STRATEGY AND STRATEGIC MANAGEMENT ACCOUNTING: AN INVESTIGATION OF ORGANIZATIONAL CONFIGURATION

Manchester Business School Research Seminar – June 18, 2008


STRATEGY AND STRATEGIC MANAGEMENT ACCOUNTING: AN INVESTIGATION OF ORGANIZATIONAL CONFIGURATIONS


Authors:

Cadez Simon Faculty of Economics University of Ljubljana  Kardeljeva ploscad 17  1000 Ljubljana Slovenija

E-mail: simon.cadez@ef.uni-lj.si 


Guilding Chris Centre for Tourism, Sport and Service Innovation Griffith University Gold Coast Campus PMB 50 Gold Coast Mail Centre Queensland 9726 Australia

E-mail: c.guilding@griffith.edu.au


ABSTRACT

The study examines the alignment and effectiveness of strategic choice and strategic management accounting (SMA) system design configurations. Six configurations were derived empirically deploying a cluster analysis of data collected from a sample of 193 large Slovenian companies. The first four clusters appear to provide some support for the central configurational proposition that higher levels of vertical and horizontal configurational alignments are associated with higher levels of performance. Evidence that contradicts the theory is also apparent, however, as the remaining two clusters exhibit high degrees of SMA vertical and horizontal alignment, but low performance levels. A discussion is provided suggesting that this latter observation might result from the final two clusters failing to develop strategic postures that are compatible to the radically changed business context that has arisen since the replacement of the centrally controlled economy with a market economy in Slovenia.  

STRATEGY AND STRATEGIC MANAGEMENT ACCOUNTING: AN INVESTIGATION OF ORGANIZATIONAL CONFIGURATIONS 

1. INTRODUCTION

The relationship between strategy and management accounting has commanded considerable attention from management accounting researchers in the last two decades (e.g., Govindarajan and Gupta, 1985; Simons, 1987; Govindarajan, 1988; Dent, 1990; Abernethy and Guthrie, 1994; Chong and Chong, 1997; Ittner and Larcker, 1997; Langfield-Smith, 1997; Perrera et al, 1997; Chenhall and Langfield-Smith, 1998; Abernethy and Brownell, 1999; Anderson and Lanen, 1999; Nyamori et al, 2001; Chenhall, 2003; Henri, 2006; Naranjo-Gil and Hartmann, 2007). The body of evidence accumulated in these studies suggests that high organizational performance may result from tailoring an organisation’s management accounting system to its strategy.   

Much of the empirical work in the area follows a contingency approach (LangfieldSmith, 1997; Chenhall, 2003). Despite the quantum of this research effort, there has been significant criticism of the manner in which it has been undertaken (see Gresov and Drazin, 1997; Chenhall, 2003; Gerdin and Greve, 2004; in press; Fiss, 2007). One particular shortcoming of contingency-based research revolves around its fragmentary nature. Typically, these studies deploy causal models where one variable or a set of variables is used as an antecedent of another variable or a set of variables (Luft and Shields, 2003). This approach treats variables as competing in explaining variation in outcomes rather then showing how variables combine to create outcomes (Fiss, 2007). 

The recent evolution of the equifinality concept in the management and organizational literature presents a challenge to contingency based modeling. The advocates of equifinality (Doty et al, 1993; Delery and Doty, 1996; Gresov and Drazin, 1997; Fiss, 2007; Marlin et al, 2007) contend that the possibility of multiple, equally effective, structures that are supportive of a given strategy undermines the contingency research design. The assumpiton of equifinality is incorporated in configurational theories. These theories suggest that organizations are best understood as clusters of interconnected structures and practices (Drazin and Van de Ven, 1985; Ferguson and Ketchen, 1999; Fiss, 2007) and degree of effectiveness can be attributed to internal consistency, or fit, among the patterns of relevant contextual, structural and strategic factors (Doty et al, 1993; Ketchen et al, 1993). 

In stark contrast to the rich vein of contingency-based accounting research, there has been very little theory development or empirical research concerning the way in which elements of management accounting combine with a variety of strategic choices to enhance performance. In fact, the work by Chenhall and Langfield-Smith (1998) stands in relative isolation. This is most likely attributable to the complexity involved in operationalizing the holistic approach. The simultaneous investigation of a variety of variables results in the problem of conflicting contingencies (Fisher, 1995; Gerdin, 2005), which makes interpretation and theory building difficult. Further, the simultaneous testing of multiple fits precludes us from using rigorous statistical methods such as regression analysis or structural equation modelling (Gerdin and Greve, 2004; Fiss, 2007). Despite this, the more holistic approach represents an acknowledgment of concerns that contingency-based research provides only a partial understanding of context-structure relationships (Chenhall, 2003; Fiss, 2007). Recognition of the paucity of prior accounting research adopting a holistic approach when examining context-structure relationships provided the broad motivation for the study reported herein.  

This study addressess a range of strategic choice configurations and two strategic management accounting (SMA) dimensions. The strategic choices explored concern: (1) type of business strategy, (2) degree to which strategy is deliberately formulated and (3) degree of market orientation. The two elements of strategic management accounting are the adoption of SMA techniques and accountant’s participation in strategic management processes. While an almost infinite number of combinations appear to be hypothetically possible, Gerdin and Greve (2004) argue that most firms can be assigned to a limited set of system states (configurations). Configurational theory asserts that not all configurations are equally effective and that the success of a particular configuration depends on its internal consistency and its appropriatness for the context in question (Ketchen et al, 1993; Gresov and Drazin, 1997; Ferguson and Ketchen, 1999). This study pursues two main aims: (1) to appraise what organizational configurations exist in a sample of large companies in a successful transitional economy, and (2) to assess the level of effectiveness and internal consistency of the configurations identified. 

The analysis is based on data collected from 193 large Slovenian companies. Slovenia was chosen following calls by Ittner and Larcker (2001) and Chenhall (2003) that research relevancy should be sought by studying novel management accounting approaches in a range of contemporary settings. Slovenia has been described as a role model of a successful transition from a socialist to a market economy (Edwards and Lawrence, 2000; Reardon et al, 2005). A mere 16 years after gaining independence from Yugoslavia and introducing a market economy, Slovenia was the first of the former East European countries to adopt the Euro currency (1st January 2007) and the first to assume the presidency of the European Union (1st January 2008). Its progress is also evident from the fact that Slovenia’s per capita GDP has surpassed Portugal and Greece, two countries whose EU membership predates Slovenia. It is also notable that Slovenia appears to have well developed SMA applications (Cadez and Guilding, 2007). 

