Knowing efficiency: the enactment of efficiency in efficiency auditing

Description
This paper reports an ethnographic study of the activities of auditors in the ®eld as they work to ful®l an eciency
auditing mandate; it analyses how auditors report on eciency in practice. Miller and Rose's work on governmentality
(Miller, P., & Rose, N. (1990). Governing economic Life. Economy and Society, 1±31; Rose, N., & Miller, P. (1992).
Political power beyond the state: problematics of government. British Journal of Sociology, 173±205) is developed as it
applies to the technologies that Miller and Rose identify as providing the means to realise programmes such as e-
ciency auditing. The study explores the operationalisation of eciency auditing through analysis of three audits as they
were conducted in the ®eld. It is argued that in the absence of detailed rules or standards practitioners themselves
developed an agreed upon knowledge and sensibility that allowed them to make eciency auditing tractable. The paper
explicates these normative guides and discusses the consequences of an apparently socially constructed form of e-
ciency in guiding auditing practice.

Knowing eciency: the enactment of eciency in
eciency auditing
Vaughan S. Radcli?e *
Department of Accountancy, Weatherhead School of Management, Case Western Reserve University, Cleveland, OH 44106, USA
Abstract
This paper reports an ethnographic study of the activities of auditors in the ®eld as they work to ful®l an eciency
auditing mandate; it analyses how auditors report on eciency in practice. Miller and Rose's work on governmentality
(Miller, P., & Rose, N. (1990). Governing economic Life. Economy and Society, 1±31; Rose, N., & Miller, P. (1992).
Political power beyond the state: problematics of government. British Journal of Sociology, 173±205) is developed as it
applies to the technologies that Miller and Rose identify as providing the means to realise programmes such as e-
ciency auditing. The study explores the operationalisation of eciency auditing through analysis of three audits as they
were conducted in the ®eld. It is argued that in the absence of detailed rules or standards practitioners themselves
developed an agreed upon knowledge and sensibility that allowed them to make eciency auditing tractable. The paper
explicates these normative guides and discusses the consequences of an apparently socially constructed form of e-
ciency in guiding auditing practice. # 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Auditing; Eciency; Practice; Governmentality; Ethnography; Canada
1. Introduction
Eciency auditing is a remarkably heroic
enterprise. Auditors charged with its execution are
called to investigate a wide array of government
practice, develop classi®cation schemes that allow
the categorisation of activities as ecient or inef-
®cient, and then publicly report their ®ndings.
These projects would be easier if eciency were a
more tangible concept, but as Hopwood notes,
this is not the case.
Just what is eciency? At best it is a concept
that can be subject to a wide variety of inter-
pretations. Although it is possible to change
the world in the name of eciency, there are
very real problems in relating the generality
of the concept to the speci®c organizational
procedures, let alone the speci®c con-
sequences in organizations, that can ¯ow
from its articulation and use. Of course that is
a very positive factor in political discourse
(Hopwood, 1988a, p. 6).
Eciency's very ambiguity clearly has been help-
ful in its insertion into public a?airs. While con-
sensus might not emerge around a more precisely
Accounting, Organizations and Society 24 (1999) 333±362
0361-3682/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0361-3682(98)00067-1
* Tel.:+1-216-468-9842; fax:+1-216-368-4776; e-mail:
[email protected]
stated objective, the idea of eciency seems to
o?er a useful ambiguity that allows the formation
of broad political coalitions. The alternative, after
all, has less rhetorical appeal: who wants to argue
for ineciency (McSweeney & Sherer, 1990)?
The potential for eciency to be asserted as a
dominant value in political debate has been dra-
matically exploited in recent years and has been
accompanied by a broad deployment of accoun-
tancy in the management of government a?airs, as
has been noted in the literature (Broadbent &
Guthrie, 1992). Indeed, there is an under-researched
drama to accountancy's interventions in these are-
nas (Gambling, 1987). The reports of government
auditors are presented with a detachment and
formality long associated with Weber's (1964,
1970) characterisation of the modern bureaucratic
form. As in many such environments, in the Oce
of the Auditor General of Alberta (the empirical
focus of the ®eld research reported here) all
correspondence was to be addressed to ``The
Auditor General'', signifying the authority of the
Oce. This Weberian approach continues in the
language of audit reports, throughout which
Auditors General write in the ®rst person, even
though they themselves may have had little direct
involvement in conducting particular audit work
(e.g. Oce of the Auditor General of Alberta,
1994).
1
Though the vocabularies expertise and
other mechanics by which eciency is ascertained in
an audit setting are little understood, the gravitas
that has been attached to such e?orts is considerable.
This paper presents an ethnographic account of
how eciency auditing is carried out in practice. It
follows earlier historical work that analysed the
broad political rationalities and administrative
programmes that jointly promoted eciency in
one speci®c jurisdiction, the Province of Alberta,
Canada (Radcli?e, 1998). Essentially the paper
moves on from the study of the conditions of
possibility that allowed for eciency auditing, and
analyses the operationalisation of that same
ecency auditing mandate in the ®eld. It examines
how auditors came, among themselves, to ``know''
what was or was not ecient. In pursuing this
analysis, the paper ultimately addresses aspects of
the e?ective operation of the concept of eciency
in the social world. Although grounded in the
particular setting of audit inquiry, I would argue
that elements of the social interpretation of e-
ciency that this study ®nds to be integral to auditors'
interpretations of what is or is not ecient, are
common to other invocations of the idea of e-
ciency. This suggests, as Hopwood argues above,
that eciency may by no means be an agreed
upon, absolute concept.
To examine the operationalisation of eciency
auditing the paper draws on archival materials
concerning agreed upon standards, audit proce-
dures, etc., and combines these with detailed eth-
nographic study of three eciency auditing
projects. These audits, which were later the subject
of published ®ndings from the Auditor General of
Alberta (Oce of the Auditor General of Alberta,
1994), provide an opportunity to match a speci®c
auditing mandate with an understanding of the
means by which that mandate is addressed in
practice.
The remainder of the paper is organised as fol-
lows. A brief discussion of the codi®ed knowledge
regarding eciency auditing in professional stan-
dards, manuals and the like reveals that this knowl-
edge is far from complete; therefore, ®eldwork was
conducted to understand how auditors approach
eciency auditing in practice. Discussion of
approaches to ethnography, and its use here pre-
cedes an account of the three audits that were
studied in the ®eld. The paper concludes with
observations of the nature of eciency as enacted
in the audit process, and suggestions for further
research and practice.
2. Approach
In previous work (Radcli?e, 1998) it was argued
that to understand the development of eciency
auditing or other professional projects it is important
to appreciate the process by which programmes for
change become meshed with the concerns of a
1
In recent years the Auditor General of Canada's Annual
reports have identi®ed those senior auditors involved in parti-
cular engagements. This presentation is a departure from the
otherwise persistent bureaucratic norm described above.
334 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
broader socio-political framework, and with speci®c
claims to professional expertise.
2
This analysis
followed Miller and Rose's (1990) and Rose and
Miller's (1992) Foucauldian inspired suggestion
that these problematisations be explored in terms
of an interaction between broadly stated political
rationalities, such as a general sentiment that gov-
ernment should be ecient, and speci®c pro-
grammes for action, such as the government
reports, bills and subsequent legislation which
map out speci®c mandates.
Political debates of both broad and local focus
combined in the Province of Alberta to provide a
climate in which eciency auditing was especially
appealing (Radcli?e, 1998). The study pointed to a
broad array of in¯uences on audit reform. Heightened
interest in the management of public a?airs, experi-
ences in other governments, and some modernist
hubris led to an eagerness in reforming manage-
ment structures. This was exacerbated in a Pro-
vince whose administration had arguably stagnated
under a long standing government. In addition to
this, the development of new centres of managerial
expertise within government challenged Albertan
auditors' then pre-eminence as management experts
overseeing Alberta's ®nancial a?airs.
This previous work follows the general concerns
of the ``new accounting history'' (Miller, Hopper
& Laughlin, 1991) in that it maps out the debates,
social environment and professional claims that
surround a given event. What it does not do is
provide an understanding of how accountants, in
working with mandates such as those regarding
eciency auditing, actually go about their work,
or how they fabricate and sustain their prac-
tices.
The more critically oriented literature has ten-
ded to shy away from professional practice as it is
worked out in accountants' everyday lives (Hop-
wood, 1994). It has instead focused on speci®c
moments of institutional and regulatory change
(e.g. Burchell, Clubb & Hopwood, 1985; Hoskin
& Macve, 1988; Loft, 1986; Preston, Cooper,
Scarborough & Chilton, 1994). In contrast, the
empirical focus of this inquiry becomes the agreed
upon knowledge with which accountants attempt
to ful®l the auditing task. Following Miller and
Rose's exposition of governmentality (Miller &
Rose, 1990; Rose & Miller, 1992) the rationalities
and practices that underpinned audit activity and
allowed eciency auditing to be conducted are of
particular interest here.
3
Because of this concern
the following account focuses on a number of
recurrent themes in eciency auditing, drawn
from ®eld observations and the kinds of ongoing
theoretical and personal observations that are
intrinsic to ethnographic research (Sacks, 1992;
Spradley, 1979). Archival material is examined in
later sections, and this is followed with a discus-
sion of the ®ndings of my ®eldwork, in which I
argue that eciency auditing was a socially
grounded practice. I elaborate this argument
through detailed review of agreed upon values
amongst auditors concerning e?ectiveness and
accountability, which for them were two central
criterion in representing their work to them-
selves and to others, and in evaluating the sig-
ni®cance and value of a particular audit project.
Before entering into the empirics of this study, I
discuss the signi®cance of ®eldwork in light of the
2
Although this paper does not explicitly deal with historical
matters, I would suggest that the research interest of eciency
auditing has increased with the seeming transference to the
broader audit function of the ideas and practices ®rst asso-
ciated with eciency audit. Moves by the American Institute of
Certi®ed Public Accountants (AICPA) and other bodies to
enhance the audit product beyond the level of ®nancial state-
ment attestation have been made in the hope of creating a
broader ``scope of service'', and raising the perceived value of
auditors' professional services in the market place (Elliott,
1994). Announcements from major ®rms that they have chal-
lenged and reviewed all aspects of the traditional attest audit
point to a similar expansion of the auditing task (Hopwood,
1998). These developments can be seen as spreading the kinds
of ideas and practices on which eciency auditors have relied.
3
This clear focus on practices stands in contrast to the
majority of studies drawing from the ``new accounting
research'' and other veins (Miller et al., 1991; Morgan & Will-
mott, 1993) which have primarily focused on the program-
matic. As Miller and O'Leary put it in their leading work, ``we
have placed greatest emphasis on what we might call program-
matic discourses as opposed to accounting as it was conducted
in particular ®rms'' (Miller & O'Leary, 1987, p. 240). Similarly,
Miller and Rose's work on governmentality has most to say on
generally stated political rationalities and least on the speci®c
technologies used in the ®eld.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 335
Foucauldian perspective which inspired this over-
all project.
4
Although archival work is well represented in
critical auditing, ®eldwork has not been an espe-
cially prevalent feature of this literature. More
conventional North American research has
involved some limited ®eld surveys in auditing
(e.g. Cushing & Loebbecke, 1986). Recent critical
work has explored aspects of attest auditing; for
example, Pentland (1993) reports a study of two
attest audits; other work attends to the auditing
work of the US Internal Revenue Service (Pent-
land & Carlile, 1996), and the interpretation of
audit working papers (Van Maanen & Pentland,
1994). Power (1991) o?ers an early example of
ethnography. Ethnography has not, however,
been a dominant feature in the new accounting
research as a whole.
5
The kinds of broadly theor-
ised ethnographies that JoÈ nsson and Macintosh
(1997) envisage have not often been developed in
the literature. There is no doubt that such studies
present particular challenges, as Morgan and
Willmott observe,
empirically based studies are rather the
exception than the rule. Considering the
demands they make, their scarcity is not dif-
®cult to understand. Getting to grips with
very challenging forms of social theory and
conducting empirical research and mobilizing
the theory to direct and reinterpret the
empirics is, to say the least, an extremely tax-
ing processÐand one that is not readily
reconciled with increasing pressures to publish
or perish (Morgan & Willmott, 1993, p. 6,
emphasis in original).
The tenure culture that dominates North
America especially discourages ®eldwork, given its
time consuming empirical and theoretical involve-
ments, and the complexity of mediating the dense
social structures in which accountancy operates
(JoÈ nsson, 1998, pp. 432±433). Yet there are special
insights that ®eldwork provides. I turn to ®eld-
work partly because the available documentary
evidence provides an empirically limited account of
eciency auditing, and partly because of theoretical
issues regarding the analysis of technologies. Miller
and Rose's (1990; Rose & Miller, 1992) work points
to the importance of technologies but tends not to
map out the operationalisation and implications of
these procedures. Indeed, the detailed practices of
accounting have emerged as a focus for inquiry, to
the extent that attempts have been made to gen-
erate new theories of the development of practices
(Hoskin, 1994). Here, I turn to ®eldwork
6
as an
additional means of mapping out what it is that
eciency auditors do. In this sense the study ®ts
broadly within what JoÈ nsson and Macintosh
(1997) refer to as the ethnomethodological genre
of ethnographic work. As they describe it,
Ethnomethodologists are more interested in
describing what actors are doing socially than
with what they are thinking (interactionists)
or how they are communicating (cognitive
anthropologists). They are interested in
understanding how actors jointly de®ne the
character of an event and how they sustain
(or change) these meanings. So they try to
capture the ways in which actors achieve
commitment, trust or whatever is ``going on''
in the social interaction under study. The aim
is to discover the situated rationality of
4
To re-iterate, this piece ¯ows directly from earlier work
concerning the general political rationalities and more speci®c
programmes that Miller and Rose present as grounding tech-
nologies such as those studied here (Miller & Rose, 1990; Rose
& Miller, 1992). In keeping with the inter-relatedness of these
features of governmentality, this study of audit technologies is
best read in the context of the earlier study of the rationalities
and programmes (Radcli?e, 1998).
