Governing culture: Legislators, interpreters and accountants

Cultural policy has become dominated by questions of how to account for the intangible value of government investments. This is as a result of longstanding developments within government’s approaches to policy making, most notably those influenced by practices of audit and accounting. This paper will outline these developments with reference to Peter Miller’s concept of calculative practices, and will argue two central points: first, that there are practical solutions to the problem of measuring the value of culture that connect central government discourses with the discourses of the cultural sector; and second, the paper will demonstrate how academic work has been central to this area of policy making. As a result of the centrality of accounting academics in cultural policy, for example in providing advice on the appropriate measurement tools and techniques, questions are raised about the role academia might take vis-a-vis public policy. Accounting professionals and academics not only provide technical expertise that informs state calculative practices, but also play a surveillance role through the audit and evaluation of government programmes, and act as interpreters in defining terms of performance measurement, success and failure. The paper therefore concludes by reflecting on recent work by Phillip Schlesinger to preserve academic integrity whilst allowing accounting scholars and academics influence and partnership in policy.


Introduction
Recent years have seen an explosion of interest in the practices associated with governing culture. Whilst questions associated with this area were traditionally the preserve of cultural studies, they have become important for scholars in a range of fields, including studies of accounting (Zan, 2006(Zan, , 2000Christiansen & Skaerbaek, 1997;Oakes & Oakes, 2012). Alongside the emerging attention given to accounting practices for cultural policy and cultural organisations, the role of accounting in restructuring the state has been an important area of study (Lapsley, 1999(Lapsley, , 2008 offer an overview). State restructuring, particularly along the lines of the New Public Management (NPM) has been an activity that provides an important contextual factor in justifying accounting's interest in the arts and culture. Finally the actual practices of accounting, what Miller (2001) terms the calculative practices found across a range of public and private sector settings, are the objects of analysis that can provide the bridge between arts management and accounting.
The rise of NPM has gone hand in hand with other changes confronting cultural organisations. Trends within government have, in turn, played out alongside trends within the cultural sector, as state-led reform of cultural funding interacted with challenges to the assumptions prevailing across the cultural sector. These assumptions include the role of the audience, the primacy of the artist, and justifications for funding culture. Moreover, the assumption within much of the literature on cultural organisations is of the distance between the accounting practices of the state and the values of the cultural organisation (although deconstructed by Zan, 2000). Even where there are overlapping discourses, for example the entrepreneurial aspects of NPM (Lapsley, 2008) and those of the cultural industries, these moments of dialogue are played down or denied.
The paper will consider these debates by placing them in the context of work on the state as an organ dominated by calculative practices. British government has attempted to fit culture into accounting frameworks by developing research partnerships with academics. The paper draws on the two reports that were the result of two phases of this academic work with the UK's Department for Culture, Media and Sport (DCMS). The paper contends that by understanding government as a site for measurement and calculation we can best make sense of the conflict between aesthetics and accounting practices that are at the heart of the modern government of culture. It presents a solution to this conflict grounded in a mixed-methods approach that accords both with British central government's demand for transparency through calculation and a recognition of the potential for the aesthetic to escape metrics.
Accounting professionals and academics not only provide technical expertise that informs state calculative practices, but also play a surveillance role through the audit and evaluation of government programmes, and act as interpreters in defining terms of performance measurement, success and failure. In this light the paper reflects on the role of the accounting academic and Zygmunt Bauman's discussion of the question of academia and intellectuals as legislators or interpreters. The paper concludes by reflecting on recent work by Phillip Schlesinger to preserve academic integrity whilst allowing academic influence and partnership in policy.

Aesthetics and accounting
Zan (2012, 3) asks why management or accounting scholars should be concerned with arts and cultural organisations. The answer to this question involves understanding what Vollmer (2003, 354, 363) describes as an 'alternative accounting discourse': a sociology of accounting created by accounting scholars that focuses on accounting, government and governance. This literature draws heavily on the idea of calculative practices developed by Rose and Miller (1992) and Miller (2001), which is in turn derived from the work of Michel Foucault on governmentality and of Bruno Latour on inscription and on networks. The central theme of this alternative accounting discourse is how economic governance and calculative practices shape social order (2003: 353). Miller (2001, 391) observes that 'The calculative practices and language of accountancy have seeped into everyday life'. The seemingly 'humble and mundane mechanisms by which authorities seek to instantiate government: techniques of notation, computation and calculation; procedures of examination and assessment' are in fact 'technologies of governance' (Rose & Miller, 1992, 183), the numerical glue that binds together the complex network of techniques and actors through which governmental programmes 'are articulated and made operable' (Miller, 2001, 239). Put simply, calculative practices, such as counting, evaluating and inscribing, are part of a common language shared across varied enterprises in the public and private sectors, and even within the domestic realm. These technologies of government impose norms by which individuals and groups regulate their actions in terms of setting goals and recording outcomes. Rose and Miller call this 'action at a distance', a cornerstone of modern government (Rose & Miller, 1992,187). Central government sets the terms by which various economic actors must calculate, and public funds are allocated or withheld within the network on this basis-most notably within neo-liberal regimes which 'seek both to create a distance between the formal institutions of the state and the other social actors, and to act upon them in a different manner ' (1992, 198, 199) so that '[c]hoosing actors may govern themselves by numbers' (Rose 1991, 691). It follows that while individual actors are apparently free to choose their actions, they are engaged within 'asymmetrical networks of influence and control' (Miller, 2001, 381).
