Management Accounting Research, 1999, 10, 363-382

Article No. mare.1999.0115

L0
Available online at http://www.idealibrary.com on IDE b
*

Performance management: a framework

for management control systems research
David Otley”

This paper proposes a framework for analysing the operation of management control
systems structured around five central issues. These issues relate to objectives,
strategies and plans for their attainment, target-setting, incentive and reward struc-
tures and information feedback loops. Their central focus is on the management of
organizational performance. Because the framework has been inductively developed,
its application is ‘tested’ against three major systems of organizational control,
namely budgeting, economic value added and the balanced scorecard. In each case,
neglected areas of development are exposed and fruitful topics for research identified.
It is believed that the framework can usefully be developed further by its use in
analysing other instances of management control systems practice, and that case-
based, longitudinal studies provide the best route to this end.
0 1999 Academic Press

Key words: performance management; management control; strategy implementa-

tion; performance measurement; balanced scorecard.

1. Introduction

The measurement of the performance of business (and other) organizations has long
been of central interest to both managers and management accounting researchers.
However, management accounting has tended to restrict itself to considering only
financial performance, and to use frameworks and theories drawn primarily from the
discipline of economics. Even the attention that has been paid to the so-called
‘behavioural aspects’ of management accounting has been incorporated into the
economic approach through the development of agency theory. However, the disci-
pline of economics does not provide a sufficiently rich picture of the internal activities
of organizations to provide reliable guidance to the designers of management control
systems. Other approaches, most notably those based on critical theory, have been
used to study other aspects of the role and use made of accounting systems, but have
tended to concentrate on sectional interest rather than on overall control. T h e
intention of this paper is to provide a perspective more focused on the operation of

* KPMG Professor of Accounting, The Management School, Lancaster University, Lancaster LA1 4YX,
U.K.
Received 5 November 1998; accepted 8 September 1999

1044-5005/99/040363 + 20 $30.00/0 01999 Academic Press

364 D. Otley

overall control systems, and to do so by looking beyond the measurement of perfor-

mance to the management of performance.
It is recognized that ‘performance’ is itself an ambiguous term, and capable of no
simple definition. In particular, the term does not specify to whom the organization is
delivering its ‘performance’. We will begin at an organizational level of analysis and
assume that an organization that is performing well is one that is successfully
attaining its objectives; in other terms, one that is effectively implementing an
appropriate strategy. Nevertheless, it will become apparent that more attention will
need to be paid to the definition of performance from the perspectives of relevant
stakeholders.’ Furthermore, rather than trying to develop a well-articulated theory
from first principles, a more inductive approach will be taken that draws upon
previous experience in studying organizational control systems to identify some key
issues that seem to be relevant to many different organizations. The usefulness of the
framework so generated will be demonstrated by using it to analyse three examples of
extant control techniques, and to identify features of their application that appear to
have been neglected by both practitioners and researchers. In other words, an
inductive approach to the generation of theory is adopted, with the methods that this
approach suggests for future research being outlined in the discussion section.

2. Management control systems

Management control systems provide information that is intended to be useful to

managers in performing their jobs and to assist organizations in developing and
maintaining viable patterns of behaviour. Any assessment of the role of such informa-
tion therefore requires consideration of how managers make use of the information
being provided to them. The traditional framework for considering these issues was
developed by Anthony (1965) at the Harvard Business School under the title of
‘management planning and control systems’. This distinguished ‘management con-
trol’ from ‘strategic planning’ and ‘operational control’.
His approach was intended to achieve two aims. First, it was intended to broaden
the scope of information being considered beyond just accounting information.
Paradoxically, it was largely unsuccessful in achieving this, mainly because of its
deliberate neglect of ‘operational control’. Operational control was neglected because
it was apparent that different organizations used very different practices at the
operating level, so Anthony concentrated on the commonalties that existed between
them. Focusing on commonalties allowed the use of a common language capable of
including all organizational activities. Accounting provided such a language and
management control became largely synonymous with management accounting at a
time when this discipline of management accounting was in almost terminal decline.’

‘Although this is manifestly a functionalist approach in that it is concerned with the achievement of
organizational objectives, it does not preclude the study of power relationships between participants.
Critical theories and institutional economics have focussed on these issues, but have tended to neglect the
issue of the overall well-being and viability of the organization. The paper is an attempt to redress this
balance by explicitly concentrating on issues of overall control and sustainable performance.
’This can be seen with hindsight. Critiques, such as that of Johnson and Kaplan (19871, suggest that the
situation had become so acute in the mid-1980s that management accounting either needed to be
abandoned as being inimical with modern management, or else required a radical overhaul and revision.
The ‘new’ techniques of the later 1980s and 1990s can be seen as a response to this situation.
 Performance Management 365

Second, it brought issues of managerial motivation and behaviour into view. Here it
was much more influential in its effects, influencing much of the behavioural
management accounting work which was to dominate the 1970s and 1980s. At the
same time, a similar approach was being developed in Europe by Hofstede (1967)
and publicized in his now famous book, ‘The Game of Budget Control’, which still
represents one of the most comprehensive studies of its kind.
A further weak link in the management control systems framework was also
intentional. Its deliberate neglect of the process of ‘strategic planning’, which at best
it took as given and, at worst, ignored completely, was intended to simplify the
research questions asked. However, such deliberate neglect inevitably led to the
specification of control systems and measures that were common to all strategies.
Again, accounting measurement was stressed and non-financial performance mea-
sures were n e g l e ~ t e dAlthough
.~ it may well have been sensible to concentrate initially
on the core area of ‘management control’, it is now necessary to pay more attention
to the neglected elements of strategy and operations. This is particularly important as
contemporary organizations are themselves changing, illustrated by such develop
ments as business process re-engineering and de-layering, where the same manager
may well be responsible for some elements of strategy, management control and
operational contr01.~This paper represents a first step towards the aim of developing
a more complete framework for analysis.

3. The performance management framework

It will be argued that there are five main sets of issues that need to be addressed in
developing a framework for managing organizational performance that are repre-
sented as a set of question^.^ The questions are phrased in a normative tone,
reflecting a managerial perspective, but can easily be re-phrased descriptively for use
as a research tool. The questions themselves appear to remain constant, but organiza-
tions need to continually develop new answers to them. This is because the context in
which the organization is set is constantly changing and new strategies need to be
developed to cope with new operating environments.
The questions are as follows:

1. What are the key objectives that are central to the organization’s overall future
success, and how does it go about evaluating its achievement for each of these
objectives?
2. What strategies and plans has the organization adopted and what are the
processes and activities that it has decided will be required for it to successfully
implement these? How does it assess and measure the performance of these
activities?

