Recent UK Central Government Accounting Reforms: Claimed Benefits and Experienced Outcomes

Focusing on the four most recent and highly publicized accounting reforms in UK central government, and utilizing discourse analysis of of ﬁ cial publications, together with semi-structured interviews, this paper examines the basis on which these four reforms were introduced, the extent to which they are considered appropriate, and the reasons for the persistence of new public management (NPM)- inspired accounting reforms in UK central government. The ﬁ ndings indicate that each was promoted in a similar manner despite limited post-implementation evidence that the supposed bene ﬁ ts of preceding reforms had not been realized to the extent originally suggested. Given NPM ’ s continued acceptance, these ﬁ ndings have international signi ﬁ cance.

Wall, 2012; Brusca et al., 2013;Christiaens et al., 2015;Gomes et al., 2015); and accounting change more broadly (Lapsley, 1999Ezzamel et al., 2013;Bracci et al., 2015;Hyndman et al., 2019). Prior research has typically concentrated on a single reform and the tensions or unintended negative consequences of its introduction, with little being known about the extent to which claimed benefits have been achieved post implementation (Hepworth, 2015;Narayan and Stittle, 2018).
This paper focuses on the four most recent high-profile accounting reforms implemented in the UK's central government to improve how the public and Parliament scrutinize the government's use of public money and hold departments accountable for the financial decisions they make (HMT, 2019a). These are the introduction of Resource Accounting and Budgeting (RAB) and Whole of Government Accounts (WGA), the adoption of International Financial Reporting Standards (IFRS) and the implementation of the Clear Line of Sight Alignment Project (CLOS). Using discourse analysis and semistructured interviews to compare the articulated pre-implementation benefits with post-implementation experiences, this paper examines the basis on which these four reforms were introduced, the extent to which they are considered appropriate, and the reasons for the persistence of NPM-inspired accounting reforms in UK central government. In addition, by exploring four reforms introduced over a period of time, each targeting different aspects of financial reporting introduced by a country that is cited as an active and early adopter of NPM ideas (Hood, 1995;Pollitt and Bouckaert, 2011), this paper enhances existing literature by developing our understanding of the pressures that cascade to organizations in an institutional field.
This research is important because, despite limited evidence of their actual benefits (Fattore et al., 2018), NPM-inspired ideas continue to be promoted in discussions about public sector modernization (Funck and Karlsson, 2019), with many early adopters (e.g., the UK and Australia) continuing to apply their tenets, and developing countries (e.g., Chile) gravitating towards them (Connolly and Hyndman, 2006;Paulsson, 2006;Heald and Georgiou, 2009;Hyndman and Lapsley, 2016;International Monetary Fund, 2016;Grossi et al., 2018;Nakmahachalasint and Narktabtee, 2019). However, questions have been raised about the extent to which recent reforms reflect core NPM principles (Steccolini, 2019). While some contend that they are aligned with NPM, with this being the new norm (Funck and Karlsson, 2019), others opine that it should not be assumed that reforms are NPM-driven (Christensen and Laegreid, 2011;Steccolini, 2019). Thus, in recognition of this debate, country-specific research that investigates whether they continue to reflect NPM ideals is important (Oulasvirta, 2014;Hepworth, 2015). Furthermore, such research is relevant and opportune given the significant investment in time and cost associated with implementing change (Chow et al., 2019). Moreover, as the four reforms examined in our study span 25 years, this can enable a broader understanding of the recursive interplay between institutional mechanisms as they unfold over time (Dillard et al., 2004).

UK PUBLIC SECTOR ACCOUNTING REFORM
Public sector reform is a compelling concept and governments from every era have sought to reform existing institutions, with those charged with introducing the change espousing particular principles for how government should be conducted and organized (Ingrams et al., 2020). In this regard, the UK's approach to planning, managing, and controlling public expenditure has changed considerably over the last 40 years. Driven by economic, ideological, and pragmatic motives (Pollitt and Bouckaert, 2011), successive governments have looked beyond the traditional bureaucratic public administrative methodologies that dominated previously in search of approaches that would drive improvements in accounting and financial reporting procedures. During the 1980s and 1990s in particular, the themes of greater accountability and improved performance dominated, with a belief that the technical language of accounting could help facilitate this. Thus, reform was initiated and guided (primarily) by NPM principles.
NPM's appeal is based upon a belief that policies based on marketization, decentralization, and risk transfer are the way to improve public sector service delivery and accountability (Humphrey et al., 1993;Dunleavy and Hood, 1994;Hood, 1995;Lapsley, 1999;Pollitt and Bouckaert, 2011;Christensen and Laegreid, 2013). As illustrated above, public sector accounting reform has received considerable academic attention, with much of this increased interest addressing NPM-related accounting issues such as the move from cash to accruals accounting, the adoption of international accounting standards, and the implementation of consolidated accounting. There is substantial literature on the theoretical benefits of adopting these reforms, with the primary contention being that they will lead to improved information which will in turn enhance accountability and allow better resource management decisions (Likierman, 1995;Ellwood, 2003;Heald and Georgiou, 2011;Hyndman and Connolly, 2011). Other notable claims include highlighting private sector accounting techniques as perceptively 'superior' to alternative systems as they reflect a focus on outputs and outcomes as well as inputs and create a more 'accurate' picture of the costs of services (Guthrie, 1998;Christensen, 2003).
Following high-profile financial and performance management changes by the Conservative Government under the leadership of Margaret Thatcher, such as the Financial Management Initiative in 1982 and the Next Steps Initiative in 1986, a groundbreaking financial reporting reform was announced in 1993. This was the proposal to adopt RAB (i.e., the replacement of cash accounting with a system that monitors and reports public expenditure using accruals accounting and budgeting principles (HMT, 1994)). Officially, RAB was promoted on the basis that it would provide better information and thus help the government plan and control expenditure, improve service delivery, and enhance decision making (HMT, 1994;Likierman, 1995;Connolly and Hyndman, 2006). During RAB's implementation, it was decided in 1998 to introduce consolidated accounting (known as WGA in the UK). WGA consolidates approximately 8,000 public organizations including central and local government departments, devolved administrations, executive agencies, the National Health Service, and nondepartmental public bodies into one set of financial statements (HMT, 2019b). In between the decision to adopt WGA and its implementation in 2010, a further reform was announced in the 2007 Budget, with IFRS replacing UK GAAP in order to follow 'private sector best practice' (HMT, 2007a, p. 154). In the same year, HMT (2007b) published The Governance of Britain, which heralded the implementation of CLOS to align the reporting of budgets, estimates, and accounts as these had developed differently over time, resulting in significant misalignment of figures and difficulty tracking authorized against actual spend (HMT, 2009).
The UK's recent approach to improving public sector financial reporting has been one of ongoing transformation, with each subsequent reform being announced as the former is still being implemented and before there is tangible evidence that its purported outcomes will be realized. Indeed, literature on the extent to which claimed benefits have been achieved is limited, with there being more focus on criticisms of the reforms, particularly in relation to the timeliness and cost of implementation (Connolly and Hyndman, 2006;Hyndman et al., 2019). Government has been reproached for not using the 'new' information made available by the reforms to assess outcomes and it is contended that, in some cases, the changes have made parliamentary scrutiny of public spending more difficult (Wynne, 2008;Chow et al., 2019). While international evidence is similar to that of the UK (Adhikari and Gårseth-Nesbakk, 2016), notable exceptions documenting experienced benefits include a comparison with what existed before (Ball, 2012), along with improvements to accountability and transparency (Kober et al., 2010;Nakmahachalasint and Narktabtee, 2019).
