Description
This paper reports on a survey of manufacturing companies, and uses structural equation modeling to examine the
relationships between the changing competitive environment, and a range of organizational variables as antecedents to
management accounting change. The results indicate that an increasingly competitive environment has resulted in an
increased focus on differentiation strategies. This, in turn, has influenced changes in organizational design, advanced
manufacturing technology and advanced management accounting practices. These three changes have led to a greater
reliance on non-financial accounting information which has led to improved organizational performance.
Antecedents to management accounting change:
a structural equation approach
Annette Baines*, Kim Lang?eld-Smith
Department of Accounting and Finance, Monash University, PO Box 197, Caul?eld East, Victoria 3145, Australia
Abstract
This paper reports on a survey of manufacturing companies, and uses structural equation modeling to examine the
relationships between the changing competitive environment, and a range of organizational variables as antecedents to
management accounting change. The results indicate that an increasingly competitive environment has resulted in an
increased focus on di?erentiation strategies. This, in turn, has in?uenced changes in organizational design, advanced
manufacturing technology and advanced management accounting practices. These three changes have led to a greater
reliance on non-?nancial accounting information which has led to improved organizational performance.
#2003 Elsevier Ltd. All rights reserved.
Introduction
Throughout the 1990s, the growing level of glo-
bal competition intensi?ed the challenges for
managers who need to consider more e?ective
ways of achieving competitive advantage and
improving organizational performance. One
means of achieving this is through the adoption of
clearly articulated strategies, ?exible organiza-
tional structures and innovative accounting sys-
tems. A common theme in normative management
accounting research is that changes in an organi-
zation’s external environment will lead to change
in an organization’s management accounting sys-
tem (Atkinson et al., 1997; Nanni, Dixon, & Voll-
man, 1992). This is based on the argument that
managers need speci?c forms of management
accounting information that support their decision
needs within increasingly uncertain environments,
and to assist them to monitor progress against
strategies. The need for an appropriate ?t between
the environment and organizational systems is an
underlying assumption of much of the empirical
contingency-style management accounting
research (see for example, Chenhall & Lang?eld-
Smith, 1998a,1998b; Gul, 1991; Perera, Harrison,
& Poole, 1997), as is the need for management
accounting systems to change to support man-
agers’ new information requirements. However,
prior empirical research in management account-
ing has examined primarily the various relation-
ships between the environment, organizational
variables, and management accounting systems at
a point in time. There has been limited empirical
research examining the nature of the changes in
management accounting systems and organiza-
tional variables made in response to environ-
mental changes, and whether or not these changes
improve organizational performance.
0361-3682/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0361-3682(02)00102-2
Accounting, Organizations and Society 28 (2003) 675–698
www.elsevier.com/locate/aos
* Corresponding author. Tel.: +61-3-9903-2792; fax: +61-
3-9903-2422.
E-mail address: [email protected]
(A. Baines).
This study uses a survey method and utilizes
structural equation modeling to identify whether
changes that have occurred within the external
environment of manufacturing organizations have
led to change in the type of management account-
ing information provided and used for decision-
making. More speci?cally, this study investigates
whether changes in the organizational environ-
ment have led to changes in the organizations’
strategy, organizational design, advanced manu-
facturing technology and management accounting
practices. These changes in turn are hypothesized
to in?uence the type of management accounting
information used by managers, which may lead to
improved organizational performance. The
remainder of this paper is organized as follows.
Section 2 summarizes the relevant prior literature,
and presents research hypotheses. The research
method and development of the survey is
explained in Section 3. Section 4 contains the
analysis of the data, and this is followed by a dis-
cussion of the results and conclusion.
Literature review and hypothesis development
E?ects of changes in the competitive environment
In this section, the literature that examines the
relationships between changes in the competitive
environment, and changes in strategy, organiza-
tion design and technology is reviewed, and
hypotheses presented.
Changes in the environment and strategy
Since the 1980s, there has been signi?cant
change in the external environment faced by ?rms
in all sectors of the economy. This includes more
active competitors, increasingly demanding cus-
tomers, and the availability of information pro-
cessing technologies (Hiromoto, 1991; Innes &
Mitchell, 1990; Shields, 1997). Such changes have
occurred as a result of market and ?nancial
deregulation, and increasingly companies com-
pete in a single global market. Increasing globali-
zation has resulted in intense and aggressive
international competition, increased customer
demands including diversi?ed customer needs,
and shorter product life cycles (Dent, 1996;
Shields, 1997).
The strategy an organization adopts constitutes
the logic underlying its interactions with its envir-
onment. It is well established that an organiza-
tion’s strategy is a response to its environment,
and that the appropriate matching of strategy and
the environment can enhance performance (Burns
& Stalker, 1961; Porter, 1980). Several empirical
studies have examined the linkage between envir-
onment and strategy. For example, Fuschs, Mif-
?in, Miller, and Whitney (2000) found that
successful ?rms aligned key elements of strategy
with the environment. Chong and Chong (1997)
found a positive association between perceived
environmental uncertainty and strategy, and
Miller (1988) found a relationship between an
unpredictable and dynamic environment and an
innovation strategy.
As the environment becomes dominated by
increasingly more demanding customers and as
competitors respond to customer demands in
increasingly sophisticated ways, a ?rm may place
greater emphasis on developing a di?erentiation
strategy that emphasizes more customer-oriented
aspects such as quality, ?exibility, innovative pro-
ducts and dependability of supply (Perera et al.,
1997; Sim & Killough, 1998). Therefore, the fol-
lowing hypothesis is proposed:
H
1a
. Firms facing a more competitive environment
will change towards a di?erentiation strategy.
Changes in the environment and organization design
Under conditions of environmental change
where markets have become more competitive,
particularly with an increased level of high quality
and competitively priced products, ?rms may
respond by reorganizing their work processes, and
adopting structures that have a stronger customer
orientation (Keidel, 1994; Miller, 1988; Partha-
sarthy & Sethi, 1993). In particular, a variety of
team-based structures have emerged, including
self-managed work teams, and cross-functional
project teams (Cohen & Bailey, 1997). Self-mana-
ging work teams are increasingly relied on in
manufacturing organisations (Anderson, Hesford,
& Young, 2002). These teams are continuing work
676 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
units responsible for producing products, often
from the initial purchase of materials, through all
stages of the production process, to the ?nal pro-
duct. Employees are empowered to make a variety
of decisions that were once the domain of super-
visors and managers, and team members are
multi-skilled. These decisions can cover the areas
of production quality, supplier and customer liai-
son, personnel management, production planning
and con?ict resolution. Cross-functional teams are
one-time team structures that are formed for a
speci?c purpose, such as developing a new pro-
duct, or improving work processes. Team mem-
bers are drawn from di?erent functional areas of
the business which allows a variety of di?erent
skills, knowledge, expertise and perspectives to
be brought to the project. These teams have the
potential to generate new ideas or derive novel
solutions, which may be essential to competing
in dynamic competitive environments (Anderson
et al., 2002). Thus, higher quality and more
timely and creative outcomes can be expected
from such teams, as their capacity to undertake
multiple activities simultaneously provides an
e?ective response to time-based competition
(Brown & Eisenhardt, 1995; Cohen & Bailey,
1997).
The adoption of teams is associated with ?atter
hierarchies and the increased empowerment of
lower-level managers and employees (Chenhall &
Lang?eld-Smith, 1998b; Elliot, 1991; Otley, 1994;
Rimmer, Macneil, Chenhall, Lang?eld-Smith, &
Watts, 1996; Shields, 1997). To ensure fast and
innovative responses in complex and dynamic
environments, there has been a move away from
hierarchical controls and centralized decision
making, towards the allocating of more responsi-
bility to lower levels of the ?rm. It has been argued
that the use of team-based structures in complex
environments enables organizations not only to
improve their speed and ?exibility of response, but
also to improve the quality of that response
(Cohen, 1993; Lawler, 1993). Therefore, the fol-
lowing hypothesis is proposed:
H
1b
. A more competitive environment will result
in changes in organizational design, with greater
use of team-based structures.
Changes in the environment and technology
An increasingly competitive environment can
lead to ?rms adopting innovative products and
production techniques to provide increased ?ex-
ibility, and to satisfy customer demands (Foster &
Gupta, 1994; Otley, 1994). In order to compete in
these highly competitive markets, many organiza-
tions have made considerable investments in
advanced manufacturing technology, such as
computer-aided design, computer-integrated man-
ufacturing and just-in-time systems. The use of
these technologies has resulted not only in
increased quality, but also in the ability to provide
di?erentiated products or services to satisfy spe-
ci?c market segments or even individual customers
(Elliot, 1991; Gosse, 1993; Otley, 1994).
The increased ?exibility provided by advanced
manufacturing technology has made it more
attractive to increase the breadth of the product
line, as making more frequent changeovers and
producing in smaller batches allows customer pre-
ferences to be satis?ed, without having to incur
high inventory storage costs (Milgrom & Roberts,
1995). Advanced manufacturing technologies
enable companies to compete on the basis of
quality, productivity and ?exibility, as well as on
cost, and have transformed the manner in which
production takes place even in relatively small
companies (Bhimani, 1994; Bruggeman & Slag-
mulder, 1995; Parthasarthy & Sethi, 1992; Pfe?er,
1994). Thus, the following hypothesis is proposed:
H
1c
. A more competitive environment will result
in increased use of advanced manufacturing
technology.
E?ects of changes in strategy
In this section, the relationships between chan-
ges in strategy and changes in organization design,
technology and advanced management accounting
practices are presented. Three hypotheses are
developed.