31t is paradoxical that this occurred at a time when control techniques were being developed in practice
which sought to include a wide range of measures, both financial and non-financial. For example, the
General Motors system of performance measures, the development of which culminated in the 1970s (see
Johnson and Kaplan (1987)) included several non-financial indicators, and can been seen as the fore-runner
of the balanced scorecard approach.
4See Otley (1994) for a more detailed account.
5These questions represent a development of the approach first put forward in Otley (1987) and utilized
and developed by Fitzgerald and Moon (1996).
366 D. Otley

3.What level of performance does the organization need to achieve in each of the
areas defined in the above two questions) and how does it go about setting
appropriate performance targets for them?
4. What rewards6 will managers (and other employees) gain by achieving these
performance targets (or, conversely, what penalties will they suffer by failing to
achieve them)?
5. What are the information flows (feedback and feed-forward loops) that are
necessary to enable the organization to learn from its experience) and to adapt
its current behaviour in the light of that experience?

These questions relate very closely to some of the central issues of modern
management and management accounting practice. The first is concerned with the
definition of goals and the measurement of goal attainment) not just financially but
also in terms of meeting all stakeholder aspirations. Clearly, the relative importance
given to different goals may well reflect the relative power of different stakeholders.
However, the issue of evaluating organizational effectiveness cannot be addressed
without confronting these issues. The second is closely connected with issues of
strategy formation and deployment) and with very practical issues of business process
and operations management. It represents the codification of the means by which
objectives are intended to be attained. The third question is more traditional and has
a long pedigree of research connected with it, but remains important) as is reflected
in the emphasis given to practices such as benchmarking. The fourth question has
tended to be neglected by those concerned with performance measurement as being
in the purview of the human resource management function. However, the inter-con-
nections between the two fields need to be better recognized to avoid the many
counter-productive examples of short-termism driven by financial incentive schemes
that are seen in practice. The final question has been considered in part by MIS and
MCS specialistsY7but still needs to be better linked to issues such as the ‘learning
organization’) employee empowerment and emergent strategy.
The remainder of the paper will first develop a theoretical justification for these
questions and present them as a framework for the study of organizational control
systems.8 It will then apply the framework to analyse three performance management
techniques) one traditional and two relatively new. This will demonstrate how the
attempt to develop answers to these questions can act as a guide to productive areas
of research and also as an aid to evaluating practical developments and suggesting
improvements. The techniques to be examined will be budgeting) Economic Value
Added, and the Balanced Scorecard.

6Rewards should be understood in the widest possible sense, and not be restricted to just short-term
financial rewards, important though these may well be.
7For an up-to-date approach, see Checkland and Holwell (1998).
‘The analysis will be conducted at an organizational level of analysis, typically that of a strategic business
unit which is relatively autonomous. However, this is only illustrative, and the framework can be applied at
the level of organizational sub-units, although some of the possibilities considered may be more restricted
because of corporate policies and decisions. Also, it is quite possible that different control configurations
and procedures may be appropriate in different organizational sub-units; in a full case analysis each of
these would need to be explored.
 Performance Management 367

4. Theoretical development

The contingency theory of management accounting suggests that there is no univer-

sally applicable system of management control but that the choice of appropriate
control techniques will depend upon the circumstances surrounding a specific organi-
zation. A central contingent variable is the strategy and objectives that an organiza-
tion decides to pursue. Not only are these objectives likely to heavily influence the
choice of performance measures to be used (i.e. the desired outcomes), but they also
must act as the criteria against which the contingent choices that have been made can
be evaluated (i.e. the ‘goodness of fit’ of the system).’ Any controlled system requires
objectives and goals against which its performance can be assessed (Otley and Berry,
1980); no specific contingent formulation is necessary to anticipate that the existence
of different goals is likely to involve the selection of different performance measures
and controls. The first question in the proposed framework is designed to explore and
elucidate organizational aims and objectives,” and the mechanisms that have been
put in place to measure and monitor goal attainment.
More recently, literature on the role of strategy as a specific contingent variable has
developed [see, for example, Simons (1995) and Langfield-Smith (1997)]. This
suggests that different types of organizational plans and strategies will tend to cause
different control systems configurations. However, the empirical studies conducted to
date have not yielded any firm conclusions about the nature of the most appropriate
connections between strategies and controls. In essence, strategies can be seen as the
means by which an organization has decided that its aims can be achieved. It is not
the purpose of this paper to elaborate upon the procedures by which strategies are
developed, nor on the relative roles of espoused (top-down) and emergent
(bottom-up) strategies [see Mintzberg (1994)]. Rather, it takes as given the organiza-
tion’s current strategies (as elucidated by means of the first question) and seeks to
investigate the means by which their achievement is pursued and monitored. How-
ever, the formal performance measurement system is seen as a major mechanism that
can be used to make explicit the set of means-end relationships that the organization
has developed as the methods it will use to implement its strategic intent. The second
question is therefore designed to explore and document the connections the organiza-
tion has made between its strategic intent (objectives) and the means by which it
hopes to realize them [see Goold and Campbell (1987)]. In practice, it is anticipated
that the process of exploring the issues raised by this question will often lead to senior
managers feeling the need to elaborate their strategies more precisely in order to
maintain a good ‘fit’ between objectives, plans and performance measures.
Strategies involve a time-frame as a key component. Many business strategies are
designed to effect ‘improvement’, often driven by competitive pressures. Even in the
public sector, the continual need to justify the use of resources produces similar
pressures for improvement and efficiency. The third question thus raises questions of
both effectiveness (to what extent are objectives expected to be achieved in a given
time-frame?) and efficiency (what level of resourcing is thought necessary to allow

