Abstract
In the recent years, there has been a considerable interest in using the innovative cost or management accounting systems worldwide that can provide relevant information to companies operating in a changing and competitive business environment. Activity-based costing (ABC) is the most important cost accounting innovation which was developed for overcoming the product cost distortion due to the use of traditional costing system.
This paper attempts to investigate the relationships between some technological and environmental factors and implementation of ABC. Data were collected through a survey questionnaire sent to Chief Financial Officer (CFO) of manufacturing companies selected from Tehran Stock Exchange (TSE). A binary logit regression analysis was used to investigate the relationships between technological and environmental factors and the implementation stages of ABC. Strong support was found for the hypothesis that the effect of the contextual factors (level of information technology quality, level of product diversity, level of overheads, perceived environmental uncertainty, and level of competition) change in ABC implementation stages. Moreover, from Miles and Snow's typology of strategy, results revealed that analyzers are more likely to be in the higher ABC implementation stages.
Key words: Activity-based costing; Implementation; Adoption, infusion
1. Introduction
Cost accounting measures the cost of products or services and its information is used for management accounting purposes in the financial reporting process as well as in decision making processes such as sell or buy decisions, transfer pricing, value inventory, cost control, and performance determination. Traditional cost accounting systems (TCA) were designed for manufacturing environments in which direct costs were a larger percentage of total costs. This system allocates overhead costs to products, using on volume-based measurement relating direct costs. Although its overhead cost allocation was not quite accurate, this system worked well, since direct labour and material, represented the majority of the total cost while overhead cost was only a small fraction. But for the past three decades, many companies have experienced significant changes in their cost structure. Overhead cost becomes the dominant cost component of many products. Many researchers highlight the change of cost structure for example Baker (1994) asserted that today direct labour cost is often less than 15% of the total cost and overhead cost may present 50% or more of total cost. During the same time period, much criticism was raised that traditional cost accounting has failed to prepare advance, related, timely, and extremely accurate information for improving management decisions. To overcome the weakness of traditional cost accounting, several management accounting innovations were introduced during the 1980s and 1990s.
Activity Based costing (ABC) is a new type of cost accounting system and it is one of the main cost and management accounting innovation systems in the twentieth century (Shields, 1995; Anderson, 1995; Gosselin, 1997; Askarany et al., 2007). Bjørnenak (1997) believed that one of the most important contemporary accounting innovations is activity-based costing. Gosselin (1997) claimed that ABC is one of the most important cost accounting innovations of the twentieth century.
The main different between ABC and TCA is related to the way of assigning overhead cost to product. ABC provides more detailed tracking and differential assignment of overhead costs, creates more costs pools, and provides more accurate product costs. Another difference is in the measurement of the component of cost. TCA measures the cost based on three components including: direct material, direct labour, and overhead but in the ABC system, the cost of products includes activities.
Krumwiede (1998) defined activity-based costing (ABC) as a costing methodology that allocates costs to individual activities based on more than one cost allocated base. The current study attempts to predict and test the influence of certain contextual factors on the implementation stages of activity-based costing (ABC) among Iranian manufacturing firms.
Activity based costing in this study is referred to as an innovation and ABC is considered as an administrative innovation (as compared to a technical innovation). However, the current study follows the theory that is used by most of the ABC adoption research, which is named the theory of organizational adoption of innovation or innovation diffusion theory. This theory is a widely accepted theoretical basis for studying ABC implementation (Anderson, 1995; Innes & Mitchell, 1995; Gosselin, 1997; Innes et al., 2000; Brown et al., 2004; Pierce, 2004; Al-Omir & Drury, 2007).
Since the emergence of activity-based costing (ABC), it has received a great deal of attention as a cost management innovation which provides more accurate product costs information than the traditional cost system (e.g. Kiani and Sangeladjiai 2003; Krumwiede, 1998). However, the existing literature shows that despite the claimed benefits of using activity-based costing; the level of implementation of this system is still lower than the traditional one. Gosselin (1997) describes this fact as the "ABC paradox". He asserted that it seems a gap exists between the great interest of management accountants in using ABC and the number of organizations that have actually implemented it. In Iran, for example, Tabrizi (1999) surveyed the CFOs of 290 manufacturing firms and found that a few Iranian manufacturers adopted ABC to calculate product costs.
Hence, the main objective of this study is to examine how certain contextual factors include information technology, product diversity, overhead level, perceived environmental uncertainty, competition, and business strategy influence the implementation stages of activity-based costing (ABC) among the Iranian manufacturing firms.
2. Literature Review
2.1 ABC implementation process
Activity-based costing is a new type of cost accounting system and it is an innovation in a firm's cost information system (Shields 1995; Anderson 1995). Askarany et al. (2007) identified an innovation as an idea, practice, or object that is perceived as new by an individual or other unit of adoption. Cooper and Zmud (1990) provided evidence from the information technology (IT) implementation, organizational change, and technology diffusion literature and developed it into an IT innovation diffusion model. In Cooper and Zmud's model, change occurs in stages and five major contextual factors influence the stages. These five categories factors included individual characteristics, organizational factors, technological factors, the task characteristics, and environmental factors.
