Accrual accounting representations in the public sector—A case of autopoiesis

Critical Perspectives on Accounting

journalhomepage: www.elsevier.com/locate/cpa

Accrual accounting representations in the public sector—A case of autopoiesis

Levi Gårseth-Nesbakk ∗ Bodø Graduate School of Business, Bodo, Norway

article info

abstract

Article history: This paper is based on a case study of an accrual accounting change endeavor at central Received 18 May 2009

government level in Norway. The case represents alternative accrual accounting solutions Received in revised form 13 March 2010

and thought. Earlier studies have discussed the (hyper) reality of accrual representations. Accepted 26 October 2010

This study contributes by examining why and in what form alternative accounting thought emerges, and what this means for accounting theory. The discussion stands out from prior

studies by relying on autopoiesis as a frame of reference and centering on public sector Accounting

Keywords: Accruals

as the empirical basis of inquiry. This enables a new way of understanding alternative Reality

accounting thought, founded on interrelated and complex organizational structures with Autopoiesis

blurred boundaries and potentially multiple identities. Such a form of autopoietic account- Entity

ing is argued to represent something new, while still maintaining and representing old Public sector

structures and identities. This paper illustrates e.g. how the debits represent something new and interrelations with other accounting environments, whereas the way of account- ing for credits helps to preserve internal authority, capital and organizational structures. Documents and interviews constitute the empirical material.

© 2010 Elsevier Ltd. All rights reserved.

1. Introduction Accruals (particularly accrual accounting) have dominated financial reporting and calculations in the private sector for

many years. One important reason seems to be its assumed superiority in calculating profit measures (on the basis of subjective assessments of expenses and revenues) and predicting future profit and cash flows (see e.g. Riahi-Belkaoui, 2004 ). By so doing, the accounts are said to represent economic activities and position, at a given time, as well as over time. Yet, researchers have also questioned accrual representations. One example is the potential for misusing the inherent subjective assessments to manipulate the profit calculation (see e.g. Chabrak and Daidj, 2007; Macintosh, 1995; McMillan, 2004; Rezaee, 2005 ).

Accrual 1 accounting has become more commonly applied in the public sector despite certain experienced difficulties with accruals in the private sector (as mentioned above). ( OECD, 2008 , p. 4) argued that:

“ . . . the emerging consensus among OECD countries is to adopt full accrual accounting and financial reporting while continuing to budget on cash, or only to apply accruals to a limited number of transactions in the budget.”

∗ Tel.: +47 755 17 695. E-mail address: levi.nesbakk@hibo.no 1 For a definition, see e.g. IPSASB (2009) .

1045-2354/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi: 10.1016/j.cpa.2010.10.004

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

This shows that accruals are much applied in the public sector, but also found difficult to employ for all purposes (see also Baker et al., 2010 ). This further relates to a critical stance in public sector accruals research literature (illustrated by Arnaboldi and Lapsley, 2009; Carlin, 2005; Carnegie and West, 2005 ), often cautioning or critically assessing the use of accruals in the public sector. It has been argued that accruals bring about unwarranted (negative) effects, whereas argued accrual benefits often do not (sufficiently) materialize (e.g. Newberry and Pallot, 2005; Paulsson, 2006 ).

Still, this matter appears as even more pressing when considering that some researchers have taken the critical discussion of accrual representations a bit further. For instance, when discussing the accounting profession, Macintosh and Shearer (2000) put forward Baudrillard’ ideas of simulacrum, implosion and hyperreality. Relying on (at least partly) the same frame of reference, but focusing more explicitly on the ontological status of information in accounting reports, Macintosh et al. (2000, p. 13) stated:

“Our major thesis is that many accounting signs no longer refer to real objects and events and accounting no longer functions according to the logic of transparent representation, stewardship or information economics.”

The debate over accrual accounting representations, not only in the private sector, but increasingly also in the public sector, signifies complexity and illustrates difficulties associated with accrual based constructions. Still, despite the criticism of accruals in both the private and the public sector, there is a paucity of research looking more closely into cases where alternative directions and accounting thought are pursued.

The purpose of this paper is therefore to explore what alternative accrual accounting representations look like. In so doing, this study briefly readdresses the referred discussion on hyperreality in accounting. Still, the particular aim of this paper is not to engage in-depth philosophical discussions, but rather to contribute by examining why and in what form alternative accounting thought emerges, and what this means for accounting theory.