’See Otley (1980) for a discussion of the issues this raises for the development of contingency theories, and
Chapman (1997) for a more recent review of the place of such theories in accounting research.
10
No assumption is made about the nature of these goals. The framework is quite compatible with a
stakeholder perspective which views organizations as seeking to meet the, potentially conflicting, require-
ments of different stakeholders.
368 D. Otley

such achievement?). There are a plethora of approaches and techniques currently

deployed in practice that have the aim of increasing goal attainment and reducing the
consumption of resources, such as continuous improvement) benchmarking) concur-
rent engineering and target-costing. Such approaches not only involve the technical
specification of goals which are required to be attained (in many cases, developed
from roots which go back to the Scientific Management movement)) but are also
necessarily concerned with issues of motivation and employee behaviour [see Mer-
chant and Manzoni (1989) for one example of such a study]. In popular management
theory, goal-setting and performance measurement play a pivotal role, expressed in
phrases, such as ‘what gets measured) gets done’.” Drawing on theories of motiva-
tion) both the process of goal-setting (e.g. employee involvement and participation)
and the content and nature of specific goals (e.g. ‘stretch’ targets; difficult but
attainable goals) have been studied in considerable detail in the psychological
literature. In addition) the encouragement of employees to suggest improvements to
production processes or product design, driven by the imperative of improvement)
has also received attention) with Japanese practices being particularly prominent.
In accounting terms, answers to all the above three questions are required in the
design and construction of budgetary planning and control systems. The first ques-
tions defines acceptable outcomes and results (often solely in profit or cost-terms);
the second is concerned with the development of plans by which the results are
expected to be delivered; and the third is concerned with the performance standards
which can be expected. A budget can only be drawn up when quantified answers to
all three questions are developed. However, the questions are not restricted to just
accounting measures and procedures. They may include many of the managerial
processes that operate within the organization and cross the functional divisions that
may exist. Many of the measurements involved may be non-financial (e.g. production
process measures) such as defect rates and cycle time; customer service measures)
such as on-time delivery statistics) etc.), or qualitative (e.g. measures of employee
morale, customer satisfaction and product innovation). The study by Coates et al.
(1992) gives many examples of the wide variety of measures that are used by
companies in different countries. Finally, some aspects of performance which are
believed to be important may not even be measured) perhaps because the level of
measurement available is inadequate or distorts the process being measured (e.g.
employee attitudes) customer preferences) R& D creativity). The framework of
questions should not be taken to imply that precise quantitative measurement and
defined target-setting procedures are universally desirable; it provides) first and
foremost) a framework within which the full range of extant practices may be
described) perhaps as a pre-cursor to attempting to evaluate their effectiveness.
Motivation and incentives are the main focus of the fourth question) which seeks to
examine the consequences that follow from the achievement (or the failure to
achieve) the performance targets which have been set. These issues are clearly related
to the structures and processes of accountability that exist within the organization)
and even to the overall governance procedures of the total organization. Increasingly
(in the U.K.) these consequences may involve explicit financial rewards (in terms of
11
However, it should not be assumed that explicit systems of performance measurement based on
common-sense design principles are always effective. See, for example, Austin and Gittell (1999) for some
interesting examples of counter-intuitive outcomes. They suggest two very different categories of models of
performance measurement, one based on compliance, the other based on ambiguity.
 Performance Management 369

salary increases or one-off bonus payments), but will also involve less tangible
consequences, such as recognition, status and reputation. Although there is a devel-
oping literature on the impact of payment schemes on employee and managerial
performance, this is dominated by U.S.-based work which may not transfer easily
across cultures." Moreover, in practice, the design of payment systems is very much
the province of the personnel or human resources function in most organizations, and
these systems may not be well-articulated with the extant performance measurement
systems. In any case, the intent of the question is to enable such connections as do
exist to be documented and, ultimately, for their motivational impact to be assessed
and evaluated.
Information is the necessary final ingredient to complete the control loop. In its
traditional feed-back form, information on actual performance is compared with
pre-set targets and standards and deviations used to signal the need for corrective
action. In addition, feed-forward (or planning) information may be used to predict
the need for corrective action before adverse consequences are observed. In both
cases, there is a role for immediate corrective action to rectify the perceived problem,
but also for double-loop learning to take place to improve the system in such a way
that errors do not occur again in the same way. Such uses of information are
ubiquitous in the contemporary organization, ranging from quality control charts to
financial budgets, and from improved training to the 'learning organization'. Nor
should the less formal uses of information be neglected; organizational cultures form
and are reproduced, at least in part, by the use of approving and disapproving
feedback signals of many types. A final key issue in documenting such feedback loops
is to distinguish the different timescales and learning processes involved. These
timescales may range from the instantaneous (in real-time production control sys-
tems) through hours, days, weeks, months, quarters, years and beyond. The learning
processes range from simple corrective action through to the revision of a corporate
strategy if it becomes apparent that the current strategy is proving ineffective.
The five areas identified by the above questions are therefore not novel. Studies
addressing aspects of them have been part of the management control and wider
management literature for many years. However, the integration of the five areas to
provide a description of the overall management control and performance manage-
ment systems of an organization is relatively novel. It is argued that the five areas are
heavily inter-connected and procedures introduced to address one question may well
impact upon the other areas identified. The framework is not intended to provide a
normative or prescriptive framework, but rather to provide a more comprehensive
descriptive framework within which the features of an overall control system can be
assessed and evaluated.
Ideally, the framework should now be applied in practice to examine the overall
control systems of an organization, and it is the author's hope that other researchers
will find it a useful tool along the lines pioneered by Fitzgerald and Moon (1996).
However, for the purpose of this paper, it will be applied to three major control
techniques. The aim is to illustrate that the questions posed give insight into the
techniques discussed, and raise a variety of interesting research questions that can be
explored. The case of budgeting is used to illustrate how a well-established area of

12
For example, the contrast between the openness about pay in the U.S. and the secrecy which exists in
the U.K., may cause significantly different reactions to performance-related rewards.
370 D. Otley

study has covered each of the areas mentioned, albeit rarely in an integrated manner.
Economic Value Added is used to demonstrate how a particular scheme of perfor-
mance measurement requires to address each of the areas mentioned, albeit with
narrow scope. Finally, the balanced scorecard is used to indicate a technique which
has addressed one or two questions in some detail, but which has also neglected other
important questions.