To conclude, the literature describing post-implementation experiences typically focuses on a single reform, with calls for further research (Oulasvirta, 2014;Hepworth, 2015;Chow et al., 2019). For many countries, the search for improved financial reporting mechanisms has been a complex, timeconsuming, and costly process, and there is evidence that the supposed benefits have not been realized to the extent originally suggested (Chan, 2003;Ellwood and Newberry, 2007;Brusca and Montestinos, 2009;Tagesson, 2009;Heald and Georgiou, 2011;Connolly and Wall, 2012;Heiling et al., 2013). Indeed, other non-NPM-inspired reforms (e.g., Public Administration and Public Governance) have travelled comparable paths, with the literature reporting similar tensions and outcomes (e.g., OECD, 2005;Andrews, 2013;Grossi and Steccolini, 2014). Nevertheless, despite the evidence that the claims that typically accompany many reforms are overly optimistic, this does not mean that they are destined to fail, but serves to emphasize the practical implementation difficulties and fundamental behavioural change required to sustain reform. Regardless, the appeal of such reforms endures and, in the context of our research, many countries continue to adopt accruals accounting, consolidated accounting, and IFRS/International Public Sector Accounting Standards (IPSAS). Consequently, using an institutional perspective to frame this issue, we consider why governments persist with (NPM) reform initiatives.

PUBLIC SECTOR ACCOUNTING CHOICE: AN INSTITUTIONAL PERSPECTIVE
While an 'institution' can be defined in different ways, it is acknowledged that it is a significant social structure that is often taken for granted and embedded in certain social settings (Scott, 2008). Although institutions are perceived as resilient and resistant to change, it is not unusual for them to transform themselves, with the reasons for this being explained through various theoretical perspectives. For example, institutional theory (IT) has been cited as 'a powerful theory' (Goddard et al., 2016, p. 11) which helps to explain why institutions choose to conform to what is deemed as the accepted rules, practices, and norms of their internal and external environments, over what is rational or needed (Meyer and Rowan, 1977;DiMaggio and Powell, 1983;Covaleski and Dirsmith, 1988;Selznick, 1996;Greenwood et al., 2008). It represents a way of thinking about the relationship between formal organizational structures and the social processes that contribute to the development of these structures (Dillard et al., 2004), providing a perspective for understanding the factors that influence change, together with how it is legitimated, in the public domain (Frumkin and Galaskiewicz, 2004;Jacobs and Jones, 2009;Jacobs, 2012;Goddard et al., 2016;Narayan and Stittle, 2018;Alvesson and Spicer, 2019). Battilana and D'Aunno (2009) distinguish between old and neo-institutional studies, with the former being prevalent in the 1950s and 1960s and viewing actors (individuals or organizations) as the primary agents of change. In contrast, the latter, which was predominant in the 1970s and 1980s, tends to assume that structural constraints shape actors' behaviours and that actors adapt to structure.
Drawing on old and neo-institutionalism, how actors embedded in institutions can change those institutions is known as the paradox of embedded agency (DiMaggio, 1988;Modell, 2009;Hiebl, 2018). Given some (institutional defenders) may seek to inhibit institutional change and retain the status quo, others (institutional entrepreneurs) have to overcome such resistance to achieve embedded agency (Battilana and D'Aunno, 2009). Factors that enable institutional entrepreneurs to implement institutional change include power and politics, with the former being 'the probability that one actor within a social relationship will be in a position to carry out his will despite resistance' (Weber, 1978, p. 53) and politics being the 'tactics and strategies actors use to articulate this power or attempt to resist it' (Fleming and Spicer, 2014, p. 238). Hiebl (2018) contends that management accounting may be at interplay with other factors in enabling embedded agency and can be an important resource not only in identifying the need for change and gaining support for change, but also for implementing change and legitimizing it ex post.
Accounting has a vital role both in organizations/states and in social relationships, with accounting knowledge affecting perceptions, changing languages, and offering new possibilities for new actions (Hopwood, 1988). IT is often used to help explain the application and development of accounting, including within the public sector (Frumkin and Galaskiewicz, 2004;Ezzamel et al., 2007;Hopper and Major, 2007;Mussari and Monfardini, 2010;Jacobs, 2012;Yang and Modell, 2012;Anessi-Pessina et al., 2016;Goddard et al., 2016;Piña and Avellanda, 2018). As the common threads running through these studies include the recognition that legitimacy lies at the heart of IT, the application of isomorphism to explain why governments adopt similar reforms and the occurrence of loose coupling, we now discuss each further in turn. Scott (2008), who views legitimacy as essentially symbolic to be displayed visibly, identifies three 'pillars' of (or ways of understanding) institutions and legitimacy: the first emphasizes a regulatory conformity to rules; the second highlights a deeper, moral base for assessing legitimacy; and the third stresses the legitimacy derived from adopting a common frame of reference. Whilst the above mechanisms and pillars are often presented separately, they may become entwined in practice.
It is argued that in their search for legitimacy, organizations adopt structures and systems similar to those of their competitors, a process referred to as isomorphism (Hannan and Freeman, 1977). If they do not do so, the contention is that they will lose their legitimacy, thus encouraging organizations to homogenize their organizational structures and practices. Isomorphism is often viewed as complementing, or augmenting IT on the basis that it extends our understanding of organizational behaviour and identifies pressures exerted upon organizations to adopt similar forms and processes (Greenwood et al., 2008). DiMaggio and Powell (1983) identify three possible isomorphic mechanisms that may induce conformity: coercive (where organizational change results from pressures exerted by other organizations, political influence, or society); mimetic (where organizations intentionally imitate each other in response to uncertainty); and normative (where organizations indirectly adopt the norms and values of other organizations due to individuals belonging to the same profession or having followed the same educational processes). Whilst these pressures are often presented as being distinct, they are not mutually exclusive and may become entwined in practice. In the context of neo-institutional theory, it is expected that organizations will react to isomorphic pressures from their institutional environment and adopt structures and practices with high social value. In this sense, the introduction of accruals accounting represents formal compliance with the expectations of the external environment and influential stakeholders (Hyndman and Connolly, 2011;Piña and Avellanda, 2018).