The remainder of the paper is organized as follows. In the next section, the concepts of strategic choices, SMA and organizational configurations are discussed. Then the research method is described, followed by an outline of the findings. The conclusion section provides an overview of the most salient issues arising from the study.  

2. STRATEGIC CHOICES AND STRATEGIC MANAGEMENT ACCOUNTING 

Strategic choices 

Strategy is generally viewed as “as a pattern of important decisions that (1) guides the organization in its relationship with its environment, (2) affects the internal structure and processes of the organization, and (3) centrally affects the organization’s pefrormance” (Hambrick, 1980, p. 567). Although Hambrick sees strategy as worthy of empirical investigation due to its linkages with many other organizational facets, he feels there is a major problem revolving around the challenge of operationalizing the concept. The focus of this study is on business strategy which refers to how firms compete in an industry or market (Slater in Olson, 2001; Olson et al, 2005).

There has been a convention in management accounting research to explore relationships between management accounting system characteristics and particular strategic types (Ittner and Larcker, 2001; Chenhall, 2003; Luft and Shields, 2003). As strategies are becoming increasingly complicated (Ketchen et al, 1993; Ittner and Larcker, 2001; Olson et al, 2005; DeSarbo et al, 2005), however, this approach appears prone to overly simplistic generalizations. Ittner and Larcker (2001) and Chenhall (2003) thus advocate that more meaningful associations may become apparent when investigating several dimensions of strategy. 

Strategy typologies constitute profiles of different strategic postures that emphasize integrative components of different strategies. These typologies have been widely drawn upon in organisational empirical research (Doty and Glick, 1994; Desarbo et al, 2005). An extensively used typology that was developed by Miles and Snow (1978) is based on four strategic types: prospector, defender, analyzer, and reactor. These types represent holistic configurations of organizational factors. Although the typology’s longevity is generally attributed to its innate parsimony and industry-independent nature (Shortell and Zajac, 1990; Desarbo et al, 2005), recent empirical validations of the typology challenge these assertions (see Desarbo et al, 2005, 2006). One particular ambiguity associated with the typology concerns the status of reactors. Some commentators see the reactor grouping as a residual category, while others see it as a particular archetype that can be effective (Doty et al, 1993). The empirical evidence concerning the effectiveness of reactors is mixed. Smith et al (1989) concluded that organizations classified as reactors were not effective, whereas Snow and Hrebiniak (1980) found that reactors were effective in highly regulated environments. A second point of contention with the typology concerns the relationships among the remaining three types. There is considerable evidence suggesting that the three types define a spectrum, with defenders and prospectors constituting the ends of the specturm and analyzers located between these two extremes (Smith et al, 1989; Shortell and Zajac, 1990; Doty et al, 1993; Anderson and Lanen, 1999). It is also notable that Olson et al (2005) feel that the Miles and Snow’s typology is limited due to its internal focus and propose a hybrid model that represents a synthesis with Porter’s (1980) low cost vs differentiation typology. This composite typology distinguishes between low cost defenders and differentiated defenders and has been supported in empirical analysis (Slater and Olson, 2001; Olson et al, 2005). It should also be noted that typologies can suffer from compromised explanatory or predictive power. This can be because they represent a theorist’s attempt to make sense out of non-quanitified observations, thus they may often not accurately reflect reality (Hambrick, 1984; Desarbo et al, 2005).     

In stark contrast to the considerable interest shown in strategic typologies, the distinction between intended and realized strategy has received negligible attention from accounting researchers. Langfield-Smith (1997) sees this to be a major weakness in management accounting research. When responding to surveys, it is unclear whether managers have reported their intended or realized strategies. Most definitions of strategy, especially in normative textbooks, imply that strategy is an outcome of a deliberate stream of decisions. Mintzberg (1987a; 1987b) counters this view by stressing the amibiguous and evolutionary nature of strategy in many organisations. He sees strategy more as a pattern or stream of actions, regardless of whether these actions are intended. In some organisations, patterns of behaviour may emerge in the absence of intentions, or in spite of them. In practice, pure deliberate and pure emergent strategies are rarely deployed, as most firms fall somewhere between the two extremes (Mintzberg et al, 1995). 

It is also notable that the quantum of attention given to market orientation by strategy and marketing researchers (Hult et al, 2005) has not been matched by management accounting researchers (Guilding and McManus, 2002). Narver and Slater (1990) see market orientation lieing at the heart of much modern management and strategy. The market orientation philosophy concerns the view that the goal of all company activities concerns satisfying customer needs (Jaworski and Kohli, 1993; Walker et al, 1998; Hult et al, 2005). In the modern business environment characterized by fast evolving customer demands, market orientation represents a prerequisite to the pursuit of competitive advantage (Perrera et al, 1997), effective creation of superior value for customers and enhanced performance (Slater and Narver, 1994; Henri, 2006). 

The examination of degree of strategy deliberation and market orientation in this study was also motivated by an expectation that both these factors carry a particular pertinence in a transitional economy context. Bogel and Hustzty (1999) argue that in response to unleashed market liberalization and privatization forces, managers in transition economies only start to think strategically once market transition commences. Further, the socialist system is characterized by a production orientation, signifying that it is only in recent years that Slovenian companies have developed degrees of market orientation (Cadez and Guilding, in press).1 These observations suggest a high degree of variability can be expected with respect to strategy deliberation and market orientation in Slovenian companies.