5
JoÈ nsson and Macintosh (1997) document some notable
exceptions, including Preston's (1986) early ethnographic work.
There has of late been increased interest in this area; recent
examples of ®eldwork in more general research includes work
by Barrett et al. (1998); Co?ey (1993, 1994); Dirsmith and
Covaleski (1985a,b, 1997), Grey (1994) and Hanlon (1994).
6
Although ethnography has primarily involved ®eld research,
it is not synonymous with such an approach. It is possible to
perform ethnographic analysis focused on written texts. For
example, Livingston uses texts to analyse the ethnomethodolo-
gical foundations of mathematical proofs (Livingston, 1985).
His later work similarly uses texts to analyse the act of reading
(Livingston, 1995).
336 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
actors' mundane, everyday social practices in
order to describe the sense-assembly equip-
ment they use to construct and to sustain
their daily procedures, practices and stocks of
social knowledge (JoÈ nsson & Macintosh,
1997, p. 373).
In essence the focus becomes the generation of
meaning among actors, as evidenced through
communication amongst them. As JoÈ nsson
argues, social facts are continually constructed
through social interaction, and this is observable
in the ®eld (JoÈ nsson, 1998, pp. 431±433). Though
JoÈ nsson's particular interest lies in management
accounting, his call to ``relate management
accounting research to managerial work'', toge-
ther with the observation that managers work
with words (JoÈ nsson, 1998, p. 411), could equally
have been made of auditors and the work of
auditing.
More generally, ®eldwork has increasingly been
read in narrative terms; Van Maanen (1988)
coined the term ``tales of the ®eld'' and re¯ected
on variations in ethnographic research:
Ethnographies are portraits of diversity in an
increasingly homogenous world. They display
the intricate ways individuals and groups
understand, accommodate and resist a pre-
sumably shared order. These portraits may
emerge from global contrasts among nations,
societies, native histories, subsistence patterns,
religions, language groups and the like. Or
they may develop from the more intimate
contrasts of gender, age, community, occu-
pation or organisation within a society (Van
Maanen, 1988, pp. xiii±xiv).
In the face of post-structuralist analysis (Sarup,
1989) the legitimacy of qualitative research and of
such concepts as ``validity,'' ``reliability'', etc. was
questioned. A variety of authors have responded
with an interpretive re-theorisation of these issues,
in response to earlier more positivist constructions
(Hammersley, 1992; Hammersley & Atkinson,
1983; Kohler Riessman, 1993; Lather, 1993;
Mishler, 1990). The kinds of local understandings
that Van Maanen envisages have taken the place of
earlier ``holistic'' e?orts to study entire cultures (e.g.
Reichard, 1939). As Van Maanen suggests, ``more
intimate contrasts'' have occurred as ethnographies
have come to deal with speci®c problems, and
attend to local concerns, particular forms of orga-
nisation and the like (Denzin & Lincoln, 1994;
Lather, 1993). Similarly, this paper presents a
localised study of eciency auditing that ®ts with
Van Maanen's latter characterisation.
In this study both ®eld and archival research are
deployed to establish the bounds of particular
bodies of knowledge and practice. This may be
seen as unusual in the context of Foucauldian
work as it has been seen in the accounting litera-
ture, and in the context of work which precedes
this present piece (Radcli?e, 1998). More gen-
erally, Foucault has been noted for his focus on
written texts in uncovering systems of thought. As
he himself put it, ``treasures lie dormant in docu-
ments'' [Foucault, quoted in (Bernauer, 1988)].
Many of Foucault's studies, such as the history of
punishment in the eighteenth and nineteenth cen-
turies, or the development of sexuality in ancient
Greece or Rome, could only be studied through
reference to documents or other artifacts. But for
more current questions the thoughts and ideas
that he sought to trace could equally be found in
regularities of discourse in contemporary state-
ments. Foucault saw the rules of discursive for-
mations as being embodied just as much in the
most recent as in the most original statement of an
idea (Gutting, 1989). His central concern was not
texts themselves, but rather the thoughts that were
recorded in them. This becomes clearer in some of
his last published work:
(Thought,) understood in this way, is not,
then, to be sought only in theoretical for-
mulations such as those of philosophy or sci-
ence; it can and must be analysed in every
manner of speaking, doing, or behaving in
which the individual appears and acts as a
subject of learning, as ethical or juridical sub-
ject, as subject conscious of himself and oth-
ers. [. . .] The study of forms of experience can
thus proceed from an analysis of `practices'Ð
discursive or notÐas long as one quali®es
that word to mean the di?erent systems of
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 337
action insofar as they are inhabited by
thought as I have characterized it here (Fou-
cault, 1984, pp. 334±335).
In his belief that speech, action and behaviour
provide windows into discourse in the ®eld, Fou-
cault's later work reveals an empirical interest that
goes beyond documents, and allows for ®eldwork
similarly concerned with the meeting of thought
and practice. Similarly Miller and Rose suggest
diverse conjunctions of thought and practice as
the foundations of the technological, including:
techniques of notation, computation and cal-
culation; procedures of examination and
assessment; the invention of devices such as
surveys and presentational forms such as
tables; the standardisation of systems for
training and the inculcation of habits; the
inauguration of professional specialisms and
vocabularies; building design and archi-
tectural forms. . . (Miller & Rose, 1990, p. 8).
The e?ects of such con¯uences are of special
importance in understanding the technologies by
which accountants in general, and auditors in
particular, accomplish their work. Organizational
sensemaking is an inevitable component in guiding
the audit task (Weick, 1995). This study examines
aspects of this sensemaking process, and uses ®eld
observation of auditors' communications as a
means of exploring thoughts and practices in
terms of their role in the social construction of
audit technologies (cf. JoÈ nsson, 1998).
At this point, analysis turns to the kinds of
programmatic (Miller & Rose, 1990) documented
knowledge that is formally held out as the basis
for professional action: standards, manuals and
the like. The signi®cance of such documents
should not be underestimated; the notion of pro-
fessionalism in accounting has increasingly been
equated with the standardisation and regulation of
professional conduct (Hopwood, 1998; Preston et
al., 1994). Standards, manuals, etc., are useful in
establishing both the professional claim and the
body of knowledge with which auditors claim to
work; this provides a starting point in under-
standing how eciency audits are conducted.
2.1. Archival knowledge
In Canada, the Public Sector Accounting and
Auditing Board
7
of the Canadian Institute of
Chartered Accountants (CICA) has developed a
series of Statements to guide government auditing.
An initial reading of these materials is suggestive
of similarity between attest auditing and eciency
auditing mandates; indeed, the standards are cou-
ched in language familiar from an attest auditing
environment. For example, PSAAC's guidance on
standards of conduct and other ethical matters is
essentially a cross reference to the ethics section of
the CICA Handbook (see PSAAC, 1985b, s. 04).
In their analysis of eciency auditing, early audit
guidelines carry on this transference of more
familiar professional knowledge, and suggest an
untroubled extension of traditional professional
roles (PSAAC, 1985a). However, as the Institutes'
work on these materials progresses, an uncertain
note enters into these descriptions:
The amount of discretion that an auditor has
in establishing the objectives and scope of a
value for money audit varies. Some value-for-
money audit mandates, such as those of legis-
lative auditors, provide only general direction
about objectives and scope. In such cases,
auditors must establish the audit objectives
and scope for particular value-for-money
audits (PSAAC, 1990).
It is a common argument of modern social and
literary theory that the meaning of a text must
inevitably be constructed by the reader (Boland,
1989; Latour & Bastide, 1986; Manninen, 1993;
Silverman, 1993). Such an observation may not be
immediately remarkable, but what is of interest
here is the way that these standards are presented
with noticeable equivocation, and with unusual
acknowledgement of the way in which their
underlying meaning must be worked through in
the ®eld. The standard setters themselves seem to
recognise that while routine attest auditing might
7
Until recently known as the Public Sector Accounting and
Auditing Committee, or PSAAC.
338 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
be reduced to a series of checklists and sampling
programmes, eciency auditors must ®rst grapple
with the most basic matters, audit objective and
scope. Auditor's mandates vary by jurisdiction,
complicating the audit task. The breadth of
potential assignments in which eciency must be
uncovered leads to a more situated understanding
of the structuring of the audit task. In the absence of
the more detailed accounting and auditing principles
that apply in the private sector (PSAAC, 1985b),
eciency auditing emerges as a more hermeneutical
process, in which theory provides little guide, and
meaning must instead inevitably be worked out in
the ®eld (Boland, 1989). This is recognised both in
the text of the guidelines and, implicitly, in their
late arrival; though audit guidelines were not
published until 1990, eciency audits had been
conducted since the late 1970s.
Professional documentation attempts a basic
de®nition of economy, eciency, and e?ective-
ness, suggesting, for example, that economy is
satis®ed by acquiring resources of the appropriate
quantity and quality at the lowest cost (PSAAC,
1988). Later material alludes to complexity in
suggesting that although the auditor should
remain at the centre of the audit process, eciency
auditing may call for auditors to seek advice from
experts in a variety of professional ®elds, including
engineering, statistics, human resource manage-
ment and economics (PSAAC, 1988).
In maintaining the centrality of the auditor in
eciency auditing the standards do more than
establish a general approach to such work. They
picture these audits as a natural part of the audi-
tor's domainÐas work which is clearly theirs. It is
these kinds of discreet, seemingly unremarkable
statements that collectively build a discursive ®eld.
The depiction of accountants as being at the cen-
tre of the audit process is in this sense an example
of the working through of competing claims to
expertise, and a more general approach in which
forms of policy, action and behaviour are eval-
uated systematically (Gordon, 1991). Expert sta-
tus is especially fragile in an environment in
which, as even the standard setters acknowledge,
``there is no body of generally accepted criteria'' to
guide audit planning (PSAAC, 1988). The stan-
dards instead suggest that auditors turn to a variety
of sources, including legislation, policy statements,
bodies of statistics, and prior audit experience.
Though the standards o?er a general window
into considerations a?ecting the audit process,
they provide little insight into how an auditor
might form an opinion on the broad array of
activities which modern government involves. The
internal manuals of government auditors begin to
o?er a more speci®c guide.
2.1.1. Manuals
In Alberta, the Oce of the Auditor General
has developed its own audit manual (Oce of the
Auditor General of Alberta, 1991). Though nor-
mally considered private, it has been made avail-
able for analysis here. As with auditing standards,
the manual exhibits signi®cant gaps in describing
audit technologies. Repeated reference is made to
the need for interpretation and judgment in the
®eld. In reviewing this material it becomes impor-
tant to appreciate that the manual's author is
immersed in the practitioner knowledge that
anchors eciency auditing as it is practiced in the
Oce. This tacit knowledge is later explored in
®eld study of the kinds of ideas, knowledge and
local cultural contexts that facilitate action
(Boland, 1989). Overall, the manual provides a
start in investigating how auditors operationalise
their mandate, since it maps out certain of the
auditors' interpretive schemes (Boland, 1993).
The manual starts with an overview of the
economy, e?ectiveness and value for money man-
dates that the Oce collectively refers to as ``sys-
tems'' auditing. In practice in the Oce of the
Auditor General of Alberta, these concerns were
addressed simultaneously, and auditors seemed to
be unconcerned with di?erences between the ideas
of eciency, e?ectiveness and value for money.
Rather than depending upon a detailed de®nition
of eciency or any other criterion, auditors pre-
ferred to talk about the vaguer notion of ``sys-
tems'' auditing (Oce of the Auditor General of
Alberta, 1991), principally because it seemed to
allow them to conduct their work while eluding
de®nition of its underlying principles
The Oce's initial experience had been that cli-
ent entities questioned auditors' categorisations of
what was and was not ecient on the basis of the
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 339
auditors' own de®nitions. Such questioning stood
to undermine the legitimacy of the eciency
auditing project. In the face of this resistance,
auditors began using the unde®ned notion of
``systems'' auditing, a di?use term drawn from the
Oce's governing legislation (Auditor General
Act, 1977). The absence of a de®nition of this term
meant that guiding concepts underlying audit
work could no longer be questioned in their own
terms, and hence intellectual threats to the legiti-
macy of the auditors' expanded mandate were cir-
cumvented.
Interestingly, the statute's reference to systems
was itself a product of political resistance to the
notion of eciency auditing and of auditors'
expertise in such areas. Ministerial concerns in this
regard were assuaged by certain of their collea-
gues, themselves chartered accountants, who in
cabinet discussions o?ered the view that accoun-
tants were quali®ed to judge ``management sys-
tems'' (Radcli?e, 1998). The acceptance of the
loosely speci®ed concept of systems auditing is a
further illustration of the productive role of ambi-
guity in the expansion of professional jurisdiction
(Abbott, 1988; Hopwood, 1998; McSweeney &
Sherer, 1990). In practice this usage successfully
evaded the problems of de®ning eciency to
external constituencies (chie¯y client entities),
while auditors agreed among themselves that it left
them free to conduct their audits for economy,
e?ectiveness, eciency and other matters as they
wished. While noting the problematic nature of
these terms, auditors' expanded mandates are
referred to here either as eciency or systems
auditing, in keeping with professional usage and
the seeming simultaneous consideration of these
concerns in the ®eld.
As discussed above, at the time it was passed
some Ministers had expressly understood Alber-
ta's governing legislation as providing a limited
eciency auditing mandate. As one Minister put it
in framing the Auditor General Act, they had been
especially concerned to avoid ``the question of
management auditing or value accounting or
value for money spentÐall those phrases having a
similar connotation'' (Alberta Hansard, 1977,
p. 1694). By contrast the audit manual suggests
that auditors have pursued a broad approach to
their mission of improving the ®nancial adminis-
tration of the province,
Improving ®nancial administration does not
mean limiting the scope of systems audits to
®nancial systems. Most management control
and information systems impact ®nancial
administration, including accountability, and
are therefore subject to audit (Oce of the
Auditor General of Alberta, 1991, s. 10.1).