This paper draws on insight and evidence from the public sector in Norway and its recent experience with exploring and developing accrual accounting solutions for subordinated central government agencies. The empirical material is based on document studies as well as interviews. The documents cover the time period of 2001–early 2009. Twenty-six interviews involving forty interviewees were conducted during the time period of late 2004–early 2007. However, they related to

a larger research project, meaning that not all of the interviews were equally relevant to the discussions in this paper. Nonetheless, they were all oriented towards the accrual accounting pilot project in Norway. This paper continues by presenting the frame of reference, and thereafter the Norwegian case. Subsequently, the empirical material is discussed, and later concluded upon. The latter also entails suggestions for further research.

2. Frame of reference Since accounting representations is the topic of discussion, the frame of reference seeks to clarify the paper’s perspective

on accounting representations. Research literature related to this topic is therefore presented in Section 2.1 . It is thereafter shown (in Section 2.2 ) how difficulties related to constructing accounting representations can be understood as a case of autopoiesis. 2

2.1. Accounting representations Accounting representations, or the perceived need for them, relate closely to the functions of accounting. The latter is

said to necessitate accounting representations to provide (especially external) decision makers with information about the organization or entity. This reduces uncertainty relating to decision making and control, herein accountability relations (see

e.g. Mellemvik et al., 1988; Oguri, 2005 ). As such, it is no surprise that representational faithfulness 3 has been regarded as a core analytical perspective on accounting standard setting (see e.g. Ruland, 1984; Shapiro, 1997 ). It has therefore been subjected to accounting theorizing, herein relating to conceptual frameworks, like that developed by FASB (see e.g. Hendriksen and van Breda, 1992; Robson, 1999; Wolk et al., 2004 ).

Macintosh and Shearer (2000) severely challenged the traditional view on representational faithfulness, namely that accounting functions as a mirror of organizational activities. Relating to Baudrillard’s work, Macintosh and Shearer (2000, pp. 612–613) , relied on three concepts in their discussion; namely simulacrum, implosion and hyperreality:

“Simulacrum refers to any sign, image, model, pretence, or shadow likeness of something else. The accounting model, for example, is a simulacrum of the entity. Implosion is any situation where the distinction, difference, or boundary between two or more entities, concepts or realms disappears, melts, dissolves, or otherwise collapses . . .

Hyperreality . . . refers to the condition where simulacra are no longer associated with any real referent and where signs, images and models circulate detached from any material reality.”

2 Autopoiesis was also discussed by Hikaka and Prebble (2010) , but primarily in relation to the (nature of) law. The focus in this paper is different, where autopoiesis relates more to ways of organizing and accounting (herein accruals). 3

The notion of representational faithfulness has also been critically discussed in relation to the use of metaphors in accounting ( Amernic and Craig, 2009; McGoun et al., 2007; Morgan, 1988; Morgan, 1997 ).

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

In addition, it is helpful to describe the simulacrum as the copy without an original, as noted by Baudrillard (1994) . Based on this, accounting representations may be seen as hyperreal if there is no correspondence between the accounting numbers and the underlying organizational activities or associated financial transactions. This interpretation is therefore slightly different than primarily considering hyperreality in relation to finance (as done by McGoun, 1997 , and later work done by Macintosh and others). Also, Robson (1999) provided a related discussion by critically commenting on the changing relationships between accounting and economics.

Moreover, years ago, the focus in accounting literature shifted away from a measurement- and technical accounting orientation. It was pointed out that accounting should not merely be prepared for its own sake, but rather relate to decision makers’ need for accounting information ( Riahi-Belkaoui, 2004 ). According to Ruland (1984, p. 224) , it is not sufficient that the accounts faithfully represent reality, rather; “they must be accurate descriptions of the economic phenomena of interest to users”. Yet, it is important to keep in mind that many standard setters have chosen to treat accounting users as

a homogeneous group of rational economic decision makers ( Young, 2006 ). These decision makers are, according to Young (2006, p. 596) , often portrayed as being;

“ . . . primarily concerned with economic events and transactions and with predicting their impact upon an entity’s future cash flows, future profitability and future financial position.”

In addition, accounting users are frequently assumed to be competent and well trained in accounting matters ( Wolk et al., 2004 ). In continuing, an alternative way of understanding accounting representations is presented, by means of the notion of autopoiesis.

2.2. Accounting representations as autopoiesis Organizations are usually thought of as being open. Yet, opposing views do exist. For example, by referring Maturana

and Varela, Morgan (1997) portrayed a more “closed” organization view. This is a useful frame of reference when discussing accounting representations and accounting entities. The content and reasoning are outlined below.

The basis for arguing the “closed” organization view is related to three core concepts, namely autonomy, circularity and self-reference, which enable systems to self-create or self-renew ( Morgan, 1997 ). The term autopoiesis refers to the ability for self-production through “closed” system relations. . . . “The aim of such systems is ultimately to produce themselves; their own organization and identity is their most important product” ( Morgan, 1997 , p. 253).