Changes in strategy and organization design
The increased emphasis on a di?erentiation
strategy may require operational employees to
adopt a stronger customer orientation, and one
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 677
way of encouraging employees to take ownership
of such a strategy is through the introduction of
team-based structures. In production teams,
workers are encouraged to pool their knowledge
of the production process and initiate innovative
approaches to improve productivity and quality,
and to reduce production lead time (Banker, Pot-
ter, & Schroeder, 1993). Teams provide the basis
for a stronger customer focus and allow the struc-
turing of work around processes rather than
functions, which has direct implications for enga-
ging in process improvement activities which may
improve an organization’s speed, ?exibility and
quality of response (Scott & Tiessen, 1999; Rim-
mer et al., 1996). Therefore, the following
hypothesis is proposed:
H
2a
. A change towards a di?erentiation strategy
will result in changes in organization design, with
greater use of team-based structures.
Changes in strategy and technology
A proper link between strategy and manu-
facturing operations is a key to developing sus-
tainable competitive advantage (Porter, 1996).
One way in which manufacturing organizations
can respond to increasing customer demands of
quality, ?exibility and dependability of supply is
through the implementation of advanced manu-
facturing technology. Schroeder and Congden
(2000), in a study of small to medium-sized man-
ufacturers, found the most ?nancially successful
?rms were those which demonstrated a tight
alignment between strategy and technology, while
Kotha and Swamidass (2000) found that for ?rms
competing on the basis of quality, customer ser-
vice, delivery reliability, product features and ?ex-
ibility, investment in advanced manufacturing
technology resulted in superior growth. A positive
association between an emphasis on di?erentia-
tion strategies and investments in advanced man-
ufacturing technologies was also found in a study
of UK ?rms (Burcher & Lee, 2000). Thus, the fol-
lowing hypothesis is proposed:
H
2b
. A change towards a di?erentiation strategy
will result in increased use of advanced manu-
facturing technology.
Changes in strategy and advanced management
accounting practices
In pursuing competitive advantage, organiza-
tions may implement manufacturing processes and
administrative functions that support their parti-
cular strategic priorities. For example, both
Chenhall and Lang?eld-Smith (1998b) and Call-
ahan and Gabriel (1998) found greater use of
advanced management accounting practices, such
as quality improvement programs, benchmarking
and activity-based management, in ?rms that
placed a strong emphasis on product di?erentia-
tion strategies, and this ultimately resulted in high
performance.
Advanced management accounting practices
can assist employees to more easily focus on
achieving di?erentiation priorities, such as quality,
delivery and customer service, compared to more
traditional ?nancially based accounting practices,
as they highlight the need to satisfy customer
requirements. For example, target costing allows
managers to focus on low cost while simulta-
neously maintaining customer expectations in
areas of quality and functionality. Also, activity-
based management focuses on maintaining or
enhancing customer value, not just controlling
costs. Thus, the following hypothesis is proposed:
H
2c
. A change towards a di?erentiation strategy
will result in the increased use of advanced man-
agement accounting practices.
E?ects of changes in technology
This section reviews the literature examining the
relationships between changes in the use of
advanced manufacturing technology, and changes
in organization design and advanced management
accounting practices.
Changes in technology and organization design
Adopting new technologies may require changes
in organizational structures and work practices to
better suit the capabilities of that technology
(Rimmer et al., 1996). Traditional manufacturing
plants tend to be laid out by machine or process
function. Line personnel, separated from their co-
workers by inventory, become specialized by
678 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
repeatedly processing large batches of similar
materials, and inventories are pushed through the
system with quality inspections conducted by
quality control personnel occurring at the end of
the production. In contrast, the successful imple-
mentation of new technology often requires the
blending of technological and social skills, which
can best be achieved through adoption of work-
based teams or production cells. These new struc-
tures emphasize products and customers rather
than mass production. A team may manage the
complete processing of products, with each
employee performing several functions. Batches
are small, there is little work-in-process inventory,
teams become responsible for production and
quality, and sometimes liaise directly with suppli-
ers and customers (Banker et al., 1993; Rimmer et
al., 1996). Therefore, the following hypothesis is
proposed:
H
3a
. The increased use of advanced manufacturing
technology will result in a change in organization
design, with greater use of team-based structures.
Changes in technology and advanced management
accounting practices
With the introduction of new technologies in
manufacturing operations, the structure of manu-
facturing costs has changed. Technologies such as
computer-integrated manufacturing and just-in-
time systems emphasize that variable direct labor
and inventory are vanishing from the factory.
Speed of operation is not determined by how fast
an operator can work, but by the type of automa-
tion and manufacturing system used (Dhavale,
1996). Consequently, traditional cost control sys-
tems which focus on variance analysis, aggregat-
ing costs and accounting for inventory, do not
e?ectively identify resources consumed, or help
managers manage those resources. In addition,
they may distort the realities of manufacturing
performance with new technological processes
(Bruggeman & Slagmulder, 1995; Gosse, 1993;
Kaplan, 1994).
Advanced management accounting practices,
such as activity-based management, life cycle cost-
ing and target costing, appear to be gaining an
increasing foothold in the modern world (Granlund
& Lukka, 1998). They provide an approach to
resource management that supports a customer
focus. For example, benchmarking provides an
important way of sensitizing the organization to
external performance standards, thus overcoming
the inward focus of cost variance analysis (Chen-
hall & Lang?eld-Smith, 1998a; Elnathan, Lin, &
Young, 1996). Quality improvement programs
have evolved as a philosophy that stresses the
notion of continuously improving manufacturing
processes by eliminating waste, improving quality,
developing people skills and at the same time
reducing costs (Sim & Killough, 1998). The fol-
lowing hypothesis is proposed:
H
3b
. The increased use of advanced manufacturing
technology will result in the increased use of
advanced management accounting practices.
In?uence of organizational changes on non-
?nancial management accounting information
In this section, the literature that examines the
relationships between changes in organizational
variables, namely advanced management accounting
practices, strategy, organization design and technol-
ogy, and changes in reliance on non-?nancial man-
agement accounting information will be reviewed.
Changes in advanced management accounting
practices and non-?nancial MAI
There is strong empirical support for the associ-
ation between advanced management accounting
practices and the increased use of non-?nancial
performance measures. Chenhall and Lang?eld-
Smith (1998b) found that ?rms which emphasized
product di?erentiation strategies bene?ted from
the use of advanced management accounting
practices and reliance on non-?nancial informa-
tion. Further, Abernethy and Lillis (1995), Banker
et al. (1993), Perera et al. (1997) and Sim and
Killough (1998) found a positive association
between the emphasis placed on various forms of
advanced management practices in an environ-
ment of manufacturing ?exibility, and the use of
non-?nancial measures such as defect rates, on-
time delivery and machine utilization. Therefore,
the following hypothesis is proposed:
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 679
H
4a
. The increased use of advanced management
accounting practices will result in greater reliance
on non-?nancial MAI.
Changes in strategy and non-?nancial MAI
An organization’s ability to survive and function
successfully in an environment of intense competi-
tion depends partially on the availability of infor-
mation upon which its managers can act (Bhimani,
1993). Therefore, accounting information should
be designed to facilitate informed judgments and
decisions that support the chosen strategy.
The development of appropriate management
accounting information requires that the strategy
be articulated so that goals and objectives are
de?ned with su?cient precision to be measurable.
For example, if quality or innovation is central to
a ?rm’s competitive strategy, then it is vital that
the ?rm’s accounting system is able to measure
events representing these factors. Measures that
focus on customer satisfaction, manufacturing
excellence, market leadership, quality, reliability,
responsiveness, and technological leadership,
innovation and human resources need to be con-
sidered (Eccles, 1991; Fisher, 1992). The failure to
include these measures in a formal management
accounting system may well hamper an organiza-
tion’s ability to achieve various di?erentiation
strategies of customer satisfaction, quality, speed,
learning and process improvement (Sim & Kill-
ough, 1998).
A change towards a di?erentiation strategy
requires a shift from cost and e?ciency-based
performance measures to ones which measure
speci?c non-?nancial aspects of strategy, such as
quality, service, ?exibility and dependability
(Cooper, Kaplan, Maisel, Morrissey, & Oehm,
1992; Samson, Lang?eld-Smith, & McBride,
1991). This leads to the hypothesis:
H
4b
. A change towards a di?erentiation strategy
will result in greater reliance on non-?nancial
MAI.
Changes in organization design and non-?nancial
MAI
A critical aspect of adopting team operations is
the process of empowerment. Teams cannot simply
be delegated responsibilities. Empowerment places
both authority and responsibility at low levels in
an organization; it implies that employees are
willing to take responsibility for their own actions,
and use their initiative to locate and solve pro-
blems to ensure that organizational goals are met,
particularly in relation to customer service (Ich-
niowski, Shaw, & Prennushi, 1997; Otley, 1994;
Rimmer et al., 1996). Changing the organization
design, including the use of teams and employee
empowerment, will result in changed employer
and employee expectations, including increased
access to relevant information (Scott & Tiessen,
1999).
The role of management accounting in this
changed organization structure is not simply to
deliver cost data, but to provide a service that
empowers team members to make the best deci-
sions in light of current conditions (Gordon &
Miller, 1976; Nanni et al., 1992). As more decision
making responsibility is passed to ‘lower’ levels of
the organization, there is an increased need for
easily accessible and relevant information at these
levels, as well as relevant information for top
management to evaluate the operations of the ?rm
and direct strategy (Banker et al., 1993). Non-
?nancial performance measures can form an inte-
gral part of the information base necessary for
team success (Scott & Tiessen, 1999). Therefore,
the following hypothesis is proposed:
H
4c
. A change in organization design, with greater
use of team-based structures, will result in greater
reliance on non-?nancial MAI.
Changes in technology and non-?nancial MAI
It has been argued that reliance on non-?nancial
information is dependent on the level of adoption
of advanced manufacturing technology (Perera et
al., 1997). If substantial changes take place in the
manufacturing processes, the management
accounting system must also change if it is to
provide relevant information for managerial deci-
sions and control (Foster & Horngren, 1988;
Kaplan, 1986; Nanni et al., 1992).