Strategic management accounting

While interest in SMA is growing (see Tomkins and Carr, 1996; Hoque, 2001; Roslender and Hart, 2003; Bhimani and Langfield-Smith 2007; Cadez and Guilding, 2007; Lanfield-Smith, 2008), there is still limited consensus with respect to what constitutes SMA. One of the more comperehensive empirical investigations of the field was conducted by Cadez and Guilding (in press). Cadez and Guilding focused on two distinct, yet complementary, dimensions of SMA. These are: (1) the adoption of strategically-oriented management accounting techniques and (2) accountant’s participation in strategic management processes. In their study, Cadez and Guilding supplemented Guilding et al’s (2000) distillation of 12 SMA techniques with four                                                          
 1 A CFO in a large Slovenian furniture company provided a very insightful comment on this issue. “In the past socialist times, all companies including ours were production oriented. First we manufactured, then we worried about seeling what we have manufactured. It is only recently that we have started emphasizing market orientation. Now we first turn to customer needs and then adjust our offerings accordingly to secure profitabilty. This is a sign of company maturity” (Cadez and Guilding, in press).  Additional techniques concerned primarily with customer accounting. These techniques are seen to manifest two orientations: environmental (outward-looking) and/or longterm (forward-looking). Cadez and Guilding subsequently classified these 16 SMA techniques according to the five categories outlined in Table 1.

Table 1: Classification of SMA techniques  

SMA categories SMA techniques Strategic costing 1. Attribute costing 2. Life-cycle costing 3. Quality costing 4. Target costing 5. Value chain costing  Strategic planning, control and performance management 1. Benchmarking 2. Integrated performance measurement Strategic decision making 1. Strategic cost management 2. Strategic pricing 3. Brand valuation  Competitor accounting 1. Competitor cost assessment 2. Competitive position monitoring 3. Competitor performance appraisal Customer accounting 1. Customer profitability analysis 2. Lifetime customer profitability analysis 3. Valuation of customers as assets

The second SMA dimension, accountant participation in strategic management processes, follows a more sociological orientation. In contemporary competitive settings, organisations are increasingly concentrating on factors that provide value to customers (Perrera et al, 1997; Slater and Narver, 2000; Henri, 2006). This customerfocus is triggering a flattening of organizational structures. The term “horizontal accounting” has evolved to reflect practices applied in companies that integrate activities across the value-chain to support a heightened customer-focussed strategy (Chenhall, 2008). In “horizontal organizations” decisions are made by cross-functional management teams, including management accountants (Scott and Tiessen, 1999; Baines and Langfield-Smith, 2003; Naranjo-Gil and Hartmann, 2007; Rowe et al, 2008). This development represents a significant change in the underlying accounting paradigm. Strategic management accountants are no longer seen as just information providers, but as active players in the strategic management process. Cadez and Guilding’s (in press) reference to the “strategic accountant” reflects the notion that modern accountants are furnishing individuals with power to achieve their own ends (Chenhall, 2003). 

Strategy and strategic management accounting configurations

Organizational configurations are sets of organizations that share a common profile with respect to key characterictics such as strategy, structure and decision processes (Ketchen et al, 1993; Ferguson and Ketchen, 1999; Moores and Yuen, 2001). The configurational approach suggests that insights into organization behaviour can be achieved by viewing organizations as clustered around particular characteristics, rather than modular entities to be viewed in isolation (Fiss, 2007). In configurational research, the focus is typically on the link between organizational configurations and performance (Ketchen et al, 1997; Ferguson and Ketchen, 1999). 

Configurational theory differs from universalistic and contingency theories because it is guided by a more holistic principle of inquiry and adopts the systems assumption of equifinality (Delery and Doty, 1996). In general, the theory is concerned with how a pattern of multiple independent variables (e.g., strategy and SMA) is related to a dependant variable (e.g., performance), rather than how individual independent variables are related to a dependant variable. 

The central assumption embodied in configurational theory is equifinality. Equifinality occurs when a sample of organizations using different strategic and structural alternatives achieve equivalent levels of performance (Payne, 2006). Gresov and Drazin (1997) suggest three forms of equifinality: suboptimal, trade-off, and configurational. A suboptimal equifinality situation arises when an organization attempts to satisfy multiple and conflicting functional demands (e.g. innovation in product design versus operating efficiency) with a limited repertoire of structural options. A trade-off equifinality situation is characterized by a single or dominant functional demand, whereas structural choice is not limited. A configurational equifinality situation is characterized by multiple and conflicting functional demands, whereby structural options avaliable to organizations are relatively unconstrained. As a consequence, there are simultaneous trade-offs between both strategies and structures that can result in the evolution of a number of effective strategic configurations (Marlin et al, 2007). In this study, a configurational form of equifinality is assumed. This is based on the expecation that organizational performance is positively affected by the selection of strategic choices that minimize functional conflict and a structural design that fits the chosen strategy. Designs that fit the chosen set of strategic choices will be equifinal relative to each other and will outperform those that do not (Gresov and Drazin, 1997). 

Organizational goal achievement is faciltated when an SMA system manifests both horizontal and vertical fit. Horizontal fit refers to the internal consistency of the organization’s SMA practices, while vertical fit refers to the congruence of the SMA system with firm strategy (Delery and Doty, 1996). The study described herein enables a consideration to be made of the extent to which SMA systems in different organisatonal configurations exhibit horizontal and vertical fit and also the degree to which well-aligned SMA systems are reflected by heightened performance. 

3. RESEARCH METHOD

A range of approaches can be taken by the researcher to capture the complexity of configurations. These can be classified according to two main categories: theoretical (deductive) and empirical (inductive). While both theoretical and empirical approaches embody strengths and weaknesses (see Ketchen et al, 1993; Doty and Glick, 1994; Bensaou and Ventkamaran, 1995; for a discussion), in this study an empirical approach has been employed. In the management literature, empirically generated classification systems are ususally regarded as taxonomies (Sanchez, 1993; Doty and Glick, 1994; Payne, 2006). 


The data analysis undertaken can be seen as comprising three steps. Firstly, an appraisal is made of what organizational configurations exist in the chosen sample. This will be achieved by way of cluster analysis, an acknowledged technique for discerning cases that exhibit similar charactersitics (Ferligoj, 1989; Ketchen and Shook, 1996; Fiss, 2007). Secondly, the organizational configurations that are effective are identified. This will be achieved by investigating the performance of the identified configurations. Thirdly, an appraisal is made of the degree to which the identified configurations are internally consistent and congruent with strategy. 

Sampling procedure

Data were collected using a mailed questionnaire survey. An initial sample was drawn from the Slovenian Chamber of Commerce and Trade disclosure of the 500 largest Slovenian companies (in terms of total revenue) which includes all industrial sectors. A second size filter was imposed to screen out companies with less than 100 employees. The sample was further reduced due to incorrect or incomplete mailing addresses for some cases. The final sample comprised 388 companies.  