The argument put forward by Maturana and Varela is based on the assumption that; “living systems strive to maintain an identity by subordinating all changes to the maintenance of their own organization as a given set of relations. They do so by engaging in circular patterns of interaction whereby change in one element of the system is coupled with changes elsewhere, setting up continuous patterns of interaction that are always self-referential. They are self-referential because

a system cannot enter into interactions that are not specified in the pattern of relations that define its organization. Thus,

a system’s interaction with its ‘environment’ is really a reflection and part of its own organization. It interacts with its environment in a way that facilitates its own self-creation; its environment is really a part of itself. . . . An understanding of the autopoietic nature of systems requires that we understand how each element simultaneously combines the maintenance of itself with the maintenance of the others. . . . changes . . . are produced by variations within the overall system that modify the basic mode of organization” ( Morgan, 1997 , pp. 253–254).

However, it should be noted that this is not so much a discussion about whether an organization is fully “closed” or open vis-à-vis its environment/surroundings. Rather, this is more about whether, how, and where, organizational boundaries may

be drawn. In an accounting context, this immediately makes the accounting entity a pivotal matter. This has to be specified in order to know what to exclude and include in financial reporting, so as to fulfil the functions of accounting (see Section

2.1 ). Nonetheless, how does autopoiesis materialize into practice; especially where alternative accounting thought is pursued? The Norwegian case is presented in the following. It serves as a basis for discussing implications of accounting change in

the public sector, with respect to the above standing theoretical frame of reference.

3. The Norwegian case The description of the Norwegian case starts with background information and the status of an accrual accounting change

endeavor at central government level. This is followed by an outline of the newly developed accrual accounting solutions.

3.1. Background and status of the accrual accounting endeavor For decades, cash accounting has been favoured at central government level in Norway. However, a commission was

established 14th September 2001 by the Norwegian central government. This commission was given a mandate to investigate the implications of using the accrual principle in central government budgeting and accounting, as well as considering the usefulness of multiyear budgeting. The commission delivered its report 27th January 2003 ( The Ministry of Finance, 2003a ).

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

3.1.1. Ex ante clarifications In association with the presentation of the central government budget documents for 2004, the central government, rep- resented by The Ministry of Finance (2003b, p. 110) , reviewed the Commission of 2001’s assessment and recommendations, as well as the related submissions, and stated that:

“The Commission of 2001 suggests that accrual accounting ought to be carried out by following a more or less pure accrual model and as far as possible by relying on private sector accounting standards. It was also proposed to be applied in all central government agencies, and both in the central government’s budget and accounts. The Ministry of Finance’s view is that a more pragmatic attitude is in order, regarding the question of to what extent and in what form accrual accounting should be used.”

This also related to The Ministry of Finance’s (2003b, p. 109) summary comments and suggestions, as presented below:

“The budget appropriations should still be made on a cash basis. A corresponding cash account is therefore needed. Accrual accounting should be implemented at central government level in a form and to an extent that derives from an overall assessment of associated costs and benefits. Accrual accounting should furthermore be implemented in several steps, so that, after each step, it is possible to assess whether it, from a cost/benefit perspective, is justifiable to continue to another step. The first step would be to develop a set of standards on an accrual accounting basis. This is going to be used in the bookkeeping for central government agencies . . . , on a trial basis. Consequently, whether to apply accruals in the central government budget and/or the accounts has not yet been decided. Rather, this will be assessed on the basis of the future experiences of preparing the government agencies’ accounts on an accrual basis.”

As such, by emphasising cost/benefit considerations and the step-by-step approach, the central government (herein the Ministry of Finance) seemed to disregard the commission’s recommendations for developing pure, consistent accounting solutions. Additionally, certain prerequisites were laid forth for the experimentation. It was made clear that appropriations were still to be made on a cash basis, and that the accrual principle was only meant to be a supplement to the current cash accounting regime. Furthermore, it was said that accruals should not imply any changes in governmental agencies’ degrees of freedom—in terms of steering and frame of constraints, herein regarding their economic or legal authority ( The Standing Committee on Finance and Economic Affairs, 2003 ).

An employee in the Ministry of Finance shed light on the recommendations given by the central government by pointing out that it is important for ministries and (especially) the Ministry of Finance, to follow up on granted appropriations. It is therefore a prerequisite that the macro level receives the same kind of information as before and that the micro level will find the new (accrual) information more useful than cash accounting.