Financially oriented accounting measurement
systems imperfectly re?ect, and often with consider-
able lags, the increase in manufacturing ef?ciency
680 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
and e?ectiveness that occurs when ?rms pursue
total quality control, just-in-time inventory sys-
tems, and computer integrated manufacturing
processes. These systems, with their emphasis on
short-term pro?ts fail to capture the bene?ts that
arise from decreased new product launch times,
reduced throughput and lead times and the
improved ?exibility which modern manufacturing
technology has made possible (Kaplan, 1986).
Poorly designed or outdated accounting and con-
trol systems can distort the realities of manu-
facturing performance. Furthermore, such systems
can place out of reach most of the promised ben-
e?ts of new methods because the performance of
individuals, production processes, and ?rms in
high technology environments are not being
assessed and evaluated appropriately (Kaplan,
1984; Young & Selto, 1991). Such technological
changes require fundamental changes to perfor-
mance measures (Bhimani, 1993; Drury & Tayles,
1995). More timely, relevant and comprehensible
information is perceived as improving managerial
decisions in areas such as cost reduction, cost
control, product quality and performance assess-
ment. For example, to evaluate the e?ectiveness of
the ?exible manufacturing systems to deliver cus-
tomer value, measures such as cycle time, delivery
performance, and the ability of manufacturing to
vary product characteristics or develop new pro-
ducts would be more appropriate to track perfor-
mance, than traditional cost metrics (Abernethy &
Lillis, 1995; Bhimani, 1993). Thus, the following
hypothesis is proposed:
H
4d
. The increased use of advanced manufacturing
technology will result in greater reliance on non-
?nancial MAI.
Change in non-?nancial MAI and organizational
performance
It has been argued that reliance on ?nancial
performance measures alone will not necessarily
produce improved ?nancial results, as ?nancial
measures only indicate the outcomes of past
activities which may be no guide to improving
future performance. Non-?nancial measures,
however, can re?ect the drivers of future ?nancial
performance. It follows, therefore, that perfor-
mance measures that exclude important aspects of
manufacturing operations will not direct managers
to areas of critical concern (Nanni et al., 1992).
However, although predictive ability is claimed
to be one of the bene?ts of non-?nancial measures
(Atkinson et al., 1997), it appears that managers
are not always sure that their e?orts are rewarded
with improvements in the ?nancial bottom line
(Fisher, 1992). While the majority of prior
research supports a positive relationship between
the increased reliance on non-?nancial informa-
tion and organizational performance, the exact
nature of the relationship is ambiguous.
Mia and Clarke (1999) found an indirect asso-
ciation between the intensity of market competi-
tion and business unit performance through the
use of management accounting information.
Davila (2000) and Chong and Chong (1997)
established that a greater use of non-?nancial
information for business units following a custo-
mer-focused or prospector-type strategy, had a
positive impact on performance. Managers in
these cases tended to place greater reliance on
non-?nancial information on the assumption that
good performance in non-?nancial areas would
drive good ?nancial performance. On the other
hand, Perera et al. (1997) found no association
between the use of non-?nancial measures and
perceived performance in an organization with a
‘customer-focused’ manufacturing strategy.
Improved performance has resulted in ?rms that
use ?exible manufacturing, and which also place
greater reliance on non-?nancial manufacturing
measures (Abernethy & Lillis, 1995; Sim & Kill-
ough, 1998). Improved performance has also
occurred when teams have been provided with
comprehensive performance measures, that is, a
combination of ?nancial and non-?nancial mea-
sures (Scott & Tiessen, 1999). Ittner and Larcker
(1995) and Sim and Killough (1998) both found a
signi?cant positive interaction between TQM
practices, management accounting information
and performance. On the other hand, Banker, Pot-
ter and Srinivasan (2000) and Ittner and Larker
(1998) found that particular non-?nancial mea-
sures were positively associated with future, rather
than current, revenues and pro?t. It is possible,
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 681
however, that organizational strategy, the type of
technology, and the structural and environmental
factors confronting the organization could also
a?ect the use and performance consequences of
reliance on non-?nancial performance measures
(Daft & Lengel, 1986; Ittner & Larcker, 1998).
Also, improvements in non-?nancial measures will
not result in improved pro?ts if management has
selected the wrong critical success factors. If, for
instance, performance measures are based on qual-
ity, but quality is not of value to the customer, the
strategy will not lead to improved ?nancial perfor-
mance. A strategy of innovation will fail unless it
addresses an environment that values novelty and is
accompanied by an organizational structure that
can produce it. Similarly, a matching of only orga-
nizational design and environment with each other
will be inadequate if either con?icts with the selec-
ted strategy (Miller, 1982; Pelham, 1999).
Therefore, the use of non-?nancial performance
measures may be just one factor in a management
accounting information system that must be con-
sidered with all other variables within the organi-
zation. Successful companies must not only track
?nancial performance and operating e?ciency,
but must also pay close attention to customer
satisfaction and human resources, and there must
be an agreement on strategy and clarity of
communication (Lingle & Schiemann, 1996).
Information is now recognized as one of the most
powerful tools that can substantially in?uence the
wealth of corporations (Mangaliso, 1995). Thus,
in response to the changed environment, the chal-
lenge for any management accounting system is to
provide up-to-date information to help managers
reach informed economic decisions, and to moti-
vate users to aim and strive for organizational
change (Horngren, 1995; Stokes & Lawrimore,
1989). Failure to rely on appropriate accounting
information may contribute to ine?ective resource
management and a gradual decline in organiza-
tional performance. This leads to the ?nal
hypothesis.
H
5
. Achange in MAI, with greater reliance on non-
?nancial management accounting information, will
result in improved organizational performance.
The hypotheses are illustrated in Fig. 1.
Method
This section contains two parts. First, there is a
description of the sample used to test the hypotheses.
Second, the development of the survey instrument
and the measurement of the variables are described.
Sample selection
A self-administered questionnaire was sent to the
general managers of 700 manufacturing organiza-
tions randomly selected from the Kompass Australia
(1999) database. A strati?ed sampling approach was
used. This involved dividing the sample into ?ve
groups based on the number of employees (a proxy
for size), and selecting ?rms proportional to the
number of companies in each group. The ?rms
selected were business units of larger ?rms, or com-
panies in their own right. The initial posting of the
survey instrument also included, (a) a letter, addres-
sed personally to the general manager of each busi-
ness unit, explaining the purpose of the study; a tear-
o? section allowing respondents to provide their
name and address for a copy of the survey results,
while ensuring anonymity, (b) a reply-paid envelope
for the return of the survey, and (c) a separate reply-
paid envelope for the return of the tear-o? section.
Three weeks after the initial posting, a reminder let-
ter was sent to all managers who had not returned
the tear-o? section. Two weeks later, each manager
who could not be identi?ed as having responded was
phoned and encouraged to complete the ques-
tionnaire. Additional questionnaires were forwarded
to those who had misplaced the original material.
A total of 155 responses (22%) were received.
Of these, 14 were unusable, mainly due to missing
data on organizational performance. This resulted
in 141 useable responses, a response rate of 20%.
The remaining sample ?rms had average sales of
$128.2 million and average assets of $88.59 mil-
lion. The majority of ?rms in the sample were
stand-alone companies (59% of the sample) or
divisions of larger companies (37%). The respon-
dents had been with their ?rm an average of 11.3
years. Most described themselves either as man-
agers (61%) or managing directors (23%).
The useable response rate of 20% is clearly
lower than anticipated. It provides a sample size
682 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
F
i
g
.
1
.
H
y
p
o
t
h
e
s
i
z
e
d
m
o
d
e
l
.
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 683
which is just large enough to undertake the struc-
tural equation analysis (discussed in a later sec-
tion), but the potential for non-response bias does
arise. The mean responses for questionnaires
received prior to the reminder phone call were
compared to those received after the phone call, to
test if responses di?ered between the two groups.
No signi?cant di?erences were identi?ed, provid-
ing some support for the absence of a non-
response bias. However, this test is not conclusive.
Follow-up calls to late and non-respondents
revealed that some managers had no time to
complete the questionnaire, while other managers
stated that they had recently received a large
number of questionnaires and their ?rm had
adopted a policy of not completing postal sur-
veys. The low response rate is clearly a limitation
of this study.
1
Survey design and variable measurement
The questionnaire was pre-tested by a small
group of academics and managers, prior to post-
age. This resulted in changes to some of the
wording and the presentation of the questionnaire.
The survey instrument consisted of demographic
data, including sales revenue, book value of assets
and employee numbers, and questions that mea-
sured each of the variables of interest. In most
cases, the measures were adapted from prior
research. All variables (except organizational per-
formance) were measured on an 11-point scale, to
capture decreased change (À5 to À1), no change
(0) and increased change (+1 to +5), over the
past 3 years. Where relevant, respondents had the
opportunity to indicate if the various practices or
items had never been used or adopted (indicated
as N/A). For analysis, scores were coded ‘0’ for N/
A while the remaining responses were coded 1–11.
Therefore, the point of ‘no change’ was coded ‘6’.
The measure of environmental change was based
on the instrument used by Tan and Litschert
(1994), developed from Khandwalla (1977). As
with other questions in this survey, managers’ per-
ceptions were relied on to measure environmental
change. Managers’ perceptions were considered
appropriate in this situation, compared to the use
of more objective measures for several reasons.
First, it is managers’ perceptions of the environ-
ment which are of interest, as it is these percep-
tions that will in?uence decisions with respect to
the choice of strategy and changes in other orga-
nizational and management accounting variables.
Second, it is di?cult to measure objectively vari-
ables such as the extent of change in the environ-
ment, or change in strategic emphasis (Duncan,
1972; Khandwalla, 1972; Lawrence & Lorsch,
1967). Finally, it has been argued that individuals
have su?cient understanding of their decision
process to be able to give relatively reliable infor-
mation (Chenhall & Morris, 1986; Shortell &
Zajac, 1990). The measure of change in the envir-
onment addressed changes in competitors’ actions
with respect to eight areas: prices and products;
customers’ expectations in regard to price, quality
and delivery; suppliers’ actions in respect to price,
quality, delivery and availability; availability of
new manufacturing technology; regulations such
as changes in tari? policies; local economic fac-
tors such as market and economic growth; inter-
national factors such as the world economy,
imports and foreign exchange; and development
of new products or services in the industry.