Anderson (1995) performed a very thorough case study of ABC adoption at General Motors (GM). She used Cooper and Zmud's stage model as a structure for describing the implementation of ABC and finds evidence supporting the theoretical model. Her theoretical model included 2 parts: it first examines the stages firms go through in ABC implementation and then considers the impact of a selected range of factors through these stages of ABC implementation.
2.2. Framing the implementation process: Stages of implementation
According to innovation diffusion theory change occurs in stages. Cooper and Zmud (1990) established a six-stage model of organizational change. Anderson (1995) adopted the Cooper and Zmud stage model and many ABC researchers followed her model. Some research models in ABC and IT stages relevant are summarized in table 1.
2.3. Explaining the implementation process: Contextual Factors
Cooper and Zmud (1990) identify five major contextual factors which influence successful transitions between the stages of implementation: individual characteristics; organizational factors; technological factors; task characteristics; and environmental factors. Four studies, Cooper and Zmud (1990), Anderson (1995), Anderson and Young (1999), and Brown et al. (2004), found 50 contextual factors which may influence the implementation process .These factors are summarized in Table 2.
2.3.1. Technological factors
It is expected that the technological factors such as: information technology quality, product diversity, and level of overhead will have a greater impact on ABC implementation than the other technological factors.
ABC System and Level of Information Technology Quality (IT)
The literature proposes two conflicting effects for the interaction level of information technology with ABC adoption. Maelah and Nasir Ibrahim (2006) argued that managers who were satisfied with information provided from the existing system might be reluctant to invest their resources in ABC, and the quality of IT is negatively related to ABC adoption. Anderson and Young (1999) found the quality of the information system is negatively related to the management's evaluation of ABC's overall value.
On the other hand, Krumwiede (1998) and Cagwin and Bouwman (2002) argued that detailed of operational data needed for activity analysis and an integrated ABC system needs high levels of information technology. An information system providing detailed historical data and easy access to users may provide much of the driver information needed by ABC. Moreover, Krumwiede (1998) found that high quality IT also appears a necessary condition to enable the achievement of the last implementation stage. Cagwin and Bouwman (2002) found that there was a positive relationship between information technology sophistication and the efficacy of ABC. Thus, it is expected that higher levels of IT quality may have a positive impact on the ABC implementation stages.
Implementation of ABC System and Level of Product Diversity (DIVER)
Product diversity relates to the variety of type and/or volume of products and/or product lines that are manufactured by a firm. Gilbert (2007) argued that the important factor affecting the adoption of ABC is the diversity of products. In high levels of diversity of products, traditional cost systems may not accurately allocate expenses to each product and may distort the cost of products. Many researchers identified product diversity as a primary reason why firms need more accurate costing systems such as ABC (e. g. Bjørnenak 1997; Clarke et al., 1999; Brown et al., 2004).Therefore, it is expected that higher levels of product diversity will be positively associated with the implementation of ABC.
Implementation of ABC System and Level of Overheads (OVER)
The early published ABC literature argued that overheads were becoming an increasingly larger component of product cost, for example, Chen (1996) found that overhead costs represent 50 per cent while direct labour cost is as low as 5 per cent of the total manufacturing cost. Because of the weakness of the traditional costing system to allocate overheads accurately, higher levels of overheads is one of the primary reasons for changing the traditional costing system to new systems such as ABC. Many researchers found that higher levels of manufacturing overhead expenses obviously pushed the firms to adopt more accurate costing systems such as ABC (Narong 2009; Xiong et al., 2008; Clarke et al., 1999; Brown et al., 2004). Cooper and Kaplan (1991) argues that the overhead allocation system distorts the cost in traditional costing systems. Therefore, it is expected that the levels of overheads will be positively associated with the implementation of ABC.
2.3.2. ABC and Environmental Factors
Krumwiede (1998) defined environmental factors as phenomena external to the organization that may influence its management accounting systems. This study will focus on three environmental factors: level of competition, business strategy and perceived environmental uncertainty. These three factors were chosen because they are closely related to the possible benefits of ABC implementation.
Implementation of ABC System and Competition (COMPT)
Level of competition refers to the degree of competition a firm faces in a particular market. Several studies have examined the relationship between the intensity of competition ABC adoption (e.g., Mia and Clarke, 1999; Dekker and Smidt, 2003; Al-Omiri and Drury, 2007). Cagwin and Bouwman (2002) showed that non-competitive situations such as a monopoly can lead to the use of traditional cost accounting rather than ABC. Bjørnenak (1997) found that competition has a positive effect on ABC adoption. Johnson and Kaplan (1987) believed that traditional cost accounting systems are obsolete in the new environment described by intensive competition. From this suggestion, it is expected that Level of competition has a positive effect on ABC implementation stage.