While the traditional view is that isomorphic pressures force organizations to conform, and once they can demonstrate that rational mechanisms are in place they are deemed legitimate, the adoption of some measures may be superficial or symbolic (DiMaggio, 1988), with rules and expected procedures being decoupled or loosely coupled to the actual practices (Covaleski and Dirsmith, 1988;Tsamenyi et al., 2006). In such instances, a two-stage diffusion process is suggested, whereby reforms are motivated by technical considerations in the beginning but then behaviour is guided by institutional rationality (Lounsbury, 2008;Scott, 2008;Goddard et al., 2016). Consistent with Meyer and Rowan's (1977) contention that organizations tend to respond to institutional pressures for change by 'ceremonial conformity', Christensen (2005) posits that, by planting 'phantom images' pre-implementation, a belief that the reforms are appropriate is embedded and post-implementation acceptance without reflecting on the evidence is facilitated. This supports Humphrey's (2005, p. 482) assertion that new public financial management reforms tend to be heavily promoted initially, but post implementation are 'presumed to work rather than being known more for what they have achieved', and Bromley and Powell's (2012) suggestion that follow-up implementation, evaluation, and monitoring processes are often absent or merely ceremonial. Greenwood et al. (2014) contend that, in highlighting the role of institutions, IT has moved away from what should be its primary focus, namely understanding how organizations are structured and managed. They argue that IT now concentrates too much on explaining institutions and institutional processes rather than how organizations work, with an over emphasis on understanding organizational similarities meaning differences have tended to be overlooked (at the expense of comparative organizational studies). Thus, an institutional logics perspective, which assumes differences between organizations and seeks to explain them, is advocated (Lounsbury, 2008;Greenwood et al., 2014). In this regard, based upon a comparison of the use of performance measurement systems in two Finnish cities, Rautiainen and Järvenpää (2012) illustrate how the response depends on the institutionalized logic in the public sector accounting context. However, while Meyer and Höllerer (2014) broadly support a logics approach, they caution that overly focusing on organizations or organizational forms may result in important developments in organizing being ignored. They encourage understanding the interactions between institutions and forms of organizing and advise that similarities are important.
Literature has increasingly identified accounting as a combination of social, institutional, and technical practice (Potter, 2005), with public sector organizations being acknowledged as highly institutionalized since they comprise 'policies, functions, programmes and procedures that are enforced by public opinion, legislation … and the views of important constituents' (Kelly et al., 2015, p. 72). This broader understanding of accounting has triggered a range of IT-related public sector accounting research to understand the conditions that influence reform (Jacobs, 2012), with studies drawing upon isomorphism to make sense of change in situations that cannot be explained by a single influence. For example, Falkman and Taggesson's (2008) investigation of the adoption of accruals accounting in Swedish municipals found that, whilst coercion provided the initial pressure, the weak anchorage of the accounting system failed to legitimize the change, resulting in minimal impact. Similarly, Pollanen and Loiselle-Lapointe's (2012) examination of the introduction of accruals accounting in Canada found that, despite strong professional influence, political factors contributed to the reforms having minimal impact, causing practices to be decoupled from the underlying ideas, values, and beliefs. Other studies have demonstrated greater structural compatibility arising from multiple isomorphic pressures. This includes Carpenter and Feroz (2001) who examine the influences behind the decision to adopt US GAAP in four states and find the pressures to be largely normative through participating with accounting professionals, together with coercive due to dependency on credit markets for financial resources.
In contrast, Hyndman and Connolly (2011) observe that whilst strong coercive and normative forces encouraged the use of accruals accounting in the UK, professional and political support in the Republic of Ireland was insufficient to initiate change, resulting in cash-based systems being retained. Mussari and Monfardini (2010, p. 491), who examine the heightened emphasis on social reporting in the Italian public and not-for-profit sectors, find that normative factors increased the pressure to use social reporting as a tool to 'signal a higher level of efficiency and socially responsible behavior'. Thus, it would appear that normative pressure is an important factor and, as highlighted by DiMaggio and Powell (1983), once norms have been endorsed by professional actors at the individual organizational level, they may be forced to conform to professional standards even if compliance is not in the organization's political interest. Despite the success of adopting GAAP in the US, Carpenter and Feroz (2001) acknowledge that its use may be more a symbol of legitimacy than an actual financial management tool as government continues to rely on cash-based budgeting for day-to-day management of government operations.
In relation to IPSAS adoption in Spain, Brusca et al. (2013) contend that this may be due to them being seen as an extension of the global accounting architecture and that there is a tendency to mimic the supposed best practices of international institutions to gain legitimacy (e.g., the EU's decision in 2005 to require listed companies in its member states to report using IFRS). Within the Portuguese context, Gomes et al. (2015) also investigate IPSAS adoption, concluding that the reform stimuli were influenced by the institutional and political context (consistent with Lapsley et al., 2009), together with a combination of coercive, normative, and mimetic forces. Relatedly, Oulasvirta (2014) points to weak coercive, mimetic, and normative forces, together with a lack of political interest, to explain the limited success of public sector accounting reforms in Finland. In the New Zealand public sector, the decision to follow sector neutral standards of IFRS was overturned in favour of IPSAS as it was believed that user needs in the public sector were not sufficiently being met using IFRS (Crook, 2014). In contrast, Piña and Avellanda (2018) find evidence that mimetic and coercive forces inadvertently led municipalities in Chile to emulate the characteristics of their peers once collaborative partnerships were entered into.
As recognized in the previous section, UK public sector accounting reform is not confined to the four specifically addressed in this paper (i.e., RAB, IFRS, WGA, and CLOS), but has spanned decades and includes myriad radical and incremental changes. In the 1980s, it (e.g., the Financial Management and Next Steps Initiatives) was driven by the rationale that better accounting information would improve the use of public expenditure and help rebuild the economy following the financial strain of the 1970s (Fry, 1988;Puxty, 1997). Moreover, there is evidence that the reforms were influenced by approaches and principles emanating from the private sector (Pollitt and Bouckaert, 2011). In the 1990s, the search for greater efficiency in the UK public sector gathered pace with accounting at the heart of the initiatives, commencing with RAB and followed by WGA, IFRS, and CLOS. A possible consequence of this is that the more private sector ideas are adopted, the more the public sector mirrors the private sector and thus the more susceptible it becomes to institutional isomorphic pressures (Covaleski and Dirsmith, 1991).
In this regard, the obligation to prepare resource accounts (subsequently under IFRS) and WGA, have them audited by the NAO and laid before Parliament under the Government Resources and Accounts Act (GRAA) 2000 is evidence of coercive mechanisms being employed to legitimize these accounting reforms. Elements of CLOS have also been mandated through the GRAA 2000 and the Constitutional Reform and Governance Act (2010). Mimetic forces are also apparent. For example, parliamentary papers highlight that, as accruals accounting had been used in UK local government since 1974 and various executive agencies within central government as part of the Next Steps Initiative in the 1980s, RAB was a logical progression in the efforts to modernize the public sector (Treasury and Civil Service Committee, 1995;HMT, 1995;Wynne, 2008). In the case of WGA, HMT (1998) legitimized its implementation by highlighting other countries that had introduced consolidated accounting. With respect to IFRS, its promotion emphasized the need to follow private sector best practice (HMT, 2007a). The influence of the accounting profession illustrates the normative forces behind the decision to adopt RAB. Indeed, its implementation was facilitated by the recruitment of private sector accountants to the public sector (Mussari and Monfardini, 2010), increased use of private sector consultants (Connolly and Hyndman, 2006), and the establishment of the Financial Reporting Advisory Board (FRAB). Moreover, despite evidence suggesting that RAB had not achieved the benefits originally purported, WGA, IFRS, and CLOS were subsequently introduced, thus suggesting that RAB contributed to the UK becoming susceptible to ideas that are the taken-for-granted best way of doing things.
To conclude, based upon the previous discussion, it is evident that many countries have implemented new financial reporting systems and practices in a bid to improve how government plans and controls public expenditure. Prior research, which is typically based upon a single reform, reports similar institutional pressures and isomorphic mechanisms driving these reforms. In contrast, utilizing discourse analysis and semi-structured interviews to examine four high-profile UK accounting reforms, a jurisdiction that is considered to be a keen adopter of such innovations, introduced over a period of time, we can observe possible trends between the reforms and subsequently consider how they may have been institutionalized.