2
Respondents were asked to indicate the extent to
which they believed the competitive environment
of their business unit had changed over the past 3
years for each of these eight areas. The anchors
of the scale were ‘signi?cantly less competitive’
(scored À5) to ‘signi?cantly more competitive’
(scored +5).
The measure of change in the use of advanced
manufacturing technology consisted of nine items
taken fromrecent management accounting literature
1
Dillman (2000) provides several measures for reducing
nonresponse bias in postal questionnaire studies. These include
the use of response-friendly questionnaires (including the design
of the letter, the length of the survey, the order of the questions),
using of multiple contacts, including a pre-paid return envelope,
personalisation of correspondence, and prepaid incentives.
While most aspects of Dillman’s ‘‘tailored design method’’ were
applied in the design of this study, on re?ection, an increase in
the number of contacts made with managers and the use of pre-
paid incentives may have increased the response rate.
2
These items are similar to those used by Gul and Chia
(1994), Govindarajan (1984), and Kren and Kerr (1993).
684 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
(see, for example, Kotha & Swamidass, 2000;
Rimmer et al., 1996). These items were just-in-time
purchasing, just-in-time production, total quality
management, ?exible manufacturing systems,
computer integrated manufacturing, computer
aided design, computer aided manufacturing,
materials requirements planning and manufactur-
ing resource planning. Respondents were asked to
indicate the extent to which their business unit’s
use of the eight technologies had changed over the
past three years, from ‘used signi?cantly less’
(scored À5) to ‘used signi?cantly more’ (scored
+5).
The strategy measure focused on the extent to
which each business unit had changed its strategic
emphasis over a range of di?erentiation aspects,
during the past 3 years. The nine items (listed in
Table A3 in the Appendix) all measured aspects of
di?erentiation, and were derived from instruments
used by Chenhall and Lang?eld-Smith (1998b),
Parthasarthy and Sethi (1993) and Perera et al.
(1997). Each element was rated on an 11-point
Likert scale ranging from ‘emphasized sig-
ni?cantly less’ (scored À5) to ‘emphasized sig-
ni?cantly more’ (scored +5).
Change in organizational design was adapted
from measures developed by Chenhall and Lang-
?eld-Smith (1998b) and Miller, DeMeyer and
Nakane (1992). Respondents were asked to indi-
cate the extent to which their use of a range of ten
organizational design practices had changed over
the past 3 years. These items were work-based
teams, manufacturing cells, cross-functional
teams, multi-skilling of the workforce, manage-
ment training, worker training, employee empow-
erment, ?attening of the formal organizational
structure, establishing a participative culture and
reorganizing manufacturing processes. The
11-point Likert scale ranged from ‘used sig-
ni?cantly less’ (scored À5) to ‘used signi?cantly
more’ (scored +5).
The measure for change in advanced manage-
ment accounting practices required respondents to
indicate the extent to which their use of a range of
nine contemporary management accounting prac-
tices had changed over the past three years. The
items were activity-based costing, activity-based
management, target costing, value chain analysis,
benchmarking, product life-cycle analysis, product
pro?tability analysis, customer pro?tability analy-
sis and quality improvement programs. The items
were derived from a measure developed by Chen-
hall and Lang?eld-Smith (1998b). The 11-point
Likert scale ranged from ‘used signi?cantly less’
(scored À5) to ‘used signi?cantly more’ (scored
+5).
Changes in non-?nancial management account-
ing information included a list of 19 items based
on items used by Abernethy and Lillis (1995) and
Stivers, Covin, Hall, and Smalt (1998). Items
included on-time delivery, customer satisfaction,
ongoing supplier evaluations, rate of new product
introductions, and measures of set-up times. (A
full list is provided in Table A6 in the Appendix.)
Respondents were asked to indicate the extent to
which their reliance on the items of information
for decision making had changed over the past
three years, on a scale ranging from ‘considerably
less reliance’ (scored À5) to ‘considerably more
reliance’ (scored +5).
The organizational performance variable was
measured using the two-part measure developed
by Govindarajan (1988) and Govindarajan and
Fisher (1990). The ?rst part of the measure asked
respondents to compare the change in their busi-
ness unit’s performance over the past three years,
relative to their competitors, over ten ?nancial and
non-?nancial dimensions of performance. These
items were return on investment, pro?t, cash ?ow
from operations, cost control, development of new
products, sales volume, market share, market
development, personnel development and poli-
tical-public a?airs. An 11-point Likert scale ran-
ged from ‘signi?cantly lower performance than
competitors’ (scored À5) to ‘signi?cantly higher
performance than competitors’ (scored +5). The
second part of the measure required respondents
to assess the same ten dimensions in terms of their
importance to the business unit, on a 5-point
Likert scale ranging from ‘no importance’ (scored
1) to ‘extremely important’ (scored 5). Final scores
for each dimension were determined by multi-
plying the respective ‘performance’ and ‘impor-
tance’ scores. A single performance score for each
?rm was calculated as the weighted average of all
10 dimensions.
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 685
Statistical analysis and results
This section includes details of the statistical
analysis undertaken using structural equation
modeling (SEM), hypotheses testing and pre-
sentation of the results. SEM was the preferred
method of analysis in this study as it allows the
analysis of multiple relationships simultaneously,
provides measures of overall model ?t, as well as
explaining the signi?cance of each of the relation-
ships between the variables (Kline, 1998). The
ability to model multiple relationships is an
advantage of latent variable SEM over multiple
regression and path analysis. In addition, unlike
regression analysis and path analysis, SEM will
account for the e?ects of measurement error in
multi-item variables.
Measurement models
Structural equation modeling (SEM) was used
to analyze the data using a two-stage process
recommended by Schumaker and Lomax (1996).
In the ?rst stage, each latent variable was modeled
as a separate measurement model. A measurement
model relates observed variables to their associated
latent variable. In this case, the latent variables were
changes in the competitive environment (environ-
ment), changes in the use of advanced manufactur-
ing technology (technology), changes in emphasis
on di?erentiation strategy (strategy), changes in the
use of organizational design practices (organization
design), changes in the use of advanced manage-
ment accounting practice (AMAP) and change in
reliance on non-?nancial management accounting
information (NF-MAI). The second stage involved
constructing the structural model by specifying the
relationships between the latent variables. In this
section, the ?rst stage will be described.
Formulating measurement models for each
variable involved using AMOS 4 to conduct con-
?rmatory factor analysis for each set of items.
3
Covariances were included between error terms in
each measurement model where suggested by
AMOS, but only where such covariances were
justi?ed theoretically.
4
Details of the ?t indices for
each measurement model are shown in Table 1.
Model ?t is de?ned by Hair, Anderson, Tatham
and Black (1998), as the ‘‘degree to which the
actual/observed input matrix is predicted by the
estimated model’’. There is no single measure of ?t
for structural equation models, so it is good prac-
tice to include a range of ?t indices. Chi-square
and the Goodness-of-Fit (GFI) index measures
overall model ?t. A non-signi?cant chi-square
indicates that the data ?t the model, while the GFI
index ranges from 0 (poor ?t) to 1 (perfect ?t) with
an acceptable minimum of 0.90 (Hu & Bentler,
1999). The GFI is considered analogous to the R
2
value reported in multiple regression models
(Hoyle & Panter, 1995; Tanaka, 1993). Model
comparison was evaluated using the Tucker–Lewis
Index (TLI) (Tucker & Lewis, 1973) and the
Comparative Fit Index (CFI) (Bentler, 1990). A
value greater than 0.90 for the TLI and 0.95 for
the CFI re?ects a good model ?t. The Adjusted
Goodness-of-Fit (AGFI) index and the Akaike
Information Criterion (AIC) were used to measure
model parsimony. The AGFI ranges from 0 (poor
?t) to 1 (perfect ?t) with a cuto? of 0.90 indicating
a good ?t, while the AIC should be less than for
the saturated model. This measure assesses whether
or not the model ?t has been achieved by ‘‘over ?t-
ting’’ the data with too many coe?cients (Hair et al.,
1998). In Table 1, in all but one case, the ?t indices
for each measurement model were better than the
recommended criteria. In the case of NF-MAI, the
?t was acceptable for all but two measures: the chi-
square was signi?cant (P=0.021) and the AGFI was
0.885. However, in the light of the other very favor-
able ?t measures, the measurement models can all be
regarded as exhibiting good ?t.
A number of items were deleted as part of the
development of the measurement models. As the
strategy item ‘provide unique product features’
had a negative factor score weight, this was
deleted from the strategy variable (Loehlin, 1998).
3
In SEM, con?rmatory factor analysis, rather than
exploratory factor analysis is used as the identi?cation of the
factors is speci?ed by theory, not by the data (Hair et al., 1998).
4
The inclusion of non-zero covariances between error terms
has been used in prior management accounting research when
semantic (or theoretical) relationships exist between items (see,
for example, Shields, Deng & Kato, 2000; Jaworski & Young,
1992).
686 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
The item ‘reorganizing manufacturing processes’
was removed from the organization design vari-
able, as it had a squared multiple correlation of
0.04, indicating that the item explained less than
5% of the latent variable. Due to the large number
of items in the NF-MAI variable, the number of
parameters to be estimated was too large in rela-
tion to the sample size, resulting in an initial poor
?t (Kline, 1998).
5
Further analysis was performed
by dividing the items into two variables, to reduce
the number of parameters to be estimated for each
variable. However, this analysis revealed the items
did not comprise more than one distinct variable.