Implementation of ABC System and Business Strategy (STRA)
Many researchers believed that firms will place more emphasis on particular accounting techniques or information, depending on which strategy they adopt. (e.g. Bhimani et al., 2005; Jusoh et al., 2006). Bhimani et al. (2005) asserted that strategy plays a key role in the diffusion process for innovation. This study applied the Miles and Snow (1978 ) typology strategy. Miles and Snow's typology identified four strategic types of organizations according to the rate at which they change their products and markets: prospectors; defenders; analyzers; reactors (Snow and Hrebiniak, 1980; Gosselin, 1997). The fundamental difference among these types is the rate of change in the organizational domain. Gosselin (1997) believed that Miles and Snow's typology is more appropriate for examining the adoption of ABC as an innovation in management accounting systems. Miles and Snow's typology has been tested in several studies in ABC adoption (e.g. Gosselin, 1997; Frey and Gordon, 1999; Bhimani et al., 2005). It is expected that from Miles and Snow's typology of strategy, prospectors are more likely to be at higher stages of ABC implementation stages.
ABC System and Perceived Environmental Uncertainty (PEU)
Perceived environmental uncertainty (PEU) is an important contextual variable in accounting information system and management information system design (Fisher, 1996). The concept has also been conducted as a variable in management accounting systems (MAS) (e.g. Gul, 1991; Lat and Hassel 1998; Jusoh, 2008) and ABC (e.g. Anderson, 1995; Gosselin, 1997). Anderson (1995) found PEU had a negative effect on the two first implementation stages of ABC. Gosselin (1997) found that competitive strategy related to ABC adoption and competitive strategy has three components: environmental uncertainty, product diversity, production process complexity. On the other hand, Iran's economy has been severely disrupted by years of uncertainty surrounding economic activities. Samimi and Motameni (2009) found evidence that high levels of inflation lead to economic activities to higher uncertainty in Iran. In current study overall mean score for PEU is more than 3.0, suggesting that Iran's environment is quite unpredictable. In these conditions, it is expected that the perceived environmental uncertainty plays an important role in the implementation stages of ABC in Iran.
2.3.3. Implementation of ABC System and Firm Size (SIZE) as Control variable
The size of the respondents' firms is measured by the level of annual sales revenue and it consider as a control variable. The literature has suggested that there is a positive relationship between firm size and ABC. Many researchers found evidence that there is a relationship between ABC implementation and SIZE as a control variable (Cagwin and Bouwman, 2002; Kruimwiede, 1998; Banker et al., 2008). Krumwiede (1998) found that SIZE was a significant control variable in the adoption analysis, but it did not appear to be significant for infusion analysis. Bjørnenak (1997) believed that large firms have the required resources, such as time, and personnel to finance the infrastructure and, therefore, are more capable of adopting ABC.
H1: The effect of the contextual factors(level of information technology quality, level of product diversity, level of overheads, perceived environmental uncertainty, level of competition and business strategy) change in ABC implementation stages.
3. Research method
3.1. Data Collection Method
Data was collected from a mailed questionnaire survey. Krumwiede (1998) asserted that because most of the data needed for statistical analysis was not available from other sources, a survey instrument is only way to collect the amount of cross-sectional information. Gosselin (1997) introduced two reasons for selecting the surveying method including: First ,the information needed for the research was not available from archival sources. Second, other data collection techniques would be too inefficient.
3.2. Sample and Questionnaire Design
The sampling firms for the survey method are manufacturing firms listed on the Tehran Stock Exchange (TSE). The TSE opened in February 1967. Today, the TSE has evolved into growing marketplace where investor trade securities of over 450 companies. Moreover, there are three reasons to choose companies listed on the Tehran Stock Exchange. First, these companies are middle and large sized firms that should have greater resources available for investment in new systems, such as ABC. Second, the data about the firms is easily available from the Tehran Stock Exchange database. Third, based on the requirement for the company to be listed on the TSE, the use of a cost accounting system is compulsory. The questionnaire in this study was first constructed in English. Considering that people in Iran speak Persian as their first language, a translation of the questionnaire was required. Sekaran et al. (2000) stated that "it is important to ensure that the translation of the instrument to local language is equivalent to the original language". In the current study, to translate the survey instrument the double translation method was applied in two stages.
3.3. Variable Measurement
3.3.1. ABC Implementation stages
Table 3 outlined the stages of ABC implementation used in the present study, which is adopted from Krumwiede and Suessmair (2005). They used this model in a study for testing the factors that affect the adoption, infusion, and achievement of high levels of German cost accounting method (GPK) implementation. They believed that because both GPK and ABC are more sophisticated costing systems with higher implementation and maintenance costs than simpler systems, they can be shared in many of the implementation issues. Also, they indicated that, in many ways, the results of their study were similar to what Krumwiede (1998) had found with ABC.
3.3.2. Contextual Variables
Information Technology Quality (IT)
The question regarding information technology quality was adopted from Krumwiede (1998). The measurements included five statements and asked the respondents to indicate the level of agreement to each statement. To measure the extent of agreement with the statements, a five-point Likert scale was used from Strongly Disagree to Strongly Agree. The mean of five items was calculated for the IT variable.
Product Diversity (DIVER)
Level of product diversity was measured by an instrument developed by Khalid (2005). With this instrument, firms were classified into five continuous groups based on the number of products. These groups included: 1 = less than 5 products, 2 = five to ten, 3 = eleven to twenty products, 4 = twenty-one to fifty products and 5 = more fifty products.