RESEARCH METHODS
For each of the four accounting reforms, the research underwent two stages ( Table 1). The first, categorized as the planning and testing stage and divided into three phases (pre-planning, planning, and dry run/testing), employed discourse analysis to identify what was being articulated as the claimed benefits of each reform. Discourse analysis is considered a beneficial way to understand the process of institutionalizing change, specifically 'how institutions are produced and maintained' (Phillips et al., 2004, p. 635). Discourse analysis was also utilized given its ability to reveal each reform's trajectory (Arnaboldi and Lapsley, 2009). Employing snowball 1 and relevance 2 sampling, 97 official government publications were identified and analyzed, with the sources including the House of Commons Parliamentary Papers database, the House of Commons Parliamentary hearings, and the National Archives (see Appendix A). The total number of observations shown in Table 1 is 132 as some of the documents examined referred to two or more of the accounting reforms.
As discussed above, the benefits of adopting the four reforms (i.e. RAB, WGA, IFRS, and CLOS) are based on the contention that they will lead to better information which will in turn improve efficiency and effectiveness and enhance accountability (Likierman, 1995;Ellwood, 2003;Heald and Georgiou, 2011;Hyndman and Connolly, 2011). Drawing on the literature and related publications, nine themes were identified (Table 2). Consistent with other public sector research (Ogden and Clarke, 2005;Lundbeg et al., 2009;Edgar et al., 2018), the documents were coded using a thematic analysis approach. 3 The coding rules were tested on a number of occasions to ensure that they enabled the authors to arrive at the same coding outcome, with the data being input and analyzed through SPSS in order to identify any trends or patterns within the themes. 4 Chisquare tests were performed to check for any statistically significant differences between observed and expected occurrences (see Appendix B for full results).
The second (post-implementation) stage combined discourse analysis with semistructured interviews (interviews hereafter), an approach adopted in other accounting (Dai et al., 2019) and public sector accounting-related (Ezzamel et al., 2007;Agasisti et al., 2008;Goddard, 2010;Harun and Robinson, 2010;   Interviews can complement and enrich the findings of discourse analysis (Gephart, 2004), allowing researchers the 'freedom to digress and capture individual perspectives and experiences that go back in time' (Narayan and Stittle, 2018, p. 508) and 'go below the radar in order to establish what's happening and how decision-makers view developments' (Heald and Hodges, 2018, p. 19). In our research, the discourse analysis enhanced our understanding of official (macro) assessment of the reforms at their pre-and post-implementation stages, with the interviews developing our appreciation of the extent to which the reforms have operated 'on the ground'. Using purposive sampling, 5 19 interviews were conducted with 23 representatives in 2014 from government departments and oversight bodies, together with academic advisors, in order to gain different perspectives on the accounting reforms (Table 3). A semi-structured interview guide was developed and used to inform the discussions, with the interviewees being asked their views on the following areas in relation to each of the reforms: the driving forces behind their introduction; their intended and actual benefits; the challenges faced during implementation; their main impact; and the extent to which they are considered a suitable way to account for public money. 6 Given the potential sensitivities of the matters being discussed, and the desire for participants to be as candid as possible, interviewees were assured that their comments would be reported in a manner where specific statements could not be attributed to particular individuals (see Table 3 for the codes used to identify each of the three categories of interviewee). Each of the interviews lasted between 45 and 90 minutes and was recorded with the interviewee's approval and transcribed. Leading to either behavioural change within the public sector; or Driven by international developments and influences Economy Addressing the wider economic context Effectiveness Achieving the intended outcomes (i.e., linking the organizational outputs and the objectives of the accounting reform) Efficiency Making the best use of resources (i.e., linking inputs to outputs) Private sector Applying private sector techniques Public sector accounting standards Amending private sector accounting standards to account for items found only within the public sector Transparency Improving the transparency of publicly available financial information Value for money Providing the best way to use public money for optimum net social benefits (i.e., benefit to the tax payer) Representatives from each interviewee group were only asked questions that were of relevance to them, an approach commonly used in interviews (Saunders et al., 2012). The questions and interview approach were designed to enable interviewees to express their views on the accounting reforms based upon their experience of working in, or engaging with, the public sector (Ezzamel et al., 2005). This is supported by (Saunders et al., 2012, p. 378) who highlight that the use of semistructured interviews can lead to 'discussions not previously considered but are significant' in gathering a better understanding of the phenomenon under investigation. Acknowledging the limitations of using semi-structured interviews and the debate over how many interviews are sufficient for the findings to be representative of the field (Dai et al., 2019), we sought to obtain interviewees with depth of knowledge. Interviewees from government departments (Table 3: D1-D9) were chosen based on their experience of the accounting reforms, including responsibility for preparing departmental resource accounts and contributing financial information for the purposes of WGA. These interviewees had an accounting background, held senior positions in their respective departments, and had been employed in the public sector for an average of 10 years. Eight of the nine departmental interviewees had 20-plus years' experience and so had been involved in the implementation of each of the four accounting reforms.
Oversight bodies typically advise government and Parliament on the implementation of (accounting) reforms, often reporting to government, Parliament, departments, and other bodies on how public money is spent. Of the 10 'oversight' interviewees (Table 3: OB1-OB10), five had 20-plus years' experience and involvement with all four accounting reforms. Of the remaining five, one of the government oversight interviewees was retired and had been involved heavily with RAB and WGA, while two each of the parliamentary and government oversight representatives had 10 to 15 years' experience and were more familiar with IFRS and CLOS. Their role as 'scrutineers' of the financial reporting procedures aids understanding of why these accounting reforms have been accepted. Lastly, academic advisors (Table 3: A1-A4) were chosen as they hold senior positions in their universities and have a strong track record of academic conference presentations and journal publications which demonstrate their indepth knowledge of the research area. Additionally, each academic interviewee has been involved (previously and currently) with various public sector committees as academic representatives and/or advisors (such as FRAB). Therefore, they were able to offer their views from both an academic and nonacademic perspective. As our research involves accounting reforms that span over 20 years, it was important to engage with well-informed respondents who could reflect on their experiences and perceptions of the accounting reforms. Thus, as illustrated above, the interviewees were holders of a long 'institutional memory' that enabled them to answer questions related to the period of the research (Narayan and Stittle, 2018, p. 509).

FINDINGS
As mentioned previously, the benefits of adopting RAB, WGA, IFRS, and CLOS are based upon the belief that they will deliver better information which will subsequently improve efficiency and effectiveness and enhance accountability. Drawing on IT, the findings from the discourse analysis and semi-structured interviews are presented below in terms of the two implementation stages referred to above (Stage One: Planning and testing; and Stage Two: Post implementation), with the implications of the main themes (Table 2) arising from the analysis being discussed.