As part of a strategy to develop an accurate mailing list and secure a high response rate, a phone call was lodged with each company and the name of the most suitable person to complete the survey was identified. These were typically the Chief Accountant, Chief Controller, or Chief Financial Officer. In most cases, the particular manager was spoken to and the purpose of the research explained. The mailed survey package included a covering letter explaining the purpose of the research, a copy of the survey with a glossary of terms used and a postage-paid reply envelope. The first mailing resulted in 124 usable responses and the second mailing one month later yielded an additional 69 responses. Thus the overall usable response rate was 49.7%. 

To investigate for possible non-response bias, Kolmogorov-Smirnoff tests of differences in the responses provided by early and late respondents (the first and last 25% of questionnaires returned) were conducted. No significant differences (p < 0.05) in the data provided by these sub-groups were noted for any questions posed. While this suggests little concern for non-response bias, it should be acknowledged that accountants in firms with relatively sophisticated accounting systems may have been more inclined to respond than those in firms with under-developed accounting systems. 

The industrial sectors represented in the sample analysed are presented in Table 2. Companies comprising the sample had an average annual sales level of €92.7 million. The smallest company’s annual sales level was €13.5 million and the largest company’s annual sales level was €1.2 billion. In terms of employees, the companies analysed had an average number of 747 employees, with the smallest company employing 104 and the largest company employing 8,765 staff. 


Table 2: Industry classification of the sampled companies

Industry Number of firms Percentage  of sample A. Agriculture  1 0.5 B. Mining 2 1.0 C. Manufacturing C1. food, beverages and tobacco C2. textile, apparel, leather, wood and furniture C3. chemicals, plastics, non-metallic products C4. metal products C5. machinery, electric, electronics and automotive 108 17 30 19 14 28 56.0 8,8 15,5 9,8 7,3 14,5 D. Public services and utilities 10 5.2 E. Construction 9 4.7 F. Wholesale and retail trade 30 15.5 G. Accommodation, food and leisure services 8 4.1 H. Transportation and logistics services 13 6.7 I. Financial intermediation and IT services   12    6.2 Total 193 100.0

Variable measurement

Prospector/defender business strategy The measure developed by Shortell and Zajac (1990) was used. This instrument assesses an organization’s overall strategic orientation on a seven-point scale, anchored at one end by a description of a defender-type organisation, and at the other end by a prospector-type organization.

Deliberate vs emergent strategy  Because this dimension of strategy has not been operationalized in any known previous work, an original measurement instrument had to be developed. In order to assess where an organisation lies on the deliberate vs emergent dimension of strategy, three statements were provided to respondents. Drawing on Mintzberg’s (1987a) terminology, these statements were: (1) “In our company, the strategic decision-makers usually think through everything in advance of strategic action” (2) “In our company, strategic intentions are seldom realized with little or no deviation”, and (3) “In our company, strategic action usually develops in the absence of strategic intention”. Next to each statement, a seven-point scale was provided, ranging from “1” (strongly disagree) to “7” (strongly agree). 

Market orientation Market orientation was measured using the same instrument applied by Guilding and McManus (2002). Using a seven-point scale ranging from “1” (not at all) to “7” (to a large extent) respondents were asked to indicate to what extent they agree with the following statements: “(1) my company has a strong understanding of our customers, (2) the functions in my company work closely together to create superior value for our customers, (3) management in my organization thinks in terms of serving the needs and wants of well-defined markets chosen for their long-term growth and profit potential for the company, and (4) my company has a strong market orientation”. 


SMA usage The degree of SMA technique usage was measured using the same approach as Cravens and Guilding (2001) and Guilding and McManus (2002). Following the question “To what extent does your organization use the following techniques?”, the 16 SMA techniques referred to in Table 1 were listed together with a Likert-type scale ranging from “1” (not at all), to “7” (to a great extent). A glossary containing definitions of the SMA techniques was provided to aid interpretation (see Appendix). 

Accountant’s participation in strategic decision making The measure used draws on Wooldridge and Floyd’s (1990) instrument designed to  assess middle management involvement in strategic decision making. Respondents were asked to record their participation with respect to five aspects of strategic management: (1) identifying problems and proposing objectives, (2) generating options, (3) evaluating options, (4) developing details about options, and (5) taking the necessary actions to put changes into place. The scale anchors ranged from “1” (not at all involved) to “7” (fully involved). 

Performance  Performance was measured using a modified version of the Hoque and James (2000) instrument. Three dimensions of performance have been appraised, namely (1) return on investment, (2) customer satisfaction, and (3) development of new products. For each of these dimensions, respondents were asked to indicate their company’s performance relative to their competitors on a scale ranging from “1” (below average) to “7” (above average). 

Data analysis 

The study addresses configurations of three strategic choices (prospector vs defender spectrum, degree of strategy deliberation and degree of market orientation), and two dimensions of SMA (SMA usage and degree of accountants participation in strategy), a total of five constructs. SMA usage, however, contrary to the other four unidimensional constructs, was specified as a multidimensional construct comprising five dimensions (the five SMA categories identified in Table 1). This signifies that the cluster analysis was based on nine constructs. The constructs were represented as composite items calculated as an average of the original items (see Table 3) to reduce measurement error (Ittner and Larcker, 2001). For example, for the market orientation construct, the composite item was calculated as the mean of four original items. The composite items include all of the variable measurement items referred in the preceding section with two exceptions. Firstly, the quality costing technique was excluded from computation of the usage of strategic costing due to low internal realiability. Similarly, the valuation of customers as assets was excluded from computation of customer accounting usage due to low internal reliability and a non-normal distribution (usage of this technique was found to be very low; a mean score of 1.97).