Overheads (OVER)
Level of overheads was measured by an instrument developed by Krumwiede (1998). This question was involved with the firm's cost structure and measured components of cost including direct material, direct labour, and production overheads. Then the percentage of overheads was calculated and applied as a measurement. Firms were classified into five continuous groups based on the percentage of overheads. The level of overhead groups included: 1 = less than 14 percent, 2 = 14 percent to 19 percent, 3 = 20 percent to 24 percent, 4 = 25 percent to 29 percent, and 5 = more than 29 percent.
Perceived Environmental Uncertainty (PEU)
With Iran economic conditions attempts were made to find a perfect instrument to cover more dimensions of PEU. Questions regarding PEU were adopted from Jusoh (2008). In this instrument, Jusoh (2008) focused on measuring the respondents' perceptions of the predictability of seven aspects which included their organization's suppliers, competitors, customers, financial/capital markets, government regulations, labour unions, and economics, politics/technology. All of the items were measured on a five-point Likert-type scale (varying from "highly predictable" to "highly unpredictable"). The mean of the components served as the overall perceived environmental uncertainty score for a firm.
Competition (COMPT)
Level of competition refers to the degree of competition a firm faces in a particular market. In this study, a single item indicator was used for the measurement of the level of competition. Level of competition was assessed using the number of competitors. This measurement was developed from Cohen et al. (2005). This measurement used five- positions of level competition from 1 to 5 (1 = Not competitors, 2 = one to three,
3 = four to ten, 4 = ten to twenty competitors, and 5 = more than twenty competitors).
Business Strategy (STRA)
This study examined the business strategy types proposed by Miles and Snow (1978 ). This typology identified four strategic types of organizations according to the rate of change of products and markets: prospectors; defenders; analyzers; reactors. The question regarding business strategy was adopted from Jusoh and Parnell (2008).This instrument included forty-eight items in twelve questions. Each question consists of four statements each of which was related to each possible strategy type. Respondents were asked to indicate agreement or disagreement with each statement concerning their organization by using a five-point Likert scale including: from strongly disagree to strongly agree. Business strategy was operated by taking the mean score across the twelve items in four strategic types. Then, for each firm the degree of the mean value which was classified into four strategic types was compared. The highest value indicated which firm emphasizes a given strategy.
3.4. Validity and Factor Analysis
In the current study based on Sekaran et al. (2000) suggestion three important types of validity were conducted including face validity, content validity and factorial validity. For face validity, a pre- testing was conducted to 30 individual financial managers for clarity and meaning. Content validity was established by a review of the instrument by 10 expert judgments comprising experienced accountants and academics. This study did not attempt to look at the dimensions of business strategy, thus factor analysis was not employed for the strategy items. For five items of (IT) factor analyses was conducted but only one factor was extracted. The cumulative percentage of variance for this factor was more than 85%.
For perceived environmental uncertainty (PEU) a confirmatory factor analysis is also applied. Of the 28 items of PEU, three items have factor loading of less than the 0.50 and one item has high loadings in more than one factor. This factor loading analysis suggests that 4 items should be deleted from the analysis. Moreover, the Scree Plot shape suggested three factors with eigenvalues exceeding 1.00 and covering a total of 62.47% of variance.
From this analysis three component factors were extracted. Factor one included 10 items including: government policies regarding financial practices, labour laws, marketing, and accounting procedures, action of labour union regarding changes in wages, hours, and working conditions, union security, grievance procedures, changes in the economy, changes in manufacturing technology, and changes in scientific discoveries. Factor one which covered 43.20% of variance were combined and named "perceived environmental uncertainty-economic (PEU-ECO).Second dimension of PEU included 11 items including: for suppliers of raw material regarding price, design, and introduction of new materials changes, for competitors regarding price, product quality, introduction of new products, and competition for manpower changes and for customers regarding demand for existing products, demand for new products, and tastes and preferences changes. These Items were combined and named "perceived environmental uncertainty -industrial "(PEU-IND). Factor 3 included 3 items which covered 7.30% of variance. Under this factor, the financial/capital markets interest rate changes, availability of credit changes, and changes in financial instruments available, were combined and named "perceived environmental uncertainty-financial (PEU-FIN).
3.5 Reliability
Reliability of the multi-item measurement scale in the questionnaire was estimated by using Cronbach's alpha, the degree of internal consistency among items in the multipoint-scaled items in the questionnaire (Sekaran et al., 2000). The coefficient alpha varies from 0 to 1 and the value of 0.60 or above indicates satisfactory internal consistency reliability (Malhotra, 2004). Table 4 presents Cronbach's alpha coefficient of the multi-items.
4. Results
4.1. Descriptive Statistics
Table 5 contains the descriptive statistics for the variables. As shown overall mean score for PEU is more than 3.0, suggesting that Iran's environment is quite unpredictable.
4.2. Hypotheses Testing
H1 states that the effects of contextual variables in this study will vary according to the ABC implementation stages. This study used the seven stages implementation model as proposed by Krumwiede and Suessmair (2005) (see Table 3). Data from this study reveal that there were no responses for stage 2, 4, and 6. The four remaining stages were stages 1, 3, 5, and 7 which are used to determine if the effects of the factors vary according to the stages (see table 5, Descriptive Statistics). For this analysis, three models are interpreted by using binary logit regression. Tables 6, 7, and 8, present the result of these three binary logit regressions models.