In this context, 'one of the principal achievements of management accounting is to link together responsibility and calculation: to create the responsible and calculating individual' (Miller, 2001, 380). Through setting standardised targets, individual actors freely pursuing economic norms have become regulated, and thus 'accounting has become a body of expertise focused on enacting responsibility from individuals rendered calculable and comparable.' For Miller, the key defining feature of the calculative practices of accountancy is 'their ability to translate diverse and complex processes into a single financial figure ' (2001, 381) with the veneer of an objective, political neutrality (2001,382). Calculative practices extend across, and make commensurable, virtually all public, private and domestic realms, no matter how disparate: 'Not only can the manager of a global corporation be governed in this manner, but so too can a doctor, a schoolteacher, or a social worker ' (2001, 238), or indeed actors within the cultural sector. As Miller puts it, 'The management of almost any organization can be transformed into a complex of incessant calculations ' (2001, 381). We now turn to the cultural sector and examples of the 'alternative accounting discourse' that builds upon this literature.
In a case study observing the implementation of a new management control system at the Royal Danish Theatre, Christiansen and Skaerbaek set out to study how 'organizations dominated by a rationality very different from an accounting and administration perspective react when a rational budgeting system is put on the agenda' (1997,406). They highlight the influence of accounting systems on individual and group goal-setting and motivation (1997,408), and that accounting systems are not neutral frameworks that just need to be implemented, but rather are given meaning in, or are constituted by, their use (1997,410). Their empirical findings support the work of Miller and O'Leary (1993) which suggest that accounting systems lead to a 'focus on calculation of abstract figures representing the production process' and so 'remove management from production'. Therefore Christiansen and Skaerbaek believe that a major reason why artists tend to resist the implementation of accounting systems is that the disciplining power of calculative practices intrudes on artists' individual liberty and disrupts collective creative processes. Like hospitals, theatres are populated by 'highly trained employees motivated by their professional standards that mostly do not include concepts from the logic implicit in accounting' (1997,432).
In their study of accounting and marketing discourses in official documents relating to widening arts engagement in England, Oakes and Oakes discuss literature relating to the behavioural aspects of accounting systems. They view accounting 'as an enforcer of colonising ideologies ' (2012, 213). Specifically, they link accounting to what they call a modernist discourse bound to 'universalising and totalising knowledge claims ' (2012, 214). They believe that, 'Through its connections with scientific knowledge, accounting information is often presented as an objective fact, ignoring the social context in which it is constructed' and that 'the purported "facts" in financial statements often gain a status that is abstracted from time so they seem to suggest something definitive and absolute about an organisation ' (2012, 214). Oakes and Oakes state that modernist accounting is an 'influential discourse' due to its 'claims to absolute truth, certainty, and objectivity'. It follows that 'attachment of these claims to accounting concepts such as cost-consciousness and efficiency make it difficult to question such concepts, and hence they become naturalised and useful for political purposes ' (2012, 214). They mark the contrast with literature that argues that 'allocation processes are arbitrary and politicised' and note the work of Chwastiak and Lehman (2008) who 'highlight the human misery in which accounting is implicated when it is linked to the de-humanising economic rationality of modernism with its emphasis on the maximisation of profit (or extraction of surplus value) at the expense of other values ' (2012, 213).
One of the central problems confronting cultural policy is how the aesthetic aspects of culture can fit into public policy. Miller describes how the calculative practices of accountancy have become closely associated with the role of management 'to the point that they have become almost indissociable' (2001,. For Zan (2000, 337;2006, 9) arts organisations, confronting the rise of the forms of accounting found in government organisations, 'are professional organizations, where a substantive-aesthetic culture prevails over the culture of management, generic and a-specific in itself. In such a context, it is management experts who find themselves in a position of alterity, being perceived as "deviant", as witnesses to the clash between these two cultures the world over.' There are three particular areas of conflict, the first being that the cultural sector tends to perceive management and accountancy practices as an attack on 'the very core of their professional identity' (Zan, 2006, viii) and to view improvements in productivity as threats to artistic quality (Christiansen & Skaerbaek, 1997, 433). Second, many cultural institutions, such as the British Museum, La Scala and Pompeii, have for centuries been successfully 'managed according to nonmanagerial ways of managing' (Zan, 2006, 14). And third that, in the UK case particularly, the accounting profession is excessively heavy-handed in 'imposing rules and norms wrought in its own image' (Zan, 2006, 11;Zan, 2000, 338) and 'does not seem to know half measures' (Zan, 2006, 10).
In this light, when accountancy attempts to 'calculate creativity' (Christiansen & Skaerbaek, 1997, 248) it is the enemy of aestheticism. As public management has adopted numeric techniques such as cost benefit analysis, a distance has been created from the aestheticism that is prevalent in many views of what publicly funded culture should be and the actual practices associated with the internal running of cultural organisations and external accountability to funders, such as the state.
Recent work on this question (O'Brien, 2013a(O'Brien, , 2013bPrince, 2013;Stanziola, 2012) has highlighted the complex interplay between the differing forms of expertise present in cultural policy making. As Prince (2013, 1) points out 'being an expert on matters of culture is popularly seen as a matter of subjective judgement'. Indeed this is the position that much of the sociology of culture has attempted to problematize. For both Bourdieu (2010) and later writers, including Chiapello (2004), the notion of expertise in aesthetic judgement was bound up with individuals' and groups' location within the stratified societies of Western Europe and also a broader ideology that demanded a specific and special social role for the artist and their work.