Table 3: Descriptive statistics for explored constructs 

Strategic choices Mean Standard  deviation Crombach alpha Prospector/defender strategy 4.62 1.42 N/A Degree of strategy deliberation 5.22 1.15 0.73 Market orientation 5.13 1.09 0.87 Accountant’s participation in strategy 4.79 1.41 0.92 SMA usage    SMA usage: strategic costing  3.51 1.40 0.77 SMA usage: planning and control  4.22 1.37 0.75 SMA usage: strategic decision making 4.28 1.59 0.85 SMA usage: competitor accounting 4.20 1.36 0.72 SMA usage: customer accounting 3.36 1.59 0.77 Performance variables    Return of investment 4.23 1.46 N/A Customer satisfaction 4.78 0.88 N/A Development of new products 4.46 1.34 N/A

Prior to cluster analysis, all variables were standardized to facilitate interpretation. In order to derive a manageable number of clusters a two step cluster analysis procedure was applied (Ferligoj, 1989; Ketchen and Shook, 1996). Firstly, Ward’s hierarchical cluster procedure was used. Under this iterative method, each subject starts as its own cluster, then iterations are run until all subjects are grouped into one large cluster (Ketchen and Shook, 1996; Hair et al, 1998). The dendrogram produced in applying this approach, together with the fusion coefficient, indicated a 6 cluster solution as a viable solution. This method suffers, however, from being biased towards the production of clusters with approximately the same number of observations (Ferligoj, 1989; Hair et al, 1998), thus the nonhierachical K-means cluster procedure was also employed. K-means is an iterative partitioning method that begins by dividing observations into a predetermined number of clusters (Slater and Olson, 2001), which, based on the hierachical procedure, was set to 6. Contrary to hierarchical methods, nonhierarchical methods allow multiple passes through the data, thus the final solution optimizes within-cluster homogeneity and between-cluster heterogeneity (Ketchen and Shook, 1996).

Next, to face validate the derived clusters, an appraisal was made to determine whether the members of each cluster correspond to the described configurations (Ketchen et al, 1993). Without validation, one cannot be assured of having derived a meaningful and useful set of clusters (Ketchen and Shook, 1996). Further, to assess the industry sector impact on derived configurations, a crosstabulation analysis was conducted that highlights the relationship between cluster membership and industry affiliation.  

4. FINDINGS

Table 4 presents the companies’ standardized mean variable scores according to the six cluster groupings. The clusters are ranked according to performance, with cluster 1 being the highest performer and cluster 6 the lowest performer.2 The first number in each cell represents a cluster’s standardized mean score for the variable in question, and the number in parenthesis indicates the rank of this mean score relative to the other clusters. The remainder of this section is devoted to a description of the six clusters, an examination of the relationship between industry classification and the clusters, a consideration of the relative performance of the clusters and an appraisal of the clusters’ SMA configurational fit.

Table 4: Cluster based cross-tabulation of mean standardised variable scores 

Cluster 1 2 3 4 5 6 Cluster name Stars Analytics First movers Protectors Laggards Socialism relics Number in cluster 29 42 20 38 49 15 Strategic choices       Strategy prospector/defender 0.90 (2) 0.40 (3) 0.96 (1) -0.45 (5) - 0.35 (4) -1.85 (6) Degree of strategy deliberation 0.54 (2) 0.01 (4) 0.58 (1) 0.17 (3) -0.41 (5) -0.91 (6) Market orientation 0.83 (1) 0.22 (3) 0.64 (2) 0.05 (4) -0.40 (5) -1.89 (6) Accountant’s participation in strategy 0.72 (1) 0.07 (4) 0.24 (3) 0.49 (2) -0.60 (5) -1.20 (6) SMA usage  SMA usage: strategic costing  1.01 (1) 0.00 (3) -0.63 (5) 0.68 (2) -0.59 (4) -0.93 (6) SMA usage: planning and control  0.85 (2) -0.16 (4) 0.89 (1) 0.55 (3) -0.72 (5) -1.41 (6) SMA usage: strategic decision making 0.94 (1) -0.34 (4) 0.31 (3) 0.76 (2) -0.60 (5) -1.23 (6) SMA usage: competitor accounting 0.74 (2) -0.08 (4) 1.16 (1) -0.04 (3) -0.55 (5) -0.85 (6) SMA usage: customer accounting 0,90 (1) 0.64 (2) 0.05 (3) -0.05 (4) -0.77 (5) -0.96 (6) Performance variables a Return of investment 0.33 (2) 0.36 (1) -0.02 (4) 0.02 (3) -0.23 (5) -0.93 (6) Customer satisfaction 0.57 (1) 0.20 (3) 0.54 (2) -0.02 (4) -0.35 (5) -1.19 (6) Development of new products 0.68 (2) 0.25 (3) 0.80 (1) -0.22 (4) -0.43 (5) -1.09 (6) a: Not used in determining clusters. Note: F-tests for all clustering variables were statistically significant at 0.01 probability level. Significant F-tests indicate that statistical differences exist for individual variables across clusters (Hair et al, 1998).

Cluster descriptions  

Following Slater and Olson’s (2001) recommendation, the quantitative findings have been synthesised into qualitative gestalts by labeling and describing the derived clusters. We have labelled cluster 1 “stars”. These businesses are prospector oriented with a deliberate approach to strategy formulation and a high level of market orientation. Accountant’s participation in strategy is highest of all 6 clusters and SMA
                                                         
 2 Allocating performance rankings to the six clusters has required the exercise of some subjectivity. The approach taken has been to aggregate the rankings of the three performance indicators appraised. This approach resulted in a tie between “analytics” and “first movers”. The decision was taken to rank the performance of first movers behind analytics because of the high ROI score recorded by the analytics and  the low ROI score recorded by the first movers. ROI is perceived by many as a fundamental performance measure that is contributed to by other aspects of performance such as customer satisfaction and development of new products (Kaplan and Norton, 1996).  Usage is high (it ranks highest for usage of three categories of SMA and second highest for two categories of SMA). The cluster comprises 29 companies. The majority of these can be referred to as Slovenian blue-chips thus the term stars appears to be a valid descriptor of the group. 

Cluster 2 has been labelled “analytics”. Organisations in this cluster score near the middle of the sample (ranking 3 or 4) for most of the variables appraised. The only SMA practice that analytics use relatively highly is customer accounting (rank 2). The group is relatively large with 42 members. As most of the businesses within this grouping do not demonstrate strong explicit strategic priorities, it seems that the label analytics is a valid descriptor for the group. 