The first model, named panel A, investigated the effect of contextual factors (independent variables) on ABC implementation stages where the dependent has two values (0 vs. 1). In panel A stage one labeled as 0 and stages 3 or 5 or 7 was recorded as 1. Besides the ABC implementation stages, panel A tests the effect of contextual variables on moving from stage one (Not consider to ABC) to stage 3 (Initiation/evaluating of ABC). The second model, called panel B, in this panel firms are recoded as 0 when they are in stages 1 or 3 and are labeled as 1 when are in stages 5 or 7. Finally, the last model is called panel C where the firms is recoded as 0 when they are in stages 1 or 3 or 5 and labeled as 1 when they are in stage 7 (0 vs. 1). These three binary logits provided evidence that any of the independent variables affect the progression from stages 1 to 3 or from, 3 to 5 or from, 5 to 7 differently.
4.3. Findings for interest in ABC Initiatives
Panel A tests the effect of contextual factors on moving from stage one (Not consider to ABC) to stage 3 (Initiation/evaluating of ABC). The odds ratio exp (B) in the binary logits analysis indicates the expected change in dependent variable when independent variables change by one unit. For testing the control role for SIZE, 2 Binary logit regression models applied including: Binary regression without control by the SIZE and with control by the SIZE. After run the model with SIZE as control variable the SIZE has significant effect on the panel A (p-values = 0.066) and changed the -2 log likelihood value indicated the control role for SIZE. This result suggested that the big size firms are more likely to be in stage 3, 5 or 7 than smaller size firms.
As shown in table 6 the effect of PEU-FIN on panel A is negative and significant (p-value = 0.011, b = - 0.43). The negative b coefficients indicate that if the PEU-FIN increases the probability of to be in stages 3 and above decreases. The odds ratio exp (B) shows that a one-unit increase in the value of PEU-FIN decreases the probability of to be in stages 3 and above from 1 to 0.65. The results suggest that firms which face lower PEU-FIN tend to be in stage 3, 5 or 7. As explained business strategy is a categorical variable comprising four types of strategy: prospectors (STRA-P); defenders (STRA-D); analyzers (STRA-A); and reactors (STRA-R) and three dummy variables were used. Reactors' strategy (STRA-R) is considered as a reference group. In panel A STRA-A is significant at p<0.01 (b = + 1.36) which indicated that the STRA-A has a more positive effect on panel A than STRA-R. The odds ratio exp (B) for STRA-A is equal to 3.92 which means the probability of to be in stages 3 and above in firms who are using STRA-A is 3.92 times bigger than STRA-R. These results show the analyzer firms are more motivated to be in stage 3, 5 or 7.
Findings for ABC Adoption
Adoption has traditionally been the central event in innovation studies. In this study adopter firms defines as firms who attain the last four stages. Panel B is a binary logit model which dependent variable has only two values, zero for firms on stage 1 or 3 and one for firms on stages 5 or 7. Thus, panel B investigate the effect of contextual factors on adopting ABC. The result for panel B shows in table 7. The result also indicated the control role for SIZE in ABC adoption. This result suggested that the big size firms are more likely to access stages 5 or 7 and to be ABC adopters than smaller size firms. As shown in Table 7 all variables are significant for Panel B (adopting ABC). More discussion about this panel B is presented as follows: IT is significant at p<0.10, the B parameter for IT is equal to - 0.44. The negative b coefficients indicate negative effect of IT on ABC adoption. The odds ratio exp (B) is equal to 0.65; this parameter indicated that one unit increase in value of IT decreases the probability of adoption taking place from 1 to 0.65. These results suggest that firms without high IT may be more able to be ABC adopters. DIVER has a positive effects on ABC adoption at p<0.01 level (b = + 0.61). The odds ratio exp (B) is equal to 1.84, this result showing that a one-unit increase in the value of DIVER increases the probability of adoption happening from 1 to 1.84. The result suggests that firms with a high diversity of products are more likely to adopt the ABC system. OVER has positive and significant on ABC adoption or panel B. (p-values = 0.002, b = + 0.78). This result suggests that firms with high OVER are more likely to adopt the ABC system. All three factors of perceived environmental uncertainty (PEU) have negative and significant effect on ABC adoption (At p<0.01 level). The result shows perceived environmental uncertainty appears to play a major role in the adoption of ABC. The results suggest that more firms which face lower PEU tend to adopt the ABC system. COMPT is significant at p<0.01 level (b = + 0.67). The positive b coefficient indicates that COMPT increases the probability of the ABC adoption. These results suggest that non-competitive situations such as a monopoly can lead to the use traditional cost accounting rather than using ABC. STRA-A has positive and significant effect on ABC adoption. Adoption results show that the analyzer firms are more motivated to adopt the ABC system.