At the time that questions of aesthetic value were central to cultural policy making, government had been reshaped by the trends of New Public Management (NPM) and its subsequent iterations of managerialism and Public Value. In the British context, of a Whitehall subject to almost 40 years of attempted reform, the effect of three influences, of Treasury, Public Choice Theory and Welfare Economics, resulted in a status quo that fixed the position of policy development within the framework of techniques such as cost benefit analysis.
In these circumstances, Whitehall policy making guidance, particularly from Treasury, framed policy as an activity of intervening and regulating markets, with the effects of such intervention to be judged via assessment of efficiency and effectiveness. In this mode of thought, the costs and benefits associated with any given policy should be compared with each other using a common standard or metric, which, in the case of cost benefit analysis was money. The emergence of the calculative device (Miller, 2001) of cost benefit analysis raised a complex question for cultural policy. If it was the case that the aesthetic aspects of culture demanded unique status, then how could culture, which in both anthropological and aesthetic senses denied its entanglements with the market, fit into public policy? To give one illustration, many cultural goods are free at the point of use in the UK, for example there is a longstanding policy of making entry into national galleries, such as the Tate, free of charge. In this context the price, paid for by general taxation, is not initially clear to the visitors. If price, or cost, is not clear to the user, the benefits, which in economic terms are expressed in the price someone is willing to pay for the ticket, are similarly obscured.
Before 2010 the state of the art in this area had used various proxies for the values of culture that were seen to be beyond market price. Beginning in the 1980s, both government and the cultural sector used estimates of social and economic impact. In those ways of thinking about culture, the value culture generated was approximated into economic outcomes, particularly in terms of tourist spend and media coverage, or social outcomes, such as community cohesion, or better educational and health outcomes.
Unease with these approximations extended across both government and the cultural sector in the UK, even as both employed these devices with increasing sophistication. Notable voices, such as former Secretary of State for culture, Chris Smith, and former head of the Barbican, John Tusa, were vocal about the potentially compromised policies that resulted from a slavish devotion to Treasury forms of decision-making. In particular, economic impact, as a way of trying to fit cultural policies into a cost benefit analysis framework, created objections from a vast range of commentators (e.g. Cowen, 2006, 15) who insisted that this risked reducing culture to a range of benefits that were better provided by other sectors of government intervention and failed to capture the full benefits of culture to individuals.
However, the quest to fit into Whitehall's NPM regime, coupled with the more recent retrenchment of public funding in the UK, has led both policy and organisations to pay closer attention to how NPM frames decision-making. In the UK this is through HM Treasury's Green Book, which provides guidance on how to conduct policy appraisal and evaluation. The text was introduced in 2003, as part of the twin pressures of the end of the first wave of New Labour's Modernising Government agenda and the continued long march of HM Treasury control over Whitehall (a long march hastened by the centralising agenda of Gordon Brown). The Green Book sets out a process for policy making to follow, moving through a series of stages to answer two fundamental questions: Are there better ways to achieve this objective? Are there better uses for these resources? These questions are answered by cost benefit analysis that seeks to translate a variety of costs and benefits, some with no market prices, into financial terms. The use of cost benefit analysis is part of the way the guidance frames decision making as an intervention into the market, whereby the decision to intervene must be to correct market failure in a way that creates more 'benefit' than any alternative (or, indeed doing nothing). The guidance is thus reflective of both an economic view of society, the application of social scientific techniques to decision making, and the difficult issue of valuing activities or practices that may resist pricing or market trades.
Thus the problem of valuing culture, in terms of how best to fit the unique aspects of culture, outside of the social and economic impacts, into the economic language of the welfare economic paradigm suggested by the guidance in government documents such as Treasury's Green Book, remained unsolved despite almost 30 years of trying.

Governing culture
The push towards forms of policy making that were made transparent by the use of social science is substantiated in the British Treasury's guidance on decision-making. The example of the British Treasury's Green Book offers a straightforward case study of how bureaucratic formations have changed since the 1970s, influenced both by market logics and economics, as well as critiques of the state. Burnham & Horton (2013) describe three elements of 'traditional' public administration: that politics and political decision making is separate from administrative work; that this administration was organised hierarchically, based on aptitude and merit, rather than personal connections, wealth or other forms of influence; and finally that this administration would have permanence, being recruited for life and exempt from the electoral cycle that inevitably led to politicians' limited time in office (Rush & Giddings, 2011). These three elements of traditional public management also encountered the application of scientific forms of management over the course of the twentieth century, as the management theories of modernity presented alternative organisational models (Boltanski & Chiapello, 2007).
The reasons for the adaptation, reform and rejection of the public management approach are highly disputed. The influential public management theorists Osborne and Gaebler (1992) described the failure of traditional public management to deliver outcomes, offering a narrative concerned with the effectiveness of this system. This goes hand in hand with narratives of the state suggesting it had become overloaded by the demands for intervention in, and management of, economy and society. There was also, from economics, a critique that suggested bureaucratic managers were just as selfinterested and just as political as their masters, seeking to maximise their bureaucratic empires at the expense of democratically elected politicians and the general public (McLean, 1987). This school of thought (Burnham & Horton, 2013) also suggested hierarchical bureaucracy was inefficient compared to market forms of organisation. Harvey (2007) and other critical scholars saw these reforms as part of the global turn to a neo-liberal state settlement in favour of already wealthy and powerful groups. However these concerns cut across the ideological spectrum, in the same way that critiques of modernity came from a range of intellectual positions. Dissatisfaction with traditional public administration was taken up widely by politicians from a range of different ideological positions during the 1980s and 1990s, figuring in right-wing critiques of the inefficiency and self-interest of the state as well as left-wing demands to powerful state institutions for greater democracy, transparency and accountability.