Cluster 3 is comprised of “first movers”. These businesses are very similar to stars with respect to strategic choices. They rank highest on pursuit of a prospector strategy and deliberate strategy formulation and high on market orientation. Differences are apparent, however, with respect to SMA attributes. Accountant’s participation in strategy is relatively modest (rank 3) and high variability is observed for SMA usage. The high prospector orientation plays out with high competitor accounting usage, and the first movers also rank highest for planning and control SMA usage. The first movers do not score highly for the remaining SMA techniques, in particular strategic costing (rank 5). The number of companies in this cluster is 20 and it appears that the label first movers is a valid descriptor of the group as most of them demonstrate a high propensity for seeking new product/market opportunities. 

Cluster 4 companies have been labelled “protectors”. These businesses have a defender strategy orientation and rank near the middle of the sample with respect to deliberate strategy and market orientation. Accountant’s participation in strategy is relatively high (rank 2) and so is the usage of two SMA categories (strategic costing and strategicdecision making). Competitor and customer accounting are not widely used in this group. The protectors group comprises 38 companies. An examination of these companies reveals that most are predominantly focused on the domestic market. 

The remaining two clusters comprise companies with similar characteristics. We label cluster 5 “laggards”. These businesses are defender type oriented, strategy is relatively undeliberate (emergent) and the level of market orientation is low. Further, accountants’ participation in strategy is low and there is low application of SMA techniques. It appears this group resembles the reactors group proposed by Miles and Snow (1978). It is the largest grouping comprising 49 of the sampled companies. 

In cluster 6, the characteristics of laggards are taken one step further. Companies in this cluster are strongly defender oriented, strategy is undeliberate (emergent) and market orientation is extremely low. This grouping also has the lowest level of accountant participation in strategy and the lowest levels of SMA usage. This group is relatively small, comprising 15 companies which are mostly government-owned and many come from the energy sector. We label this cluster “socialism relics”. 

An investigation for a relationship between industry affiliation and cluster membership has been conducted. The sample represents nine primary industry sectors (Table 2). In

12
light of the small representation of agriculture and mining, companies in these sectors have been assigned to other industry groups. The single representative of the agricultural sector is essentially a poultry producer and was reassigned to group C1. As the two mining companies are government-owned coal mines which sell all their output to government-owned coal power plants, they were reassigned to industry group D. Further, since the manufacturing group is relatively diverse (Table 2 highlights 5 subgroups), this group has been consolidated into two subgroups. The first group, comprising C1 and C2, has been labelled ‘CI’ and can be denoted as ‘traditional manufacturing’. The second group, labeled ‘CII’, comprises groups C3-C5 and can be denoted ‘contemporary manufacturing’. A crosstabulation analysis comprising the refined 8 main industry groups and 6 strategic configurations is presented as Table 5. 

Table 5: Group membership and industry sector crosstabulation

Cluster name Stars Analytics First movers
Protectors Laggards Socialism relics Number in cluster 29 42 20 38 49 15 Industry sector (number in sector)       CI. Traditional manufacturing (48) 4/7.2 11/10.4 2/5.0 13/9.5 15/12.2 3/3.7 CII. Contemporary manufacturing (61) 13/9.2 17/13.3 5/6.3 12/12.0 12/15.5 2/4.7 D. Public services and utilities (12) 0/1.8 0/2.6 0/1.2 3/2.4 3/3.0 6/0.9 E. Construction (9) 0/1.4 2/2.0 0/0.9 2/1.8 4/2.3 1/0.7 F. Wholesale and retail (30)  7/4.5 5/6.5 5/3.1 3/5.9 9/7.6 1/2.3 G. Accommodation, food, leisure services (8) 0/1.2 3/1.7 2/0.8 1/1.6 2/2.0 0/0.6 H. Transportation and logistics services (13) 3/2.0 4/2.8 1/1.3 2/2.6 2/3.3 1/1.0 I. Financial intermediation and IT services (12) 2/1.8 0/2.6 5/1.2 2/2.4 2/3.0 1/0.9 Legend: the first digit in each cell represents observed frequencies/the second digit in each cell represents expected frequencies. Cells where observed frequencies are 0 are underlined. Cells with  oberved/expected frequencies ratio above 2 or below 0.5 are highlighted in bold.   Note: Chi square test indicates that observed frequencies are significantly different to expected frequencies, however this should be interpreted with care due to very small expected counts in most cells.

Table 5 suggests a relationship between the industry sectors and cluster affiliation. Viewed from the industry perspective, it is notable that the public services and utilities sector does not feature any stars, analytics or first movers configurations, rather 50% of the companies in this sector manifest a socialism relics configuration. A similar finding holds for construction sector which also has no stars or first movers configurations, but nearly half feature in the laggard cluster. Viewed from the cluster membership perspective, it is notable that 40% of the socialism relics group are public services and utilities. Also, the stars group includes no representation of three industry sectors (public services and utilities, construction, and accomodation, food and leisure services). Aside from these observations, the industry sectors are relatively spread across the six configurations.  


Cluster Performance 

The performance of the six clusters was assessed on three dimensions: return on investment, customer satisfaction and new product development. As is evident from Table 4, performance of the clusters varies substantially along these dimensions. Overall, the most successful cluster is the stars, where performance on the three dimensions is either the highest or second highest of the six clusters. In terms of financial return on investment the analytics cluster is most successful, in terms of new product development the first movers cluster is most successful. These observations carry an intuitive consistency. The first movers’ high performance with respect to new product development and customer satisfaction appears consistent with their high prospector orientation. Unlike first movers, the analytics group attaches high importance to financial performance. The stars and analytics can be viewed as relatively successful, scoring above the mean on all three performance dimesnions appraised. The first movers are viewed as less successful due to their below average ROI.  

Next, the protectors group is close to the overall sample average in terms of financial performance and customer satisfaction. Only with respect to new product development does it score markedly below the average which appears consistent with the defender strategic orientation of this group. 

The remaining two groups are less successful and score relatively lowly across all of the three performance dimensions appraised. This is particularly true for the socialism relics group, where scores rank the lowest for every performance dimension appraised.  

Appraisal of SMA configurational fit

Configurational theory suggests that superior performance is attributed to internal consistency in the pattern of relevant strategic and structural factors, whereby both horizontal (internal consistency of SMA practices) and vertical fit (congurence of the SMA system with strategy) are important. We now turn to assess the SMA internal consistency and congruence for each configuration. 