While politicians had sought to limit the audi-
tors' gaze, the auditors' own interpretations of
their mandate defeated this aim. The intentions of
certain of the Act's authors were circumvented, and
the scope of the auditing enterprise widened, by
this kind of expert re-interpretation of legislative
authority.
The manual describes the audit process in terms
of a series of stages. The audit is ®rst de®ned in
terms of issues,
8
signi®cant aspects of an entity's
operations which if not managed well could jeo-
pardise the success of the entity or programme
(Oce of the Auditor General of Alberta, 1991, s.
20.1). In a way that presages auditors' focus on
particular conceptions of the tractability and
``e?ectiveness'' of their work,
9
audit issues are
understood to followfromthese, as those issues likely
to result in reportable audit ®ndings and asso-
ciated recommendations for improvement (Oce
of the Auditor General of Alberta, 1991, s. 20.2).
Systems auditing is presented as a selective pro-
cess; only those issues which seem likely to result
in reportable audit ®ndings and suggestions for
improvement are pursued. Under the Auditor
General Act, 1977, the Auditor General chooses
the areas of investigation, both for regular annual
reports, and for any special reports that he/she
deems necessary.
10
To qualify as an audit issue a
8
At this point I italicise certain key terms from the Audit
Manual to stress their special meaning.
9
See discussion regarding ``e?ectiveness'' and the tractability
of eciency auditing below, in material regarding ®eldwork.
10
The only element of direction allowed is a request from
government that the Auditor General investigate speci®c issues;
the Auditor General may choose not to pursue such requests,
although this refusal would be dicult to make in the face of a
request from the highest levels of government (Radcli?e, 1997).
340 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
matter must relate to issues important to manage-
ment success, must be likely to produce reportable
®ndings, and must be matched with speci®c audit
criteria against which auditors may compare
management's handling of the issue.
Auditor interpretations such as these e?ectively
blur distinctions between ``systems'' and ``opera-
tions'' even though politicians developing the Act
had hoped that such matters would be separated.
Instead, the manual suggests that ``a system is
rarely the audit issue. More often the issue is the
management concern that the system is designed
to address'' (Oce of the Auditor General of
Alberta, 1991, s. 20.4), a view which opens the way
to policy concerns.
11
In this sense, auditors estab-
lish broader management concerns as audit issues,
an approach which directs auditing away from
mechanics and towards wider social objectives.
The concept of signi®cance becomes important
in understanding the focus of eciency audits as
they have developed in the Canadian context.
Distinct from the concept of materiality in ®nan-
cial statements, signi®cance is a term embodied in
legislation and in government practice (PSAAC
1988). Signi®cance includes a concern with ®nan-
cial magnitude,
12
but encompasses far more. The
audit manual suggests consideration of an issue's
importance to government policy; its potential
economic, social or environmental impact; inadequate
managerial action on previous audit issues; interest
expressed in the matters by the Legislature, govern-
ment or public; and the impact of a centralised func-
tion, which might extend throughout government
(Oce of the Auditor General of Alberta, 1991, s.
20.4). Importantly, the signi®cance of an issue is
in¯uenced by a far greater range of concerns than
the monies immediately involved. In de®ning sig-
ni®cance, auditors make it quite plain that in the
eciency auditing arena an item may become sig-
ni®cant because of political interest, whether
through questions in the Legislature, news stories
or other means. This written acknowledgment of
the role of the political in framing questions of
accountability is con®rmed and elaborated in ®eld
analysis, which is addressed later.
The manual makes it quite clear that it is audi-
tors' ideas of what should be critical to manage-
ment, rather than management's ideas, that should
prevail in de®ning audit issues. This e?ectively
places auditors in a super-ordinate position in
de®ning the scope both of management and
auditing activity (Oce of the Auditor General of
Alberta, 1991, s. 20.4).
13
Audit criteria are them-
selves the product of a variety of exposures, includ-
ing auditors' training and consequent outlook on
best management practice, professional standards,
similar prior audits, and more general manage-
ment knowledge [as the audit manual notes, ``the
Oce library and commercial bookstores contain
many books on contemporary management theory
and style'' (Oce of the Auditor General of
Alberta, 1991, s. 20.5)].
The manual's diagrammatic representation of
the audit process (Fig. 1) attempts to show a for-
mal audit process, with discrete steps, standards of
evidence, and objectives. In this light, initial
knowledge of business is collected as a series of
facts, held to be beyond reasonable doubt. When
compared with agreed upon audit criteria these
facts result in ®ndings, which after a process of
selection are seen as being veri®ed, developed, and
®nally published as audit recommendations.
This formal representation is at odds with the
manual's other observations of the audit process.
11
For example, in one audit that was followed in the ®eld, the
``system'' concerned performance measurement systems
designed to redeploy resources to those hospitals that were seen
as more ecient. The management issues that were involved
went far beyond this system, however, to include overall gov-
ernment policy in the ®nancing of health care and of govern-
ment services in general.
12
Financial magnitude is the usual interpretation of materi-
ality in an attest audit. This is formally stated as an item of
information, or an aggregate of items, which if they were
omitted or misstated would in¯uence or change a decision
(CICA, 1991, s. 1000 para. 17), Amounts over a cut o? point,
e.g. 0.5±2% of total revenues, might be termed material and
worth further investigation; amounts under this point would be
termed ``immaterial.''
13
Others have noted the superior position often ascribed to
management and to managerial knowledge; managers are rou-
tinely presented as having the expertise, vision and power to
control the social world (Macintyre, 1985). The interpretation
of ``systems'' auditing seen in the manual instead places audi-
tors in a super-ordinate position, above management and other
groups, as auditors form opinions regarding management
behaviour.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 341
In formulating initial knowledge of the client
entity's business, the manual suggests team meet-
ings to ``brainstorm'', interviews with client man-
agement, and the investigation of ``hunches''. In
moments such as these the manual's depiction of
auditing suggests an underlying tension. These
statements provide the important insight that
such activity is seen as a legitimate method of
audit inquiry, and that it may provide the foun-
dations for later work. Yet though the informal
knowledge gathered in the initial stages of an
audit may ®nd formal expression as audit issues,
memoranda, etc., the manual does not indicate
the kinds of knowledge or norms that inform
``brainstorming'', ``hunches'' and other pre-
liminary analysis.
In such cases when the manual fails to map out
the roles of potential sources of information,
modes of analysis, or the kinds of qualitative
environmental information to which the manual
refers, the reader is assumed to be party to com-
mon practitioner knowledge. While the manual
details certain mechanics of an audit, such as entry
and exit conferences, the preparation and
approval of management letters and other prac-
tices drawn from attest auditing, it leaves less for-
malised practices and practitioners' tacit
knowledge to be worked through in the ®eld.
Having exhausted archival knowledge, I now turn
to ®eldwork as a means to explore the situated
rationality of auditors' everyday practices in per-
forming eciency audit work (JoÈ nsson & Macin-
tosh, 1997).
2.2. Fieldwork
The mechanics of a technology rely both on the
practices that are involved (e.g. auditing tests and
procedures) and the ideas and knowledge which
underlie these practices, such as conceptions of
what is or is not ecient, or of which techniques
are appropriate. My exploration of these issues
was facilitated through research access with the
Oce of the Auditor General of Alberta gained in
a process of negotiation which followed their
chance exposure to a conference paper which I
Fig. 1. The audit process (Oce of the Auditor General of Alberta, 1991).
342 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
had presented. It was agreed that I could follow a
number of audits. The audits were all in the health
area,
14
partly because senior members of the Oce
felt that some of their most advanced work was
being done in this ®eld. I found a focus on one
area helpful since I had to develop specialised
knowledge in only one area of government activ-
ity. My initial plan was to focus on two audits,
one routine, the other less so, in an attempt to use
these as counterpoints to each other, drawing out
di?erences between them and so testing the
boundaries of systems audit as understood by
practitioners. Accordingly the two audits were to
be selected in a process of negotiation with the
auditors themselves.
In this process of negotiation it became clear
that eciency audits di?ered in terms of the extent
of ®eldwork, audit meetings, etc., that were con-
ducted. Further, the more ``complex'' audits were
counter intuitive, in that they often involved far
less ®eldwork and meetings than other audits
which, though seen as conceptually simple, were
logistically complex. Learning from this, I was
able to study three audits of varying complexity.
This choice allowed me to draw out di?erences
between these audits, and to study signi®cantly
more auditing work. In later interviews, auditors
have con®rmed that they believe these audits are
typical of their work. I now provide brief details of
each audit.
The ®rst audit, seen by the auditors as simpler
than the others, concerned a project to enhance
management information in the Department of
Health. The project centred on a large scale fed-
eral initiative, known as the management infor-
mation systems (MIS) project, and involved
electronic gathering of information from all pro-
vincial hospitals. Though chosen as a routine or
simple audit to assess the implementation of the
MIS project it became apparent that the audit was
considerably more logistically demanding than the
other audits examined. It involved some twenty
®eld visits, interviews and planning meetings, over
the course of ®ve months.
One aim of the MIS project as applied in
Alberta had been to re-allocate funding away from
inecient hospitals and towards ecient ones,
thereby rewarding eciency. This had however
been overtaken by the government's de®cit reduc-
tion plans, which called for a 20% reduction in
total government expenditures over three years
(Dinning, 1993). As a consequence, MIS funding
adjustments became overshadowed by system-
wide cuts to government funding. The situation
grew more complex as the MIS audit
15
continued,
and as auditors added more references to the
social environment, including ongoing develop-
ments in the policy arena. What was ostensibly
simple was in fact embedded in a complex web of
social relations.
The second audit (hereafter, the ``subsidies
audit'') concerned government subsidies of senior
citizens' health care premiums. These premiums
would otherwise have been paid directly by retir-
ees to ®nance the provincial health plan. This
audit contrasted with the MIS audit in logistical
terms, since it involved largely oce based analy-
sis and review, with information gathered through
telephone conversations, and documents gathered
by courier or fax; audit ®eldwork was quite mini-
mal.
The third audit (hereafter, the ``physicians'
audit''), was seen by auditors as the most complex
and politically sensitive. It centred on a $14 mil-
lion liability to provincial doctors that arose from
a fee for service agreement between the province
and physicians. The agreement capped total pay-
ments by relating fee levels to patient volumes; if
patient volumes fell below budget, unit fees rose; if
volumes rose over budget, fees fell. The agreement
was predicated on the belief that physicians were
able to in¯uence demand for their services by
recommending certain courses of medical treatment.
14
Under the Canada Health Act all provincial governments
provide medical services, with partial federal government
funding.
15
The terms ``MIS audit'', ``subsidies audit'' and ``physician's
audit'', are the titles that were used within the Oce when talking
of these projects. I use themhere as useful shorthand in describing
my work, although I would emphasise that the names assigned to
each audit should themselves be re¯exively examined. Such titles
can provide important signs of how these audits were understood.
Interestingly, the names of these audits remained ®xed, even as
the subject of the audits varied, sometimes greatly.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 343
The physicians' audit similarly involved oce
based work.
In examining these audits I practiced a form of
triangulation, the use of more than one method,
observer or site to provide further checks on a
single observer
0
s ®ndings (Sommer & Sommer,
1991). In this study triangulation occurs in track-
ing three audits, and through the use of a variety
of sources of information, including interviews,
passive observation and documentary analysis. I
collected copies of relevant working papers,
including, for example, successive drafts of
recommendations, which developed from day to
day. In capturing audit planning meetings, infor-
mation gathering meetings, etc., I engaged in pas-
sive observation, since my aim was to learn about
systems auditing, and I did not have the detailed
prior knowledge that would have allowed me to
act as an informed participant observer (Sommer
& Sommer). I taped research interviews and made
condensed notes during ®eld visits and other
meetings, turning these notes into expanded
accounts directly after the meetings, following
Spradley's (1979) suggestions for data collection. I
prepared analytical notes and memoranda so as to
systematise analysis and preserve analysis from
the ®eld, as and when it occurred.
16
To check my interpretations of the data
obtained through documentary review and ®eld
observation, I conducted respondent validation
(also referred to as participant veri®cation) (Wer-
ner & Schoep¯e, 1987) through a series of inter-
views, held over the course of the audit, and after
the auditors' review of this paper. I reviewed most
of my re¯ections on their work in this way.
Respondent validation is useful in highlighting
those matters which actors regard as being of greater
or lesser signi®cance, and can ¯esh out under-
standings gained from passive observation, parti-
cularly in areas where the researcher e?ectively
has to deduce understandings and ideas from
behaviour.
17
I believe that this triangulation and use of mul-
tiple methods provided broader and more reliable
information than any one approach alone (Oswald
et al., 1987). Indeed, this use of multiple methods
is characteristic of ethnographic research. This
combination of methods may be regarded as bri-
colage, ``a pieced together, close-knit set of prac-
tices that provide solutions to a problem in a
concrete situation'' (Denzin & Lincoln, 1994,
p. 2). An understanding of the qualitative
researcher as bricoleur suggests that the ®eldwor-
ker uses methods as tools, choosing whichever is
appropriate to (and available for) the task in hand
(Becker, 1989).
In addition to the use of multiple methods, this
research had the signi®cant advantage of being
conducted in real time, since my observations were
made throughout the course of the audit. My
initial research interviews began in January 1993;
audit ®eldwork commenced in August 1993, with
recommendations ®nalised in December 1993.
Over the period September±November 1993 I
was based in the Audit Oce, enabling close study
of the audit process and a variety of informal
interaction. This rich, contemporaneous observa-
tion provided for more accurate and thorough
analysis than would be possible through retro-
spective interviews or the analysis of working
papers alone.
16
Spradley (1979) notes the importance of such notes since in
®eldwork the researcher becomes a major research instrument;
because of this it is important to capture the researcher's ana-
lysis and re¯ection on events, a suggestion that is endorsed by
others (Hammersley & Atkinson, 1983; Oswald, Schoep¯e, &
Ahern, 1987; Sommer & Sommer, 1991).