Therefore, to decrease the number of free para-
meters and improve the stability of the measure-
ment model, seven items with low squared
multiple correlations (
This paper reports on a survey of manufacturing companies, and uses structural equation modeling to examine the
relationships between the changing competitive environment, and a range of organizational variables as antecedents to
management accounting change. The results indicate that an increasingly competitive environment has resulted in an
increased focus on differentiation strategies. This, in turn, has influenced changes in organizational design, advanced
manufacturing technology and advanced management accounting practices. These three changes have led to a greater
reliance on non-financial accounting information which has led to improved organizational performance.
Antecedents to management accounting change:
a structural equation approach
Annette Baines*, Kim Lang?eld-Smith
Department of Accounting and Finance, Monash University, PO Box 197, Caul?eld East, Victoria 3145, Australia
Abstract
This paper reports on a survey of manufacturing companies, and uses structural equation modeling to examine the
relationships between the changing competitive environment, and a range of organizational variables as antecedents to
management accounting change. The results indicate that an increasingly competitive environment has resulted in an
increased focus on di?erentiation strategies. This, in turn, has in?uenced changes in organizational design, advanced
manufacturing technology and advanced management accounting practices. These three changes have led to a greater
reliance on non-?nancial accounting information which has led to improved organizational performance.
#2003 Elsevier Ltd. All rights reserved.
Introduction
Throughout the 1990s, the growing level of glo-
bal competition intensi?ed the challenges for
managers who need to consider more e?ective
ways of achieving competitive advantage and
improving organizational performance. One
means of achieving this is through the adoption of
clearly articulated strategies, ?exible organiza-
tional structures and innovative accounting sys-
tems. A common theme in normative management
accounting research is that changes in an organi-
zation’s external environment will lead to change
in an organization’s management accounting sys-
tem (Atkinson et al., 1997; Nanni, Dixon, & Voll-
man, 1992). This is based on the argument that
managers need speci?c forms of management
accounting information that support their decision
needs within increasingly uncertain environments,
and to assist them to monitor progress against
strategies. The need for an appropriate ?t between
the environment and organizational systems is an
underlying assumption of much of the empirical
contingency-style management accounting
research (see for example, Chenhall & Lang?eld-
Smith, 1998a,1998b; Gul, 1991; Perera, Harrison,
& Poole, 1997), as is the need for management
accounting systems to change to support man-
agers’ new information requirements. However,
prior empirical research in management account-
ing has examined primarily the various relation-
ships between the environment, organizational
variables, and management accounting systems at
a point in time. There has been limited empirical
research examining the nature of the changes in
management accounting systems and organiza-
tional variables made in response to environ-
mental changes, and whether or not these changes
improve organizational performance.
0361-3682/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0361-3682(02)00102-2
Accounting, Organizations and Society 28 (2003) 675–698
www.elsevier.com/locate/aos
* Corresponding author. Tel.: +61-3-9903-2792; fax: +61-
3-9903-2422.
E-mail address: [email protected]
(A. Baines).
This study uses a survey method and utilizes
structural equation modeling to identify whether
changes that have occurred within the external
environment of manufacturing organizations have
led to change in the type of management account-
ing information provided and used for decision-
making. More speci?cally, this study investigates
whether changes in the organizational environ-
ment have led to changes in the organizations’
strategy, organizational design, advanced manu-
facturing technology and management accounting
practices. These changes in turn are hypothesized
to in?uence the type of management accounting
information used by managers, which may lead to
improved organizational performance. The
remainder of this paper is organized as follows.
Section 2 summarizes the relevant prior literature,
and presents research hypotheses. The research
method and development of the survey is
explained in Section 3. Section 4 contains the
analysis of the data, and this is followed by a dis-
cussion of the results and conclusion.
Literature review and hypothesis development
E?ects of changes in the competitive environment
In this section, the literature that examines the
relationships between changes in the competitive
environment, and changes in strategy, organiza-
tion design and technology is reviewed, and
hypotheses presented.
Changes in the environment and strategy
Since the 1980s, there has been signi?cant
change in the external environment faced by ?rms
in all sectors of the economy. This includes more
active competitors, increasingly demanding cus-
tomers, and the availability of information pro-
cessing technologies (Hiromoto, 1991; Innes &
Mitchell, 1990; Shields, 1997). Such changes have
occurred as a result of market and ?nancial
deregulation, and increasingly companies com-
pete in a single global market. Increasing globali-
zation has resulted in intense and aggressive
international competition, increased customer
demands including diversi?ed customer needs,
and shorter product life cycles (Dent, 1996;
Shields, 1997).
The strategy an organization adopts constitutes
the logic underlying its interactions with its envir-
onment. It is well established that an organiza-
tion’s strategy is a response to its environment,
and that the appropriate matching of strategy and
the environment can enhance performance (Burns
& Stalker, 1961; Porter, 1980). Several empirical
studies have examined the linkage between envir-
onment and strategy. For example, Fuschs, Mif-
?in, Miller, and Whitney (2000) found that
successful ?rms aligned key elements of strategy
with the environment. Chong and Chong (1997)
found a positive association between perceived
environmental uncertainty and strategy, and
Miller (1988) found a relationship between an
unpredictable and dynamic environment and an
innovation strategy.
As the environment becomes dominated by
increasingly more demanding customers and as
competitors respond to customer demands in
increasingly sophisticated ways, a ?rm may place
greater emphasis on developing a di?erentiation
strategy that emphasizes more customer-oriented
aspects such as quality, ?exibility, innovative pro-
ducts and dependability of supply (Perera et al.,
1997; Sim & Killough, 1998). Therefore, the fol-
lowing hypothesis is proposed:
H
1a
. Firms facing a more competitive environment
will change towards a di?erentiation strategy.
Changes in the environment and organization design
Under conditions of environmental change
where markets have become more competitive,
particularly with an increased level of high quality
and competitively priced products, ?rms may
respond by reorganizing their work processes, and
adopting structures that have a stronger customer
orientation (Keidel, 1994; Miller, 1988; Partha-
sarthy & Sethi, 1993). In particular, a variety of
team-based structures have emerged, including
self-managed work teams, and cross-functional
project teams (Cohen & Bailey, 1997). Self-mana-
ging work teams are increasingly relied on in
manufacturing organisations (Anderson, Hesford,
& Young, 2002). These teams are continuing work
676 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
units responsible for producing products, often
from the initial purchase of materials, through all
stages of the production process, to the ?nal pro-
duct. Employees are empowered to make a variety
of decisions that were once the domain of super-
visors and managers, and team members are
multi-skilled. These decisions can cover the areas
of production quality, supplier and customer liai-
son, personnel management, production planning
and con?ict resolution. Cross-functional teams are
one-time team structures that are formed for a
speci?c purpose, such as developing a new pro-
duct, or improving work processes. Team mem-
bers are drawn from di?erent functional areas of
the business which allows a variety of di?erent
skills, knowledge, expertise and perspectives to
be brought to the project. These teams have the
potential to generate new ideas or derive novel
solutions, which may be essential to competing
in dynamic competitive environments (Anderson
et al., 2002). Thus, higher quality and more
timely and creative outcomes can be expected
from such teams, as their capacity to undertake
multiple activities simultaneously provides an
e?ective response to time-based competition
(Brown & Eisenhardt, 1995; Cohen & Bailey,
1997).
The adoption of teams is associated with ?atter
hierarchies and the increased empowerment of
lower-level managers and employees (Chenhall &
Lang?eld-Smith, 1998b; Elliot, 1991; Otley, 1994;
Rimmer, Macneil, Chenhall, Lang?eld-Smith, &
Watts, 1996; Shields, 1997). To ensure fast and
innovative responses in complex and dynamic
environments, there has been a move away from
hierarchical controls and centralized decision
making, towards the allocating of more responsi-
bility to lower levels of the ?rm. It has been argued
that the use of team-based structures in complex
environments enables organizations not only to
improve their speed and ?exibility of response, but
also to improve the quality of that response
(Cohen, 1993; Lawler, 1993). Therefore, the fol-
lowing hypothesis is proposed:
H
1b
. A more competitive environment will result
in changes in organizational design, with greater
use of team-based structures.
Changes in the environment and technology
An increasingly competitive environment can
lead to ?rms adopting innovative products and
production techniques to provide increased ?ex-
ibility, and to satisfy customer demands (Foster &
Gupta, 1994; Otley, 1994). In order to compete in
these highly competitive markets, many organiza-
tions have made considerable investments in
advanced manufacturing technology, such as
computer-aided design, computer-integrated man-
ufacturing and just-in-time systems. The use of
these technologies has resulted not only in
increased quality, but also in the ability to provide
di?erentiated products or services to satisfy spe-
ci?c market segments or even individual customers
(Elliot, 1991; Gosse, 1993; Otley, 1994).
The increased ?exibility provided by advanced
manufacturing technology has made it more
attractive to increase the breadth of the product
line, as making more frequent changeovers and
producing in smaller batches allows customer pre-
ferences to be satis?ed, without having to incur
high inventory storage costs (Milgrom & Roberts,
1995). Advanced manufacturing technologies
enable companies to compete on the basis of
quality, productivity and ?exibility, as well as on
cost, and have transformed the manner in which
production takes place even in relatively small
companies (Bhimani, 1994; Bruggeman & Slag-
mulder, 1995; Parthasarthy & Sethi, 1992; Pfe?er,
1994). Thus, the following hypothesis is proposed:
H
1c
. A more competitive environment will result
in increased use of advanced manufacturing
technology.
E?ects of changes in strategy
In this section, the relationships between chan-
ges in strategy and changes in organization design,
technology and advanced management accounting
practices are presented. Three hypotheses are
developed.