4.5. Findings for ABC infusion
In this study, firms were labeled as infusers if they attain stage seven. Some researchers named the infusion of ABC as ABC success, while adoption means applying ABC and is a starting point for implementing ABC.Panel C is a binary logit analysis that dependent variable value is zero for firms on stage 1 or 3 or 5 and one for firms on stage 7. Thus, panel B investigate the effect of contextual factors on infusion ABC. The result for r panel C is reported in Table 8. As shown, Under the Model Summary, the -2 Log Likelihood statistics is equal to 28.18. The Cox and Snell R2 is equal to 0.14 that can be identified as R2 in a multiple regression as showing goodness of fit in the model. The Nagelkerke R2 parameter is equal to 0.35. The Hosmer and Lemeshow test was also conducted to ensure the goodness of fit of the model. The finding shows non significance of the p-value for Hosmer and Lemeshow (p-value = 0.676) indicating that the predicted model is not significantly different from observed values and model fits well. The result for panel C also indicated the control role for SIZE. This result suggested the big size firms are more likely to be in the last stage (stage 7). As shown in Table 8, DIVER, PEU- IND, and STRA-A are significant for panel C. The results suggest that firms which face higher DIVER and lower PEU-IND tend to be ABC infuser. And the analyzer firms are more motivated to access stage 7.
4.6. Discussion of results
H1 states that the effects of contextual variables in this study will vary according to the ABC implementation stages. As the result shows, PEU- FIN and STRA-A, are significant for panel A (moving from stage one: Not consider to ABC to stage 3: Initiation/evaluating of ABC). All variables are significant in panel B (ABC adoption), and DIVER, PEU- IND and STRA-A are significant for panel C (ABC infusion). The results of these three panels are summarized in Table 9. As a shown, of the 8 variables only 1 variables (STRA-A) have the same effect on the various stages. So H1 is partially supported.
As shown, each factor affects the progression from panel A to the others differently. Based on the overall findings in this study, Figure 1 shows the key factors for each of the ABC implementation stages. Thus the model is adequate based on the residual regression analysis for supporting the organizational change theory. Cooper and Zmud (1990) believed that change occurs in stages and the significance effects of factors might vary across the stages.
Not considered
Initiation/evaluating
Used occasionally
Used extensively
STRA-A (+) PEU-IND (-)
COMPT (+) DIVER (+) OVER (+) STRA-A(+) PEU-IND (-) PEU-FIN(-)
PEU-ECO(-) IT(-)
DIVER (+) PEU-IND (-)
STER-A (+)
Figure 1: Key Factors for ABC Implementation Stages
5. Conclusions
In the current study test the effects of several contextual factors on the implementation of ABC. These contextual factors were classified as technological, environmental factors. Three models were developed to capture the effect of certain factors (independent variables) on different stages of ABC. All three models are interpreted by using binary logit regression.
The first model, named panel A. tests the effect of contextual factors on moving from stage one (Not consider to ABC) to stage 3 (Initiation/evaluating of ABC). The second model, called panel B, investigate the effect of contextual factors on adopting ABC. Finally, the last model is called panel C investigates the effect of contextual factors on infusing ABC.
The findings show that the effect of certain contextual factors change according to the ABC implementation stages. This finding is consistent with some ABC studies (e.g., Anderson, 1995; Krumwiede, 1998; Gosselin, 1997) which found that the effects of selected factors vary from stage to stage. Although they use a different stage model and different factors, they found the same result that different factors are associated with the different stages. Moreover, this finding is consistent with information systems (IS) innovation theory which is widely accepted by most of the ABC researchers. The theory suggested that change occur in stages and the degree of importance for each contextual factor differs in several ABC implementation stages.
This study found evidence that perceived environmental uncertainty (PEU) plays an important and negative role in the implementation stages of ABC. It is also important to note that Iran is confronted with an unpredictable environment. The findings suggest that under uncertain conditions, the manager does not provide resources for implementing an innovation system such as ABC. Moreover, the results suggested that the big size firms are more likely to be at higher stages of ABC implementation stages than smaller size firms. Furthermore, interesting results to emerge from this study were: the negative effect of IT on the adoption of ABC, and from Miles and Snow's typology of strategy, analyzers firm are more likely to be at higher stages of ABC implementation stages.
Table 1: ABC Implementation Models
Author ,Title
Source
Implementation stages
Cooper and Zmud (1990) "Information Technology Implementation Research: A Technological Diffusion Approach,
MS*,
Vol.36, pp, 123-139.
(1)Initiation(2)Adoption
(3) Adaptation(4)Acceptance
(5) Routinization(6)Infusion
Anderson (1995) ''A framework of assessing cost management system changes: the case of activity-based costing implementation at General Motors
JMAR*, Vol. 7,
pp.
1-51.
(1)Initiation(2)Adoption
(3)Adaptation(4)Acceptance
Gosselin (1997) ''The effect of strategy and organizational structure on the adoption and implementation of activity-based costing''
AOS*, Vol. 22, No. 2, pp.
105-122
(1)Adopted AM(2)Adopted AA
(3)Adopted ACA(4)Adopted ABC
Krumwiede (1998) ''The implementation stages of activity-based costing and the impact of contextual and organisational
factors''
JMAR*, Vol. 10, pp.