17
In performing respondent validation it becomes important
to sort through actors' representations so as to bring together
elements of both their and the researcher's understandings.
That an actor might disagree with an observation need not
mean that the observation is incorrect, it might simply mean
that it touched on an area of sensitivity. In making these and
other judgments an ethnographer's own subjectivity comes into
play and the researcher as author clearly takes responsibility
for the text. But this awareness of the subjectivity that the eth-
nographer embraces need not mean that a greater level of
judgment is involved than in, say, a capital markets empirical
piece, or an experimental study. It is perhaps more accurate to
see these and other instances as providing qualitative research-
ers with a greater awareness of the role that subjectivity neces-
sarily brings in research, and a consequent introspection and
sensitivity to these issues, especially in recounting events (Van
Maanen, 1982, 1988).
344 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
3. Knowing eciency: auditing in the ®eld
The writing of ethnography is always a delicate
process, if not a struggle (Morgan & Willmott,
1993). The use of ``thick'' description o?ers a
richness of experience drawn from the ®eld
(Geertz, 1988; Van Maanen, 1988; Werner &
Schoep¯e, 1987), while running the risk of losing
theoretical insightÐa key interest of critical
research (JoÈ nsson & Macintosh, 1997). Here I
have elected to move away from a narrative of the
stories of each audit, and instead present a the-
matic discussion of the ways in which I found
auditors conducted eciency audits in practice. The
identi®ed themes are drawn out of my research
materials and represent recurrent dimensions of this
audit work (Hammersley & Atkinson, 1983; Oswald
et al., 1987). Following earlier discussion, I focus on
the mechanics which allowed this audit technology
to operate (Rose & Miller, 1992). In particular, I
depict auditing as socially grounded, and elabo-
rate on the role of the social through close exam-
ination of two concepts, e?ectiveness and
accountability, that seemed central to auditors'
conceptions of their work. While recognising that
technologies may be widely heterogeneous it is
through their use that the objects of audit inquiry,
and audit inquiries themselves, become tractable
(Miller & Rose, 1990, p. 5). Accordingly, I attend
to the means by which eciency auditing was
operationalised, mindful that, as Foucault put it,
``the main problem is not whether [. . .] [people]
conform to principles of rationality, but to dis-
cover which kind of rationality they are using''
(Foucault, 1988b, p. 59). Discussion of ®eld
observations continues in this spirit.
3.1. Socially grounded auditing
In most signi®cant respects, eciency audits
were quite di?erent to attest work. Perhaps the
most fundamental di?erence lay in the actual
choice of audits. In an attest audit, standard proce-
dures and tests would guide much of the audit work.
The auditor is, in most cases, expected to issue a
standard opinion on the ®nancial statements.
But compare this to the ®rst question that the
eciency auditor has to face, namely, ``what shall
I audit?'' The Alberta mandate, together with
those of other jurisdictions, calls for eciency
audits, but does not require the kind of compre-
hensive audit opinion associated with an attest
environment. Auditors do not audit everything
each year; instead they get to choose what they
look at in systems audits, and then construct audit
activities around these choices.
In this section I argue that this dimension of
choice has important in¯uences on eciency
audit, and that these choices invoke a broader
social awareness which grounds much of the audit
technologies that are used in this work. In parti-
cular, auditors seek out a series of organisational
and cultural guides, one of the central concerns
being to establish which kind of recommendations
would be found credible by government adminis-
trators, managers, and, to a certain extent, politi-
cians. This credibility was important because the
e?ectiveness of audit recommendations was, in
large part, understood by whether they were later
implemented. In this sense eciency auditing was
socially grounded, because of initial choices, the
environmentally driven technologies that auditors
used to aid their initial choices and subsequent
navigation of the audit, and through their sensi-
tivity to what others might potentially implement,
since implementation vindicated audit work as
``e?ective''. Audit technologies (Miller & Rose,
1990) involved sophisticated study of the social
world as part of an approach designed to promote
what auditors saw as an e?ective audit.
The idea of eciency auditing as the socially
grounded, strategic process that I have outlined
18
is
in some ways a surprising contrast to the seemingly
absolute manner in which eciency is often dis-
cussed. Indeed, this was a process which auditors
themselves did not expect others in their profession
to immediately understand. As one senior auditor
explained the initial choice of audit topics:
18
I would distinguish this view of socially grounded auditing
from that of other researchers who have termed attest auditing
``strategic'' as part of their very di?erent theoretical concerns to
understand auditing as a game motivated by discernible self
interest (e.g. Morton, 1993). Equally, although the word
grounded seems most appropriate for events as I understood
them, use of this word is not an e?ort to invoke the tenets of
grounded theory (Glaser & Strauss, 1967).
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 345
People will ®nd this an unusual conception of
what auditing isÐof the auditing process.
This is one of the things that green-horns in
the oce often get wrong. They think that
audits arise out of nothing, in a scienti®c
tabla rasa kind of way. But that is not how
things work.
Instead, audits arose out of information that the
auditors gathered from a variety of sources, both
formal (attest audit work, government statements,
etc.) and informal (conversations, news stories, a
sense of ``what was going on'').
The same auditor suggested that attempts to
®nd the start of the idea for an audit would be
futile, since these were often ideas that developed
through years of experience in the activities of
Departments, and through knowledge of the peo-
ple and projects involved. Accounting is often
suggested as being about information; and this
was certainly true of eciency audit, although not
in a dry, technical sense. The information that
auditors sought out was often concerned with
organisational and other social knowledge. For
example, auditors acknowledged amongst them-
selves that certain audit topics or inquiries were
facilitated by growing political interest, or by
changing views among senior civil servants. An
audit whose seeming e?ectiveness might be com-
promised by resistance would be avoided or at
least modi®ed so as to more directly ®t with cur-
rent governmental concerns. Auditors understood
themselves as operating in dynamic surroundings
in which environmental sensitivity was of over-
riding importance. This had clear consequences
for the types of knowledge and skills which were
valued as auditing resources, as another auditor
commented on the value of chartered accountancy
(CA) training and attest audit knowledge in e-
ciency audit:
Standard CA training, does it help? Yeah, a
little bit it does. But [. . .] you really need to
have powerful analytical skills. You need to
have the ability to come to terms with com-
plex issues, rather than just simple things. A
lot of people, when we discuss some of these
things, they tell me something that's very
nice, but its still at a super®cial level. I think
its really, in terms of the skills that you need,
unless the person has the ability to relate dif-
ferent pieces of information, has the ability to
know what is happening, and put it in a con-
text, because things can easily be taken out of
context, in terms of what you might end up
recommending.
[. . .] A lot of it is through experiences,
through mistakes. A lot of it is through my
continuing involvement in the health sector.
And that sort of contributes to it.
One of the recurring features of this work was
its dependence on a highly developed environ-
mental knowledge, both within government depart-
ments, and within the wider social environment.
Auditors believed that their knowledge of the
environment, usually referred to (after the audit
manual's terminology) as ``knowledge of busi-
ness'', was fundamental to their work. Auditors'
knowledge of business came to bound what was
for them the organizational reality within which
they had to work. This knowledge was accumu-
lated chie¯y through informal, tacit means.
In the course of my research it became clear that
the existence of papers in a ®le was no sure guide
to their signi®cance in the conduct of an audit.
The subsidies audit working papers included
copies of numerous Alberta statutes providing for
health care premium subsidies; however, this
knowledge was only a formal background to the
audit. As the auditor conducting this work sug-
gested above, prior knowledge regarding premium
subsidies had ``caused the recommendation to
arise''; the copies of Alberta statutes, ®nancial
analyses and similar provided support to this
initial conception.
In the audits as a whole hardly any of the audit
working papers resembled the ``T'' accounts, ®nan-
cial accounting analyses, checklists and schedules
associated with attest auditing work (Buttery,
Hurford & Simpson, 1993; Turley & Cooper,
1991); discussions of such detail, and even of more
abstract policies such as revenue recognition, or
the determination of liabilities, was largely rele-
gated to attest work. Instead, systems auditors
346 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
gathered organisational knowledge regarding
departmental policy and intent. The formal
aspects of this could be written down and analysed
in working papers. But many of the more sig-
ni®cant in¯uences in this auditing did not appear
in working papers. Knowledge of business was
signi®cant in the establishment and pursuit of
auditing tasks, but was largely unrecorded in con-
ventional terms. Rather than being a part of writ-
ten records, knowledge of business often gained its
signi®cance from its incorporation into auditors'
collective memory. What is incorporated into col-
lective memory as it appears in discursive forma-
tions is particularly signi®cant in attempting to
track knowledge and practice (Foucault, 1991,
pp. 59±60).
The socially grounded nature of eciency
auditing had consequences for the kinds of
recommendations that it was possible to make. As
the same auditor remarked, an audit might sug-
gest:
that, yeah, you have people signing things
three times because that's what a good con-
trol system asks you to do. [. . .] Like I learned
a lot of it through bitter experiences, a lot of
this is because you go to a client and they say,
[. . .] it doesn't make any sense. Here is the
real world; you're not in the real world. And
so that's the kind of thing.
Client interactions had in¯uenced more than an
auditor's unwillingness to suggest a stylised,
``textbook'' system of accounting control (signing
things three times). It is clear that systems work
was understood in terms of its di?erence to what
was seen as simple, ``lower level'' (and less con-
sequential) book-keeping. One auditor would go
so far as to describe some audits as low prestige
``accounting'' jobs and others as higher level,
``non-accounting'' work, joining with colleagues in
seeing such systems work as requiring more intel-
lectual depth and a di?erent skill set to that
required for attest work. Systems work was largely
reserved to senior auditors (manager and above),
with audit assignments being carefully matched to
individual auditors' abilities, for example data
processing expertise, or unusually adept diplomacy
in client relations. Overall, systems audits were
seen as occupying a ``higher level'' than attest
work, with gradations between various types of
systems work. A recommendation to improve sys-
tems regarding data quality in MIS would, for
example, be seen as a low level, almost ``book-
keeping'' type of recommendation. More valued
recommendations focused on ``higher level''
issues, such as recommendations to evaluate whe-
ther the information gathered was in fact the kind
of information that the Department needed. I
return to these issues of categorisation later, but
would note the theoretical signi®cance of such
interpretive schema in de®ning a body of knowl-
edge (Foucault, 1977, 1988a).
As auditors understood it, the systems auditing
environment called for a di?erent mode of audit-
ing, in contrast to the standardisation of attest
audit, and a shared sensitivity as to what it was
possible to recommend. Non-standardisation pro-
vided dear areas of di?erence, and led to audit
work which was considered more challenging. As
one auditor described it,
If I were auditing receivables, I would know
what 1 had to do. If I was auditing inventory,
I would know what I had to do, but on this
one I'll have to decide, [. . .] In some you don't
go to [a] detailed level because it's, you know,
you can say that the issue and the criteria
would drive the extent of the work that you
do. And they have to be, if they are going to
be meaningful, they have to be related to
individual entities to that particular issue that
you are looking at. Whilst on attest there is a
general menu that you can go to.
There might be some variations based on
the particular systems that they have in place.
But you have the CICA Handbook, which
gives you all the standards that are already
articulated. And here there are none of those
things, so that's why its more challenging,
and maybe more fun, because you are really
using your intellect more.
To auditors the power of the wider social and
normative environment was self-evident. They
invested signi®cant amounts of time studying it.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 347
Of course, a great deal of information was gath-
ered through conversations, contacts with depart-
mental ocials, and a general knowledge of ``what
was going on.'' But besides this there were a vari-
ety of formal and informal procedures: Hansard,
the record of Legislative proceedings, was manda-
tory reading for all senior auditors, and local news
outlets were carefully studied (indeed, newspaper
stories had signi®cant e?ects on two of the three
audits studied). Though this knowledge was said
to aid attest work as well, all ``knowledge of busi-
ness'' work was charged against systems work in
the Oce's internal control statements;
19
scanning
the environment was ocially a cost of doing e-
ciency audit work.
Knowledge of business helped in the framing of
audit issues, procedures and eventual outcomes,
but it could also foreclose an audit issue at the
outset. For example, auditors were aware of gov-
ernment discussions regarding substantial reform
of the provincial health care system, including
consideration of payment systems for physicians
other than the traditional fee for service method.
In meeting to plan their audit on the physician's
agreement
20
they made it dear that wider norms of
what was within the range of acceptable debate
would often limit the kinds of things that they
would express at a given time.
Brian:
21
What we're getting at is a need to
fundamentally review the way that physicians
are compensated. There are things in process
that may go that wayÐnot the capping, but
moving away from fee for service. We have
some important things to say, but we don't
want to say things about what's already going
on.
Sam: There are ongoing plans to reform
health care. Leonard [senior health ocial]
says we should replace fee for service [. . .]. He
says we are working on this and maybe you
guys should have been on the boat then.
Fred: We couldn't have talked about fee for
service then.
Brian: No, it wouldn't have been strategically
wiseÐit would have been impossible to say
that then.
Knowledge of business, and of the allowable
discourses of the day, e?ectively mapped out the
range of what it was possible for auditors to
recommend. This had important e?ects in the
operationalisation of eciency auditing, and in
the eventual outcomes of this work. But this was
not a slavish attendance to the concerns of those
in positions of power. Instead, auditors limited
themselves by attempting to make recommenda-
tions that would lead to action; an important
guide to this was provided by their ideas of e?ec-
tiveness, understood in terms of e?ecting change.
3.2. E?ectiveness
The idea of e?ectiveness underlay much of the
auditors' work. E?ectiveness was tied closely to
the Oce's mission statement goal of the
improvement of the ®nancial administration of the
province. Recommendations and audits were seen
as ine?ective if they did not lead to action.
Knowledge of the normative environment, and in
particular knowledge of what others might accept,
conditioned what the auditors believed to be
worth while, with worth usually equated to accep-
tance and action (``e?ectiveness''), as this inter-
view excerpt demonstrates:
Brian: If you [the auditors] were talking about
restructuring the health system a few years
ago. You would not have had any success.