In the New Public Management the dominant form of organisation was the market as the type of social organisation that would not be subject to the problems associated with traditional public management. The management practices of market organisations were applied to public services at the same time as large swathes of public service provision were opened up to competition from private sector provision. As Osborne and Gaebler (1992) describe, administration was concerned with 'steering, not rowing' the state, in an effort to lower costs. The state retained control of this 'contracted-out' system of service provision by applying forms of measurement, considered against performance standards and objectives (Burnham & Horton, 2013). The tools and techniques of measurement were used to make the world of public administration as it responded to the contradictions, ambivalences and ambiguities of modernity.
Thus under NPM financial controls over the dispersal of public money shaped the environment in which economic actors operated. The government's ability to withhold funds prompted actors to adopt common calculative practices, maintain networks, and to demonstrate shared interests and strategies (Rose & Miller, 1992, 198). Throughout the 1990s, arts organisations dependent on full or partial public funding became 'an additional territory for the seemingly inexorable diffusion of managerial rhetoric', and so came under pressure to adopt the language and practices of NPM and to demonstrate 'value for money'. Albeit that this 'colonization' or 'contamination' happened 'reluctantly rather than willingly' and at a slower pace than within, for example, the education and health care sectors (Ferri & Zan, 2014, 368;Zan, 2006, viii, 11-12;Zan, 2000, 336-7).

Measuring cultural value
While this reading of public administration in modernity has important implications for culture, it also plays out in other areas of public life where the adoption of calculative practices has been contested. For example, there have been related debates about introducing an assessment of the economic and social impact of university research in the UK as part of the 2014 Research Excellence Framework (REF). This exercise decided the allocation of around £2 billion per year in block funding to higher education institutions, and impact criteria counted for 20% of the evaluation. The initial proposal from the Treasury was to use a metrics-only approach to gauge the impact of university research upon the economy and industry. However the metrics proposed, such as grants from business, and numbers of patents and spin-out companies created, were found to be very low order measures of such impact. The proposed model also centred on the natural and physical sciences and so neglected the humanities, arts, and social sciences, and had little relevance for gauging the wider social, cultural and economic benefits of academic research (Donovan, 2007).
After public consultation, and discussions with the academic community, the metrics-only model was abandoned. The Higher Education Funding Councils commissioned a report on international best practice (RAND Europe, 2009) and adopted a narrative approach to assessing research impact based on the design of Australia's proposed equivalent to the UK's REF, the Research Quality Framework (RQF), which used case studies informed, where appropriate, by robust indicators or metrics (Donovan, 2008). In this respect calculative approaches were adapted to employ mixed methods to capture more meaningful descriptions of impact. For example, broader social and cultural impacts were initially disregarded as being 'intangible' through the lens of economic calculation, yet were made tangible by adopting both quantitative and qualitative approaches. Critically, the RQF model on which impact assessment for the 2014 REF was based, was strongly influenced by lobbying from academics and their professional associations in the humanities, arts, and social sciences, arguing that robust evidence of the social, economic, environmental and cultural impacts of research could best be found using a combination of narratives and appropriate indicators or metrics (CHASS, 2005). The cases of the RQF and 2014 REF represent a moment in modernity where academics engaged with the policy making process and disrupted, opened up, and reshaped the use of calculative practices within government. One may, however, question whether the rejection of a metrics-only approach constitutes an escape from the calculative practices of government, as a mixed methods approach expands the terrain made calculable.
In light of these developments, we now turn to explore the limitations of the British approach to measuring culture, and then discuss an alternative holistic approach that synthesises capturing the unique aspects of culture with the calculative practices embodied in the Treasury's Green Book. Stated preference, revealed preference, and subjective wellbeing approaches are the government's preferred methods for assessing the economic value of culture (Fujiwara and Campbell, 2011;H.M. Treasury, 2003, 57-8), yet these are associated with various technical, practical and philosophical limitations when applied to the cultural sector (Donovan, 2014(Donovan, , 2013. We now describe these techniques and their key strengths and weaknesses. Stated preference and revealed preference techniques are used to estimate the non-market values of cultural goods and services, and have for some time been an important component of cost-benefit analysis (Fujiwara & Campbell, 2011, 7). Stated preference techniques are based on hypothetical future markets, and questionnaires are constructed to elicit people's maximum willingness to pay for particular cultural goods. A key strength is that in addition to use value, stated preference approaches also estimate non-use value, such as a person's willingness to pay to support a local library even though they may never intend to use it but gain satisfaction from the fact that other people may do so now and in the future. It follows that money values can be established for otherwise unobservable cultural benefits. This is of particular importance for the cultural sector, where in terms of the total economic value of a good required for social cost-benefit analysis, its non-use value may greatly exceed its use value. Relatively few stated preference studies have been applied to the cultural sector, particularly in the UK. There are concerns that respondents may overstate their willingness to pay in a hypothetical situation. Revealed preference techniques infer people's willingness to pay for a cultural good by observing actual behaviour in related markets (e.g. the differential cost of rental or housing prices related to proximity to cultural goods) or in consuming the good itself (e.g. the cost of theatre tickets and travel). But revealed preference techniques do not capture the non-use value of cultural goods and so may severely undervalue the total economic value of a cultural good.