5. Budgeting

Budgeting has traditionally been a central plank of most organizations’ control

mechanisms, as it is one of the few techniques capable of integrating the whole gamut
of organizational activity into a single coherent summary. Performance is defined
essentially as profitability; in a profit centre, the overall measure of performance
combines an output measure (revenue) with an input measure (cost) and the
budgeting process seeks to keep the two elements in balance. Cost centres are more
problematic as results can no longer be measured in financial terms, and thus cannot
be directly compared with costs. The budgeting process tends to assume a given level
of output or sales and attempts to determine an appropriate level of ~ p e n d i n g . ’In~
order to develop a budget there is a need for an underlying plan by which the
organization’s objectives are expected to be achieved and which serves as the basis for
the cost structure underlying the budget.14
Target setting has long been seen as an important part of budgeting, with both the
process (e.g. participation) and the outcome (e.g. target difficulty) being the subject
of behavioural accounting research. Although reward structures and incentive schemes
have received some attention (particularly in the agency literature), the topic has been
more implicit than explicit in the budgeting 1 i t e r a t ~ r e . IHowever,
~ clearly much of
the influence of performance evaluation is driven by the desire of managers to please
their superiors and thus gain some form of reward (financial or otherwise). Finally,
feedback loops are assumed to exist because of the regular reporting of (typically
monthly) accounting variances and the consequences that these engender. The virtue
of the budgetary control process is that it provides an encompassing framework by
means of which all aspects of an organization’s activity are encapsulated into a single
set of financial statements against which actual outcomes can be monitored. How-
ever, the downside of the narrowness of the budgetary process has been likened to
driving a motor car solely by looking through the rear view mirror (and a mirror that
provides only an imperfect reflection, at that).
There appears to be a growing dissatisfaction amongst practitioners with current
budgeting practice [see, for example, Bunce et al. (1995)]. The rate of change in the
current environment for many businesses means that the annual budget process is too

13
This is especially true of discretionary cost centres; in production units, some surrogate measure of
output is often constructed (number of units; standard direct labour hours, etc.) and a linear ‘flexed’
budget developed. The literature on ABC also represents an attempt to develop more sophisticated ‘cost
drivers’ which will allow better budgets to be constructed.
14
However, it has to be recognized that the phenomenon of the ‘planless budget’ also occurs, with budget
numbers merely being extrapolated from past experience.
15
The literature on performance evaluation and style of budget use is also relevant, although the
connection with rewards is still usually implicit or expressed in terms of obtaining a good evaluation from a
superior [see Briers and Hirst (1990) for a useful overview].
 Performance Management 37 1

infrequent; but if frequent budget revisions are undertaken, they prove to be time-
consuming and can lead to control loss. The essentially hierarchical nature of
budgetary control is in stark contrast to the focus on value chains and business
processes that many organizations are adopting. The budget focuses only on financial
results and, worse, does not necessarily pay sufficient attention to the means by which
those results are to be achieved. Valid as these criticisms undoubtedly are, the
budgeting process still represents the central co-ordinating mechanism (often the only
co-ordinating mechanism) that most organizations have. It is therefore not to be
discarded lightly, but the key areas needing improvement must be addressed. Some
questions that arise from the preceding framework include:

How can budgeting be better tied into the achievement of strategic goals?
How can resource allocation be matched to strategic imperatives?
How can budgeting be adapted to monitor and control the business processes
along the value chain running from the extraction of raw materials through to
the delivery of products to the final consumer?
Are there better ways of setting budgetary targets than the usual incrementalism
based on historic achievement?
Can we avoid the distorting effects that arise when managers are given a reward
for achieving budget targets?
Can variances be used in processes of learning and adaptation rather than in the
apportionment of blame?
Above all, can the budget process be harnessed to add value to organizational
activities rather than representing a drain on organizational and managerial
resources?

The performance management framework thus flags up some vital issues for
studying and revising budgetary practice. These issues will be considered in more
depth by looking in some detail at two currently popular techniques that have been
more recently developed to improve organizational control. The first is a purely
financial performance measure, Economic Value Added [EVATM],I6 which it is
argued can avoid some of the performance measurement problems currently experi-
enced with other financial performance measures. The second is the Balanced
Scorecard approach developed by Kaplan and Norton (1992, 1996)) which explicitly
adopts a multi-dimensional framework. Although these are sometimes seen as com-
peting approaches, they will be regarded here as complementary, for reasons that will
become apparent. Both have been explicitly devised to allow a more structured
approach to performance management and to avoid some of the problems associated
with more traditional control methods, such as budgeting.

6. Economic value added

Economic Value Added has been developed by the Stern Stewart Corporation as an
overall measure of financial performance that is intended to focus managers’ minds
on the delivery of shareholder value.I7 The aim of the stock market quoted organiza-
16
EVATM is a trademark of the Stern Stewart Corporation.
372 D. Otley

tion is seen as maximizing shareholder value and it is therefore argued to be

important that the main measure of financial performance used within such organiza-
tions should be congruent with this objective. As is well known, most measures of
financial performance, such as profit or return on investment, suffer from inherent
defects that may cause dysfunctional decision-making on the part of managers.
EVATM,which is defined as accounting profit less a charge for capital employed, is
claimed to be less problematic in this respect. Indeed, the debate on this issue is well
known in the management accounting journals of the 1970s) where the concept of
residual income was extensively debated." On the face of it, EVATM is little more
than a new acronym for old-fashioned residual income.
In technical terms such a view is correct, although it understates the significance of
the development work undertaken by Stern Stewart. They state, for example, that
they have developed nearly 200 accounting adjustments that may need to be
undertaken to convert conventional accounting profit into a sound measure of
EVATM, although most organizations will need only to use about ten of these.
Although the reasons behind many of these adjustments are technical (for example,
to preserve the articulation between the balance sheet and the profit and loss account,
in the maintenance of a 'clean surplus' view of accounting profit), others are clearly
designed to be motivational in their impact. O'Hanlon and Peasnell (1998) demon-
strate how the Stern Stewart formulation comes to a balance between the extremes of
a cash flow (objective, but historic) measure of profit and a net present value
(subjective, but future-oriented) measure. Stern Stewart attempt to motivate the
increase of shareholder value (which is only subjectively assessable) by developing a
more objective measure that is less prone to managerial manipulation. In so doing,
they demonstrate formidable technical expertise in addressing one of the fundamen-
tal issues that has long concerned management accountants.
The first question in the framework for analysis is thus answered by the assertion
that the single objective of such a stock market quoted organization is to deliver as
much value as possible to the shareholders. The surrogate for measuring this value is
defined in historic accounting terms in a way that minimizes the dysfunctional
consequences inherent in the use of any such measure. However, it needs to be
recognized that the measure does not anticipate the earning of future income, despite
the existence of predictions based on stock market valuations, but it remains an
historic income measure. Stern Stewart would argue the EVATM;is the best surrogate
for or predictor of future share price performance, but this remains an open empirical
question.
The second question is largely unanswered. The means by which such outcomes
are to be attained are not specified. The most charitable interpretation of this lacuna
is that a very decentralized organization is envisaged which allows subordinate
mangers a great deal of discretion over their operating activities. In such circum-
stances, it could be argued that it is inappropriate to constrain subordinates by
insisting they operate to plan; rather, they have to deliver value by whatever means
they can invent.