Stage One: Planning and testing Accruals accounting is deeply anchored in the way we traditionally think about accounting. On the basis that those who fall short of the demands of logic are generally considered to be rationally defective, RAB, which was the first of the four reforms, sought to instruct how we ought or ought not to account and was the vehicle for a normative logic that has come to dominate financial reporting principles in UK central and local government. Our discourse analysis indicates that RAB (2,690) was the most frequently communicated reform in official publications at Stage One of its implementation (refer to Table 4 7 ). This was followed by CLOS (834), WGA (763), and IFRS (420). Our analysis also reveals that similar themes were prevalent at Stage One for each of the reforms, with rhetoric around Efficiency (1,965) dominating, followed by Effectiveness (1,115), and then Accountability (658) (with Culture (213) ranked third for WGA only).
We now consider each of these four themes. In this paper, we define Efficiency as the extent to which the public sector makes the best use of its resources (i.e., linking inputs and outputs) ( Table 2), with there being more of an emphasis on pursuing greater efficiency (Hopwood, 1988)    The shaded themes are the three identified as occurring most frequently for each of the accounting reforms.

A C C O U N T I N G R E F O R M S
Financial Management Initiative and Next Steps in the 1980s. Hence, it is not surprising that Efficiency dominates the discourse for all four accounting reforms at Stage One (Table 4), reflecting the underlying rationale and expectations for the changes prior to their introduction. For example: It has been a fundamental objective throughout that RAB should be developed in a way which leads to efficiency improvements. (Treasury Committee, 1997, p. 92) Effectiveness is the second most dominant theme in official discourse for RAB, WGA, and CLOS (and the third most dominant theme for IFRS) ( Table 4). This represents the extent to which the accounting reform achieves its intended outcome(s) (i.e., linking the organizational outputs and the objectives of the accounting reform) ( Table 2), with there being an increased emphasis on outcomes as justification for change more generally from the mid-1980s onwards. An example of Effectiveness at Stage One is: [RAB] offers the possibility of making effective decisions. (Procedure Committee, 1998, p. 17) Accountability is the third most prevalent theme in official discourse relating to RAB and CLOS (second with respect to IFRS and fourth for WGA) (Table 4). This theme captures how official publications report the manner in which the accounting reform will help to improve departmental accountability (Table 2). An example of Accountability at Stage One is: The move to IFRS from the 2008-09 financial year shows the Government's commitment to maintain high standards in public accountability and will place it at the leading edge of financial reporting internationally. (FRAB, 2007, p. 1) Finally, Culture is the third most dominant theme identified in official discourse for WGA (and fourth for RAB, IFRS, and CLOS). This captures references to the reform leading to either behavioural change within the public sector or having been driven by the mimetic forces (including those influenced by the wider institutional and political context) ( Table 2). An example of Culture at Stage One is: Experience from other countries suggests that there is a general, if gradual, move in the direction of preparing information on a whole of government basis, and that those countries which have taken this step have seen considerable advantages. (HMT, 1998, p. 5) Chi-square tests were performed to check for any statistically significant differences between observed and expected occurrences (see Appendix B for full results), with the results for Efficiency, Effectiveness, Accountability, and Culture summarized in Table 5. As indicated in Table 4, Efficiency dominates official discourse for all four accounting reforms at Stage One, with the highest number of occurrences relating to RAB (1,047). With respect to Efficiency, Table 5 shows that when comparing RAB with WGA (row 2, column (2)), the references to Efficiency in the documents analyzed are significantly greater (at the 5% level) for RAB. This suggests that official discourse promoting the implementation of RAB referred to the benefits of Efficiency more than what would be expected when compared with the introduction of WGA. As RAB is often cited as among the most radical of accounting reforms to have been implemented within UK central government (HMT, 1994), this reveals the emphasis that was placed on initially promoting RAB as the 'logical' way to account, perhaps in an effort to legitimize this 'first' accounting change. With RAB 'promising' to overcome the inefficiencies of the prior system of cash accounting and facilitate better decision making, it paved the way for future accounting-relating reforms.
When comparing RAB to both IFRS (Table 5, row 2, column (3)) and CLOS (Table 5, row 2, column (4)), the number of observations communicating Efficiency is significantly greater (at the 1% level) than expected in publications documenting IFRS and CLOS. Efficiency is considered a key characteristic of NPM ideas (Lapsley, 1999) and these results suggest that the subsequent introduction of IFRS and CLOS was guided by an institutional rationality that emphasized their potential Efficiency benefits. However, given that the Stage One discourse relating to IFRS and CLOS was published after the introduction of RAB, this justification for adopting IFRS and CLOS may have been on a more ceremonial basis given that evidence that RAB (which had been fully implemented by this stage) and WGA (which was intended to be operational for the 2005/06 financial year but was delayed to 2009/10) were not delivering such benefits to the extent originally intended was materializing. Indeed, their  (2)), the references to Efficiency in the documents analyzed is significantly greater (at the 5% level) for RAB.
introduction may be indicative of institutional pressures for further accounting change that had cascaded across the UK central government following RAB's introduction, with the legitimacy of RAB and WGA being supported through coercive mechanisms such as the Finance Act 1998 and GRAA 2000. Table 4 shows that promoting Effectiveness was the second most dominant theme within official discourse for RAB, WGA, and CLOS, and third for IFRS. While in numerical terms, Effectiveness receives the most attention in relation to RAB (645) compared with the subsequent reforms, Chi-square tests indicate that the actual number of occurrences for Effectiveness is significantly greater (at the 1% level) than expected with respect to WGA (Table 5, row 3, column (2)). Thus, given that Efficiency was not a 'significant' theme (at the 1% level) for WGA (Table 5, row 2, column (2)), claims of improved Effectiveness dominated. As RAB was going through its testing stage at this time, evidence was emerging that it had not improved efficiency to the extent originally thought. Hence, in order to prevent this adversely impacting on WGA's legitimacy, official discourse may have emphasized how it could enhance effectiveness (rather than efficiency).
Accountability has the second highest number of observations in official discourse for IFRS and the third highest for RAB and CLOS (Table 4). Like Efficiency and Effectiveness, Accountability features heavily in RAB (455), with Chi-square results indicating that the actual number of occurrences of Accountability is significantly greater than expected with respect to RAB when compared to WGA (Table 5, row 4: column (2)), IFRS (Table 5, row 4: column (3)) and CLOS (Table 5, row 4: column (4)).
Numerically, Culture, which represents references to the reform resulting in either behavioural change or having been driven by mimetic influences (Table 2), is referred to most frequently in official discourse with respect to RAB (121), receiving the third highest number of observations for WGA (79) (Table 4). However, Chi-square results reveal that the actual number of references to Culture is significantly greater than expected with respect to WGA in comparison to RAB (Table 5, row 5, column (2)), IFRS (Table 5, row 5, column (5)) and CLOS (Table 5, row 5, column (6)). As WGA was initially discussed in the 1994 Green Paper which addressed the introduction of RAB, but not progressed as HMT believed it would not be of 'practical value' (HMT, 1994, p. 11), it is conceivable that mimetic forces were employed to promote WGA by highlighting the benefits of consolidated accounting experienced in other countries (many of which were active adopters of NPM ideas) (HMT, 1998;Chow et al., 2007).

Stage Two: Post implementation
While RAB dominated official publications pre implementation, WGA received the most attention within official discourse in Stage Two (Table 6 8 -1,026). Indeed, with the exception of WGA, reference to each of the reforms declined post implementation. Given that discourse is often viewed as a transmitter of  The shaded themes are the three identified as occurring most frequently for each of the accounting reforms.