The stars group features high SMA adoption rates and high accountant participation in strategy. Given that all SMA techniques exhibit a relatively ‘strategic’ orientation compared to conventional accounting techniques that tend to have a historical, shortterm and inward focus (Guilding et al, 2000), an equivalent degree of usage across the SMA techniques appraised would appear to signify internal consistency. Further, strong accountant’s involvement in strategy process also appears consistent with high SMA usage. Greater involvement in strategy can be expected to inculcate accountants with an appreciation of information needs posed by strategic management, thus it is likely to result in accoutants instigating novel techniques (Abernethy and Bouwens, 2005), such as SMA. It therefore appears that the stars group manifest high SMA horizontal fit. With respect to vertical fit, a deliberate prospector market-oriented configuration implies broad inter-functional discussion (Perrera et al, 1997; Chenhall, 2008), thus greater accountant participation in strategy decision making appears warranted. Also, given the outward and forward focus of SMA techniques, it appears that high SMA usage levels are congruent with a dynamic prospector type strategy (Guilding, 1999), deliberate prospecting decisions (Simons, 1987; Davila, 2000), and high market orientation (Guilding and McManus, 2002; Hult et al, 2005). Overall, the observations for stars suggest a high degree of consistency on both the horizontal and vertical dimensions. This view is further supported by the high performance achieved by the stars.  

The analytics group demonstrates a fairly consistent (mid-range) level of SMA usage across the techniques appraised, although there is fairly high use of customer accounting. Accountant’s participation in strategy is moderate for this grouping. These observations signify fairly high SMA horizontal fit. This cluster also falls into the moderate range with respect to the prospector, deliberate strategy and market orientation measures. It would appear the analytics exhibit characteristics resembling the analyzers strategic archetype described by Miles and Snow, ie, characteristics of both prospectors and defenders, seeking effectiveness through both efficiency and a quest for new products/markets (Doty et al, 1993; Olson et al, 2005). Moderate levels of SMA adoption and participation appear to be congruent with the selected strategic choices, hence implying fairly high vertical fit which is manifested by a high ROI performance.  

Of the six clusters, the first movers exhibit the highest degree of variability with respect to usage of the 5 SMA dimensions. While competitor accounting and planning and control forms of SMA are most extensively used by this group, this is not the case for the remaining three dimensions. In particular, strategic costing stands out with a low level of usage. The level of accountant participation in strategy is in the mid-range to high level. With respect to strategic choices, the first movers exhibit a similar configurational pattern to the stars, ie., scoring highly on the prospector, deliberate approach to strategy and market orientation measures. This signifies relatively low alignment between strategy and SMA. A case could be argued, however, that the high use of competitor accounting and planning and control appears consistent with the first movers high prospecting orientation (Shortell and Zajac, 1990; Doty et al, 1993), and that their low use of strategic costing may relate to the type of muted concern for cost efficiency that one might expect in a highly prospector oriented enterprise (Miles and Snow, 1978; Doty et al, 1993; Langfield-Smith, 1997). Considered collectively, the moderate level of accounting participation in strategic management, low use of strategic costing and low return on investment performance may signfiy a limited accounting vein running through first movers, expecially when compared to the stars grouping that has a similar strategic posture. Despite this conjecture, we conclude that the first movers variable level of SMA adoption signfies a low level of SMA horizontal and vertical fit. This low level of fit has not been strongly manifested by low performance measures that range from moderate to high.   

Protectors also exhibit some variability with respect to SMA adoption rates. While strategic costing, planning and control and strategic-decision making SMA techniques are relatively widely used, this is not the case for competitor and customer accounting dimensions. This grouping has the second highest level of accountant participation in strategy. The protectors are defender oriented with a moderate level of strategy deliberation and market orientation. The variability in levels of SMA techniques usage suggest that this configuration comprises mainly low cost defenders. Such a conclusion was drawn from the fact that strategic costing is widely used in this group whereas competitor accounting as an essential attribute of differentiation is not widely used (Guilding, 1999; Olson et al, 2005). Following the arguments presented in the description of the stars group relatively high SMA adoption and participation in strategy are not consistent with the defender strategy, a view supported also by Cadez and Guilding (in press). With respect to accounting participation, Porter (1996) argues that defenders are primarily concerned with operational effectiveness where efficencies tend to be sought with an intra-departmental philosophy, contrary to prospectors who are more concerned with strategic positioning which requires broad inter-functional discussion (Nyamori et al, 2001). Further, there is evidence that broad scope information systems such as SMA are of limited value to defender type companies (Abernethy and Guthrie, 1994; Guilding, 1999) as their focus is primarily on cost efficiency. Overall, it appears that the relationships in the protectors configuration are somewhat inconsistent both along the horizontal and vertical dimension.   

Laggards group features low SMA adoption rates and also accountant’s participation in strategy. These relations appear consistent, signfiying high horizontal fit. With respect to strategic choices, laggards are defender type oriented, strategy is relatively undeliberate (emergent) and the level of market orientation is relatively low. Building on the arguments already outlined in this section, the laggards SMA configuration appears to be consistent with its strategy, signifying high vertical fit. 

Finally, the socialism relics can be interpreted as a more extreme case to the laggards. Hence, the configuration of underdeveloped SMA system and a defender type emergent non market-oriented strategy is again consistent. It is notable that the socialism relics have the lowest level of performance.  

Table 6: Summary of the 6 clusters’ configurational consistency and performance   

Configuration Horizontal fit degree Vertical fit degree Performance level Stars High  High  Highest Analytics Fairly high Fairly high High First movers Lowest  Low Fairly high Protectors Low Low Moderate Laggards High High Low Socialism relics Highest  Highest Lowest

Table 6 provides a summary of the observations made in this section.  It is evident from this table that the observed configurations constitute varying degrees of fit. The most effective configuration stars is also highly consistent. It comprises prospector type businesses who form strategy in a deliberate manner, are highly market oriented and exhibit developed SMA systems. Also of note, the least effective configurations socialism relics and laggards have highly consistent SMA configurations. These observations are interpreted in the next section. 