17
Stakeholders other than shareholders are not explicitly considered in the EVA framework. At best, other
stakeholders are seen in an instrumental manner as parties with whom contracts are entered into, as the
means by which the objective of increasing shareholder value is attained.
18
See, for example, Otley and Emmanuel (1976).
 Performance Management 373

With regard to target setting, at first glance one would imagine that this is a simple
issue, because the implicit standard against which residual income or EVATM is
assessed is zero.” However, this only holds true where valuations are conducted on
an NPV basis. Because EVATM takes a more historic view and only uses accounting
rather than economic valuations, there can be an ‘inheritance effect’ whereby
managers can benefit from or be penalized by the past history of the organization.
For example, in the ‘rust bowl’ industries, the depreciated historic cost of assets may
exceed their market valuation; here, EVATMis expected to be negative and a target of
zero would be over-ambitious. Thus, the EVATMapproach pays particular attention
to the setting of appropriate targets. The objective appears to be the traditional one of
attempting to ensure that targets are ‘tip-toe’, ‘stretch’ or challenging, whilst still
being regarded as realistic by those who will be required to attain them.
Reward structures are another major focus of the EVATM approach. Not only is
this a by-product of the Stern Stewart history (it was originally a compensation
consultant), but reflects an awareness of the imperfect nature of the EVATMmeasure,
even after all the recommended adjustments have been made. The central suggestion
is that, although bonus calculations should be based on the attainment of target levels
of EVATM, such a bonus should not be paid immediately in cash, but should be
subject to smoothing over a 3-year period, and payable in full only if performance is
maintained into the future. The reason for using this method is explicitly motivatio-
nal and designed to avoid potentially dysfunctional short-term behaviour. Stern
Stewart thus recognize the potential dysfunctional effects of short-term performance
targets coupled too closely to financial rewards, and have developed a scheme to
reduce the worst such effects.
Finally, Stern Stewart also briefly discusses the process by which future targets
should be adjusted in the light of actual results. Here again they are at pains to avoid
the circularity that can result from a strong management team being expected to
perform well and thus inflating stock market expectations. If this were to be allowed
to inflate EVATM targets, a good management team would be penalized purely
because they were expected to deliver shareholder value. A compromise is therefore
proposed which is again based on an historic approach, perhaps benchmarked against
what is being achieved elsewhere.
The EVATM literature is therefore a good example of the performance manage-
ment framework being used in practice, albeit to just a single over-arching measure of
financial performance. Each of the five major questions is addressed, although the
second question is touched upon only implicitly. By concentrating on a single
over-arching financial performance measure, the EVATM approach is able to disre-
gard questions of strategy, although only at the cost of spending a great deal of effort
on discussing the capitalization of items, such as R & D expenditure. That is,
strategies which require current spending to produce expected future benefits must be
‘properly’ accounted for. In summary, the approach has been well worked through,
and represents one of the most coherent performance management systems currently
on offer. Nevertheless, even under its own assumptions concerning organizational
objectives, it is clearly not as comprehensive as it claims, and is particularly weak in
measuring and monitoring the means by which managers have adopted to achieve
their overall objectives.
19
There are some other sufficient conditions for such a relationship to hold, discussed in O’Hanlon and
Peasnell (1998).
374 D. Otley

Financial I

f'
processes must
we excel at?"

L
Learning and Growth
"To achieve our
wesustainour jj d Q' &'
abilityto 0 -? 4
' %a
/ihangeand
improve?"
I I I I I I I
Figure 1. The balanced scorecard: a framework to translate a strategy into operational terms. Source:
Robert S. Kaplan and David P. Norton, "Using the balanced scorecard as a strategic management
system," Harvard Business Review Oanuary-February 1996): 76. Reprinted with permission.

7. The Balanced Scorecard Approach

The Balanced Scorecard Approach has been developed at the Harvard Business
School by Kaplan and Norton since the early 1990s. It is an essentially multi-dimen-
sional approach to performance measurement and management that is linked speci-
fically to organizational strategy." It suggests that as well as financial measures of
performance) attention should be paid to the requirements of customers, business
processes and longer-term sustainability. Thus four areas of performance are defined
(now labelled as financial) customer) internal business and innovation and learning))
and it is suggested that up to four measures of performance should be developed in
each area. These (potentially 16) performance measures are not necessarily compre-
hensive) but should represent the critical success factors necessary for continued
organizational success or, minimally) survival. Thus, there is intended to be a close
link between the business unit strategy adopted and the performance measures
selected.
In the following discussion, the performance management framework will be
applied to analyse the balanced scorecard approach and to suggest some extensions
and improvements that might be made to the approach. These are quite tentative)
and meant primarily to illustrate the power of the framework both to make practical
recommendations) and to provide a structure for empirical research and analysis.
A major strength of the balanced scorecard approach is the emphasis it places on

20
Compare Kaplan and Norton (1992) and Kaplan and Norton (1996) to see how the approach has been
developed during that period.
 Performance Management 375