Additionally, the following reasons were given: “The Swedes at first had plans to implement the accrual principle in their central government budget, but later recon-

sidered. Elsewhere, the experiences have also been mixed. For example, Australia still struggles, even though the country has worked for a long time on implementing the accrual principle in the public sector. Because of this, we wish to obtain more information before making such a far-reaching decision. . . . It does not seem as though the Com- mission of 2001 focused much on cost/benefit, making it thereby even more necessary to include such considerations in the pilot project . . . Moreover, if the interest is limited and the reform initiative is not necessarily perceived as important, then it is even more urgent to carry out cost/benefit considerations, so that one does not incur unnecessary expenditures by implementing a superfluous reform.”

Hence, also the international applications of accrual accounting made it attractive to stall a formal implementation in Norway, so that more information could be gathered. Lack of interest may also slow down or stop the implementation or initiation, and can in itself be regarded as a need for incorporating a cost/benefit approach.

Soon after these clarifications had been put forward, the accrual accounting project was established. Ten (later one more—totalling eleven) central government agencies were subsequently selected to form a pilot project, dedicated to devel- oping and experimenting with accrual accounting. The preparations for the pilot project started in 2004, followed by new accounting solutions being developed and experimented with (i.e. being tested). The experiences were then summarised during the first half of 2006, ending with the conclusion that too little was known in regards to matters like the perceived usefulness of the new accounting model and solutions ( The Ministry of Finance, 2006 ). Consequently, the piloting continued throughout 2008, ending with a new evaluation in 2009 (see the Government Agency for Financial Management, 2009 ). A decision in terms of the future role of the developed and tested accrual accounting solutions was disclosed in October 2009, as a part of the proposed central government budget outline for 2010. It was then concluded that no mandatory changes would result from the pilot project, except aiming at making the developed and tested chart of accounts from the project obligatory for all central government agencies within the next few years (see, The Ministry of Finance, 2009 ). This decision was made (at least partly) because of there being mixed experiences amongst users concerning the extent and form they found accrual accounting to be useful. On the basis of the formal cost/benefit orientation taken in the policy making, it then became difficult to make the accounting solutions compulsory for all central government agencies.

Two types of central government agencies participated in the accrual accounting pilot project, namely gross- and net budgeted agencies. Certain distinctions between them led to the development of different accounting solutions (as described below and in the appendices).

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

3.1.2. Key distinctions between the pilots—gross vs. net budgeted agencies There are noteworthy distinctions between different organizational forms in place at the central government level, which also impacted on the pilot project. The two most important groups in this respect are the gross vs. the net budgeted central government agencies. The gross budgeted agencies resemble traditional service producing units in the public sector, not yet being subject to any larger steering or accountability changes. Conversely, the net budgeted agencies have come closer to the private sector in their way of accounting, and with respect to their organizational autonomy. For instance, they have for some years applied modified accruals. Also, they are allowed to transfer a surplus from one year to the next and to take on externally funded projects (so as to increase their revenues and profits). In similar situations, gross budgeted agencies must normally settle with reducing their activities to meet the budget constraints.

The empirical description continues with an outline of some of the developed accrual accounting solutions. They also contain a few additional illustrations of differences between gross and net budgeted agencies.

3.2. Illustrative accrual accounting solutions The description starts with objectives, before moving on to the structure of developed accounting statements. The empir-

ical illustrations of idiosyncratic traits with the accounting model continue with the liability model, followed by particular calculations made for gross budgeted agencies, certain transparency motivated solutions, and thereafter certain areas of exception. Some bookkeeping illustrations are also given in the appendices, showing how some of the accrual constructions are tackled in practice.

3.2.1. Objectives The objectives of the development and testing of the accrual accounting solutions were said to be to ( The Ministry of Finance, 2008 ):

• Provide accounting solutions that would contribute to efficient resource utilization (herein by displaying information about the costs of service production), • Enhance comparability (across government agencies, sectors and countries), and • Provide a better overview of the central government’s assets.

3.2.2. Overview of the income statement and the balance sheet The first part of the income statement resembles similar reports in the private sector, but is distinguishable in certain areas (like appropriations being a part of the reported revenue). Also, the asset side of the balance sheet corresponds to the accounting items and reporting structure of (the Norwegian) private sector accounting regulation.

Conversely, the credit side of the balance sheet is distinct by means of employing the term agency capital rather than equity. The prior is generally not applicable to central government agencies, except those granted the opportunity to earn agency capital (i.e. a surplus). This possibility is for all practical purposes restricted to the net budgeted central government agencies. On the liability side of the balance sheet, one can also find a settlement section with the central government’s cash register, as well as an item called “non-recognized appropriation associated with fixed assets”. This constitutes a kind

of deferred revenue recognition (see Section 3.2.3 for details). Moreover, there is no pension liability in the balance sheet. These accounting items located at the credit side of the balance sheet statement stand out as being distinct from the private sector accounts and unique to the (Norwegian) public sector. This means that virtually the entire credit side of the balance sheet is public sector specific.