Now you will have success. If you are talking
about the salary restraint programme, what-
ever, it depends on the context, and so there-
fore you have to be aware of the reality.
R [Researcher]: Success is a recommendation
that helps further the administration of the
province.
Brian: Absolutely.
R: That is in fact accepted, put into practice.
Brian: Right, I mean to us unless something
is, you know, we have to draw a line in terms
19
See below for discussion of the Oce's ``income state-
ments'' and their role in generating accountability.
20
See earlier descriptions of each audit.
21
All auditor's names have been changed for reasons of con-
®dentiality.
348 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
of long term things that you realise you really
have to keep on measuring, like the pensions
issue,
22
until it happens. So it's kind of a hard
rock, and you're trying to make it fertile. It's
real hard work to do that, so it's a long term
view. But there are other views as well, [. . .] I
suppose the analogy is a clock which is not
working, but it's right two times a day.
Would you have that or a clock which is
running ®ve minutes late. You know, [. . .] it
might not be the Cadillac, but, whatever,
Chevy, whatever, it is e?ective in terms of
what it does. So what we want is something
that is going to be e?ective. And the type of
recommendations that were made in the
Auditor General's report this year,
23
if they
had been made before, they would just have
had no impact. . .
R: Right.
Brian: So you need to be cognisant of what
the context is. You need to be cognisant of
the. . . realities that drive. . . I mean people
drive the, you know, I mean, it doesn't matter
how good you are. So, yeah, so the pointer I
use is whether they are going to he e?ective or
not.
This conception of e?ectiveness was a central
guide in the auditor's work. It operated both in
terms of the wider environment, and in terms of
speci®c, local knowledge of departmental person-
nel and practices. One senior auditor would speak
of a client department's ``maturity'' in terms of the
kinds of recommendations which could be made,
and which the client would understand and follow
through on. More complex recommendations
demanding greater ``maturity'' could be made. but
if they were ignored they would ultimately be
ine?ective. The result was a tailoring of recom-
mendations to individual clients' perceived matur-
ity in ®nancial administration.
24
The audits occurred in a time of signi®cant
public debate regarding government de®cits, with
a drastic government plan to balance the budget
through major spending reductions (Harrison &
Laxer, 1995; Taft, 1997). This background
brought out interesting dimensions of audit stra-
tegies. Auditors universally referred to the envir-
onment as a ``climate of restraint'', and would
sometimes refer to ``what had to be done'' often
expressing things as imperatives. Yet at other
times they would voice their concerns that deci-
sions might be made with inadequate considera-
tion, that appropriate management information
was not available to guide government action, and
that inadequate consideration of the consequences
of action could produce unforeseen failures. Such
issues would, for example, ground recommenda-
tions addressing whether management had appro-
priate information to measure the impact of
proposed health cuts. The audits were not an
abject capitulation to the ideals of the present.
Instead, auditors were tailoring their recommen-
dations to the environment in which they found
themselves, believing, as Brian expressed it, that
attainable ``Chevy'' eciency was preferable to an
apparently unattainable ``Cadillac'' standard.
E?ectiveness was a fundamental concern in the
auditors' daily work. It operated both in terms of
the wider environment, and in terms of speci®c,
22
The Oce of the Auditor General had made a series of
recommendations that the government address unfunded
liabilities in the pension plans of public sector employees, esti-
mated in 1992 to be some $6000 million (Oce of the Auditor
General of Alberta, 1992a). The government subsequently
began to re-negotiate with employees in order to revise con-
tributions to these plans and address this liability.
23
A series of demanding recommendations, including calls
for consolidated budgets and further e?ectiveness reporting,
had been made in that period's Auditor General Report (Oce
of the Auditor General of Alberta, 1993) which had been
released in an election year. The government
0
s ®nancial com-
petence had been seriously questioned in the light of rising
de®cits and the failures of government attempts to diversify the
provincial economy. The most prominent failure had been the
collapse of NovAtel, a cellular phone company, which resulted
in a conservatively estimated loss of some $544 million. The
government had requested a report into the issue from the
Auditor General. The eventual report (Oce of the Auditor
General of Alberta, 1992b) found governance irregularities in
NovAtel were symptomatic of those in a variety of government
agencies and called, amongst other things, for board appoint-
ments to be made on merit rather than patronage, a recom-
mendation which the government seemed to accept, under
signi®cant political pressure (Radcli?e, 1997).
24
There are parallels between this assessment of client
``maturity'' and the balancing that management consultants
must engage in to sense which recommendations a client might
be willing, or able, to implement.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 349
local knowledge of departmental personnel and
practices. As a consequence of this view of e?ec-
tiveness auditors understood that in their report
they would often refrain from making ambitious
``Cadillac'' recommendations; in the interests of
e?ectiveness, they would instead propose a
``Chevrolet'' solution. They recognised that the
eciency that they reported was not of an abso-
lute, publish and be damned variety. Rather, the
eciency audit reports that they made were tied to
the social climate of the time. Certain recommen-
dations might simply not be made in a certain cli-
mate, for fear that they would not be adopted, and
might sour client relations (as is witnessed by ear-
lier quotes regarding physician payment systems,
and auditors' belief that to question these some
years earlier ``would not have been strategically
wise''). Still other recommendations would be
carefully phrased and prepared so as to appeal to
popular or local discourse, regarding, for example,
contemporary government policy, popular man-
agement literature or ideas known to be in play
locally, within client departments.
25
Because of these concerns the communication of
audit ®ndings was attended to with great care. The
phrasing of a recommendation was seen by audi-
tors as another potential area in which to enhance
the prospect of acceptance, enactment and hence
e?ectiveness. This produced great attentiveness in
the development of recommendations; meaning
and rhetoric were both attended to in detail in a
series of meetings in which draft recommendations
were edited word by word. As one auditor
expressed it to a colleague, ``We need to commu-
nicate a message.'' To illustrate the detail of this
consideration, and the speci®c in¯ections in
meaning that were purposively introduced by the
auditors as they worked, I now reproduce my
notes from one of these editing meetings concerning
the MIS audit recommendations. In the following
notes quotations of auditors' comments are inter-
spersed with my observations of their activities in
editing recommendations.
Audit Manager (AM): I don't think there are
any speci®c guidelines.
Senior Audit Manager (SAM): [Explaining to
me] We want to put MIS in context to say
that really you are receiving all this information.
[Substitutes ``monitor'' for ``evaluate'' hospi-
tal performance, p. 4.]
AM: When you see the numbers even the
nursing cost doesn't make sense.
SAM: We need to have something here to say
why.
[Types: ``due to the long period of imple-
mentation''.]
We're really saying that you missed the boat
because the hospitals have changed or the
infrastructure has changed.
[Adds, ``for example ****Hospital X,''
26
p. 4.]
We're covering a fair bit of stu? on MIS
which leads me to say why don't you bullet it,
and show it under headings?
[Types: ``Problems:
. Period
. Data Quality
. Lack of programme de®nition
. Dynamic Nature''].
It's changing all the time.
AM: Actually that's another point, the fact
that with the reduction in the budget the
availability of resources to do this is changing.
SAM: This is a di?erent point.
AM: No it isn't. . .
[Moves recommendation up to the top.]
[Text Reads: ``The Department will not use
data until. . . 1994/95''].
SAM: We have to be careful because that
assumes it is used. But based on the data
quality [. . .] based on what we know they are
unlikely to be able to use this data.
25
Conventional ideas of audit e?ectiveness as being a ful®l-
ment of objectives set by the client stand in contrast to ``e?ec-
tiveness'' as it was employed by the auditors here. This sense
focuses much more on the positioning of recommendations in
the role of ``e?ect''. For auditors it was important that their
work be seen to encourage the client entity to change its prac-
tices. As discussed above, such change was seen by them as
establishing the success of an audit.
26
A major hospital in the City of Edmonton.
350 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
[. . .]
SAM: [Text Reads: ``Hospitals are not using
information for decision-making.'']
We have no evidence. We have to say they are
unlikely to use the information. They might
be using it, but its unlikely to in¯uence deci-
sion making.
We just want to make sure we communicate
that message, and that it's the right message
that's communicated.
[Edits text.]
A number of streams of discourse in¯uenced the
interpretations that are seen in this meeting.
E?ectiveness is a major theme, attended to as
auditors focus in detail on the communication of
their report material. In some ways this focus was
much like that of other authors, trying to express
themselves as dearly as is possible. But this com-
munication is signi®cant because of its purposive-
ness, understood at the time of its development by
its authors. The auditors approach their work with
the aim of increasing the e?ectiveness of their
recommendations. The particularities of their
conception of e?ectiveness lead them to consider
how the recommendations might be made as per-
suasive as possible, so as to enhance acceptance
and enactment. It is for this reason that the audi-
tors speak of the importance of communication:
SAM: We just want to make sure that we
communicate a message, and that it's the
right message that is communicated.
Of course, in showing the recommendations at a
somewhat nascent stage the meeting indicates
some uncertainty in the development of audit
recommendations. Although they had been pur-
sued purposively over the investigative stage of the
audit, the meaning of recommendations could be
signi®cantly altered in the detail of their expres-
sion (as the auditors seemed to be aware). In the
presentation of their recommendations to the
management of particular government entities,
however, it was important that such suggestions of
uncertainty be erased. The auditors' activity was
again oriented to acceptance and enactment of
recommendations, and hence con®rming their
work as e?ective. Elements of the tailoring process
that this involves can be seen later in the same
meeting as auditors work with the language of
contemporary popular management literature.
Osborne and Gaebler's (1992) work on ``reinvent-
ing'' government was much cited in Alberta, and
elsewhere, at the time. A series of similar terms
focused on popular management concerns for
reform in practice: auditors were aware of these
appeals, and deployed these in the text of their
recommendations. The appeal may have been self-
consciously made, with some elements of irony,
but it was clearly purposive: the language of this
literature was used to promote the auditors'
recommendations as a modern and attractive
approach to management, as evidenced in this
later exchange from the same meeting:
SAM: Restructuring, re-engineering, re-
inventing, these are the words involved, eh?
AM: Right-sizing.
I would point to the simultaneous operation of
the various aspects of the knowledge and practice
that I outline here. As noted earlier, it is analyti-
cally interesting to strip out dimensions of knowl-
edge and practice, and to analyse each in turn. But
in the ®eld particular discursive representations
and the practices that they supported acted in
conjunction with each other. Thus in the editing
session above, audit behaviour is seen to draw on
these guides. Signi®cantly, the strategies of this
auditing are shown directly as auditors speak
overall of what they are trying to achieve in their
recommendation. Their reports are not a dis-
passionate, uninterested account of activity in the
world; nor do auditors understand them as such.
Rather, the report material is itself a key element
in the auditors' active pursuit of change, as they
themselves make clear later in their meeting to
draft report material:
SAM: The MIS thing [. . .] is not making a
di?erence in terms of costs.
We have to make this a loud message that
this is costing you a bundle and that you're
not going to get what you wantÐthat there is
no pot of gold at the end of the rainbow.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 351
AM: We're really focusing in now on MIS.
Their programmes are anything from youth
shelters to drugs for aids.
[. . .]
SAM: They're really in the dark.
It's like a manufacturer making a product
without thinking of the consumer. It won't
work.
This purposive approach to auditing is again
seen to rely on auditors' detailed knowledge of
their client departments' a?airs, the interests of
senior Departmental ocials, and the more gen-
eral normative environment. The recommenda-
tions considered in relation to MIS were themselves
built upon a more general impression of the his-
tory and character of the Department, as auditors
themselves interpreted it. This more general
knowledge set the scene for the recommendations
that auditors would eventually make, as one audi-
tor commented, ``They collect information without
a clear understanding of what they want.''
Auditors could not be similarly accused of
¯oundering around in a sea of information; rather,
the purposiveness with which they approached
eciency auditing provided a coping mechanism
to deal with the prospect of information overload.
In e?ect, ideas of e?ectiveness extended to the
collection of evidence. Once an initial under-
standing had been developed amongst the audit
team, inquiry shifted from a general prospecting
for data to acquisition of material that upheld
their shared view. At an early stage in the MIS
audit, for example, an auditor spoke of the
remainder of the audit as being a process of
``gathering evidence to support the recommenda-
tion.'' The auditor further argued that,
To be successful, and to have the recommen-
dation adopted, you have to come up with
numbers, these are far more convincing than
generic opinion. I will get contacts for mag-
netic media, and test the data for anomalies
and errors.
The Oce had signi®cant computer resources,
and auditors were able to download data ®les
from client departments to their own machines for
later analysis. The Oce's Annual Report made
much of this capability as a sign of expertise
(Oce of the Auditor General of Alberta, 1994).
Yet even in these most overtly technological
aspects of auditing, the gathering and testing of
data was predicated on a need for persuasive
material.
27
Later, the same auditor spoke in an
audit planning meeting of the prospect of using
the Oce's custom designed audit software. The
auditor introduced the idea of using computer
audit software, ®rst speaking of it in purely tech-
nical terms, and then explaining that the data
would not particularly be used in a process of dis-
covery or invention, rather it would be used to
illustrate recommendations to the Department
using the Department's own data. The ``data'' and
the testing was useful because of its rhetorical
powers, and a belief that information could be
found to support a recommendation, as the audi-
tor remarked:
once we get the issue con®rmed and what we
want to do, we'll say that, given that, what
information do we need to convince manage-
ment that in fact they need to change?
Thus such apparently formally rational techno-
logies as computer auditing were themselves envi-
saged as part of a wider approach to the
production of a persuasive and ultimately ``e?ec-
tive'' recommendation.