linking performance measures with business unit strategy. This appears to be a very
weak area in many organizations and the technique provides a practical approach to
addressing this issue. At first sight, it would appear to be a stakeholder approach.
Two of the major areas defined represent major stakeholders, namely the providers of
finance and the customers. Employees also figure, although they seem to have
migrated from the business process box to the long-term ‘Innovation and Learning’
box during the development of the technique. Other stakeholders, such as suppliers,
governments, local communities and the environment receive only passing mention.
However, it must be pointed out that in the introduction to their 1996 book, Kaplan
and Norton explicitly state that the balanced scorecard is not a stakeholder approach;
the shareholders are still the dominant group. This appears to be a reflection of the
cultural environment of the U.S.A. and possibly also a consequence of the rise in
popularity of EVATM.From the perspective of this paper, the Balanced Scorecard
approach is clearly a stakeholder approach, and this represents one of its major
advantages. It also suggests other boxes that might usefully be incorporated into the
development of balanced scorecards for specific organizations in both the public and
private sector.
There is also rather little detail given of how to select specific performance
measures to be placed in the balanced scorecard boxes. Clearly some of these must
represent key result areas; but others need to incorporate the strategic plans of the
organization in reflecting the choices that have been made as to how those results are
to be achieved. For example, customer service levels and satisfaction may be a key
result area, but strategic choices have to have been made concerning the means of
delivering such service (e.g. speed of delivery, product quality, technical advice etc.).
There is much work to be done in determining how to map the necessary pattern of
means-end relationships onto the balanced scorecard boxes. Clearly, for example, a
customer objective may well be attained by a business process means. Much of this
knowledge is probably tacit in the practices of the management consultants who
implement the Balanced Scorecard and could usefully be more formally explicated.
It is sometimes suggested that the upper left-hand boxes (financial and customer)
represent results measures, whereas the bottom right-hand boxes (business process
and innovation and learning) represent the means by which the desired results will be
obtained. However, this is clearly true only in the most simple-minded terms. Little
or no guidance is given in the Balanced Scorecard literature on how means and ends
should be linked analytically. A further area of ambiguity is the way in which the
‘balanced’ scorecard appears to have become more ‘linear’ in its approach. In the
1996 book, a linear chain is suggested whereby better trained employees (now in the
Innovation and Learning box) will lead to better business processes being designed
(one input to such changes, but surely by no means the only one); these in turn will
lead to more satisfied customers and then to happier shareholders. Although a
plausible chain of events, it is again very much a simplification of reality. It can be
argued that the original exposition of the approach in the 1992 article better preserves
the ‘balanced’ nature of the scorecard; all the performance objectives need to be
achieved in order for the organization to have been successful. However, the mapping
of means-end relationships for a given organization is of crucial importance for the
development of a meaningful Balanced Scorecard, and is worthy of much greater
attention.
The topic of setting performance targets is not much discussed in the balanced
376 D. Otley

scorecard literature. This is unfortunate, as it provides one response to the often-

voiced criticism of the approach, namely that it does not specify how trade-offs are to
be made between the difference measures used. Indeed, others would view this as the
strength of the approach; all the areas identified should only be on the scorecard if
they are crucial to the success of the organization. One means of resolving this
conflict lies in the setting of targets. Here, the level of difficulty of attaining the
required level of performance in different areas essentially defines the relative levels of
attention that managers need to pay to them. Thus, target setting is a crucial feature
of well-implemented balanced scorecard, and worthy of much more detailed atten-
tion.
Reward structures are also little mentioned, although they have the potential to
destroy the impact of an otherwise well-designed scorecard. The most obvious
example of this would be where a scorecard has been implemented, but bonuses are
still given based on the achievement of budget targets. By contrast, one organization
has reinforced its commitment to the balanced scorecard approach by linking
bonuses specifically to scorecard measures. In particular, if threshold levels of
performance were not achieved on every scorecard measure, then no bonus would be
paid despite high achievement elsewhere. Whilst not necessarily recommending this
approach, it does seem to be congruent with the balanced scorecard philosophy. If
each measure represents a critically important area of performance, not achieving any
one represents an inadequate standard of overall performance.
Finally, the role of feedback from the balanced scorecard has similarly had little
attention paid to it. Most obviously, feedback provides information about the extent
to which a company is achieving its key strategic aims. However, it is perhaps the
‘double loop’ learning that is even more important. Is the strategy working as
expected? If not, is this because of inadequate implementation or because the strategy
itself is faulty? Furthermore, only part of a company’s activities are driven by the
deployment of espoused strategy; is the organization picking up the emergent
strategies that are being developed at grass roots level and incorporating these into
the scorecard? What becomes evident from this analysis is that the balanced score-
card is clearly a dynamic tool, the contents of which will change over time as
strategies develop and key success factors change. This is recognized in the scorecard
literature, but there is little guidance as to how it should be managed. Lastly, this
approach also makes it clear that the scorecard does not stand in isolation; rather, it is
underpinned by the traditional measurement systems present in all organizations.
Just because an item does not appear on the scorecard does not imply that it is no
longer measured and reviewed. Perhaps the scorecard can be seen as an embodiment
of Simon’s (1995) interactive control systems; that is, it reports those measures which
senior managers have decided should be emphasized for a period of time.
The balanced scorecard is designed to be at the centre of an organization’s control
mechanisms to effectively deploy strategy and to link operational practices with
strategic intent. However, it cannot stand alone and its links with more traditional
control systems need to be reviewed. For example, in Kaplan and Norton (1996, p.
197) there is a diagram with the balanced scorecard placed at the centre, illustrating
the benefits that can flow from its adoption. By contrast, a very similar diagram (p.
192) with the budget at the centre is used to illustrate the barriers to action that
follow from the use of budgetary control systems. However, it seems unlikely that an
organization can survive using just the balanced scorecard without the normal
 Performance Management 377

budgetary apparatus. What is needed is an examination of how the two techniques

can be linked in a productive and complementary manner. Furthermore, both the
balanced scorecard and budgeting tend to be deployed through the organizational
hierarchy; some attention also needs to be paid as to how to best link measures of
process effectiveness along the value chain to this type of hierarchical approach.
The Balanced Scorecard is thus a potentially powerful tool by which senior
mangers can be encouraged to address the fundamental issue of effectively deploying
an organization’s strategic intent. It focuses on establishing links between strategic
objectives and performance measures; it also pays some attention to measuring the
achievement of the components of the strategic plan the organization has espoused
(i.e. the means that it is believed will lead to the desired ends). The balanced
scorecard literature also indicates that it as much the process of establishing a
scorecard that yields benefit as the resultant measurement schema. However, the
literature is remarkably silent on this point. Procedures for mapping means-end
relationships are not explicated. In addition and surprisingly, target-setting is not
mentioned despite its central role; the links with reward structures are neglected; and
the establishment of information systems and feedback loops are taken for granted.
All of these neglected areas provide opportunities for further research.