One can also find an unfamiliar section in the income statement that stands out. The income statement is namely con- structed in a way that yields a zero result as the end (bottom line) figure. When asked why this is so, one interviewee stated:

“This relates to accounting standard no. 10 (SRS 10). This is a technical exercise (carried out) to provide a complete model.”

The section (technically) deals with the settlement of appropriations. This basically means that the result is set to zero 4 and any difference (seemingly a surplus or deficit) is charged against the “settlement with the central government’s account (or simply provision for liabilities)” at the liability side of the balance sheet. Yet, one can ask; what is the underlying rationale for this accounting construction? Another interviewee elaborated as follows:

“Still, there is a principle trait with this solution as well. It is namely important to keep in mind that traditionally organized central government agencies, because of formal reasons, cannot earn a profit. The appropriation is after all granted with a specific amount, for a specific purpose, and relates to a specific year. It is a fundamental precondition that they (i.e. the central government agencies) are not to earn a profit . . . To the extent that they end up using

4 An exception is made to this adjustment if the surplus derives from extra revenue generated in relation to externally funded activities in net budgeted central government agencies.

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

more or less than appropriated, then this is regarded as accounts receivable or accounts payable vis-à-vis the central government’s cash register.”

Nonetheless, this was not the only specific accounting construction in the project. The liability model as described forthwith is another such example.

3.2.3. The liability model What is called the liability model holds an important position in the developed accrual accounting model. The liability model relates to the financing of the opening balance sheet that was prepared for the central government agencies early on in the pilot project. It basically says that the appropriate counter item for non-current assets (excluding financial ones) is a liability (as opposed to equity) representing deferred appropriations. Specifically, the item at the liability side, called “non-recognized appropriation associated with fixed assets” shall exactly equal the sum of fixed assets, except non-current financial assets. Whenever there is a change in non-current assets (like when such assets are sold, new investments are being made or in cases of depreciation and impairment), an equal amount is entered against the liability and revenue items

(in addition to the traditional income statement item entry). 5 This means that the model is result-neutral, thereby evading discussions of the appropriate funding level of central government agencies (caused by matters like a negative bottom line figure due to depreciation or impairment of non-current assets).

The underlying assumption behind the model is thus that, when central government agencies are granted appropria- tions that are used to buy non-current assets (except financial assets), the part of the assets not yet being consumed during that accounting period, constitutes a liability to the central government, as some kind of deferred service production. Sub- sequently, an equal amount is entered as appropriation based revenue, and the liability is similarly reduced by the same

amount, whenever a related expense 6 is recognized. In effect this represents a reversed matching principle. Notwithstanding the aforementioned, there are additional specific calculations being made with respect to gross budgeted agencies.

3.2.4. Specific calculations for gross budgeted agencies The accounts prepared by gross budgeted central government agencies (within the developed accrual accounting model) contain additional calculative accounting items. The pension area is one example. Whereas net budgeted agencies pay a pension premium to fund their involvement in the central government’s pension scheme, an exception is made for gross budgeted agencies. For them, a simplified solution has been put in place where a ministry pays the pension premium on behalf of the gross budgeted agencies (as a group). Despite not paying a pension premium themselves, the gross budgeted agencies are nonetheless a part of the same pension scheme. To restore comparability, they report a calculative charge, intended to approximate each gross budgeted agency’s pension premium. This simulates how the pension premium would have affected the agencies’ accounts, if they had paid for the pension scheme they are a part of. A calculative revenue is also reported (with the same amount) in order to avoid impacting the result figures, and to create a consistent model. Exactly the same is being done in regards to the associated employer fee on labor.

3.2.5. Transparency-oriented reporting Further down in the income statement one encounters a settlement section containing collection activities and adminis- tration of central government grants, both being cash oriented and so-called result neutral. These items represent third-party transactions—such as money collected from others to be given to the central government’s cash register. The same applies to grants administered by a central government agency, where it pays money to other parties on the basis of approved applications for such funds. Thus, despite not relating to the economic interest of the agency in question, as well as being result neutral, this section is nonetheless included.

Moreover, the abovementioned discussion (in Section 3.2.4 ), concerning how to account for pension schemes, also represents an attempt to increase transparency, by visualizing activities undertaken in central government agencies.