A C C O U N T I N G R E F O R M S
institutional change, caused by action and requiring constant maintenance (Phillips et al., 2004), the fall in emphasis post implementation supports assertions by Bromley and Powell (2012) that after adoption, evaluation and monitoring processes are often lacking or ritualistic. Moreover, as the reforms, or at least the notion of accounting reform, had become embedded, this potentially lessened the need for justification and made it easier to overcome any resistance to change.
Within official discourse, the three most dominant themes in Stage Two were Efficiency, Effectiveness, and Accountability (Table 6). Similar to Stage One, Efficiency dominated the discussions, with it being the primary theme for RAB, WGA, and IFRS and the second most prevalent for CLOS (Table 6) (Effectiveness is the most prevalent theme for CLOS at Stage Two-see below). Official discourse at Stage Two, admittedly on the basis of reduced references in comparison to Stage One, continued to reinforce the legitimacy of these reforms albeit without providing evidence. For example: Accruals-based information in providing more comprehensive and complete cost and performance information is a powerful tool to identify scope to improve efficiency and make better use of resources [in reference to RAB]. (Comptroller and Auditor General, 2003, p. 13) Again, Chi-square tests were performed to check for any statistically significant differences between the observed and expected occurrences (see Appendix C for full results). A summary of the Chi-square test results for Efficiency, Effectiveness, and Accountability (the three dominant themes) is presented in Table 7.
Although there was an increase in the number of documented references to Efficiency in the case of WGA at Stage Two compared with Stage One (Table 6: from 260 to 401), Chi-square tests reveal that, in comparison to the other reforms, this was not statistically significant (Table 7). With the exception of WGA, given the reduction in Efficiency-related references between Stage One and Stage Two, this suggests that the changes brought about by the reforms were more symbolic than real and the desire or need to continue pushing the same rhetoric in official discourse had lost its appeal by this stage or was no longer considered necessary given the reforms were in place. If so, this illustrates how once (accounting) actors become embedded in an institution, resistance to change can be more easily overcome (Battilana and D'Aunno, 2009), especially if reformers (e.g., institutional entrepreneurs) have power and politics on their side (Weber, 1978;Fleming and Spicer, 2014). This is supported by the fact that the overwhelming consensus across each interviewee group was that, of the four reforms, whilst RAB had improved Efficiency this was definitely not to the extent suggested in official publications. For example, perhaps given their day-today hands-on experiences, departmental interviewees were most vocal in contesting whether it facilitated better use of resources.
RAB was meant to help us [departments] tie our inputs to outputs to outcomes so that we could make better decisions on how to prioritize our funds in a more efficient manner. While it's helped in some way, because resource accounts are backward  For each of the accounting reform comparisons, the accounting reform indicated is the reform for which the differences between observed and expected occurrences of the theme is significantly greater at the level stated. For example, when comparing RAB with WGA at Stage One, the references to Efficiency in the documents analyzed is significantly greater (at the 5% level) for RAB.
looking, it doesn't achieve that objective. That comes more from the internal management accounting side and even then there's more that could be done. (D7) However, most departmental interviewees did concede that RAB was delivering more than just ceremonial or symbolic benefits, especially if compared to cash accounting.
With cash accounting it was just money in and money out. It never drove you to think about the future. But with accruals, it drives us to think about the implications of the decisions we make. (D2) Similar acknowledgements were forthcoming from the academic advisor interviewees group. For example: I believe that as a result of RAB, there is now a fresh impetus to look to the medium and long term and have some measure of how these obligations such as investing and revenue expenditure are affecting the organization. (A3) Likewise, albeit without the benefit of evidence from post-implementation reviews, oversight interviewees generally spoke positively of RAB. Reflecting on their experiences, the perception was that evidence of greater Efficiency in decision making was emerging, and that it would encourage key decision makers to think beyond the term of the Spending Review.
Back in the 1950s when we built nuclear power stations, nobody considered that in the future they would have to be decommissioned and made safe. The latest estimate puts that cost at about £70 billion. Had we been doing resource budgeting back then, we would have asked that question and it would have influenced our nuclear strategy. But because it was cash, all they cared about was how much it was going to cost to build the nuclear power station. The PAC [Public Accounts Committee] regularly highlight this example to demonstrate how accruals has helped us be more efficient. (OB1) As mentioned above, Effectiveness was found to be the most dominant theme in official discourse for CLOS at Stage Two (Table 6) (and the second most prevalent for RAB, WGA, and IFRS). An example of Effectiveness at Stage Two is: Just over a quarter of the 46 departments and agencies we examined had made good progress in implementing accruals-based accounting and were using it effectively. (Comptroller and Auditor General, 2003, p. 6) While Effectiveness occurs most frequently in official discourse for WGA (Table 6: 245), when compared against CLOS (Table 7, row 3, column (11)), the number of actual occurrences is significantly higher than expected for CLOS (at the 1% level). As CLOS is the only accounting reform to have undergone a formal post-implementation review (HMT, 2012), the relative emphasis on CLOS may be the result of government needing to reinforce its legitimacy post implementation. Like Efficiency, the consensus among interviewees from each group was that the accounting reforms had been effective, but again not to the extent originally claimed. Of the four reforms, RAB and CLOS were highlighted as being the most effective, with improvements in professionalism being the reason cited most frequently by the majority of interviewees from each group: I would say RAB has been the most effective so far, mostly because of the professionals it brought into the department. Having those skill sets in the department has been a huge benefit. (D3) This is consistent with Falkman and Taggesson's (2008) post-implementation observation that, had Sweden gained sufficient professional support, its move to accruals accounting would have had a greater impact. Another area of improved Effectiveness which was acknowledged by the majority of interviewees in each group, but particularly oversight interviewees, was the availability of 'better' information, something that was claimed for all reforms in official discourse at Stage One. For example: These accounting reforms, particularly RAB and CLOS, have contributed in their own way to giving us better information. For example, we now have a better understanding of the public sector's asset base and short-and long-term obligations like clinical negligence and PFI. Departments are now more aware of what's going on in their arm's length bodies and how their money is being spent. (OB8) However, reflecting on original claims of the extent to which RAB would enhance Effectiveness, all interviewees considered them unrealistic in practice (perhaps with the benefit of hindsight) given the complexities of the public sector, thus further supporting the contention that the benefits were more ceremonial than functional: Does it [RAB] help us to achieve our outcomes? Maybe a little, but nowhere near to the extent that was expected when it was being pushed through. In the public sector, outcomes are very complex and not related to finance. For example, here, one of our outcomes is maintaining political stability. Yes, some of our measures are finance based, but this can't capture everything. (D5) Across each of the interviewee groups, there was clear consensus that WGA's legitimacy remains the most questionable of the four accounting reforms. Promoted upon technical considerations, such as facilitating departments in their capital planning and resource allocation (HMT, 1998), departmental interviewees were WGA's strongest critics, with all contending that it offers little or no benefit in achieving departmental outcomes. These findings are consistent with earlier studies which report that WGA has not delivered to the extent originally intended (Chow et al., 2019). This suggests that, from a departmental perspective, the institutionalization of WGA has failed as it is deemed inappropriate from a technical and ceremonial perspective and continues to be legitimized through coercive pressure only. However, while