8. Discussion

The above analysis is intended to give an indication of how the proposed framework
can be used to analyse control techniques that are being used in many organizations.
An overall summary is given in Table 1 below.
It can be seen that no single technique has developed answers to all five of the
questions posed, although the application of such techniques requires attention to be
paid to them all. It may therefore be beneficial for both practitioners and researchers
to consider all of the aspects identified in evaluating and refining the above tech-
niques. However, there is a significant shortcoming in analysing single techniques in
this way. It may be that, although a particular technique does not address all of the
issues identified, such matters are dealt with by a combination of a variety of
techniques in an overall organizational control system. A more holistic approach is
clearly appropriate, with the unit of analysis being the organization. Here an attempt
should be made to analyse the totality of the control systems being utilized against
the issues defined above.
It is important to be clear at this point that the purpose of the framework outlined
is not intended to serve as a prescription for ‘best practice’ on the assumption that
explicit procedures for developing a performance measurement and management
system which emphasize the establishment of comprehensive and quantitative mea-
sures of performance, explicitly linked to financial rewards, is the universally best way
forward. At the very least, there is insufficient evidence available to substantiate such
a position, and the distinct possibility that in many circumstances such an approach
may be counter-productive. However, there is a dearth of information available
concerning what is current practice in major business, and what is the impact of
different configurations of controls. Rather, the intention has been to develop a
framework which can provide a structure for examining extant practice in a more
holistic way than has previously been the case.
 w
4
00
Table 1
Comparison of the three control techniques analysed using the pe fomance management framework

Question Budgetary control EVA^^ Balanced scorecard P

0
1. Objectives Financial objectives: Single financial Multiple objectives
-profit; objective. based on strategy. e
Y
-cash flow; and
-ROCE

2. Strategies and plans Means/end Delegated to Implicit in selecting

relationships not responsible managers. some performance
formally considered, May be considered measures; no formal
although budget is when setting targets. procedures suggested.
based on a plan of
action.

3. Targets Best estimates for Some guidance is Not considered,

financial planning; given with respect to despite being central
literature on target- ‘inheritance effect’. to ‘balance’.
setting gives some
guidelines for control

4. Rewards Not addressed, despite Appropriate incentive Not addressed.

many rewards now schemes a central part
being made contingent of the methodology.
upon budget
achievement.

5. Feedback Short-term feedback Some discussion of Reporting of

of budget variances. longer-term impact. performance assumed,
Incremental budgeting but no explicit
from year to year. guidance given.
 Performance Management 379

Although these three performance management techniques have been reviewed

from a practical point of view, there are evidently some important research questions
that arise out of this analysis. Given the paucity of studies of control systems
operation, we could start with description and work our way towards explanation.
Why have organizations changed their control systems to incorporate some of these
techniques? What are the antecedents that caused them to make changes, and what
have the consequences of the changes been? Only by asking such questions can we
begin to develop a soundly based understanding of the merits and disadvantages of
the approaches currently on offer from the management consultants
This paper does not necessarily espouse the highly rationalistic approach to
developing performance measures and attaching monetary rewards directly to the
attainment of performance targets by individual managers that may be attributed to
some proponents of approaches, such as EVATM.It is an open research question as to
how effective such techniques are and whether their long-term ‘side effects’ outweigh
the possible short-term benefits achieved. Other, ‘softer’ and more ‘visionary’ ap-
proaches, involving the encouragement of shared values may be seen as being more
appropriate to some circumstances, and less destructive of organizational culture and
ethos. However, even such gentler approaches can be analysed using the preceding
framework, and the assessment of the benefits and weaknesses of alternative ap-
proaches is essentially an empirical question that can be resolved only by the detailed
observation and analysis of practice.
Indeed, to talk of the ‘management control system’ may itself be misleading. Some
20 years ago, an article on the contingency theory of management accounting (Otley,
1980)) initially labelled one of the boxes in a diagram the ‘organizational control
system’, but this was later changed it to the ‘organizational control package’ because
the term ‘system’ seemed to imply too rational a perspective. Organizational control
systems are more like packages. Different elements are added by different people at
different times. Studying such systems is perhaps more akin to archaeology (see
Hopwood (1987)) than anything else, although we have the advantage of being able
to talk to the current operatives of the ‘systems’. It is therefore misleading to assume
that the study of performance measurement and management methods will result in
a totally coherent outline of a rational set of control mechanisms well-suited to the
purposes for which they have been designed. Although individual component parts of
such systems may approach this degree of rationality, it is unlikely that the total
package of control measures that are in place at any point in time will possess such a
degree of coherence. This situation has a number of implications for research
methods.
Without wishing to rule out the use of a wide range of approaches in studying an
essentially very complex topic, some approaches are likely to more productive than
others, and should be emphasized. These would firstly involve a longitudinal element,
and study the operation of a set of control systems over a period of time. This would
enable the inter-relationship between different control systems elements to be ex-
amined and explained. It would also permit exploration of both the antecedents and
consequences of control systems choices. The second element would involve the
study of a single organization in some depth, to appreciate the context in which it
operates and the reciprocal impact of context and organization. The third element
would allow for a survey component within a case study of a single organization, so
that the inductive generalizations made can be tested, at least within the organization
380 D. Otley

studied. That is to say, a case study of a single organization can also usefully include
a survey of (say) 100 managers. Thus, one major methodological approach can be
adopted to develop explanations of control practices within single organizations and
result in inductive generalizations. This proposed methodology is not new, although
few examples of its application occur outside of Scandinavia and the U.K." How-
ever, the methods have been applied to research on management control systems and
performance management issues in only a few instances." Given the significant
changes that have taken place in management and control systems practice over the
past decade or so, there is clearly scope for a great deal of important research to be
undertaken.
The framework outlined above can therefore be seen as a template against which
extant practice can be both described and assessed. A complete control system
involves each of the five elements identified both separately and in combination.
Weaknesses in one area can be, at least partly, compensated for by strengths in other
areas. For example, EVATM uses the 'bonus bank' as a means of avoiding the worst
effects of an imperfect performance measure. It is therefore misleading to look at
only some of the areas identified as weaknesses there may well be balanced by
strengths elsewhere. T h e framework can provide a checklist to help ensure that a
more complete picture of control systems operation is observed. Practice can be
assessed in terms of the behavioural consequences that are observed to occur when a
particular system is operated in a specific context. Furthermore, as evidence is
amassed concerning the effects of different control system configurations, it may
become possible to assess the appropriateness of a particular system to the circum-
stances in which it is implemented.