28
I have earlier argued that the programmatic
(such as the legislated mandate, PSAAC state-
ments, etc.) provided auditors with some initial
framework, but that substantial areas were left to
27
Indeed, the auditors showed a keen awareness of the
rhetorical power of accounting numbers that has only lately
entered the literature
28
As the MIS audit developed, it became apparent that
computer auditing would not be used. The auditor noted in a
research interview that examination of committee minutes and
investigation of matters through interviews with departmental
ocials accounted for the bulk of his work, and speculated that
this was due to the more ``macro'' level that the audit had
taken, focusing more on the overall goals of the system and its
capacity to meet them. As an increasingly ``higher level'' audit,
it had moved further away from what were seen as more mun-
dane, ``book-keeping'' issues, hence large scale data analysis
was seen as unnecessary.
352 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
be worked through in the ®eld. That is what is seen
here, as auditors addressed their mandate, apply-
ing the audit technologies that they had matched
with their work. These technologies themselves
in¯uenced the nature and character of this audit
work, as auditors strove to report in ways which
would be e?ective, mindful of departmental and
governmental norms and values, knowledge of
which was essential in determining which recom-
mendations had a realistic hope of e?ecting
change, and hence being construed as worthwhile.
Audit technologies involved such formal devices
as audit memoranda, planning meetings, ®eldwork
and so on, but beyond this the social grounded-
ness of eciency auditing, and the tacit knowledge
that this involved, was itself a technology that
allowed eciency audits to function. All audit
procedures were approached with a carefully
developed sense of the normative climate in which
this work was conducted. Auditors formally
recognised the importance of this climate through
their attempts to track it in their ``knowledge of
business work''. But knowledge of business was
not simply a discrete stage; rather, it infused the
audit process as a whole, and had important
e?ects on the kinds of recommendations which
auditors felt would be e?ective. This dense and
inter-related social detail contrasts with the rela-
tively discrete stages that might ®rst be read into
Miller and Rose's analysis of the technological,
creating a sense of technologies providing a simple
working through of broader political programmes
(Miller & Rose, 1990). Instead a fuller apprecia-
tion of the technological suggests that in their
workings audit technologies involve ongoing
negotiation and recon®guration of their subject
and ends.
3.3. Accountability: sorting the good from the bad
The operationalisation of eciency auditing can
be more closely studied through an examination of
conceptions of accountability, both of the auditee
and the auditor, as these were worked through in
the ®eld. Ideas of accountability were formalised
in the audit report, and in internal ®nancial state-
ments, which auditors used to track their perfor-
mance. I track accountability to provide detailed
illustration of the complexity of some of the
environmental considerations that fed into con-
ceptions of what was or was not ``ecient''.
Auditors believed accountability was at the
centre of their work; as one senior auditor put it,
``everything we do here is about accountability.''
We might expect such notions, but perhaps would
not appreciate the speci®cities of this concern and
its application. In establishing recommendations
auditors constantly mediate a line between audit-
ing and management, trying to manage their own
position and that of others. I address some of
these issues in illustrating the complexities of
operationalising eciency audit. These issues
in¯uenced auditors' strategies for change.
The dilemma that an auditor must recommend
action, while not usurping management control,
was brought out in a government proposal that
the Oce report on newly devised performance
indicators for all government agencies and
departments (Dinning, 1993). At the time of these
audits it was unclear what the auditor's eventual
role would be, whether they would develop perfor-
mance measures or simply report on them. They
approached this new task with wariness. One used
the example of management's responsibilities for
®nancial statements: management would establish
standards and set goals, and auditors would report
on whether management met those goals. His rea-
soning was simple, ``we are not managementÐwe
don't want to be management. That's what you're
in danger of getting into with this. We are audi-
tors.'' In practice this divide was more complex.
The whole idea and purpose of the audits that I
followed was to e?ect change, to improve the
®nancial administration of the province. This was
the criterion of e?ectiveness. It inevitably involved
recommending courses of action, however general,
and considering alternatives to present practice.
Always the aim was to set the stage for change; in
speaking of client departments one of the most
recurrent phrases used was that the audit should
``get them to change the way that they do busi-
ness.'' Auditors, then, would consider the poten-
tial of di?erent methods of hospital organisation,
physician payment, and provision of health care
for welfare recipients. Their consideration of these
potential courses of action guided their work.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 353
Auditors understood accountability in terms of
this purposive approach: they compared practice
with their ideas of the ideal and wanted to e?ect
change (Miller & Rose, 1990).
Auditors addressed issues with widely varying
public pro®les. Sometimes they would question
programmes which were of public concern, as
envisaged in their manual (Oce of the Auditor
General of Alberta, 1991), while being wary of an
appearance of what was referred to as ``jumping
on the bandwagon''Ða loss of apparent detach-
ment and hence legitimacy. But their attention to
accountability brought them into direct contact
with political debates. The subsidies audit, for
example, highlighted the costs of providing health
care to welfare recipients and premium free cov-
erage to senior citizens.
29
Ultimately the auditors
believed that fuller disclosure would facilitate
accountability and control of these programmes:
at the time the relevant ®nancial disclosures
occurred in a series of footnotes to departmental
statements. One auditor addressed accountability
in terms of these programmes:
There may be some inequities here. This is
legislation, so we want to steer away from
that, but our intent is to provide information
to decision makers so that prudent decisions
can be made.
Accounting could be used to make things visible
(Burchell et al., 1985) so as to bring attention to
issues without directly prescribing action. Of
course, in the midst of an attempt to reduce gov-
ernment expenditures by 20%, revelation of the
full ®nancial costs of such programmes would be
more likely to promote their reduction, a prob-
ability of which the auditors were aware in their
quest for ``prudent decisions''. In this and other
attempts to avoid prescription while e?ecting
change, auditors engaged in a dicult process of
mediation.
Accountability applied both to the auditee and
the auditor, an impression that was felt keenly in
the ®eld. While the audit report was a basic
accountability document for management, it was a
report not just on management's performance, but
also on the auditor's. For auditors, the report was
``a showcase of our work''. Their rewards as indi-
viduals working within the Oce were determined
by their colleagues' perceptions of how they had
enhanced this ``showcase.'' Similarly, it was
believed that for the Oce as a whole being seen to
produce timely and e?ective recommendations was
critical in securing its rewards as an institution,
including its budget, in¯uence and prestige. These
conceptions suggest that in its ``showcase'' dimen-
sions the audit was performative for auditors
themselves. As a consequence, auditors were con-
cerned to avoid the use of their time and the O-
ce's resources in audit projects that might not
generate publishable recommendations. At one
stage, for example, the physicians audit was nearly
cancelled, largely because the auditor charged with
the project feared that a viable recommendation
might not result. Similarly, work that was expec-
ted to produce recommendations would be step-
ped up and expanded.
Auditors' strategies in conducting audit work
were in¯uenced by their own accountability. In
this setting, accountability operated at multiple
levels: individual auditors might feel accountable
to their colleagues or superiors, while auditors as a
whole might feel pressures ¯owing from the per-
ceived accountability of the Oce, and the judg-
ments that outside parties might make on the
quality of their work. Throughout this, their work
would continue to be seen by others as a central
element in the accountability frameworks of gov-
ernment, even as it was constructed by auditors as
an accountability mechanism for themselves. This
re¯exive accountability, in which audits were
understood as a test both of auditor and auditee,
and the varying arenas for accountability that it
presented, would produce a variety of interactions
and in¯uences for action.
29
It became apparent at an exit conference that there were a
number of political initiatives attending to the issue of senior
citizen's coverage. The government subsequently proposed
ending their premium free status as part of a wide-ranging
review of health care issues which also proposed deductibles on
``non-essential'' medical services, and funding much of the
health care system through such deductibles and premiums
rather than with general revenue funding (Mirosh & Oberg,
1993). These proposals were later only partially enacted.
354 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
Signi®cant resources were expended in mon-
itoring and evaluating audit e?ectiveness, and in
identifying the most valuable audit recommenda-
tions. As discussed earlier, auditors distinguished
``good'' recommendations from simple book-
keeping or other recommendations concerning
basic ®nancial control. This categorisation was
highlighted in the Oce's own internal account-
ability and control documents, not least in a
recently designed ``income statement.''
Like private ®rms, the Oce tracked employee
time, charging this out against individual audits.
In the income statement these costs were then
matched against simulated income. In the case of
attest work estimates were made of private sector
prices for each audit job. But for systems work
income was decided upon by a committee of the
most senior auditors. They would analyse those
major recommendations set out in the report as
major numbered issues (``type 1'' recommenda-
tions) together with those secondary recommen-
dations in surrounding text (``type 2''
recommendations). Each type was divided into
three classes of quality, with the most income for
the best type 1 recommendations, and the least for
the lowest category type 2 recommendations. The
most aggregated income statement showed the
major audit groups, with their costs and incomes,
and an Oce total. These statements were then
broken down into various levels, with employees
down to the manager level being treated as indivi-
dual quasi-pro®t centres.
The kinds of audit recommendations that were
valuedÐthose with a potential for high level
change, in changing the way that client depart-
ments ``did business'', were at the forefront of
concerns to demonstrate the Oce's ongoing
worthiness to maintain its mandate. The auditors'
activities would be evaluated in a world in which
government expenditures were to be substantially
reduced (Dinning, 1993; Taft, 1997). Auditors
attended to accountability both in their desire to
e?ect change in government administration and in
attempts to justify and sustain their professional
standing. These dual concerns can be seen in the
income statements' in¯uence. The audits that I
studied occurred in the ®rst full year of the state-
ment's operation. A senior auditor con®rmed that
proposals for systems audits (and potential
``income'') had doubled over the previous year.
It is analytically interesting to examine the con-
ceptions of accountability that the Oce's internal
income statements represented. My ®eldwork
occurred during the ®rst full year of the state-
ments' operation; at the time a discussion paper
regarding the income statements had been circu-
lated within the Oce. It presented income state-
ments and similar accountability measures as a
logical extension of the work that the auditors had
themselves conducted. In calling for enhanced
accountability for departments, and an attentive-
ness to such issues as economy, e?ectiveness and
eciency, the auditors were aware of the internal
logic of their recommendations. Moreover, they
felt compelled to apply this logic to their own
environment, in a context of de®cit reduction:
30
The world does not owe us a living; our stat-
utory appointment must be supported by the
merits of the work done. The challenge is to
demonstrate that we deserve to continue to
have the scope of work given to us.
We believe that when considering the
actions necessary to correct Alberta's operat-
ing de®cit, members of the general public
should ®rst expect the government to assess
the scope for delivering existing programmes
at less cost. If programmes have to be cut,
they should expect that the least e?ective be
cut ®rst. And if the general public is asked to
contribute signi®cantly more by way of taxes
and fees, then they need to be convinced that
their funds are being used e?ectively. We
cannot exclude ourselves from the rigour of
the analysis that we have recommended.
Therefore, we are, in all practical respects, in
competition with anyone who feels that they
can do better.
A number of themes intersect in the discussion
paper, as auditors consider their own accountability.
30
As discussed, the government has planned overall expendi-
ture reduction of 20% to achieve this end (Dinning, 1993;
Government of Alberta, 1994).
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 355
The auditors understood the accountability that
they had promoted for others as re¯exive (Gid-
dens, 1991). Having promoted enhanced scrutiny
for others they felt that they could not hold o?
similar scrutiny of their own a?airs. Auditors'
strategies in the conduct of their audits were
altered as a result, as they balanced consideration
of various audits with the likelihood of publish-
able ®ndings. Once started, an audit conducted in
an environment of income statements seemed
more likely to produce recommendations; in the
past not reporting might have been valued (as
suggested by the Auditor General Act of 1977,
which allowed the Auditor General to omit material
on which managerial action had been taken). Once
an auditors' ``income,'' and hence credibility to
his/her peers and superiors became dependent on
publication, these considerations materially chan-
ged. One can understand how proposals for report
material might have doubled in the ®rst full year
of the system's operation, with auditors develop-
ing the kind of routinely strategic behaviours that
they hoped to alter in others (cf. Munro, 1993).
Though the focus of audit recommendations
had been altered as a result of a changing climate
of accountability, it was by no means clear that
this di?erent emphasis had involved substantial
change in the ideas and practices that grounded
audit inquiry. In the way that auditors spoke of
and approached their work it seemed that con-
cerns to pursue their audits strategically, mindful
of generally speci®ed environmental knowledge,
had long been present. Similarly, e?ectiveness as it
was conceived in the Oce was a longstanding
concern. In this sense the income statements
seemed to provide a more ampli®ed statement of
what had long been the auditors' discursive
understanding of their role.
Income statements gave quantitative expression
to the knowledge with which auditors worked. In
this way, it seemed that the statements were an
artifact of this knowledge and served to reinforce
and reproduce related audit practices. In the face
of systems designed to control and scrutinise audit
activities so as to maximise publishable ®ndings
and therefore income, auditors seemed to become
more purposive. The administrative technology of
the income statement was informed by wider oce
discourse regarding ``e?ective'' recommendations
and the productive use of time, and the statements
in turn served to reproduce and elaborate this
understanding. Though ostensibly a simple inter-
nal control structure, a detailed technological
response operating at the level of data gathering,
codi®cation and calculation that Miller and Rose
(1990) envisage, the income statements had sig-
ni®cant social consequences in the conduct of
audit activity. Auditors themselves understood the
consequences of income statements and the
accountability that they represented. In a memo to
sta?, a very senior auditor noted that the state-
ments were themselves outgrowths of the Oce's
wider concern to emphasise a relation of cost to
inputs. The auditor was, ``con®dent that applying
[this] concept to our Oce will encourage
improvements in our work and will identify, and
therefore discourage, unproductive work.''
An awareness of ideas and practices regarding
accountability o?ers illustration of the complex
considerations feeding into eciency auditing.
In ®lling out the ambiguities left by rationalities
and programmes, and in developing their own
answers to how eciency auditing could be con-
ducted, auditors developed and used normative
frameworks regarding appropriate practice (Ber-
nauer, 1988; Foucault, 1972, 1977). These dis-
courses were technological resources, enabling
their work to take place. They informed practice,
providing the intellectual machinery with which
auditors worked. The shared knowledge with
which auditors worked became embedded into
such highly speci®c practices as the Oce's inter-
nal control systems, as is seen in the Oce's
income statements.