Benefits transfer is an approach to economic valuation that takes preference-based economic values from a previous study (the study site) and then uses these values in a similar decision-making context (the policy site) and calculates new preference-based economic values to aid social cost-benefit analysis (Pearce et al., 2002, 16, 35). Its major appeal lies in the fact that the opportunity costs of conducting original studies are high and so the need to conduct preference-based studies 'would be vastly reduced' (Pearce et al., 2002, 35, 135). Cultural assets tend to be unique as they display 'complex characteristics, which may make value transfer subject to considerable error' (Provins et al., 2008, 142) because this heterogeneity clashes with the fact that benefits transfer relies on a stock of homogeneous goods. In other words, benefits transfer depends on the strength of similarity between the original study and the new policy site, yet unique cultural goods have no comparisons. For benefits transfer to be valid there must be a strong similarity between the study site and the policy site, and a sufficient pool of relevant preference-based studies to draw from, particularly when considering the diverse range and scale of goods and services the cultural sector offers across the arts, built heritage, galleries, libraries, and museums. Provins et al. (2008, 166) found worldwide 33 examples of preference-based studies applied to the valuation of cultural heritage, 11 of which were in the UK and relate to fairly distinct areas and had little consistency in types of data gathered. Yet this small number of cultural heritage studies represented the bulk of all such UK studies in the cultural sector. The fact that there are a very small number of preference-based studies applied to the UK cultural sector, and the fact that these are in distinct areas, effectively rules out the possibility of benefits transfer.
Subjective wellbeing approaches attempt to infer value from the relationship between wellbeing and income, and are an attempt to 'measure people's experiences rather than expose their preferences' (Fujiwara & Campbell, 2011, 7) and thus avoid having to ask people their willingness to pay. In terms of culture the aim is to be able to understand the value of cultural engagement through the relationship between wellbeing and income. It is, however, hard to claim that an increase in reported wellbeing is linked to cultural engagement or linked to a particular policy. And this is very much viewed as 'an evolving methodology and existing valuations are not sufficiently accepted as robust enough for direct use in Social Cost Benefit Analysis' (H.M. Treasury, 2003, 58;Fujiwara & Campbell, 2011, 5).
In practical terms, these approaches collectively require high levels of expertise, and are complex, time-consuming and expensive. The financial and opportunity costs required are beyond the means of the majority of cultural enterprises and projects, and so run against the principle of proportionality, i.e. the notion that the cost and effort required to supply data should be relative to the size of public funds requested. While it is appropriate to use economic valuation techniques to assess the benefits of large capital projects or national policies, this is impracticable for the majority of cultural sector organisations. In philosophical terms, the wholesale use of economic valuation techniques is unpopular as this grates against the cultural sector ethos, and more specifically there are concerns that the uniqueness and quality of the cultural goods in question cannot be represented by economic measures. However, appropriate or not, these techniques allow DCMS to render culture calculable and hence governable.
It follows that there is a paradox in DCMS's work, in the need to fit culture into the calculative practices of the Treasury, while facing resistance to these practices and the reality of their practical inappropriateness for smaller organisations. This paradox is related to DCMS's comparatively weak position in Whitehall (O'Brien, 2013a, 2013b), whereby capability reviews highlighted the lack of evidence-based policy making and capacity within the department. Moreover it is linked to DCMS's role as a distributor of funding to policy making quangos, in particular Arts Council England, that are more directly responsible for dealing with arts organisations. The arms-length principle (Hewison, 1995) that DCMS does not interfere in cultural or artistic decisions means the department is removed from the policy concerns of smaller-or medium-sized arts organisations whilst at the same time having to represent those policy concerns in the language of the Treasury.

Alternative approaches to valuing culture
The results of this paradox point towards the need for alternatives to a solely economic framework for valuing culture. There have been two academic projects with DCMS on measuring cultural value, and the Phase One project reviewed various measures of cultural value endorsed by the Treasury's Green Book (O'Brien, 2010). The Phase Two project tested the principle of adopting an additional range of quantitative, qualitative and narrative techniques that were accessible to the whole cultural sector (Donovan, 2013(Donovan, , 2014 and the result was unanimous cultural sector support for a holistic approach to valuing culture that used mixed methods which could vary according to context. The holistic approach was seen as being responsive to scale; and the possible combination of economic, other quantitative, and qualitative approaches was readily accepted on the grounds of being cheaper and more accessible, and also because this combination of methods captured a broader range of cultural value than standard government calculative practices. Treasury representatives in the Phase Two project workshops also agreed that in addition to economic valuation techniques, alternative methods, including narrative approaches, could enhance large bids for public funds by demonstrating a layer of 'added cultural value'. In this respect the recent £2 million investment by the UK's Arts and Humanities Research Council in its Cultural Value Project represents a timely initiative to develop and test a wide range of techniques for capturing the value of culture (AHRC, 2015).