Changes in strategy and organization design
The increased emphasis on a di?erentiation
strategy may require operational employees to
adopt a stronger customer orientation, and one
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 677
way of encouraging employees to take ownership
of such a strategy is through the introduction of
team-based structures. In production teams,
workers are encouraged to pool their knowledge
of the production process and initiate innovative
approaches to improve productivity and quality,
and to reduce production lead time (Banker, Pot-
ter, & Schroeder, 1993). Teams provide the basis
for a stronger customer focus and allow the struc-
turing of work around processes rather than
functions, which has direct implications for enga-
ging in process improvement activities which may
improve an organization’s speed, ?exibility and
quality of response (Scott & Tiessen, 1999; Rim-
mer et al., 1996). Therefore, the following
hypothesis is proposed:
H
2a
. A change towards a di?erentiation strategy
will result in changes in organization design, with
greater use of team-based structures.
Changes in strategy and technology
A proper link between strategy and manu-
facturing operations is a key to developing sus-
tainable competitive advantage (Porter, 1996).
One way in which manufacturing organizations
can respond to increasing customer demands of
quality, ?exibility and dependability of supply is
through the implementation of advanced manu-
facturing technology. Schroeder and Congden
(2000), in a study of small to medium-sized man-
ufacturers, found the most ?nancially successful
?rms were those which demonstrated a tight
alignment between strategy and technology, while
Kotha and Swamidass (2000) found that for ?rms
competing on the basis of quality, customer ser-
vice, delivery reliability, product features and ?ex-
ibility, investment in advanced manufacturing
technology resulted in superior growth. A positive
association between an emphasis on di?erentia-
tion strategies and investments in advanced man-
ufacturing technologies was also found in a study
of UK ?rms (Burcher & Lee, 2000). Thus, the fol-
lowing hypothesis is proposed:
H
2b
. A change towards a di?erentiation strategy
will result in increased use of advanced manu-
facturing technology.
Changes in strategy and advanced management
accounting practices
In pursuing competitive advantage, organiza-
tions may implement manufacturing processes and
administrative functions that support their parti-
cular strategic priorities. For example, both
Chenhall and Lang?eld-Smith (1998b) and Call-
ahan and Gabriel (1998) found greater use of
advanced management accounting practices, such
as quality improvement programs, benchmarking
and activity-based management, in ?rms that
placed a strong emphasis on product di?erentia-
tion strategies, and this ultimately resulted in high
performance.
Advanced management accounting practices
can assist employees to more easily focus on
achieving di?erentiation priorities, such as quality,
delivery and customer service, compared to more
traditional ?nancially based accounting practices,
as they highlight the need to satisfy customer
requirements. For example, target costing allows
managers to focus on low cost while simulta-
neously maintaining customer expectations in
areas of quality and functionality. Also, activity-
based management focuses on maintaining or
enhancing customer value, not just controlling
costs. Thus, the following hypothesis is proposed:
H
2c
. A change towards a di?erentiation strategy
will result in the increased use of advanced man-
agement accounting practices.
E?ects of changes in technology
This section reviews the literature examining the
relationships between changes in the use of
advanced manufacturing technology, and changes
in organization design and advanced management
accounting practices.
Changes in technology and organization design
Adopting new technologies may require changes
in organizational structures and work practices to
better suit the capabilities of that technology
(Rimmer et al., 1996). Traditional manufacturing
plants tend to be laid out by machine or process
function. Line personnel, separated from their co-
workers by inventory, become specialized by
678 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
repeatedly processing large batches of similar
materials, and inventories are pushed through the
system with quality inspections conducted by
quality control personnel occurring at the end of
the production. In contrast, the successful imple-
mentation of new technology often requires the
blending of technological and social skills, which
can best be achieved through adoption of work-
based teams or production cells. These new struc-
tures emphasize products and customers rather
than mass production. A team may manage the
complete processing of products, with each
employee performing several functions. Batches
are small, there is little work-in-process inventory,
teams become responsible for production and
quality, and sometimes liaise directly with suppli-
ers and customers (Banker et al., 1993; Rimmer et
al., 1996). Therefore, the following hypothesis is
proposed:
H
3a
. The increased use of advanced manufacturing
technology will result in a change in organization
design, with greater use of team-based structures.
Changes in technology and advanced management
accounting practices
With the introduction of new technologies in
manufacturing operations, the structure of manu-
facturing costs has changed. Technologies such as
computer-integrated manufacturing and just-in-
time systems emphasize that variable direct labor
and inventory are vanishing from the factory.
Speed of operation is not determined by how fast
an operator can work, but by the type of automa-
tion and manufacturing system used (Dhavale,
1996). Consequently, traditional cost control sys-
tems which focus on variance analysis, aggregat-
ing costs and accounting for inventory, do not
e?ectively identify resources consumed, or help
managers manage those resources. In addition,
they may distort the realities of manufacturing
performance with new technological processes
(Bruggeman & Slagmulder, 1995; Gosse, 1993;
Kaplan, 1994).
Advanced management accounting practices,
such as activity-based management, life cycle cost-
ing and target costing, appear to be gaining an
increasing foothold in the modern world (Granlund
& Lukka, 1998). They provide an approach to
resource management that supports a customer
focus. For example, benchmarking provides an
important way of sensitizing the organization to
external performance standards, thus overcoming
the inward focus of cost variance analysis (Chen-
hall & Lang?eld-Smith, 1998a; Elnathan, Lin, &
Young, 1996). Quality improvement programs
have evolved as a philosophy that stresses the
notion of continuously improving manufacturing
processes by eliminating waste, improving quality,
developing people skills and at the same time
reducing costs (Sim & Killough, 1998). The fol-
lowing hypothesis is proposed:
H
3b
. The increased use of advanced manufacturing
technology will result in the increased use of
advanced management accounting practices.
In?uence of organizational changes on non-
?nancial management accounting information
In this section, the literature that examines the
relationships between changes in organizational
variables, namely advanced management accounting
practices, strategy, organization design and technol-
ogy, and changes in reliance on non-?nancial man-
agement accounting information will be reviewed.
Changes in advanced management accounting
practices and non-?nancial MAI
There is strong empirical support for the associ-
ation between advanced management accounting
practices and the increased use of non-?nancial
performance measures. Chenhall and Lang?eld-
Smith (1998b) found that ?rms which emphasized
product di?erentiation strategies bene?ted from
the use of advanced management accounting
practices and reliance on non-?nancial informa-
tion. Further, Abernethy and Lillis (1995), Banker
et al. (1993), Perera et al. (1997) and Sim and
Killough (1998) found a positive association
between the emphasis placed on various forms of
advanced management practices in an environ-
ment of manufacturing ?exibility, and the use of
non-?nancial measures such as defect rates, on-
time delivery and machine utilization. Therefore,
the following hypothesis is proposed:
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 679
H
4a
. The increased use of advanced management
accounting practices will result in greater reliance
on non-?nancial MAI.
Changes in strategy and non-?nancial MAI
An organization’s ability to survive and function
successfully in an environment of intense competi-
tion depends partially on the availability of infor-
mation upon which its managers can act (Bhimani,
1993). Therefore, accounting information should
be designed to facilitate informed judgments and
decisions that support the chosen strategy.
The development of appropriate management
accounting information requires that the strategy
be articulated so that goals and objectives are
de?ned with su?cient precision to be measurable.
For example, if quality or innovation is central to
a ?rm’s competitive strategy, then it is vital that
the ?rm’s accounting system is able to measure
events representing these factors. Measures that
focus on customer satisfaction, manufacturing
excellence, market leadership, quality, reliability,
responsiveness, and technological leadership,
innovation and human resources need to be con-
sidered (Eccles, 1991; Fisher, 1992). The failure to
include these measures in a formal management
accounting system may well hamper an organiza-
tion’s ability to achieve various di?erentiation
strategies of customer satisfaction, quality, speed,
learning and process improvement (Sim & Kill-
ough, 1998).
A change towards a di?erentiation strategy
requires a shift from cost and e?ciency-based
performance measures to ones which measure
speci?c non-?nancial aspects of strategy, such as
quality, service, ?exibility and dependability
(Cooper, Kaplan, Maisel, Morrissey, & Oehm,
1992; Samson, Lang?eld-Smith, & McBride,
1991). This leads to the hypothesis:
H
4b
. A change towards a di?erentiation strategy
will result in greater reliance on non-?nancial
MAI.
Changes in organization design and non-?nancial
MAI
A critical aspect of adopting team operations is
the process of empowerment. Teams cannot simply
be delegated responsibilities. Empowerment places
both authority and responsibility at low levels in
an organization; it implies that employees are
willing to take responsibility for their own actions,
and use their initiative to locate and solve pro-
blems to ensure that organizational goals are met,
particularly in relation to customer service (Ich-
niowski, Shaw, & Prennushi, 1997; Otley, 1994;
Rimmer et al., 1996). Changing the organization
design, including the use of teams and employee
empowerment, will result in changed employer
and employee expectations, including increased
access to relevant information (Scott & Tiessen,
1999).
The role of management accounting in this
changed organization structure is not simply to
deliver cost data, but to provide a service that
empowers team members to make the best deci-
sions in light of current conditions (Gordon &
Miller, 1976; Nanni et al., 1992). As more decision
making responsibility is passed to ‘lower’ levels of
the organization, there is an increased need for
easily accessible and relevant information at these
levels, as well as relevant information for top
management to evaluate the operations of the ?rm
and direct strategy (Banker et al., 1993). Non-
?nancial performance measures can form an inte-
gral part of the information base necessary for
team success (Scott & Tiessen, 1999). Therefore,
the following hypothesis is proposed:
H
4c
. A change in organization design, with greater
use of team-based structures, will result in greater
reliance on non-?nancial MAI.
Changes in technology and non-?nancial MAI
It has been argued that reliance on non-?nancial
information is dependent on the level of adoption
of advanced manufacturing technology (Perera et
al., 1997). If substantial changes take place in the
manufacturing processes, the management
accounting system must also change if it is to
provide relevant information for managerial deci-
sions and control (Foster & Horngren, 1988;
Kaplan, 1986; Nanni et al., 1992).