Of course, audit practices were themselves sub-
ject to quiet reform over time. The Oce's internal
procedures had developed over an extended per-
iod, with the introduction of such controls as
income statements being one aspect of agreed
procedure. The detail of these operations, while
material to practitioners, and indicative of the
ideas guiding their work, would rarely concern
outside parties. The speci®c practices of eciency
audit, like much else in governmentality, had been
left in the hands of experts (Foucault, 1978; Rose
& Miller, 1992).
356 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
4. Conclusions
Auditors General provide a tone of detachment
and calm to their reports. In the Audit report
material which has been worked on by hundreds
of people is expressed as the action of one Ocial,
the Auditor General [``I am not yet satis®ed that
expenditures. . .'' (Oce of the Auditor General of
Alberta, 1993)]. These reports are persuasive
documents; indeed, they are treated with rare
reverence by the press. Perhaps this is unsurprising
given that they are presented with a powerful
combination of sober analysis and informed
authority. But the detachment of these reports,
along with their symbolic investment of author-
ship in a single auditor, is a useful ®ction designed
to maintain the appearance of the kind of formal
rationality that Weber observed in the modern
bureaucratic form (Weber, 1964). Auditors are
very much aware of their social and organisational
context, and their appreciation of these factors is a
central part of the technology that allows them to
operationalise eciency auditing.
The concepts of eciency around which audi-
tors build their reports draw on accounting and,
importantly, other professional knowledge. They
are grounded in conscientious ®eldwork and
information gathering. But their work does not
present an eciency that ®nds universal applica-
tion. In what they understand as their e?orts to
improve government's ®nancial administration,
auditors pay very signi®cant attention to the social
world in which they work. They map out the
dynamics of political and administrative policy,
they track the norms of government and of
departments, and they tailor their work for per-
suasiveness mindful of these norms.
For these auditors the social and normative
environments were facts which they had to
negotiate: social concerns were a ``reality'' with
which they had to work. Their attention to context
resulted in recommendations which took account
of numerous speci®cities: the ®nancial and opera-
tional circumstances of a client provided an
important base, but this was joined with an atten-
tion to the language, norms and ideas which circu-
lated both in the client department and in the wider
audit environment. Their ability to track shifting
normative contexts was a crucial technology that
enabled them to work e?ectively, since failure to
appreciate the rami®cations of such shifts would
result in their being viewed as irrelevant, as one
auditor made clear in anticipating client comments
(``Here is the real world, you're not in the real
world').
Auditors both reacted to and acted upon their
environment; this dialectical relationship illus-
trates the mutual constitution of organisations
and their environments (Neimark & Tinker, 1986).
We can see this in auditors' attempts to e?ect
change through repeated recommendations (in some
cases highlighting issues for a decade or more),
even though they recognised the diculty of such
a course. But as I quoted one auditor earlier, ``it's
kind of a hard rock, and you're trying to make it
fertile. It's real hard work to do that, so it's a long
term view. But there are other views as well. . .''
These other views again involved auditors in the
constitution of their environment, although in a
more selective and strategic sense. They attempted
to estimate the kind of change which was attain-
able at a given time, and to encourage that, even
though the auditors might themselves prefer a
higher ideal or standard. The result was a very
practical sort of eciency audit, one which relied
on a sophisticated reading of the environment so
as to gauge attainable change. In these cases one
year's audit had to be understood in the context of
other audits, with knowledge that the auditors
were, in many cases, aiming to progressively
change their client institutions. In appreciating
these stratagems I have pointed to some of the
particular and local considerations that the audi-
tors themselves faced, both as individual accoun-
tants with their own career aims and expectations,
and collectively as an Oce which wanted to
maintain constructive relations with its overall
government client. The accountability of these
audits was re¯exive.
We should also remember that auditors are
themselves socially situated, and that their own
normative context is subject to ongoing shifts.
Though auditors might themselves envisage their
strategies as leading ultimately to stable goals as
clients reached greater ``maturity'' or willingness
to change, it would he a mistake to see auditors'
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 357
own aims as a static or immutable model of gov-
ernment administration. Instead, their own ideas,
practices and values are modi®ed over time, and as
a consequence, their strategies undergo multiple
amendment. Eciency auditors face many of the
same problems of other professionals seeking to
conduct their work and maintain their status in an
uncertain and variable environment. For those
studying such problems the analytical task is very
much one of combining a concern with systematic
patterns of e?ects and intentions with the day to
day calculations, confusions, actions and inactions
which are part of these professionals' organisa-
tional lives (Radcli?e et al., 1994).
This paper examined the conduct of eciency
auditing. In addition to providing insights into the
way that such work is conducted, the paper o?ers
contributions in two other signi®cant areas: the nat-
ure of technologies, and the nature of eciency itself.
As JoÈ nsson and Macintosh (1997) argue, eth-
nography can o?er a way of extending knowledge
of theory, and managing theoretical change. While
ethnography has often been presented as being in
opposition to more overtly theorised approaches,
in this study it is approached with broad theore-
tical foundations, drawn from Miller and Rose's
discussion of technologies (Miller & Rose, 1990;
Rose & Miller, 1992), and with consideration of
prior Foucauldian work in accounting, and aspects
of Foucault's own work (Foucault, 1991; Gordon,
1991). In dealing with the operationalisation of
eciency auditing in the ®eld, ideas of the techno-
logical come to the fore. Yet although Miller and
Rose emphasise the importance of technologies
and other aspects of the ``micro-physics'' of power,
they do not o?er a detailed speci®cation as to
what constitutes a technology, nor do they discuss
how technologies might be examined or recog-
nised. Instead their work highlights the role of
technologies in relation to the kinds of pro-
grammes and rationalities that have been more
often examined in the new accounting research,
and o?ers a broad description of what technolo-
gies might involve, together with some examples
of modes of calculation, notation and the like
(Miller & Rose, 1990, p. 8).
One of the signi®cant bene®ts of ethnography
is its capacity for discovery and surprise. While
ethnographers may come to the ®eld with prior
conceptions as to what might be found, the
experience of extended immersion in the ®eld can
erode these perceptions (JoÈ nsson & Macintosh,
1997). I found that in conducting ®eldwork my
observations combined to form a powerful repre-
sentation of how auditors went about their
workÐone that was di?erent to my own pre-
conceptions, drawn from the literature. This
forced signi®cant re-interpretation of my theore-
tical framework. In particular, much of the formal
elements of the technologicalÐthe kinds of speci-
®ed procedures, documentary analysis and other
trappings associated with the popular image of
auditing as a bureaucratically rational practice
proved to be of relatively minor importance in the
®eld. As an example, the one time that I saw an
auditor refer to a manual was early in my ®eld-
work, when in an interview I asked a clumsily
worded question that appeared to the auditor to
be a test of his knowledge. The manual was dug
o? the shelf.
For some time I was concerned that in not see-
ing formal, routinised structures (audit tests, stan-
dardised working papers, etc.) I must have been
missing important elements of the work, my belief
being, in line with most written accounts of the
audit process, that such overtly technological fea-
tures dominated auditing. In the ®eld I realised
that auditors had themselves turned away from
such tools, having found them to be of limited
value. Rather, if technologies were addressed more
generally as that which allowed the ful®lment of
more general programmes, such as a mandate to
conduct eciency audits, then the ideas and prac-
tices that I observed that auditors relied on in
conducting their work were themselves technolo-
gical. This realisation led me to understand how
auditors came to see the local discourse and tacit
knowledge that they had accumulated as being the
essential technological resource that allowed them
to make eciency auditing tractable. In this light,
while much energy has been expended in the
``new'' accounting research in tracing broader,
more publicly available discourses regarding the
programmatic (Miller & O'Leary, 1987; Morgan &
Willmott, 1993), it may be that to uncover what it
is that makes the craft of accountancy feasible
358 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
(Hopwood, 1988b) research will have to develop
an altogether more localised focus, attending to
detailed practitioner understandings and practices
(cf. JoÈ nsson, 1998).
While analysis of the role and nature of tech-
nologies is signi®cant, in mapping out the way
that eciency auditing was conducted this work
presents a questioning of over-arching concepts
such as ``eciency.'' Eciency audit reports are
strategic reports; they are written to promote
action. Yet the Auditor cannot compel change;
they merely advise, persuade and monitor. The
audits on which this work is grounded shared the
aim of promoting change (auditors themselves
construe the e?ectiveness of their work in terms of
achieving change). The result of this strategic
auditing is an eciency that is very much for a
time and place: a contingent eciency which
attends to organisational and social dynamics as
much as to technical procedures of ®nancial
administration.
Though auditors' work did not involve the kind
of calculations or deliberations that an economist
might prefer, the widespread acceptance of the
authority of reports made by Auditors General
and others charged with eciency audit work
suggests that this activity is nevertheless legit-
imate. Eciency audit reports often frame public,
political and administrative conceptions of what is
or is not ecient in government, and so in¯uence
ideas of government business and of eciency
itself. Earlier work questioned the basis of e-
ciency auditing, casting doubts on its objectivity,
the universality of its rules, etc. (McSweeney &
Sherer, 1990). But such a critique depends on a
belief in the centrality of such formalised rules and
procedure, which were found to be peripheral
here. Indeed, such arguments are predicated on a
general appreciation of economic and bureau-
cratic rationalities, and an expectation that audi-
tors should work within them. What this research
shows is that there are other, less formalised ways
of thinking of, and approaching, eciency, and
that these undermine a sense of the univocality of
such concepts. As Hopwood argued eciency, ``is
a concept that can be subject to a wide variety of
interpretations'' (Hopwood, 1988a, p. 8). That, in
many ways, is the point that is underlined by these
ethnographic inquiries. Eciency auditors may
not analyse ``eciency'' through the more calcu-
lative, rational economic approaches that might
®rst be expected (Mankiw, 1997; McCloskey,
1994), but their work is accepted. It presents a
view of eciency that in¯uences action and frames
debate, and is as a result socially signi®cant. In
response to this, appreciations of eciency in both
research and practice may be more accurate if
recon®gured so as to treat social constructions of
eciency as seriously as more formally rational
operationalisations of this concept.
Eciency is not the only concept that may be
revisited through a realisation of its multiple
potentialities. Corvellec (1995, 1997) analyses
``performance'' from a literary and narrative
standpoint, pointing to the diculties of translat-
ing the idea of performance from one language to
another. Just as general ideas of eciency may
diverge from the particular operationalisations
seen here in eciency auditing, so ideas of perfor-
mance seem to be speci®c to given languages,
times and other semantic networks. The e?ect of a
questioning and reconsideration of terms such as
eciency or performance is to undermine the
modernist attractions of much modern manage-
rial, accounting and auditing knowledge, and tar-
nish the veneer of formal rationality that seems to
make these areas attractive (Weber, 1970). In a
time in which ideas of rationality and eciency
seem to enjoy broad political support it may be
that such questioning is not fatal to the usefulness
of such conceptions as motivation for action, but
a more socially sophisticated understanding may
lead to their being used with greater reticence, self-
awareness and care.
Acknowledgements
Dean Neu, Chris Poullaos, Michael Power,
Steve Salterio, Mike Sherer and the anonymous
reviewers of the Fifth Interdisciplinary Perspec-
tives on Accounting Conference, the 1998 Amer-
ican Accounting Association Conference (New
Orleans, LA) and AOS have made helpful com-
ments on this paper. Further helpful comments
have been made by participants in seminar series
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 359
at The Weatherhead School of Management, Case
Western Reserve University, the University of
Alberta, McGill University, the University of New
Brunswick at Saint John, the University of North
Texas, the University of Toronto, the University
of Western Ontario and Wilfrid Laurier Uni-
versity; I am grateful for their guidance. Further
constructive comments are welcome. I am grateful
to my dissertation committee, David Cooper
(Chair), Mike Gibbins, Bob Hinings, Barbara
Townley and Jere Francis (external examiner), for
their support, encouragement and advice. My ®eld
research was made possible through the generous
co-operation of the Oce of the Auditor General
of Alberta. I thank all those members of the
Oce, both past and present, who have helped me
in the course of this research. The Oce does not
necessarily endorse the opinions expressed herein.
I gratefully acknowledge the support of the Ph.D.
programme and Faculty of Graduate Studies,
both of the University of Alberta, and of the
Research Committee and my colleagues at the
Weatherhead School of Management.
References
Abbott, A. (1988). The system of the professions: An essay on
the division of expert labour. Chicago: The University of
Chicago Press.
Alberta, H., (1977). Legislative Assembly of Alberta. Edmon-
ton, Alberta: Queen's Printer.
Barrett, M., Cooper, D. J., & Jamal, K. (1998). Managing
across the `Friction of space': transformations in the audit
planning process (working paper) Edmonton, AB: University
of Alberta.
Becker, H. S. (1989). Tricks of the trade. Studies in Symbolic
Interaction, 10, 481±490.
Bernauer, J. W. (1988). Michel Foucault's ecstatic thinking. In J.
W. Bernauer, & D. Rasmussen (Eds.), The ®nal Foucault
(Cambridge, MA: MIT Press).
Boland, R. J. (1989). Beyond the objectivist and the sub-
jectivist: learning to read accounting as text. Accounting,
Organizations and Society, 14, 591±604.
Boland, R. J. (1993). Accounting and the interpretive act.
Accounting, Organizations and Society, 18, 125±146.
Broadbent, J., & Guthrie, J. (1992). Changes in the public sec-
tor: a review of recent `alternative' accounting research.
Accounting, Auditing and Accountability Journal, 5, 3±31.
Burchell, S., Clubb, C., & Hopwood, A. G. (1985). Accounting in
its social context: towards a history of value added in the United
Kingdom. Accounting, Organizations and Society, 10, 381±413.
Buttery, R., Hurford, C., & Simpson, R. K. (1993). Audit in the
public sector. Hemel Hempstead, UK: ICSA Publishing.
CICA (1991). CICA handbook. Toronto, ON: The Canadian
Institute of Chartered Accountants.