239-277.
(1)Not considered(2)Initiation/ evaluating (3) Evaluated then rejected(4) Evaluated and approved for implementation(5)Analysis(6) Gaining acceptance (7) Implemented then abandoned(8) Restricted Use(9) Used somewhat(10) Used extensively
Innes et al. (2000) ''Activity-based costing in the UK's largest companies: a comparison of 1994 and 1999 survey results
MAR*, Vol. 11, pp.
349-362.
(1)No consideration of (2)Rejected after assessment(3)Currently considering (4)Currently using
Joshi (2001) "The international diffusion of new management accounting practices: the case of India"
JIAAT* Vol. 10, pp.58-109.
(1) Low adoption (2) Moderate adoption(3) High adoption
Pierce (2004) "Activity-Based costing"
AI*,
Vol.36,
pp,28-33.
Innes et al. (2000) stages model.
Brown et al. (2004) "Technological and organizational influences on the adoption of activity-based costing in Australia"
AAF*,
Vol. 44
Pp,
329-356.
Krumwiede (1998) stages model.
Khalid (2005)"Activity-Based Costing in Saudi Arabia's Largest 100 Firms"
JAABC*,
Vol. 6, Pp,
285-292.
(1) Never considered(2)Presently considering (3)Reject after evaluation
(4)Abandoned after using (5)Presently apply .
Cohen et al. (2005)" ABC: adopters, supporters, deniers and unawares"
MAJ*, Vol.20,Pp,981-100,
(1)ABC unawares(2)ABC deniers
(3)ABC supporters(4)ABC adopters
Maelah and Ibrahim (2006) "Activity Based Costing (ABC) Adoption among Manufacturing Organizations - The Case Of Malaysia"
IJBS*, Vol.7,
Pp,
70-101.
Krumwiede (1998)stages
*Note:
JMAR: Journal of Management Accounting Research; MAR: Management Accounting
Research; AOS: Accounting, Organizations and Society; MS : Management Science;
JIAAT: Journal of International Accounting, Auditing and Taxation; AAF: Accounting and Finance; AI: Accountancy Ireland; JAABC: Journal of American Academy of Business, Cambridge. MAJ: Managerial Auditing Journal; IJBS: International Journal of Business and Society.
Table 2: Factors and ABC Implementation
1-Individual Characteristic
Job tenure
Education
Critical mass
Role involvement
Cosmopolitanism
Disposition to change
Informal support
Unit professionally orientated
2-Organizational factors
Top management support
Local management support
Local union
support
Vertical differentiation
Champion
Consultants
Informal network
Formalization
Size
Other firms adopting
Fit
Specialization
Internal communications
Formal support in accounting function
Extrinsic reward systems
Centralization
3-Task characteristics
Task uncertainty
Training
Resource adequacy
Job shop
Autonomy
Variety
Responsibility
Lean production
4-Technological factors
Dominance of overheads
Product line complexity
Relevance to decision making
New application portfolio
Decreasing price
IT quality
Decision usefulness
Trainability
Relative advantage
Product diversity
Compatibility
Absorbability
Complexity
Accuracy
5-Environmental factors
Likelihood of
layoffs
External communicant
Inter-organizational dependence
Importance of site to company
Competition
Heterogeneity
Concentration
Growth opportunities
Strategy
Labor relations
Uncertainty
Table 3: Seven Stage Model of Study
Stages
Krumwiede and Suessmair (2005)
The present study
1
Not considered: GPK has not been seriously considered. We use either single or departmental /multiple plant-wide allocation methods only
Not considered: ABC has not been seriously considered. We use either single or departmental / multiple plant-wide allocation methods only
2
Considered then rejected: GPK has been considered but later rejected as a costing method.
Evaluated then rejected: ABC has been evaluated (but not implemented) and was later rejected as a cost assignment/ management method.
3
Considering: GPK Implementation is possible in the future but has not been approved
Initiation/evaluating: ABC is being
evaluated and implementation is possible, but implementation has not yet been approved.
4
Implemented then abandoned: GPK was previously implemented but is not currently being used.
Implemented then abandoned: ABC was previously implemented but is not currently being used.
5
Used occasionally: Occasionally used by non-accounting management or departments for decision making.
Used occasionally: Occasionally used by non-accounting management or departments for decision making.
6
Used frequently: Frequently used for management decision making; considered normal part of information system
Used frequently: Frequently used for management decision making; considered normal part of information system
7
Used extensively: Used extensively for management decision making; clear benefits of GPK can be identified.
Used extensively: Used extensively for management decision making; clear benefits of ABC can be identified.