Financially oriented accounting measurement
systems imperfectly re?ect, and often with consider-
able lags, the increase in manufacturing ef?ciency
680 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
and e?ectiveness that occurs when ?rms pursue
total quality control, just-in-time inventory sys-
tems, and computer integrated manufacturing
processes. These systems, with their emphasis on
short-term pro?ts fail to capture the bene?ts that
arise from decreased new product launch times,
reduced throughput and lead times and the
improved ?exibility which modern manufacturing
technology has made possible (Kaplan, 1986).
Poorly designed or outdated accounting and con-
trol systems can distort the realities of manu-
facturing performance. Furthermore, such systems
can place out of reach most of the promised ben-
e?ts of new methods because the performance of
individuals, production processes, and ?rms in
high technology environments are not being
assessed and evaluated appropriately (Kaplan,
1984; Young & Selto, 1991). Such technological
changes require fundamental changes to perfor-
mance measures (Bhimani, 1993; Drury & Tayles,
1995). More timely, relevant and comprehensible
information is perceived as improving managerial
decisions in areas such as cost reduction, cost
control, product quality and performance assess-
ment. For example, to evaluate the e?ectiveness of
the ?exible manufacturing systems to deliver cus-
tomer value, measures such as cycle time, delivery
performance, and the ability of manufacturing to
vary product characteristics or develop new pro-
ducts would be more appropriate to track perfor-
mance, than traditional cost metrics (Abernethy &
Lillis, 1995; Bhimani, 1993). Thus, the following
hypothesis is proposed:
H
4d
. The increased use of advanced manufacturing
technology will result in greater reliance on non-
?nancial MAI.
Change in non-?nancial MAI and organizational
performance
It has been argued that reliance on ?nancial
performance measures alone will not necessarily
produce improved ?nancial results, as ?nancial
measures only indicate the outcomes of past
activities which may be no guide to improving
future performance. Non-?nancial measures,
however, can re?ect the drivers of future ?nancial
performance. It follows, therefore, that perfor-
mance measures that exclude important aspects of
manufacturing operations will not direct managers
to areas of critical concern (Nanni et al., 1992).
However, although predictive ability is claimed
to be one of the bene?ts of non-?nancial measures
(Atkinson et al., 1997), it appears that managers
are not always sure that their e?orts are rewarded
with improvements in the ?nancial bottom line
(Fisher, 1992). While the majority of prior
research supports a positive relationship between
the increased reliance on non-?nancial informa-
tion and organizational performance, the exact
nature of the relationship is ambiguous.
Mia and Clarke (1999) found an indirect asso-
ciation between the intensity of market competi-
tion and business unit performance through the
use of management accounting information.
Davila (2000) and Chong and Chong (1997)
established that a greater use of non-?nancial
information for business units following a custo-
mer-focused or prospector-type strategy, had a
positive impact on performance. Managers in
these cases tended to place greater reliance on
non-?nancial information on the assumption that
good performance in non-?nancial areas would
drive good ?nancial performance. On the other
hand, Perera et al. (1997) found no association
between the use of non-?nancial measures and
perceived performance in an organization with a
‘customer-focused’ manufacturing strategy.
Improved performance has resulted in ?rms that
use ?exible manufacturing, and which also place
greater reliance on non-?nancial manufacturing
measures (Abernethy & Lillis, 1995; Sim & Kill-
ough, 1998). Improved performance has also
occurred when teams have been provided with
comprehensive performance measures, that is, a
combination of ?nancial and non-?nancial mea-
sures (Scott & Tiessen, 1999). Ittner and Larcker
(1995) and Sim and Killough (1998) both found a
signi?cant positive interaction between TQM
practices, management accounting information
and performance. On the other hand, Banker, Pot-
ter and Srinivasan (2000) and Ittner and Larker
(1998) found that particular non-?nancial mea-
sures were positively associated with future, rather
than current, revenues and pro?t. It is possible,
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 681
however, that organizational strategy, the type of
technology, and the structural and environmental
factors confronting the organization could also
a?ect the use and performance consequences of
reliance on non-?nancial performance measures
(Daft & Lengel, 1986; Ittner & Larcker, 1998).
Also, improvements in non-?nancial measures will
not result in improved pro?ts if management has
selected the wrong critical success factors. If, for
instance, performance measures are based on qual-
ity, but quality is not of value to the customer, the
strategy will not lead to improved ?nancial perfor-
mance. A strategy of innovation will fail unless it
addresses an environment that values novelty and is
accompanied by an organizational structure that
can produce it. Similarly, a matching of only orga-
nizational design and environment with each other
will be inadequate if either con?icts with the selec-
ted strategy (Miller, 1982; Pelham, 1999).
Therefore, the use of non-?nancial performance
measures may be just one factor in a management
accounting information system that must be con-
sidered with all other variables within the organi-
zation. Successful companies must not only track
?nancial performance and operating e?ciency,
but must also pay close attention to customer
satisfaction and human resources, and there must
be an agreement on strategy and clarity of
communication (Lingle & Schiemann, 1996).
Information is now recognized as one of the most
powerful tools that can substantially in?uence the
wealth of corporations (Mangaliso, 1995). Thus,
in response to the changed environment, the chal-
lenge for any management accounting system is to
provide up-to-date information to help managers
reach informed economic decisions, and to moti-
vate users to aim and strive for organizational
change (Horngren, 1995; Stokes & Lawrimore,
1989). Failure to rely on appropriate accounting
information may contribute to ine?ective resource
management and a gradual decline in organiza-
tional performance. This leads to the ?nal
hypothesis.
H
5
. Achange in MAI, with greater reliance on non-
?nancial management accounting information, will
result in improved organizational performance.
The hypotheses are illustrated in Fig. 1.
Method
This section contains two parts. First, there is a
description of the sample used to test the hypotheses.
Second, the development of the survey instrument
and the measurement of the variables are described.
Sample selection
A self-administered questionnaire was sent to the
general managers of 700 manufacturing organiza-
tions randomly selected from the Kompass Australia
(1999) database. A strati?ed sampling approach was
used. This involved dividing the sample into ?ve
groups based on the number of employees (a proxy
for size), and selecting ?rms proportional to the
number of companies in each group. The ?rms
selected were business units of larger ?rms, or com-
panies in their own right. The initial posting of the
survey instrument also included, (a) a letter, addres-
sed personally to the general manager of each busi-
ness unit, explaining the purpose of the study; a tear-
o? section allowing respondents to provide their
name and address for a copy of the survey results,
while ensuring anonymity, (b) a reply-paid envelope
for the return of the survey, and (c) a separate reply-
paid envelope for the return of the tear-o? section.
Three weeks after the initial posting, a reminder let-
ter was sent to all managers who had not returned
the tear-o? section. Two weeks later, each manager
who could not be identi?ed as having responded was
phoned and encouraged to complete the ques-
tionnaire. Additional questionnaires were forwarded
to those who had misplaced the original material.
A total of 155 responses (22%) were received.
Of these, 14 were unusable, mainly due to missing
data on organizational performance. This resulted
in 141 useable responses, a response rate of 20%.
The remaining sample ?rms had average sales of
$128.2 million and average assets of $88.59 mil-
lion. The majority of ?rms in the sample were
stand-alone companies (59% of the sample) or
divisions of larger companies (37%). The respon-
dents had been with their ?rm an average of 11.3
years. Most described themselves either as man-
agers (61%) or managing directors (23%).
The useable response rate of 20% is clearly
lower than anticipated. It provides a sample size
682 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
F
i
g
.
1
.
H
y
p
o
t
h
e
s
i
z
e
d
m
o
d
e
l
.
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 683
which is just large enough to undertake the struc-
tural equation analysis (discussed in a later sec-
tion), but the potential for non-response bias does
arise. The mean responses for questionnaires
received prior to the reminder phone call were
compared to those received after the phone call, to
test if responses di?ered between the two groups.
No signi?cant di?erences were identi?ed, provid-
ing some support for the absence of a non-
response bias. However, this test is not conclusive.
Follow-up calls to late and non-respondents
revealed that some managers had no time to
complete the questionnaire, while other managers
stated that they had recently received a large
number of questionnaires and their ?rm had
adopted a policy of not completing postal sur-
veys. The low response rate is clearly a limitation
of this study.
1
Survey design and variable measurement
The questionnaire was pre-tested by a small
group of academics and managers, prior to post-
age. This resulted in changes to some of the
wording and the presentation of the questionnaire.
The survey instrument consisted of demographic
data, including sales revenue, book value of assets
and employee numbers, and questions that mea-
sured each of the variables of interest. In most
cases, the measures were adapted from prior
research. All variables (except organizational per-
formance) were measured on an 11-point scale, to
capture decreased change (À5 to À1), no change
(0) and increased change (+1 to +5), over the
past 3 years. Where relevant, respondents had the
opportunity to indicate if the various practices or
items had never been used or adopted (indicated
as N/A). For analysis, scores were coded ‘0’ for N/
A while the remaining responses were coded 1–11.
Therefore, the point of ‘no change’ was coded ‘6’.
The measure of environmental change was based
on the instrument used by Tan and Litschert
(1994), developed from Khandwalla (1977). As
with other questions in this survey, managers’ per-
ceptions were relied on to measure environmental
change. Managers’ perceptions were considered
appropriate in this situation, compared to the use
of more objective measures for several reasons.
First, it is managers’ perceptions of the environ-
ment which are of interest, as it is these percep-
tions that will in?uence decisions with respect to
the choice of strategy and changes in other orga-
nizational and management accounting variables.
Second, it is di?cult to measure objectively vari-
ables such as the extent of change in the environ-
ment, or change in strategic emphasis (Duncan,
1972; Khandwalla, 1972; Lawrence & Lorsch,
1967). Finally, it has been argued that individuals
have su?cient understanding of their decision
process to be able to give relatively reliable infor-
mation (Chenhall & Morris, 1986; Shortell &
Zajac, 1990). The measure of change in the envir-
onment addressed changes in competitors’ actions
with respect to eight areas: prices and products;
customers’ expectations in regard to price, quality
and delivery; suppliers’ actions in respect to price,
quality, delivery and availability; availability of
new manufacturing technology; regulations such
as changes in tari? policies; local economic fac-
tors such as market and economic growth; inter-
national factors such as the world economy,
imports and foreign exchange; and development
of new products or services in the industry.