A key argument against the use of non-economic approaches to valuing culture is that while this information can be considered within the context of social cost-benefit analysis, it ultimately sits outside the valuation framework: noneconomic approaches are less 'calculable' because they do not offer a simple monetary value. However, both Phase One and Phase Two of the academic projects with DCMS found that multi-criteria analysis (MCA) provides an alternative framework to guide decision-making processes for funding non-marketed cultural goods and services, and can be informed by a range of economic and non-economic data. Multi-criteria analysis (MCA) proceeds by creating a performance matrix, where rows describe options, and columns rate the performance of options against various criteria; and this rating can be numerical or even colour-coded. In a simple format, this matrix in itself may be the final product, and decision-makers are given this information on which to base their deliberations. More complex versions of this approach such as multi-criteria decision analysis (MCDA) use a specialised computer programme to convert the performance matrix into consistent numerical values (DCLG, 2009, 21). It is important to consider the viewpoints of various stakeholders or affected groups of people when deriving criteria, either through involving those parties directly, or conducting research into, for example, policy statements provided by interest groups (DCLG, 2009, 33).
The strengths of MCA for valuing culture are that it makes economic and non-economic data commensurable; it can be used to appraise the benefits of purely non-economic data; it can involve expert judgement on the selection and weighting of criteria; it can involve the opinions of various stakeholders and affected groups on the weighting of criteria; and it can be relatively simple to perform if complex weighting and scoring is excluded. Its limitations are that using weighting and scoring can be quite complex, and so needs relevant expertise, and may be more technical and time-consuming than social cost-benefit analysis; disputes may occur over the relative weights to be assigned to the criteria; it may not necessarily include data derived from people's preferences; and there is a need to ensure that weighting and scoring are not arbitrary.
The purpose of MCA within government is to allow a small team of people to assess options and provide advice to decision-makers. Like cost-benefit analysis, MCA is a decision-making aid that allows the appraisal of various options, however MCA differs in that unlike social cost-benefit analysis it can directly incorporate economic and non-economic data and make these commensurable, and it can also proceed solely with non-economic data. Crucially, MCA is endorsed by the Green Book for use when valuing non-marketed goods, and is increasingly being used to inform government decision-making in areas such as transport and environment policy, which take into consideration a combination of monetary and nonmonetary data, including qualitative information on impacts on the landscape, and 'as an "alternative" to defining monetary values for all the major costs and benefits when this is impractical' (DCLG, 2009, 5, 9). MCA provides a single framework within which it is possible to deal consistently with large amounts of complex information derived from using diverse methods. In this way, 'MCA techniques can be used to identify a single most preferred option, to rank options, to short-list a number of options for subsequent detailed appraisal, or simply to distinguish acceptable from unacceptable possibilities' (2009,19). In the UK the Department of the Environment, Transport and the Regions used MCA for the appraisal of 67 schemes in the Roads Review, including options for Stonehenge (DETR, 1998).
In terms of its application to the cultural sector, Provins et al. (2008, 133) discuss the use of MCA in the context of attempting to place a value on the non-use benefits of the preservation of cultural heritage assets when there is a lack of information on the monetary value of this preservation. They recommend that appraisal can nonetheless be undertaken: in its simplest form expert opinion would be sought, which would then feed into quantitative and qualitative assessments of the relevant issues using weighting and scoring. Examples of the application of MCA internationally also apply to heritage in the context of the sustainability of various proposed projects for the redevelopment of Venice Arsenale, where assessment criteria included intrinsic sustainability, context sustainability, and economic-financial feasibility (Giove et al., 2010); and the grading of 69 heritage sites in Calcutta, where assessment criteria included historical value, architectural value, sociocultural value, signs of deterioration, accessibility, integrity, public opinion, local response and usability (Dutta & Hussain, 2009).
The option of using MCA allows a broad articulation of the value of culture that is nonetheless endorsed by the Treasury. This removes the paradox in DCMS's work: there is general acceptance of mixed-methods approaches throughout the cultural sector and MCA can be applied by DCMS to large-scale cultural investments and by arms-length organisations to smaller cultural enterprises. The Phase One and Phase Two projects' recommendations that DCMS adapt its calculative practices to use mixed methods to assess the value of culture has clear parallels with the case of academic lobbying to use mixed methods to assess the impact of research in the RQF and the 2014 REF. In both cases academics engaged with the policy making process, with the potential to decentre and open up the use of calculative practices within government. In the DCMS's case, cultural value was rendered 'intangible' through the lens of calculative practices, yet could be made tangible by the use of MCA as an valuation framework that incorporates economic and non-economic data. It is ironic that DCMS's need to fit into the calculative practices of the Treasury has led to policies wedded to economic calculation, while a more fitting approach for assessing the value of culturemulti-criteria analysisis firmly embedded in the Treasury's Green Book and provides a compelling solution to this impasse. Yet while this compromise may be compelling from the viewpoint of arts organisations, DCMS, as a relatively small and vulnerable government department, has to date displayed a preference for economic valuation techniques that produce a monetary value, rendering culture calculable and governable within the network of government.