academics proffered similar views to departmental interviewees, they were more inclined to acknowledge its potential to offer better information at a macro level and thus be more beneficial at governmental and parliamentary levels. For example: WGA has had a bad reputation from the start because of the time it took to get it implemented, the time it takes to actually prepare and the audit qualifications. I think it was a difficult exercise that Treasury didn't put enough resources into. WGA didn't have the strict schedule that RAB had, which sent out a signal that it wasn't terribly serious. For departments, it's nothing more than a box ticking exercise. But WGA will be of benefit to government and Parliament. (A1) As a group, oversight body interviewees were also divided on WGA's effectiveness. For example, some claimed HMT over-endorsed WGA's benefits and that a clearer, more realistic perspective was preferable: I question how effective WGA has been so far. It has been a massive operation and has taken considerable effort and money. PAC's feeling is that its purpose and who would use it should have been more clearly thought through. HMT highlighted certain issues like long-term liabilities and here it does have a value. But they [HMT] need a more refined focus of what they are trying to do with it. (OB4) However, perhaps not unsurprisingly, the body responsible for preparing WGA contended that concerns over WGA are being addressed: We were able to foresee some issues, but others were unexpected. For example, a major issue is the time we allow departments to submit their information and the quality of it when received. But it's getting better. We have reduced the production time from 19 to 15 months and we are down to three audit qualifications, at the start we had five. Now that we have a few years of information, we can do some trend analysis to better understand what's going on with our overall financial picture. It's getting better but there's more to do. (OB3) Consistent with Stage One, Accountability is a dominant theme in official discourse for all four reforms at Stage Two. An example of Accountability at Stage Two is: In general, respondents said that the relationship between departments and arm's length bodies was strengthened because the project [CLOS] had brought to the surface a number of unclear accountabilities and relationships which had gone unchallenged. (HMT, 2012, p. 7) Similar to Efficiency and Effectiveness, Accountability was discussed predominantly in official discourse for WGA (243) ( Table 6). Moreover, the findings of the Chi-square test results show that the actual number of occurrences of Accountability was significantly greater than expected (at the 1% level) with respect to WGA when compared to RAB (Table 7, row 4, column (7)) and CLOS (Table 7, row 4, column (11)). These results suggest that, given the negativity surrounding WGA (Committee of Public Accounts, 2012; HMT, 2016a), official discourse at Stage Two focused on emphasizing WGA's potential Accountability attributes in order to sustain legitimacy for this reform. While all interviewees acknowledged that the accounting reforms had enhanced Accountability, this was attributed mostly to RAB and CLOS, with departmental interviewees acknowledging that these two reforms had fostered better departmental accountability to Parliament: Prior to CLOS, we were budgeting for the likes of foundation trusts as if they were in our boundary but financial reporting as if they weren't, which meant that financial reporting systems were not supporting our budgeting requirements. Arm's length bodies had an element of financial independence as they were only controlled by cash. But now we have to consolidate them into our departmental accounts. This has brought more accountability to Parliament as we are now responsible for them. (D5) Each of the interviewees from oversight bodies expressed similar views.
Departments are now required to have more control over their arm's length bodies. As they are now consolidated, they have to ensure that financial management is spot on. The finance officer in every department is now accountable to Parliament for their performance whereas in the past they weren't. We have a much better grasp of government as a whole, of what people are doing. Through CLOS, departments are now held to account. (OB9) Despite consensus that the accounting reforms have not delivered the benefits to the extent originally claimed at their pre-implementation stage, consistent with Scott's (2008) cognitive-cultural pillar which emphasizes the legitimacy stemming from implementing a common frame of reference, the majority of interviewees from each group acknowledged that the reforms were appropriate (or legitimate) as they had instilled a set of ideas and values on which behaviour and decisions were now based.
Communication is key. There will be different terminology used between the public and private sectors. But accounting is just a technically minded language and IFRS is just as good as any language to communicate with. So, in that respect, it has most certainly been an appropriate move. (OB7) Departmental interviewees in particular commented that more work needs to be done as users of departmental annual reports and financial statements often find the documents difficult to understand, limiting their use from an external accountability perspective and also for improving financial management more broadly. However, most added that this was a general problem and was not confined to public sector accounting and reporting.
The accounts are too complex. A small number of people actually read and, more importantly, understand them. Something is not quite right with how we do the accounts and it's not confined to the public sector. (D8) The most common issue referred to was in relation to IFRS disclosures, specifically, that they are too extensive, with the requirements needing to be applied more sensibly: We have gone overboard with disclosures. For example, with financial instruments, departments that don't have any are still hammering out 10-15 pages, but to Treasury [HMT] that makes sense. I'm not sure that's a problem with IFRS, I think it's how we apply it. (OB2) To address this, HMT (2013) published a consultation paper with the aim of simplifying and streamlining departmental annual reports and financial statements. Consequently, to 'better meet the needs of the users of the accounts' (FRAB, 2014, p. 2), departments are required to structure the traditional 'fronthalf' annual report and 'back-half' financial statements into three integrated sections based on: performance, which includes a clear statement of the purpose and activities of the organization, high-level financial information with cross references to the audited accounts, and trend information with commentary against trends and performance against policy; accountability, which includes a governance statement and information on strategic risks to the entity, the remuneration report and information on parliamentary accountability; and financial statements, the format of which is based on the WGA reporting structure (HMT, 2016b).

DISCUSSION
To date, research has tended to examine a single reform and/or the microinstitutional effects before or during implementation, albeit with the focus more recently on questioning whether NPM remains the central motivation for reform (Funck and Karlsson, 2019). Our findings support the proposition that, from a macro position, NPM-inspired reform remains a desired approach for 'improving' public sector accounting (Hyndman and Lapsley, 2016). We reveal that official discourse institutionally embeds the need for reform on the basis of technical considerations by applying a distinct rhetoric of Efficiency, Effectiveness, and Accountability. This pattern of promotion continues into the post-implementation stages contrary to emerging evidence that the reforms are not delivering improvements to the extent officially purported. Moreover, while the same themes are prevalent post implementation, the extent to which they arise in official discourse fades considerably, suggesting the follow-up processes were more ceremonial than real. However, if it is assumed that the ultimate goal was to reform public sector accounting and reporting then evaluating the outcome of the reforms may lead to a different conclusion, with there being fairly widespread support and arguably only WGA receiving general criticism.
To demonstrate, RAB, the first of the four reforms, was institutionalized through technical considerations on the basis that it would improve the Efficiency, Effectiveness, and Accountability of the public sector. Post implementation however, the extent to which official discourse referred to these benefits declined considerably. Indeed, whilst documented departmental experiences are sparse, the evidence indicates that the intended benefits did not materialize to the extent originally suggested. While it might be expected that the experiences with RAB would dampen expectations for subsequent reforms, when WGA, IFRS, and CLOS were announced, these reforms promised similar improvements in Efficiency, Effectiveness, and Accountability (at Stage One), with the emphasis in Stage Two discourse diminishing for each subsequent reform (with the exception of WGA).