2
Respondents were asked to indicate the extent to
which they believed the competitive environment
of their business unit had changed over the past 3
years for each of these eight areas. The anchors
of the scale were ‘signi?cantly less competitive’
(scored À5) to ‘signi?cantly more competitive’
(scored +5).
The measure of change in the use of advanced
manufacturing technology consisted of nine items
taken fromrecent management accounting literature
1
Dillman (2000) provides several measures for reducing
nonresponse bias in postal questionnaire studies. These include
the use of response-friendly questionnaires (including the design
of the letter, the length of the survey, the order of the questions),
using of multiple contacts, including a pre-paid return envelope,
personalisation of correspondence, and prepaid incentives.
While most aspects of Dillman’s ‘‘tailored design method’’ were
applied in the design of this study, on re?ection, an increase in
the number of contacts made with managers and the use of pre-
paid incentives may have increased the response rate.
2
These items are similar to those used by Gul and Chia
(1994), Govindarajan (1984), and Kren and Kerr (1993).
684 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
(see, for example, Kotha & Swamidass, 2000;
Rimmer et al., 1996). These items were just-in-time
purchasing, just-in-time production, total quality
management, ?exible manufacturing systems,
computer integrated manufacturing, computer
aided design, computer aided manufacturing,
materials requirements planning and manufactur-
ing resource planning. Respondents were asked to
indicate the extent to which their business unit’s
use of the eight technologies had changed over the
past three years, from ‘used signi?cantly less’
(scored À5) to ‘used signi?cantly more’ (scored
+5).
The strategy measure focused on the extent to
which each business unit had changed its strategic
emphasis over a range of di?erentiation aspects,
during the past 3 years. The nine items (listed in
Table A3 in the Appendix) all measured aspects of
di?erentiation, and were derived from instruments
used by Chenhall and Lang?eld-Smith (1998b),
Parthasarthy and Sethi (1993) and Perera et al.
(1997). Each element was rated on an 11-point
Likert scale ranging from ‘emphasized sig-
ni?cantly less’ (scored À5) to ‘emphasized sig-
ni?cantly more’ (scored +5).
Change in organizational design was adapted
from measures developed by Chenhall and Lang-
?eld-Smith (1998b) and Miller, DeMeyer and
Nakane (1992). Respondents were asked to indi-
cate the extent to which their use of a range of ten
organizational design practices had changed over
the past 3 years. These items were work-based
teams, manufacturing cells, cross-functional
teams, multi-skilling of the workforce, manage-
ment training, worker training, employee empow-
erment, ?attening of the formal organizational
structure, establishing a participative culture and
reorganizing manufacturing processes. The
11-point Likert scale ranged from ‘used sig-
ni?cantly less’ (scored À5) to ‘used signi?cantly
more’ (scored +5).
The measure for change in advanced manage-
ment accounting practices required respondents to
indicate the extent to which their use of a range of
nine contemporary management accounting prac-
tices had changed over the past three years. The
items were activity-based costing, activity-based
management, target costing, value chain analysis,
benchmarking, product life-cycle analysis, product
pro?tability analysis, customer pro?tability analy-
sis and quality improvement programs. The items
were derived from a measure developed by Chen-
hall and Lang?eld-Smith (1998b). The 11-point
Likert scale ranged from ‘used signi?cantly less’
(scored À5) to ‘used signi?cantly more’ (scored
+5).
Changes in non-?nancial management account-
ing information included a list of 19 items based
on items used by Abernethy and Lillis (1995) and
Stivers, Covin, Hall, and Smalt (1998). Items
included on-time delivery, customer satisfaction,
ongoing supplier evaluations, rate of new product
introductions, and measures of set-up times. (A
full list is provided in Table A6 in the Appendix.)
Respondents were asked to indicate the extent to
which their reliance on the items of information
for decision making had changed over the past
three years, on a scale ranging from ‘considerably
less reliance’ (scored À5) to ‘considerably more
reliance’ (scored +5).
The organizational performance variable was
measured using the two-part measure developed
by Govindarajan (1988) and Govindarajan and
Fisher (1990). The ?rst part of the measure asked
respondents to compare the change in their busi-
ness unit’s performance over the past three years,
relative to their competitors, over ten ?nancial and
non-?nancial dimensions of performance. These
items were return on investment, pro?t, cash ?ow
from operations, cost control, development of new
products, sales volume, market share, market
development, personnel development and poli-
tical-public a?airs. An 11-point Likert scale ran-
ged from ‘signi?cantly lower performance than
competitors’ (scored À5) to ‘signi?cantly higher
performance than competitors’ (scored +5). The
second part of the measure required respondents
to assess the same ten dimensions in terms of their
importance to the business unit, on a 5-point
Likert scale ranging from ‘no importance’ (scored
1) to ‘extremely important’ (scored 5). Final scores
for each dimension were determined by multi-
plying the respective ‘performance’ and ‘impor-
tance’ scores. A single performance score for each
?rm was calculated as the weighted average of all
10 dimensions.
A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698 685
Statistical analysis and results
This section includes details of the statistical
analysis undertaken using structural equation
modeling (SEM), hypotheses testing and pre-
sentation of the results. SEM was the preferred
method of analysis in this study as it allows the
analysis of multiple relationships simultaneously,
provides measures of overall model ?t, as well as
explaining the signi?cance of each of the relation-
ships between the variables (Kline, 1998). The
ability to model multiple relationships is an
advantage of latent variable SEM over multiple
regression and path analysis. In addition, unlike
regression analysis and path analysis, SEM will
account for the e?ects of measurement error in
multi-item variables.
Measurement models
Structural equation modeling (SEM) was used
to analyze the data using a two-stage process
recommended by Schumaker and Lomax (1996).
In the ?rst stage, each latent variable was modeled
as a separate measurement model. A measurement
model relates observed variables to their associated
latent variable. In this case, the latent variables were
changes in the competitive environment (environ-
ment), changes in the use of advanced manufactur-
ing technology (technology), changes in emphasis
on di?erentiation strategy (strategy), changes in the
use of organizational design practices (organization
design), changes in the use of advanced manage-
ment accounting practice (AMAP) and change in
reliance on non-?nancial management accounting
information (NF-MAI). The second stage involved
constructing the structural model by specifying the
relationships between the latent variables. In this
section, the ?rst stage will be described.
Formulating measurement models for each
variable involved using AMOS 4 to conduct con-
?rmatory factor analysis for each set of items.
3
Covariances were included between error terms in
each measurement model where suggested by
AMOS, but only where such covariances were
justi?ed theoretically.
4
Details of the ?t indices for
each measurement model are shown in Table 1.
Model ?t is de?ned by Hair, Anderson, Tatham
and Black (1998), as the ‘‘degree to which the
actual/observed input matrix is predicted by the
estimated model’’. There is no single measure of ?t
for structural equation models, so it is good prac-
tice to include a range of ?t indices. Chi-square
and the Goodness-of-Fit (GFI) index measures
overall model ?t. A non-signi?cant chi-square
indicates that the data ?t the model, while the GFI
index ranges from 0 (poor ?t) to 1 (perfect ?t) with
an acceptable minimum of 0.90 (Hu & Bentler,
1999). The GFI is considered analogous to the R
2
value reported in multiple regression models
(Hoyle & Panter, 1995; Tanaka, 1993). Model
comparison was evaluated using the Tucker–Lewis
Index (TLI) (Tucker & Lewis, 1973) and the
Comparative Fit Index (CFI) (Bentler, 1990). A
value greater than 0.90 for the TLI and 0.95 for
the CFI re?ects a good model ?t. The Adjusted
Goodness-of-Fit (AGFI) index and the Akaike
Information Criterion (AIC) were used to measure
model parsimony. The AGFI ranges from 0 (poor
?t) to 1 (perfect ?t) with a cuto? of 0.90 indicating
a good ?t, while the AIC should be less than for
the saturated model. This measure assesses whether
or not the model ?t has been achieved by ‘‘over ?t-
ting’’ the data with too many coe?cients (Hair et al.,
1998). In Table 1, in all but one case, the ?t indices
for each measurement model were better than the
recommended criteria. In the case of NF-MAI, the
?t was acceptable for all but two measures: the chi-
square was signi?cant (P=0.021) and the AGFI was
0.885. However, in the light of the other very favor-
able ?t measures, the measurement models can all be
regarded as exhibiting good ?t.
A number of items were deleted as part of the
development of the measurement models. As the
strategy item ‘provide unique product features’
had a negative factor score weight, this was
deleted from the strategy variable (Loehlin, 1998).
3
In SEM, con?rmatory factor analysis, rather than
exploratory factor analysis is used as the identi?cation of the
factors is speci?ed by theory, not by the data (Hair et al., 1998).
4
The inclusion of non-zero covariances between error terms
has been used in prior management accounting research when
semantic (or theoretical) relationships exist between items (see,
for example, Shields, Deng & Kato, 2000; Jaworski & Young,
1992).
686 A. Baines, K. Lang?eld-Smith / Accounting, Organizations and Society 28 (2003) 675–698
The item ‘reorganizing manufacturing processes’
was removed from the organization design vari-
able, as it had a squared multiple correlation of
0.04, indicating that the item explained less than
5% of the latent variable. Due to the large number
of items in the NF-MAI variable, the number of
parameters to be estimated was too large in rela-
tion to the sample size, resulting in an initial poor
?t (Kline, 1998).
5
Further analysis was performed
by dividing the items into two variables, to reduce
the number of parameters to be estimated for each
variable. However, this analysis revealed the items
did not comprise more than one distinct variable.
Therefore, to decrease the number of free para-
meters and improve the stability of the measure-
ment model, seven items with low squared
multiple correlations (