Culture and expertise
When applied to the cultural sector, the 'alternative accounting discourse' tends to regard arts organisations as 'fragile institutions' under the dual threat of the calculative practices of NPM and the excesses of management and accounting culture (Zan, 2012, 5). There is, however, a desire to find solutions to accounting for the arts 'but in modest doses' (Zan, 2000, 337) that balance planning and artistic freedom (Christiansen & Peter Skaerbaek, 1997, 429) and in 'more creative and less modernistic ways in order to encourage more emancipatory forms of accounting' (Oakes & Oakes, 2012, 221). While multi-criteria analysis presents a possible solution, this nonetheless remains a discussion framed by calculative practices: 'Even when individuals seek to subvert or avoid calculations made of them, their actions still take place in reference to an economic norm based on accounting numbers' (Miller, 2001, 393-4). Within sociological literature on states and calculative devices (e.g. Scott, 1998;Mitchell, 2002) there is a straightforward critique of the preceding discussion. The application of devices of calculation, such as multi-criteria analysis, is an application of the existing power of states to open up new territories to be governed, in this example the territory of the subsidised cultural sector. Alongside this new territory for government comes the commensurative effects of comparing different values within the same analytical framework, and subsuming the non-financial into the calculable. A critical position drawing on this literature might raise the spectre, as Bauman (2004) has done, of the inherent violence of reducing culture, the sphere that offered alternative visions of the social in the work of thinkers as diverse as Keat (1999) and Adorno, to both the realm of the bureaucratic and the realm of the market in one fell swoop. Bureaucratic control runs the risk of removing what is specifically beyond the state from within culture, whilst market production risks only offering the homogeneity of what will command an exchange value. The risk of submitting culture to multi-criteria analysis may then be to see it enveloped by the realms from which it is both distinct and might thus offer critique.
However, seeing the state as an aggregate of calculative practices, which is one reading of the tradition within which Scott (1998) and Mitchell (2002) write, may go too far from the specificity of individual departmental traditions and experiences that shape the impact and use of calculative devices. As Prince (2013, 14) argues 'blanket assessments that calculative reason instrumentalises or disciplines cultural activity conceded too much to it . . . it has not necessarily colonised the subsidised cultural sector'. To address this point, and by way of conclusion, as well as to offer a response to the risks of commensuration posed by the recommendation of multi-criteria analysis, the paper will now consider the specificities of cultural policy.
Whilst there is an extensive literature ranging across various issues in cultural policy (for an overview see O'Brien, 2013a, 2013b) the most pertinent work for the purposes of this discussion focuses on the role of accounting experts in cultural policy. This is because the recommendation as to the usefulness, or not, of multi-criteria analysis is bound up with the position of academic expertise within cultural policy making. It is also bound up with a series of structural factors, not least of which is the peripheral and weaker status of the UK's ministry of culture vis-à-vis the Treasury, which make the application of that expertise worthy of consideration.
By way of an example it is worth considering the contradictory status of expertise and evidence within British cultural policy. In the first instance cultural policy is, perhaps more than any other area of policy, the clearest example of policy-based evidence making (Slater, 2012). Selwood (2002) identifies the irrelevance of data collection and the use of statistics to the practices of cultural policy, whereby an industry of Whitehall civil servants, consultants and academics has been bound up with the production of data that has served little or no purpose in decisions over cultural policy. Indeed the fate of the work of Miles and Sullivan (2012), which was directly critical of foundational cultural policies including the subsidising of free entry to national organisations, support Selwood's analysis. Miles and Sullivan's report, written as part of the same stream of DCMS research as the Phase One and Phase Two measuring cultural value projects, did little to alter the insistence by both Labour and subsequently the Coalition administration on a 'deficit model' of cultural policy aimed at those from marginalised social and economic groups.
If cultural policy is, therefore, an area of government activity that has shown little concern with allowing evidence to influence policy making, why is the role of expertise of interest? It is important because, despite the problems associated with the effectiveness of evidence and expertise, cultural policy has been highly dependent on evidence and expertise for a new form of legitimacy as an area of government policy in the UK, then subsequently worldwide, since 1997. This is because the notion of creative industries has been the driving force for the reorientation of cultural policy as potentially central to policymakers' attempts to found new forms of economic activity that were not associated with Fordist manufacturing that had prevailed in the twentieth century.
Statistical expertise, driven by the academic status of many of the individuals who contributed to initial mapping documents that helped to define creative industries, has been crucial. Campbell (2013) has shown how both the definition of what a creative industry is, or is not, was totally dependent on the ability of, in the UK, the Office for National Statistics and other government data-gathering agencies to monitor and measure specific forms of economic activity. Moreover the global career of creative industries and the associated success of the term in inserting itself into a range of policy contexts that had little or nothing to do with the British, Whitehall, model of politics and public administration, was directly dependent on the ability of creative industries to offer growth for the British economy. This offer of growth was bound up in definitional debates that yoked software design to more traditional cultural activity such as the visual arts. Thus cultural policy's story in recent years is of a contradictory rejection of large amounts of data produced by the application of governmental monitoring systems to institutions such as museums and galleries at the same time as a dependence on highly questionable data on the economic efficacy of the role of culture and creativity.
What then should be the role of the accounting academic if their recommendations are both demanded and ignored? One way of conceiving of this is to follow recent ethnographic work in Whitehall (Rhodes, 2011;Stevens, 2011) that has shown how uneven and almost random the take-up and use of academic evidence in government can be. Moreover, one might observe the dynamics of policy making, particularly its speed and need for clarity and certainty, and conclude that any settlement with policy is to subordinate and thus damage the function of the academic in contemporary society.