This divergence between pre-and post-implementation discourse may be explained by decoupling, where institutional pressures to adopt an approach conflict with what is actually experienced in practice (Piña and Avellanda, 2018). These findings indicate that official discourse post implementation was distanced, but not completely decoupled from institutional support for these NPM-based reforms given the actual outcomes do not appear to have met initial expectations. Pre-implementation publications legitimized the introduction of these accounting reforms by focusing on the potential for improved Efficiency, Effectiveness, and Accountability and thus, consistent with Christensen's (2005) notion of 'phantom images', contributed to a post-implementation belief that the reforms were acceptable and appropriate without necessarily being based upon evidence. Given the lack of post-implementation reviews for RAB, WGA, and IFRS 9 to ascertain the extent to which the anticipated benefits had been achieved, this supports Humphrey's (2005) contention that while such reforms are strongly endorsed at the beginning, this fades post implementation.
While interviewees endorsed this view, they acknowledged that the reforms had impacted positively, albeit not to the extent claimed in pre-implementation discourse. Departmental interviewees, as preparers with first-hand experience, considered RAB and CLOS to have delivered the most by way of improved Efficiency, Effectiveness, and Accountability, notably that these reforms had: (i) provided better information on areas that were once opaque; (ii) introduced a new level of professionalism to the public sector; and (iii) encouraged further consideration of the longer-term implications of decisions. However, departmental interviewees in particular were unconvinced about WGA. The discourse analysis 9 A post-implementation review of CLOS reported that it had improved the quality of available information and contributed to a better understanding of financial activities between departments and their arm's length bodies (HMT, 2012).
revealed that, unlike the other reforms, WGA was more heavily documented at Stage Two, perhaps reflecting a need to continue to push its potential benefits in the face of ongoing audit qualifications and skepticism over its widespread appeal. This will likely continue until the surrounding issues are resolved and there is greater clarity and acceptance on how it is being used. This has implications for policymakers. Given the current economic uncertainty amidst the ongoing COVID-19 pandemic, reporting government expenditure and liabilities at the macro level is arguably more important than ever, as evidenced by the release of a COVID-19 recovery strategy report (HM Government, 2020). In this regard, WGA has the potential to support longer-term fiscal analysis given that its data support the Office for Budget Responsibility's annual fiscal sustainability report and provides a reconciliation with the national accounts (HMT, 2019b). Therefore, policymakers should exercise care to ensure that those involved (either directly or indirectly) perceive the changes as beneficial.

CONCLUSION
By combining an analysis of official discourse on RAB, WGA, IFRS, and CLOS with the post-implementation experiences of representatives from government departments and oversight bodies, together with academic advisors, this paper examines the basis on which these four reforms were introduced, the extent to which they are considered appropriate, and the reasons for the persistence of NPM-inspired accounting reforms in UK central government. Our research complements existing literature from this perspective by exploring the pre-and post-implementation institutional effects of accounting reform over a 25-year period. In doing so, the study makes a four-fold contribution. First, by comparing official discourse before and after the implementation of the reforms, it appears that UK central government legitimized their introduction by promoting their anticipated benefits, even when there was strong evidence that previous reforms had not (yet) delivered the positive outcomes originally suggested. This supports Lounsbury's (2008) contention that while reforms may initially be driven by technical considerations, over time institutional reality and stability guide behaviour.
Second, this paper extends Lapsley's (2008) assertion that reform leads to more reform, irrespective of the success of previous related reforms. This can be explained by the interdependency of coercive, mimetic, and normative isomorphic forces which help to drive through and maintain legitimacy for NPM ideas. In the UK, mandated procedures impose the application of accounting reforms at parliamentary, governmental, and departmental levels. Global support for public sector accruals and consolidated accounting, together with the adoption of IFRS or IPSAS, reflects mimetic forces that will remain prevalent while NPM concepts are perceived as appropriate, with the transition from RAB to WGA using IFRS indicative of increasing normative support and reliance on accounting and accountants in the public sector. While the institutional and political context in some countries does not provide similar support and commitment (Oulasvirta, 2014), these findings are still relevant for audiences beyond the UK given that NPM ideas remain widely accepted. Relatedly, the findings also have application beyond NPM-inspired reforms as (public sector) modernization often begins by offering new inspiration for managers but then encounters resistance when the principles are put into practice (Nabatchi, 2017), with this perhaps explaining the challenges that programs suffer in delivering their original ideals. For example, in a public sector context, those promoted to alleviate public administrative bureaucracy or facilitate new public governance openness may experience a similar trajectory whereby a vision is initially heavily endorsed but then experiences (unintended) implementation problems and questionable levels of success (Ingrams et al., 2020). Further research into the causes of this disparity and what works best in specific circumstances is encouraged.
Third, this paper contributes to theory by providing potential evidence of isomorphism in longitudinal contexts as it helps to address possible temporal variation while providing a more complete understanding of the factors associated with support for NPM reform. Most prior research has explored the isomorphic tendencies of single public sector reforms. Our findings, which explore four reforms pre and post implementation over a 25-year period, offer further insight into how reform, induced by isomorphic mechanisms, develops over time. For example, we find that in the UK there is a tendency for accounting reform to develop along an institutional maturity curve beginning with coercive and mimetic pressures through to normative acceptance. This is evidenced from our discourse analysis which revealed that the need for the rhetoric of Efficiency at the postimplementation stage declined, perhaps indicating that NPM ideas had become embedded in the roots of UK public sector accounting. However, it is accepted that this conclusion is based purely on the four reforms included in this study and that wider research would be beneficial.
Lastly, this paper addresses calls for further empirical research that explores the perceptions of these accounting procedures in practice. As highlighted by Hammerschmid et al. (2019, p. 401), reporting reform outcomes tends to be founded on generalizations about 'broad trends rather than empirically informed observations'. Furthermore, prior literature has typically emphasized the negative aspects of NPM-related policies on public sector practice. We shed light on what is, or has the potential to be, positive aspects of these reforms. Whilst acknowledging the UK focus of our research, with the findings showing the persistence of NPM, this empirical evidence 'may help to sustain the momentum for change' for other countries going through similar reforms to their financial reporting (Hyndman et al., 2019, p. 606).
This study is not without limitations. It does not evaluate whether the reforms have been successful (however that may be defined). While it is accepted that the reforms have not accomplished what was claimed at their pre-implementation stages, it is acknowledged that they have been beneficial in certain respects. In response to the Swedish and Canadian cases, the UK situation indicates that the practice of introducing (accounting) reform, when coupled with coercive accountability and a strong support for normative professionalism, has an impact upon the public sector's cognition that goes beyond the ceremonial, the ritual, and de-coupling (Falkman and Taggesson, 2008;Pollanen and Loiselle-Lapointe, 2012). But, as highlighted in the Finnish case, where reform of this nature was resisted, 'if the governmental accounting culture is deeply rooted, it is difficult to change it' (Oulasvirta, 2014, p. 282).
We acknowledge that IT and isomorphism may not fully explain why government persists with its support for NPM ideas, nor do the theories necessarily offer a complete understanding of the impact of the reforms at the micro level. Therefore, care and caution should be exercised when conducting research on how institutionalization is diffused, particularly at the micro level. Furthermore, as official discourse is written in a different context to this research, the use of themes to analyze how the reforms were discussed is necessarily subjective (Krippendorff, 2013). Moreover, the discourse analysis only draws upon official publications and therefore excludes other sources, including academia and the media. As this research views the implementation of the accounting reforms through the eyes of technically proficient and knowledgeable actors (accountants/ oversight bodies/academics), further research engaging with additional numbers of similar actors and private sector consultants who have been involved in the implementation of accounting reforms, together with Members of Parliament, could be advantageous and lead to additional insights.