3.2.6. Areas of exception The description earlier (in Section 3.2.4 ), regarding the handling of pension premiums in gross budgeted agencies, rep- resents a deconsolidation case where the aggregated pension premium is deconsolidated from the central government’s accounts and to each agency’s accounts. This is only one out of several areas where central governmental organizing affected the accounts by making them less comparable (if not various calculative adjustments had been made to them).

Another illustration of the perceived need to incorporate calculative components into the developed accrual model relates to the central government’s decision to not engage in any insurance contracts. Rather, it will cover its losses when they occur. This created a hypothetical, technical accounting challenge in the pilot project in terms of comparability. It meant that during regular accounting periods, comparable organizations (e.g. in the private sector or in other countries), would (all other things

6 See appendices for bookkeeping illustrations. For example, when the assets are depreciated over the succeeding reporting periods, the assets are facilitating service production. The deferred appropriation based revenue is therefore recognized accordingly, along with the reduction of the associated liability (related to non-current assets).

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

held equal) have a higher expense level since they frequently engage in insurance schemes. As such, the Norwegian central government agencies could easily be perceived as being unjustified efficient, thereby eroding the comparability objectives of the accrual accounting project. This matter was debated in the project, but such a calculative insurance charge has not yet been developed and reported.

Furthermore, there is no pension liability in the balance sheet statement. Much of the reasoning made in relation to the insurance area applies also to this matter. Another area of exception was library collections (books, journals, reports, etc.). On the basis of materiality considerations, the standard developers ended up with excluding this area from being entered into the balance sheet as an asset.

Having presented various calculative charges, adjustments and report structures, as well as some exceptions, it is timely to ask what they represent and why they have taken a particular form.

4. Discussion The discussion starts off with a brief summary discussion of alternative accounting thought underlying the developed

accrual accounting solutions. From there, the discussion moves on to illustrating how challenging accounting representations can be seen as a case of autopoiesis, which is followed by contemplation about associated implications, centering on what the aforementioned signify for accounting theory.

4.1. Alternative accounting thought underlying the accounting model Different factors described in the empirical section above played a role in bringing forth alternative accounting thought.

Examples include (the strive for) comparability, the ambition to avoid steering effects (i.e. only desiring technical account- ing changes) and (different) ways of organizing. The accounting solutions were constructed so as to bridge the goal of

comparability and the disrupted relation between transactions 7 and activities 8 related to central government agencies (see for instance the description of the gross budgeted agencies). At the same time efforts were made (e.g. by the Ministry of Finance) to ensure that the accounting solutions would not bring about any steering effects. However, this was not feasible without developing alternative accounting thought, herein creating certain central government specific accrual accounting solutions.

One may question though what the various accounting constructions (illuminated and discussed in the empirical Section

3.2 ) implicate for accounting at a more abstract and theoretical level. Do they e.g. represent hyperreality, or merely more accurate descriptions of a complex public sector? When discussing representations in dispute between accounting and production, ( Lowe and Koh, 2007 , p. 952) stated that:

“It is these sometimes messy and inaccurate representations which enable control of complex and heterogeneous activities at a distance.”

The discussion continues by considering reasons that make accrual accounting representations challenging in the public sector, where insight is obtained via the concept of autopoiesis (as outlined in the frame of reference).

4.2. Challenging accounting representations—the case of autopoiesis Hitherto, this paper has illustrated and discussed alternative accounting thought. In the frame of reference, the concept

and meaning of autopoiesis was described. The aforementioned will be used when discussing how accounting represen- tations became difficult on the basis of the defined accounting objectives, ways of organizing (herein accounting entity structures) and governmental identities. Together these factors induced what may be coined a disruption between debits and credits. Illustrations and arguments follow below.

4.2.1. Autopoietic objectives and identities—disruption between debits and credits The empirical illustrations above show how attempts were made to create the accounts to be similar to other organizations (herein private firms) in terms of information about expenses in the income statement, and the asset side of the balance sheet. In short, one sought to achieve comparability in terms of debits (but not credits). On the other hand, it became important to portray the government as being dissimilar (others) in terms of capital structure, or at least to preserve its existing structure. Additionally, also revenues were (in part) made distinct from others (as evident by various charges discussed in this paper, herein as a part of the description of the liability model and pension revenue calculations). Consequently the credit side (as opposed to the debit side) became unique to the government in the developed accrual accounting model.

7 The term “transaction” basically refers to financial transactions, that is, amounts paid or received by an agency. These may be paid immediately, or deferred, i.e. to be paid or received later on. Until they are paid or received, they turn up in the balance sheet statement as accrued items. 8

The term “activities” connotes daily operations carried out by employees (in central government agencies), such as attending meetings, answering questions from the public, treating applications for funds, carrying out various controls and filling out forms.