5. CONCLUSION AND DISCUSSION

Following a call to empirically examine the way in which elements of management accounting combine with a variety of strategic choices to enhance performance (Chenhall and Langfield-Smith, 1998), this study is believed to be the first to attempt an empirically based configurational analysis of strategic management accounting. It has achieved this by drawing from configuration based analyses found in the management literature (Doty et al, 1993; Ketchen et al, 1993; Bensaou and Venkatraman, 1995; Delery and Doty, 1996; Gresov and Drazin, 1997; Ketchen et al, 1997; Ferguson and Ketchen, 1999; Payne, 2006; Fiss, 2007; Marlin et al, 2007). The study provides several contributions. First, it provides insights into business configurations with respect to strategic choices and SMA system designs in a successful transition economy. Second, it provides some support for the view that organizational configurations that exhibit internally consistent SMA alignments and SMA systems that are well aligned to strategic choices will demonstrate higher performance. It also provides some support for the view that similar levels of performance can be achieved using different strategic and structural alternatives. 

Six configurations were derived using cluster analysis. The first four clusters appear to provide some support for the central configurational proposition that high vertical and horizontal configurational alignments are associated with high performance. Of the four clusters, the stars, has the highest degrees of alignment and the highest levels of performance. The second cluster, the analytics, has fairly high degrees of alignment and the second highest levels of performance. The third and fourth clusters, the first movers and protectors, both have lower degrees of alignment and lower levels of performance. 

The final two clusters counter the central configurational proposition, however. This is because they (laggards and socialism relics) exhibit high degrees of SMA horizontal and vertical fit, yet they are the two lowest performing groups. This contradictory observation beckons further scrutiny. While Miles and Snow (1978) assert that all their proposed strategic types can be effective, this is only true if they are well alligned with the environment confronted (Desarbo et al, 2005) and if the business functions are supportive of the grand strategies selected (Slater and Olson, 2001). Olson et al (2005) provide evidence that a base level of marketing is required in every company to be effective, whereas Desarbo et al (2006) find that for defenders marketing and marketlinking capabilities are most significantly related to profit. This consideration is particularly pertinent given the particular context of the country surveyed in this study. Less than two decades ago, all Slovenian companies were production oriented and strategic thinking was minimal. It appears plausible that the laggards and socialism relics have yet to evolve strategies that are appropriately aligned with their radically changed environment. Ketchen et al (1993) and Ferguson and Ketchen (1999) note that this type of misalignment is not uncommon, as evolution from a mis-aligned poorperforming configuration can require strategic and structural change that is time consuming and expensive, thus disparities in fit are bound to persist. 

This is a very important consideration in this study, as we have a somewhat idiosyncratic country context under examination. The dramatic changes experienced in the workings of the Slovenian economy can be expected to have resulted in a wide variation in the extent of adaptation achieved across the country’s businesses during the same time period. As a result, even though the socialism relics and laggards exhibit consistent SMA systems that are congruent with their strategies, it may well be that their strategic choices are not compatible with their business context, an over-riding factor contributing to low performance. Stated alternatively, underdeveloped SMA systems seem consistent with the strategic inertia of these two groups, however this particular consistency is not a sufficiently strong factor to counter the implications of the pursuit of inapporpriate strategies. 

Some limited support is also provided for the validity of the equifinality concept. The cluster analysis yielded two clusters (i.e. stars and first movers) that are very similar in terms of selected strategic choices and in terms of non-financial performance, yet quite different with respect to SMA system design. A similar observation occurs when comparing stars and analytics. While these configurations yield very similar financial performance levels, both the selected strategic choices and SMA system designs vary considerably. This interpretation has to be qualified, however, as the stars configuration is successful both along financial and non-financial dimensions of performance, while the first movers and analytics are successful on either the financial or non-financial dimensions of performance, but not both. 

While configurational theory has received considerable attention in management and organizational literature, this has not been mirrored in the management accounting field as very few studies employing this approach exist (e.g, Chenhall and Langfield-Smith, 1998; Moores and Yuen, 2001). Most of the empirical work in the area follows a contingency approach (Langfield-Smith, 1997; Chenhall, 2003), despite the fact that it has been criticized for its piecemental and often contradictory nature over an extended period of time (Fisher, 1995; Gresov and Drazin, 1997; Chenhall and Langfield-Smith, 1998; Chenhall, 2003; Gerdin and Greve, 2004; in press). Fiss (2007) feels there is a clear need to move beyond simple contingency approaches. 

We believe that the reasons for the continuing dominance of the contingency approach, relative to the configurational approach, in management accounting research are many and often of a practical nature. Firstly, the simultaneous investigation of a variety of variables that is involved in the configurational approach inevitably leads to the problem of conflicting contingencies (Fisher, 1995; Gerdin, 2005). This makes interpretation and theory building difficult. Secondly, testing multiple fits simultaneously precludes the use of rigorous statistical methods (Gerdin and Greve, 2004; Fiss, 2007). Thirdly, Ittner and Larcker (2001) contend that managerial accountants exhibit an innate interest in providing insights to which management accounting practices impact positively on organizational performance. This quest is consistent with contingency theory. Following Fiss’ (2007) pronouncement of a need to move beyond contingency approach, this study is one of the first management accounting investigations that attempts to address these weaknesses. 

As in any other study, a number of limitations exist that need to be borne in mind when intrepreting the study’s findings. Many limitations are associated with the cluster analytical statistical technique. Perhaps the major weakness relates to the reliance placed on the researcher’s judgement. For example, determining the number of clusters


requires the exercise of subjectivity, although an attempt was made to mitigate this weakness by deploying a recommended two step clustering approach and a cluster validation procedure (Ketchen and Shook, 1996). Further, cluster analysis does not provide statistical rigour, nor a clear notion of fit. Another problem is the potential for multicollinearity between clustering variables. In this study, the correlations were not excessively high (the highest recorded correlation was 0.51), hence the decision was taken not to correct for multicollinearity due to the many pitfalls associated with the correctional methods (Ketchen and Shook, 1996). A further shortcoming of the study relates to the use of a single item measure of business strategy. Single item measurement is likely to result in measurement error (Ittner and Larcker, 2001). Another limitation arises from the fact that the study did not examine the moderating effects of other environmental and organizational variables, except for company size. 

Despite these limitations, it is believed the study provides some insight into the way that strategic choices and SMA practices can combine to affect performance. It also represents an important demonstration of how a holistic configurational approach may be applied in management accounting research.