Co?ey, A. J. (1993). Double entry: The professional and orga-
nizational socialization of graduate accountants. Ph.D. the-
sis, University of Wales College.
Co?ey, A. J. (1994). Timing is everything: graduate accoun-
tants, time and commitment. Sociology, 281, 943±956.
Corvellec, H. (1995). Translating management accounting
terms: the case of performance. Advances in International
Accounting, 8, 129±149.
Corvellec, H. (1997). Stories of achievements: Narrative features
of organizational performance. New Brusnwick, NJ: Trans-
action Publishers.
Cushing, B. E., & Loebbecke, J. K. (1986). Comparison of audit
methodologies of large accounting ®rms. Sarasota, FL:
American Accounting Association.
Denzin, N. K., & Lincoln, Y. S. (1994). Introduction: entering
the ®eld of qualitative research. In N. K. Denzin & Y. S.
Lincoln (Eds.), Handbook of qualitative research (pp. 1±17).
Thousand Oaks, CA: Sage.
Dinning, J. (1993). A ®nancial plan for Alberta: Budget '93.
Edmonton, AB: Queen's Printer.
Dirsmith, M., & Covaleski, M. (1985a). Informal communica-
tions, nonformal communications and mentoring in public
accounting ®rms. Accounting, Organizations and Society, 22,
149±170.
Dirsmith, M., & Covaleski, M. (1985b). Practice management
issues in public accounting ®rms. Journal of Accounting,
Auditing and Finance, 5±21.
Dirsmith, M. W., Heian, J. B., & Covaleski, M. A. (1997).
Structure and agency in an institutionalized setting: the
application and social transformation of control in the big
six. Accounting, Organizations and Society, 22, 1±17.
Elliott, R. K. (1994). The future of audits. Journal of Accoun-
tancy, 178, 74±82.
Foucault, M. (1972). The archeology of knowledge. New York:
Basic Books.
Foucault, M. (Ed.) (1975). I, Pierre RivieÁ re, having slaughtered
my mother, my sister and my brother. . . A case of parracide
in the 19th century. F. Jellinek (trans.).
Foucault, M. (1977). Discipline and punish. New York: Vintage
Books.
Foucault, M. (1984). Preface to the history of sexuality, Vol. II.
In P. Rabinow (Ed.), The Foucault reader (pp. 333±339).
New York: Pantheon Books.
Foucault, M. (1988a). Critical theory/intellectual history. In L.
D. Kritzman (Ed.), Michel Foucault: Politics, philosophy,
culture: Interviews and other writings 1977±1984 (pp. 17±46).
New York: Routledge.
Foucault, M. (1988b). Politics and reason. In L. D. Kritzman
(Ed.), Michel Foucault, Politics, philosophy, culture: Inter-
views and other writings 1977±1984 (pp. 57±85). New York:
Routledge.
Foucault, M. (1991). Politics and the study of discourse. In G.
Burchell, C. Gordon, & P. Miller (Eds.), The Foucault e?ect:
360 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362
Studies in governmentality (pp. 53±72). Toronto, ON: Har-
vester Wheatsheaf.
Gambling, T. (1987). Accounting for rituals. Accounting,
Organizations and Society, 12, 319±329.
Geertz, C. (1988). Works and lives: The anthropologist as
author. Stanford, CA: Stanford University Press.
Giddens, A. (1991). Modernity and self-identity: self and society in
the late modern age. Stanford, CA: Stanford University Press.
Glaser, B. G., & Strauss, A. L. (1967). The discovery of groun-
ded theory: Strategies for qualitative research. Chicago, IL:
Aldine Publishing Company.
Gordon, C. (1991). Governmental rationality: an introduction.
In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault
e?ect: Studies in governmentality (pp. 1±51). Toronto, ON:
Harvester Wheatsheaf.
Government of Alberta. (1994). A better way: Securing Alber-
ta's future, Edmonton, Alberta: Queen's Printer.
Grey, C. (1994). Career as a project of the self and labor pro-
cess discipline. Sociology, 28, 479±497.
Gutting, G. (1989). Michel Foucault's archaeology of scienti®c
reason. New York: Cambridge University Press.
Hammersley, M. (1992). What's wrong with ethnography? New
York: Routledge.
Hammersley, M., & Atkinson, R. (1983). Ethnography: princi-
ples in practice. New York: Routledge.
Hanlon, G. (1994). The commercialisation of accountancy:
Flexible accumulation and the transformation of the service
class. Basingstoke, UK: Macmillan.
Harrison, T., &Laxer, G. (Eds.). (1995). The Trojan horse: Alberta
and the future of Canada. New York: Black Rose Books.
Hopwood, A. G. (1988a). Accounting and the domain of the
public: some observations on current developments. In A. G.
Hopwood (Ed.), Accounting from the outside. New York:
Garland.
Hopwood, A. G. (Ed.). (1988b). Accounting from the outside.
New York: Garland.
Hopwood, A. G. (1994). Accounting and everyday life: an intro-
duction. Accounting, Organizations and Society, 19, 299±301.
Hopwood, A. G. (1998). Exploring the modern audit ®rm: an
introduction. Accounting, Organizations and Society, 23, 515±
516.
Hoskin, K. (1994). Boxing clever: for, against and beyond
Foucault in the battle for accounting theory. Critical Per-
spectives on Accounting, 5, 57±85.
Hoskin, K. W., & Macve, R. H. (1988). The genesis of
accountability: the West Point connections. Accounting,
Organizations and Society, 13, 37±74.
JoÈ nsson, S. (1998). Relate management accounting research to
managerial work! Accounting, Organizations and Society, 23,
411±434.
JoÈ nsson, S., & Macintosh, N. (1997). Cats, rats and ears: mak-
ing the case for ethnographic accounting research. Account-
ing, Organizations and Society, 22, 367±386.
Kohler Riessman, C. (1993). Narrative analysis. Newbury Park,
CA: Sage Publications.
Lather, P. (1993). Fertile obsession: validity after post-
structuralism. Sociological Quarterly, 34, 673±693.
Latour, B., & Bastide, F., (1986). Writing scienceÐfact and
®ction: the analysis of the process of reality construction
through the application of socio-semiotic methods to scien-
ti®c texts. In M. Callon, J. Law, & A. Rip (Eds.), Mapping
the dynamics of science and technology. London: Macmillan.
Livingston, E. (1985). The ethnomethodological foundations of
mathematics. Boston, MA: Routledge & Kegan Paul.
Livingston, E. (1995). An anthropology of reading. Bloo-
mington, IN: University of Indiana Press.
Loft, A. (1986). Towards a critical understanding of account-
ing: the case of cost accounting in the UK. Accounting,
Organizations and Society, 11, 137±170.
Macintyre, A. (1985). After virtue. London: Duckworth (®rst
published 1981).
Mankiw, G. N. (1997). Principles of economics. New York:
Dryden Press.
Manninen, A. (1993). The problem and the solution of
accounting in a text. Accounting, Management, and Informa-
tion Technologies, 3, 33±49.
McCloskey, D. N. (1994). Knowledge and persuasion in eco-
nomics. Cambridge: Cambridge University Press.
McSweeney, B., & Sherer, M. (1990). Value for money auditing:
some observations on its origins and theory. In D. J. Cooper, &
T. M. Hopper (Eds.), Critical accounts (pp. 294±312). London:
Macmillan.
Miller, P., Hopper, T., & Laughlin, R. (1991). The new
accounting history: An introduction. Accounting, Organiza-
tions and Society, 16, 395±403.
Miller, P., & O'Leary, T. (1987). Accounting and the construc-
tion of the governable person. Accounting, Organizations and
Society, 12, 235±265.
Miller, P., & Rose, N. (1990). Governing economic life. Econ-
omy and Society, 19, 1±31.
Mirosh, D., & Oberg, L. (1993). Starting points: Recommenda-
tions for creating a more accountable and a?ordable health
system. Edmonton, AB: Alberta Health Planning Secretariat.
Mishler, E. G. (1990). Validation in inquiry-guided research:
the role of exemplars in narrative studies. Harvard Educa-
tional Review, 60, 415±442.
Morgan, G., & Willmott, H. (1993). The `new' accounting
research: on making accounting more visible. Accounting,
Auditing and Accountability Journal, 6, 3±36.
Morton, S. (1993). Strategic auditing for fraud. The Accounting
Review, 68, 825±839.
Munro, R. (1993). Accountability and the new commercial
Alberta. Critical Perspectives on Accounting, 4, 339±359.
Neimark, M. D., & Tinker, A. M. (1986). The social construc-
tion of management control systems. Accounting, Organiza-
tions and Society, 11, 369±395.
Osborne, D. E., & Gaebler, T. (1992). Reinventing government.
Reading, MA: Addison-Wesley.
Oce of the Auditor General of Alberta. (1991). Sta? manual
volume 2 section 18: systems auditing. Edmonton, AB: Oce
of the Auditor General of Alberta.
Oce of the Auditor General of Alberta. (1992a). Annual
report of the Auditor General 1990±91. Edmonton, AB:
Queens' Printer.
V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362 361
Oce of the Auditor General of Alberta. (1992b). Report of the
Auditor General on NovAtel Communications Ltd. Edmonton,
AB: Queens' Printer.
Oce of the Auditor General of Alberta. (1993). Annual report of
the Auditor General 1991±92. Edmonton, AB: Queen's Printer.
Oce of the Auditor General of Alberta. (1994). Annual report of
the Auditor General 1992±93. Edmonton, AB: Queen's Printer.
Oswald, W. G., Schoep¯e, G. M., & Ahern, J. (1987). Sys-
tematic ®eldwork. Newbury Park, CA: Sage.
Pentland, B. T. (1993). Getting comfortable with the numbers:
auditing and the micro-production of order. Accounting,
Organizations and Society, 18, 605±620.
Pentland, B. T., & Carlile, R. (1996). Audit the taxpayer not
the return: tax auditing as an expression game. Accounting,
Organizations and Society, 21, 269±287.
Power, M. (1991). Educating accountants: towards a critical
ethnography. Accounting, Organizations and Society, 16,
333±353.
Preston, A. (1986). Interactions and arrangements in the pro-
cess of informing. Accounting, Organizations and Society, 11,
521±540.
Preston, A. M., Cooper, D. J., Scarborough, R., & Chilton, R.
(1994). Transformations in the code of ethics in the U.S.
accounting profession, 1917 & 1988: changing narratives of
legitimation. Presented at the Annual Conference of the
Canadian Academic Accounting Association, Calgary, AB.
PSAAC. (1985a). Public sector auditing statement 1: Auditing in
the public sector. Toronto, ON: Public Sector Accounting
and Auditing Committee of The Canadian Institute of
Chartered Accountants.
PSAAC. (1985b). Public sector auditing statement 2: Audit of
®nancial statements in the public sector. Toronto, ON: Public
Sector Accounting and Auditing Committee of The Cana-
dian Institute of Chartered Accountants.
PSAAC. (1988). Public sector auditing statement 4: Value for
money auditing standards. Toronto, ON: Public Sector
Accounting and Auditing Committee of The Canadian
Institute of Chartered Accountants.
PSAAC. (1990). Public sector auditing guideline 1: Planning
value-for-money audits. Toronto, ON: Public Sector
Accounting and Auditing Committee of The Canadian
Institute of Chartered Accountants.
Radcli?e, V. S. (1997). Competing rationalities in `Special'
government audits. Critical Perspectives on Accounting, 8,
343±366.
Radcli?e, V. S. (1994). Eciency audit: an assembly of ration-
alities and programmes. Accounting, Organizations and
Society, 23, 377±410.
Radcli?e, V. S., Cooper, D. J., & Robson, K. (1994). The
management of professional enterprises and regulatory
change: British Accountancy and the Financial Services Act,
1986. Accounting, Organizations and Society, 19, 601±628.
Reichard, G. A. (1939). The Navajo medicine man. New York:
J. J. Augustine.
Rose, N., & Miller, P. (1992). Political power beyond the state:
problematics of government. British Journal of Sociology, 43,
173±205.
Sacks, H. (1992). Lectures on conversation: Volumes 1 and 2.
Oxford: Blackwell.
Sarup, M. (1989). An introductory guide to post-structuralism and
postmodernism. Athens, GA: The University of Georgia Press.
Silverman, D. (1993). Interpreting qualitative data: methods for
analysing talk, text and interaction. Thousand Oaks, CA: Sage.
Sommer, B., & Sommer, R. (1991). A practical guide to beha-
vioural research. New York: Oxford University Press.
Spradley, J. P. (1979). The ethnographic interview. New York:
Holt, Rinehart and Winston.
Taft, K. (1997). Shredding the public interest: Ralph Klein and
25 years of one-party government. Edmonton, AB: University
of Alberta Press.
Turley, S., & Cooper, M. (1991). Auditing in the United King-
dom: A study of development in the auditing methodologies of
large accounting ®rms. New York: Prentice Hall.
Van Maanen, J. (1982). Fieldwork on the beat. In J. Van
Maanen, J. M. Dabbs, & R. R. Faulkner (Eds.), Varieties of
qualitative research. Beverly Hills: Sage.
Van Maanen, J. (1988). Tales of the ®eld: on writing ethno-
graphy. Chicago, IL: University of Chicago.
Van Maanen, J., & Pentland, B. T. (1994). Cops and auditors:
the rhetoric of records. In S. Sitkin, & R. J. Bies (Eds.), The
legalistic organization (pp. 103±151). Thousand Oaks, CA:
Sage.
Weber, M. (1964). The theory of social and economic organiza-
tion. New York: Free Press.
Weber, M. (1970). From Max Weber. London: Routledge.
Weick, K. E. (1995). Sensemaking in organizations. Thousand
Oaks, CA: Sage.
Werner, O., & Schoep¯e, G. M. (1987). Systematic ®eldwork:
foundations of ethnography and interviewing. Newbury Park,
CA: Sage.
362 V.S. Radcli?e / Accounting, Organizations and Society 24 (1999) 333±362

doc_915123025.pdf