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

Therefore, the newly developed accrual accounting solutions may be seen as palimpsest, rather than something com- pletely new or opaque. That is, despite the image of (representing) something new, the accrual accounting solutions also clearly represent something old and well established. The palimpsest metaphor 9 illuminates the way in which accrual accounting technology inscribes the Norwegian central government’s identify and the subordinated agencies’ (limited) degrees of freedom, as well as their responsibilities.

In this way, the accounting solutions may have given off mixed signals; one representing change (i.e. via debits), another depicting stability and fortification of the established (via credits). Change was envisaged via emphasizing efficiency as a goal (by means of pursuing improved decision making, founded on having information about expenses and a more comprehensive overview of the assets). Yet, it was made clear that no (formal) steering changes were going to take place. Rather, the accrual accounting development was only destined to represent technical accounting changes (as illuminated by the ex ante

clarifications above, in Section 3.1.1 ).

On such a basis, it seems demanding to embark on change in the handling of the debits, whereas the credits (like the capital, liabilities and revenues) remain unchanged or subject to government specific calculations and (relatively) strict control. This also ended up materializing the accrual accounting project. After all, it basically ended with little or no official changes being made, beyond making the developed accrual accounting standards freely available for central government agencies to implement on a voluntary basis. One reported reason for this outcome was that the project ended up with somewhat mixed experiences that did not convincingly justify making extensive mandatory changes in reporting practices amongst Norwegian central government agencies.

However, on a more theoretical level, the aforementioned clearly constitutes a case of autopoiesis. In the way described above, debits represent attempts to stabilize relations with the (country’s) “external” environment, like the private sector and the international accounting community (herein public sector accounting). This is clear by means of making the accounting reports very similar to these relational associates (consider e.g. the asset side of the accounting report or the effort made to provide information about expenses). Besides, this is made explicate by stating comparability across sectors and countries as being one overall objective.

On the other hand, the particular accounting constructions related to the credits represent ways of ensuring stable “internal” patterns, and control. This was for instance done by ensuring that the central government remains in control over the appropriations. Likewise, the same understanding emerges when studying the way in which calculative revenue charges were organized. They were namely set up to cancel out transactions—thereby ensuring result neutral accounts—and little or no changes. But why was control needed and how can autopoiesis shed light on the matter?

This aforementioned point relates closely to established traditions and practices at different government levels—namely the macro vs. the micro steering levels. Firstly, the government has many duties and roles. Secondly, the Norwegian central government holds a long tradition for its macroeconomic counter cyclical steering where control over cash is an essential steering tool ( The Ministry of Finance, 2003a ). The accounting solutions therefore reflect the need to preserve the macroe- conomic steering tradition, while equipping the micro steering level with more information (about the debits—like assets and expenses). Hence a multi-task government structure and set of objectives needed to be reflected and preserved. The accounting technology had to be constructed in ways that reflected this, and so it was. This finding resembles somewhat the discussion by Caccia and Steccolini (2006) . They argued that externally imposed changes hardly will be able to change routines, which are more easily changed when the accounting changes proposed are consistent with traditional values.

As for the Norwegian case, the accounting solutions constitute some kind of compromise between; on the one hand, the country’s relations with its environment, and on the other hand, the way in which this challenged “internal” and “external” identity orientations. In this way, the new set of accrual accounting solutions actually functions as a stabilizing mechanism, and is therefore; de facto, circular and self referential (see also Morgan, 1997 ). Given this and considering the autonomy aspect as well (herein the voluntarism basis relied on when constructing and applying the accrual accounting solutions), it is not very surprising that status quo remained.

Moreover, in this way, the accrual accounting solutions represent reality by means of representing the country’s relations with its (more traditionally viewed inner and outer) environment. However, the case for autopoiesis does not rest on governmental identity alone. This can namely also be related to ways of organizing and the structuring of accounting entities. This is illustrated in the next section.

4.2.2. Autopoietic objectives, organizing and entities—disruption between debits and credits

A standard view on accounting entity proclaims that the accounting entity first and foremost ought to represent its transactions (see e.g. Hendriksen and van Breda, 1992; Wolk et al., 2004 ). This was the image portrayed by the (old) cash accounts. Conversely, the developed accrual accounting solutions represent something else, namely operational activities carried out by government agencies. This was brought forward by the wish for also obtaining comparability between central government agencies—especially across net- and gross budgeted entities. This was a stated objective from the outset. The empirical material sheds light on the differences between the net- and gross organizing forms, illustrating for instance differences in terms of payment structures, even though activities carried out may be more or less the same in gross and net budgeted agencies.