However, the latter position is unsatisfactory. In a recent paper Schlesinger (2013) draws on the work of Zygmunt Bauman to discuss the role of the academic in cultural policy making. For Bauman (1992) the academic (or more specifically the intellectual) is caught between the promise of their expertise to grant them the power to make definitive pronouncements on social issues, to legislate,and the postmodern suspicion of legislative certainty that offers the academic the opportunity to offer interpretations, particularly of competing narratives. Schlesinger (2013, 34) finds a path between these options in two ways, which are worth directly quoting at length: 'First, under ideal, unconstrained conditions, our research practice would lead us, as a matter of course, to contribute to discussion and deliberation through various forms of public engagement. This might and does include the production and publication of independent research. It might also entail academics joining boards and commissions, supplying expert advice to governments and agencies, advising parliamentary committees, making submissions to public inquiries, offering assistance to civil society organizations, contributing to media and so forth' And another model that reflects the auditing regime of British universities, the REF and the so-called impact agenda, whereby academics are forced into a 'necessity-driven, demand-led model. Ideal-typically, you produce research and engage in knowledge exchange in line with what is requested in order to justify your existence. Public intellectuality, therefore, is wanted but only on certain, quite instrumental, terms' In both regimes the academic is still potentially acting as legislator. However it is an action constrained by both the question of the legitimacy of the academics to act in this fashion, pace Bauman, and the question of motivation. If academics are driven only to contribute expertise to policy because they are expected to do so as part of funding agreements and bureaucratic monitoring, then by definition their findings may be suspect. This is not to comment on the robustness of any given academic interaction with government. Rather, it is to point to the need for a continued space for academia to create understandings of the social world that deal with the dilemmas facing policymakers. The discussion of multi-criteria analysis is one such example, grounded as it is in a recognition of the structural constraints facing policy making, for example Treasury power and the role of a specific form of balancing the costs and benefits of interventions.
Such considerations resonate with discussions of the role of expertise, calculative reason, and practices of accounting in the governmentality literature (Miller & Rose, 2008;Rose, 2001). On the one hand, accounting professionals and academics provide the technical expertise that informs state calculative practices, the effects of which extend beyond government to invisibly permeate the everyday lives of citizens. It is this 'know how' that makes government possible through the 'calculated administration of diverse aspects of conduct' (Rose & Miller 1992, 175-7). On the other hand, accounting scholars provide a surveillance role through audit and evaluation of government programmes and (potential) recipients of government funding. For example, the expert has gained an important social role, and is perceived as 'embodying neutrality, authority and skill in a wise figure, operating according to an ethical code' and being 'beyond good and evil'; yet the figure of the expert is held to have evolved as a bridge between political will and regulating private actions (1992,187). Expertise and accounting is therefore not neutral, but provides a means of inscribing activities so that they can be evaluated and governed. Accounting scholars and academics also act as interpreters in, for example, framing and reframing the terms of performance measurement and what counts as success or failure, and in the context of this paper what constitutes cultural value. In this sense, academics can create 'enclosures' of concentrated skills and authority, and thus challenge political authority (1992, 188).

Conclusion
The above discussion returns us to the question of whether cultural policy is unique or special, or are the dynamics under consideration in this paper merely reflections of how states and their governments function under our current modernity? Comparative work across other areas of public sector life is clearly necessary. But taken as a case study, cultural policy points to the reality of calculative practices as both useful for decision-makers, bearers of risks and threats to the objects of calculation, and finally sites whereby the complicated and uncertain position of the modern accounting academic can be worked out. It is in the position of the academic, their role and function with regard to state power, which we can hope to monitor the effects of calculation, as Prince (2013), Stanziola (2012) and Schlesinger (2013) have done with regard to cultural policy, whilst opening important reflexive positions for the profession. The future, under REF, will, as Schlesinger suggests, be one of an enforced usefulness for academics working with policy. The task for the accounting academic is to create this relationship on terms of their own, by applying their expertise as we have sought to do in this paper, as opposed to being captured or ignored by the ideologies and the techniques of government. Martin (2010) outlines the mutual benefits and risks that can emerge when academics engage in research with practitioners in government. He presents a spectrum of research that is co-produced by academics and practitioners, ranging from practitioners being relatively passive informants to being highly active in the commissioning, conduct and oversight of research projects. While higher levels of engagement increase the likelihood that research will be utilised (or have 'impact') there is also a higher risk of the research process becoming politicised. With regard to academic engagement with cultural policy making, Donovan elsewhere reflects on how the findings of the Phase One and Phase Two measuring cultural value projects were received within DCMS (Donovan, 2014, 24-6). While both projects recommended the use of multi-criteria analysis, Phase One was a review of potential techniques for measuring cultural value, and was perceived to support DCMS policy. Phase Two focused on translating Phase One recommendations into policy, and was initially criticised on the grounds that its recommended holistic approach ran counter to Green Book requirements, although a subsequent change in the cultural policy environment led to the view that the holistic approach was, in fact, Green Book compliant. As our discussion of the work of Miller and Rose (2008), Bauman (1992) and Schlesinger (2013) demonstrates, the role of the expert in government, including that of the accounting scholar, is neither disinterested nor value neutral. We have shown that the role of the academic expert within governmentacting as legislator and/or interpreteris inherently political. Yet, as the example of the Phase One and Phase Two cultural value projects illustrates, a reflexive approach can be used to decentre calculative practices while also opening up possibilities for deeper engagement with cultural sector values and retaining the patina of Green Book legitimacy.