (Eds.), 2006. Effektivitet och styrning i statliga myndigheter (Performance Management in Central Government). Studentlitteratur, Lund. Molander, P., et al., 1999. Vem styr? Relationen mellan Regeringskansliet och myndigheterna (Who is in Control? The Relationship between Government Offices and the Agencies). SNS F¨orlag, Stockholm. Moynihan, D.P., 2006. Managing for results in state government: evaluating a decade of reform. Public Admin. Rev. 66, 77–89. National Council for Quality and Development, 2003a. Processer och processorientering i 10 myndigheter (Processes and Process Orientation in Ten Central Government Agencies). National Council for Quality and Development, Stockholm. National Council for Quality and Development, 2003b. Processorientering i staten—en studie av hur statliga myndigheter arbetar med processorientering av verksamheten (Process Orientation in Central Government—A Study of Process Orientation in Central Government Agencies). National Council for Quality and Development, Stockholm. National Council for Quality and Development, 2004. Processorienterad verksamhetsstyrning i staten—fr˚an a¨ rende till medborgare (Process-orientated Control in Central Government—From Errands to Citizens). National Council for Quality and Development, Stockholm. National Financial Management Authority, 1999a. Myndigheternas syn p˚a resultatstyrningen (The Views of Agencies Concerning Managing for Results). National Financial Management Authority, Stockholm. National Financial Management Authority, 1999b. Informella kontakter i samband med regleringsbrev och a˚ rsredovisningar—en studie f¨or att utveckla m˚al- och resultatstyrning i staten (Informal Contacts Related to Annual Appropriation Directives and Annual Reports—A Study for Developing Managing for Results). National Financial Management Authority, Stockholm. National Financial Management Authority, 2000. Redovisning av enk¨atunders¨okning g¨allande myndigheternas interna styrning (Findings from a Survey of Government Agencies’ Internal Control Practices). National Financial Management Authority, Stockholm. National Financial Management Authority, 2003. En kartl¨aggning av anv¨andningen av myndigheternas a˚ rsredovisningar i Regeringskansliet (A Survey of the Use of Agency Annual Reports by Government Offices). National Financial Management Authority, Stockholm. National Financial Management Authority, 2004. Konsten att styra—en analys av dokument i den statliga budgetprocessen (The Art of Control—An Analysis of Documents in Government Budgetary Control). National Financial Management Authority, Stockholm. National Tax Board, 2002. I medborgarnas tj¨anst—v¨agarna dit (In the service of citizens—the means to the ends). National Tax Board, Solna. Neale, A., Pallot, J., 2001. Frontiers of non-financial performance reporting in New Zealand. Aust. Account. Rev. 11 (3), 27–43. Norman, R., Gregory, R., 2003. Paradoxes and pendulum swings: performance management in New Zealand’s public sector. Aust. J. Public Admin. 62 (4), 35–49. North, D.C., 1990. Institutions, Institutional Change and Economic Performance. Cambridge University Press, Cambridge.

S. Modell et al. / Management Accounting Research 18 (2007) 453–475

475

Osborne, S.P., 2006. The new public governance. Public Manage. Rev. 8, 377–387. Pallot, J., 1998. The New Zealand revolution. In: Olson, O., et al. (Eds.), Global Warning! Debating International Development in New Public Financial Management. Cappelen Akademisk Forlag, Oslo. Pierre, J., 2004. Central agencies in Sweden. A report from Utopia. In: Pollitt, C., Talbot, C. (Eds.), Unbundled Government. A Critical Analysis of the Global Trend to Agencies, Quangos and Contractualisation. Routledge, London. Pierson, P., 2000a. Increasing returns, path dependence, and the study of politics. Am. Polit. Sci. Rev. 94, 251–267. Pierson, P., 2000b. The limits of design: explaining institutional origins and change. Governance 13, 475–499. Pollitt, C., Talbot, C. (Eds.), 2004. Unbundled Government. A Critical Analysis of the Global Trend to Agencies, Quangos and Contractualisation. Routledge, London. Quist, J., 2003. Att o¨ vers¨atta TQM—en longitudinell studie kring reflekterande akt¨orer (Translating TQM—A Study of Reflexive Actors). Karlstad University Studies, Karlstad. Radin, B.A., 1998. The Government Performance and Results Act (GPRA): hydra-headed monster or flexible management tool. Public Admin. Rev. 58, 307–316. Seo, M.-G., Creed, W.E.D., 2002. Institutional contradictions, praxis, and institutional change: a dialectical perspective. Acad. Manage. Rev. 27, 222–247. Siti-Nabiha, A.K., Scapens, R.W., 2005. Stability and change: an institutionalist study of management accounting change. Account. Audit. Account. J. 18, 44–73. Sundstr¨om, G., 2003. Stat p˚a villov¨agar. Resultatstyrningens framv¨axt i historiskt-institutionellt perspektiv (The State at a Loss. The Development of Managing for Results from a Historical-Institutional Perspective). Stockholm University, Stockholm. Svara, J.H., 2001. The myth of the dichotomy: complementarity of politics and administration in the past and future of public administration. Public Admin. Rev. 61, 176–183. Swedish Agency for Public Management, 2005. Service, o¨ ppenhet och effektivitet. En modell f¨or att leda, utveckla och f¨olja upp den moderna f¨orvaltningen (Service, Openness and Efficiency. A Model for Managing, Developing and Evaluating Modern Government). Swedish Agency for Public Management, Stockholm. Swedish Parliament, 1987/1988. F¨orslag till slutlig reglering av statsbudgeten f¨or budget˚aret 1987/1988 (The Government Budget Bill of 1987/1988). Swedish Parliament, Stockholm. Swedish Parliament, 1997/1998. Statlig f¨orvaltning i medborgarnas tj¨anst (Government in the service of citizens). Swedish Parliament, Stockholm. Swedish Tax Agency, 2004. Grundl¨aggande uppf¨oljningsinformation (Basic Information for Performance Evaluation). Swedish Tax Agency, Solna. Swedish Tax Agency, 2005. a˚ rsredovisning f¨or budget˚aret 2004 (Annual Report 2004). Swedish Tax Agency, Solna. Talbot, C., 2004. The agency idea. Sometimes old, sometimes new, sometimes borrowed, sometimes untrue. In: Pollitt, C., Talbot, C. (Eds.), Unbundled Government. A Critical Analysis of the Global Trend to Agencies, Quangos and Contractualisation. Routledge, London. van Helden, J., et al., 2006. Exploring the public administration and accounting divide of public sector performance measurement research. Paper presented at the Fourth International Conference on Accounting, Auditing and Management in Public Sector Reforms, Siena. Watkins, A.L., Arrington, C.E., 2007. Accounting, new public management and American politics: theoretical insights into the National Performance Review. Crit. Perspect. Account. 18, 33–58. Wiesel, F., 2006. Kundorienterad effektivitetsm¨atning i Banverket och V¨agverket (Customer-orientated performance measurement in the Railroad and Road Authorities). In: Modell, S., Gr¨onlund, A. (Eds.), Effektivitet och styrning i statliga myndigheter (Performance Management in Central Government). Studentlitteratur, Lund.