9 For a discussion, see Joerges and Czarniawska (1998) .

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258

In the Norwegian case one therefore had to choose between making the financial transactions transparent or making the operational activities more transparent. The choice fell on the latter. Consequently, the accounting solutions were constructed to achieve this. This also meant that the newly developed accrual accounting solutions were self-referential, by means of their basis in existing organizational relations and structures. Rather than changing the existing ways of organizing, one tried to build accounting solution in relation to them, alternatively, around them. Since organizational differences existed

e.g. between gross and net budgeted agencies, additional and tailor-made debits and credits charges were constructed. Such accounting solutions were therefore autopoietic in their nature, since the new element (i.e. the accrual accounting) was made to fit with the existing (entities and ways of organizing).

4.3. Implications and insight from viewing accounting as a case of autopoiesis On the basis of the prior discussion, some deriving implications are pointed out forthwith. Firstly, the debate over whether

accrual accounting representations reflect reality is readdressed. From there the focus shifts to implications relating to what government level and accounting users to focus on.

4.3.1. Accounts do still reflect reality—but beg for holism In accounting literature it has been discussed to what extent accrual accounting represents reality (see e.g. Macintosh et al., 2000; Macintosh, 2002; Mattessich, 2003; Vollmer, 2007 ). For example, if applying the view 10 held by Macintosh et al. (2000) on the case presented in this study, some might infer that parts of the developed accrual accounting solutions are artificial—when considering the underlying transactions. This may happen because accounting constructions are charged to an accounting entity that has not paid an amount as specified, and is not expected to do so either. It can be argued that (at least some of) the accounting solutions become self-referential in the sense that they do not relate to any “real” referents, but are merely justified on the basis of each other. A good example is the calculative pension revenue. This does not refer to a real transaction, but is merely included on the basis of other constructions (in this case, particularly the calculative pension expense). The rationale given is that this (i.e. adding the calculative pension revenue) provides a complete model (in the world of debits and credits)—and was regarded as a necessary step to achieve comparability.

On the other hand, the empirical material in this study (see also the reasoning in Section 4.2.2 ) provides a basis for arguing against the notion of hyperreality in accounting. The latter namely emphasizes the importance of considering how accounting solutions relate to accounting objectives, organizing and accounting entity structures. In short this means that a discussion of reality vis-à-vis accounting representations is less meaningful (at best) without considering for what purpose the accounting solutions are being prepared. For instance, if the accounts had been prepared to only or primarily reflect financial transactions, then the reality of some of the accounting entries may be put into question. However, this is not so in terms of the described and discussed accounting solutions. They purport to represent activities and make them more transparent. After all, a frequently cited benefit of applying accrual accounting in the public sector is said to be increased transparency ( Carlin, 2005 ).

A discussion of the reality basis of accrual accounting representations therefore requires considerations of the larger picture. Hence, accounting solutions must be appraised with respect to the accounting objectives they are meant to fulfil and the circularity of accounting entries. That is; both the debits and the credits—as well as how and why they interrelate in the way they do, for different accounting events, need to be understood.

This means that the concept of autopoiesis again becomes helpful in understanding government accounting. It this case it points to the bigger picture, rather than fractions of it. It is not hard to justify that the accounting solutions, when considering the reporting package as a whole, represents a new and more comprehensive image of the agencies’ economic position and performance (e.g. by means of providing more information about expenses). In this case therefore, any discussion of hyperreality may only be reasonable (and perhaps not even then) when considering fractions of the accounts—as opposed to the larger picture and holism.

4.3.2. What government level and accounting users to account for? Despite the aforementioned struggle for comparability, adjustments to make the accounts as comparable as possible were not done in all areas. Areas of exception included e.g. pension liability and insurance expenses. The challenge in this case was determining to which accounting entity different charges or expenses should be allocated. For example, it was (for some time) undetermined whether it was more appropriate to treat pension liability as a consolidated or deconsolidated matter. The same matter concerned insurance-like payments. In other words, these matters were (at least partially) brought about by unclear accounting entity boundaries. This, together with materiality considerations (and some uncertainty in terms of who the primary users were supposed to be) instigated another area of exception, namely library collections (such as books, journal and reports). Standard setters decided, after some consideration, to treat library collections as expenditures rather than assets. This underlines the need to carefully consider what government level one is aiming for in terms of providing useful information. For instance, if emphasizing the agency level as the unit of analysis, and e.g. agency managers as primary

10 See referred text in Section 1 .

L. Gårseth-Nesbakk / Critical Perspectives on Accounting 22 (2011) 247–258