9. Conclusions

By means of the three application areas reviewed, it has been shown how the wider
perspective of performance management and strategy implementation can be used to
analyse the working of practical control systems. In each case, there are suggestions
for improving business practices and issues raised for academic research. For exam-
ple, what is the role of budgeting in the modern organization? How is it now being
used in practice, and what changes are being made to traditional practices? New
financial performance measures, such as EVATMare being adopted by many organi-
zations. How do they link with currently used measures, and how are they integrated
into an overall control system? In what circumstances do they seem to be appropriate
and where may they need to be amended? What are the contextual factors that affect
an organization's likely interest in such matters? The balanced scorecard is also
proving to be a very popular tool, but how are organizations actually using it in
practice? Does it deliver the benefits claimed for it, and how might it be most
effectively be combined with existing control systems?
Performance management therefore provides an important integrating framework,
both academically and practically. It goes well beyond the traditional boundaries of
management accounting, and will require the skills of management accountants (and
management accounting researchers) to be developed in at least three areas. First, the
21
The pages of Management Accounting Research represent one of the prime sources of this type of work.
22
See Otley and Berry (1 994) for one example.
 Performance Management 38 1

management accountant needs to understand the operational activities of the organi-

zation. This was a traditional skill of the old ‘cost and works’ accountant, but one
that has been neglected more recently. Attempting to design control systems without
having a detailed knowledge of how the business works is likely to prove a recipe for
disaster. Second, there is a need to connect control systems design with issues of
strategy, both espoused and emergent. Control systems need to reflect the aims of an
organization and the plans that have been developed to achieve those aims. T h e
‘strategic management accounting’ movement has recognized this challenge, but has
been more concerned to develop new techniques than to design overall control
systems. Third, there is a need to focus on the external context within which the
organization is set, rather than just being concerned with internal activities. Competi-
tor analysis is clearly important, but even more central is the value that an organiza-
tion is delivering to its customers. A process orientation that focuses on value chains
is required to complement the vertical and hierarchical approach to control that has
long dominated the literature.
Furthermore, the developments outlined so far cannot be treated as purely techni-
cal matters that can be analysed from an economic perspective alone. The intention
in using performance measures is to influence managerial behaviour, so that man-
agers have the knowledge and motivation to act in the organization’s best interests.
This is an area where there are likely to be very different approaches that are
dependent upon national and organizational culture. Interestingly, this is exactly the
field to which that early pioneer of budgetary behaviour, Geert Hofstede, devoted
much of his later career.
The conclusion is straightforward. Although individual techniques of management
accounting and control have been studied individually within a restricted context,
they need also to be studied as part of a wider organizational control system. The use
of management accounting and control systems can be fruitfully analysed from the
framework of performance measurement and performance management. This makes
it clear that management accounting and other performance measurement practices
need to be evaluated not just from an economic perspective, but from a social,
behavioural and managerial perspective, within an overall organizational context. It is
these social, cross-national and cultural aspects that make the study of control
systems such a fascinating topic for academic research and such a challenge to the
practitioner. This paper has attempted to provide an outline framework that will help
both academics and practitioners to more fully understand the context in which they
are working, and to help develop control practices that are well-suited to the contexts
in which they are applied.

References

Anthony, R. N., 1965. Management Planning and Control Systems: A Framework f o r Analysis,
Harvard Business School Press.
Austin, R. D. and Gittell, J. H., 1999. Anomolies of High Performance: Reframing Economic
and Organizational Theories of Performance Measurement. Working Paper, Harvard
Business School.
Briers, M. and Hirst, M., 1990. The Role of Budgetary Information in Performance Evalua-
tion, Accounting Organizations and Society, 373-398.
Bunce, P., Fraser, R. and Woodcock, L., 1995. Advanced Budgeting: a journey to advanced
management systems, Management Accounting Research, 253-265.
382 D. Otley

Chapman, C. S., 1997. Reflections on a Contingent View of Accounting, Accounting Organisa-

tions and Society, 189-205.
Coates, J. et al., 1992. Multinational Companies’ Performance Measurement Systems: Inter-
national Perspectives, Management Accounting Research, 133- 150.
Checkland, P. and Holwell, S., 1998. Infomation, Systems, and Infomation Systems, Wiley.
Fitzgerald, L. and Moon, P., 1996. Perfomance Measurement in Seruice Industries: Making it
Work, CIMA.
Goold, M. and Campbell, A., 1987. Strategies and Styles, Blackwell.
Hofstede, G., 1967. The Game of Budget Control, Tavistock.
Hopwood, A. G., 1987. The Archaeology of Accounting Systems, Accounting Organisations
and Society, 207-234.
Johnson, H. and Kaplan, R., 1987. Relevance Lost: The Rise and Fall of Management
Accounting, Harvard Business School Press.
Kaplan, R. and Norton, D., 1992. The Balanced Scorecard-Measures that Drive Perfor-
mance, Haruard Business Review, Jan/Feb, 7 1-79.
Kaplan, R. and Norton, D., 1996. The Balanced Scorecard, Harvard Business School Press.
Langfield-Smith, K., 1997. Management Control Systems and Strategy: A Critical Review,
Accounting Organisations and Society, 207-232.
Merchant, K. and Manzoni, J.-F., 1989. The Achievability of Profit Center Budget Targets: A
Field Study, The Accounting Review, LXIV, 539-558.
Mintzberg, H., 1994. The Rise and Fall of Strategic Planning, Prentice-Hall.
Otley, D. T., 1980. The Contingency Theory of Management Accounting: Achievement and
Prognosis, Accounting, Organizations and Society, 194-208.
Otley, D. T., 1987. Accounting Control and Organizational Behavionr, CIMA.
Otley, D. T., 1994. Management Control in Contemporary Organizations: Towards a Wider
Perspective, Management Accounting Research, 5, 289-299.
Otley, D. T. and Berry, A. J., 1980. Control, Organisation and Accounting, Accounting
Organizations and Society, 23 1-244.
Otley, D. T. and Emmanuel, C. R., 1976. The Usefulness of Residual Income, Journal oj
Business Finance and Accounting, Winter, 43-5 1.
Otley, D. T and Berry, A. J., 1994. Case Study Research in Management Accounting and
Control, Management Accounting Research, 5, 45-65.
O’Hanlon, J. and Peasnell, K., 1998. Wall Street’s Contribution to Management Accounting:
The Stern Stewart EVA@Financial Management System, Management Accounting Research,
9, 421-444.
Simons, R., 1995. Levers of Control: How Managers Use Innovative Control Systems to Drive
Strategic Renewal, Harvard Business School Press.