Table 4: Reliability of the Multi-Items
Factor
Alpha
Number of Item
Business strategy: Prospector
0.88
12
Business strategy: Defender
0.81
12
Business strategy: Analyzer
0.85
12
Business strategy: Reactor
0.80
12
Perceived environmental uncertainty-Industrial
0.91
11
Perceived environmental uncertainty-Financial
0.83
3
Perceived environmental uncertainty- Economic
0.94
10
Level of information technology quality
0.96
5
Table 5: Descriptive Statistics
Competition status
Frequency
Percentage
Cumulative %
No Competitors
31
16.50
16.50
1-3 Competitors
37
19.70
36.20
4-10 Competitors
37
19.70
55.90
11-20 Competitors
44
23.40
79.30
More than 20 Competitors
39
20.70
100.00
Number of products:
Frequency
Percentage
Cumulative %
Less than 5
41
21.80
21.80
5-10
57
30.30
52.10
11-20
30
16.00
68.10
21-50
28
14.90
83.00
More than 50
32
17.00
100.00
Business Strategy
Frequency
Percentage
Cumulative %
Analyzers
40
21.30
21.30
Defenders
51
27.10
48.40
Prospectors
36
19.10
67.60
Reactors
61
32.40
100.00
Level of overhead
Frequency
Percentage
Cumulative %
Less than 14%
41
21.80
21.80
14% -19%
48
25.50
47.30
20% - 24%
43
22.90
70.20
25% - 29%
31
16.50
86.70
More than 30%
25
13.30
100.00
Level of IT
Frequency
Percentage
Cumulative %
1.00-1.99
19
10.10
10.10
2.00-2.99
48
25.53
35.63
3.00-3.99
45
23.93
59.56
4.00.5.00
76
40.44
100.00
ABC Implementation Stages
Frequency
Percentage
Cumulative %
Stage one-Not considered
112
59.58
59.58
Stage Two-Evaluated then rejected
0
0.00
59.58
Stage Three -Initiation/evaluating
43
22.87
82.45
Stage Four -Implemented then abandoned
0
0.00
82.45
Stage Five -Used occasionally
20
10.64
93.09
Stage Six -Used frequently
0
0.00
93.09
Stage Seven -Used extensively
13
6.91
100.00
Environmental uncertainty- Factors
Min
Max
Mean
PEU-ECO
1
5
3.50
PEU-IND
1
5
3.62
PPEU-FIN
1
5
3.39
Table 6: Binary Logistic Result: Panel A Implementation Stages
Factor
B-Estimate
Std.Error
Wald Chi-Square
df
p-value
Exp(B)
Constant
- 2.36
0.84
7.96
1
0.005
0.09
IT
-0.05
0.14
0.15
1
0.703
0.95
DIVER
0.19
0.12
2.60
1
0.106
1.21
OVER
0.07
0.13
0.33
1
0.568
1.08
PEU-IND
- 0.03
0.17
0.05
1
0.830
0.96
PEU- FIN
- 0.43
0.17
6.50
1
0.011
0.65
PEU-ECO
- 0.08
0.16
0.24
1
0.626
0.92
COMPT
0.13
0.12
1.17
1
0.280
1.14
STRA-A
1.36
0.46
8.99
1
0.003
3.92
STRA-D
0.15
0.42
0.13
1
0.717
1.16
STRA-P
- 0.52
0.52
1.04
1
0.307
0.59
SIZE
0.24
0.13
3.37
1
0.066
1.27
Test
Wald Chi Square
df
p
Omnibus test of model coefficients
29.77
11
0.02
Goodness-of-fit test(Hasmer & Lemeshow)
10.02
550
0.264
Note: -2 log likelihood=223.91, Cox and Snell R2 = 0.15, Nagelkerke R2 =0.20
Table 7: Binary Logistic Result: Panel B Implementation Stages
Factor
B-Estimate
Std.Error
Wald Chi-Square
df
p-value
Exp(B)
Constant
- 11.70
2.10
29.52
1
0.001
0.00
IT
- 0.44
0.24
3.09
1
0.078
0.65
DIVER
0.61
0.23
7.07
1
0.008
1.84
OVER
0.78
0.25
9.33
1
0.002
2.19
PEU-IND
-0.98
0.33
8.83
1
0.003
0.37
PEU- FIN
- 0.91
0.30
8.83
1
0.003
0.40
PEU-ECO
- 0.60
0.28
4.39
1
0.036
0.55
COMPT
0.67
0.22
8.69
1
0.003
1.96
STRA-A
3.45
0.91
14.15
1
0.001
31.67
STRA-D
0.84
0.95
0.77
1
0.377
2.33
STRA-P
0.91
0.85
1.15
1
0.283
2.49
SIZE
0.95
0.26
12.77
1
0.001
2.59
Test
Wald Chi Square
df
p
Omnibus test of model coefficients
89.89
11
0.001
Goodness-of-fit test(Hasmer & Lemeshow)
10.46
8
0.230
Note: -2 log likelihood=84.78, Cox and Snell R2 = 0.38, Nagelkerke R2 =0.63
Table 8: Binary Logistic Result: Panel C Implementation Stages
Factor
B-Estimate
Std.Error
Wald Chi-Square
df
p-value
Exp(B)
Constant
-7.95
2.10
14.35
1
0.000
0.00
IT
-0.29
0.31
0.86
1
0.352
0.74
DIVER
0.50
0.26
3.81
1
0.051
1.66
OVER
0.09
0.27
0.10
1
0.744
1.09
PEU-IND
- 0.68
0.36
3.57
1
0.059
0.50
PEU- FIN
-0.10
0.35
0.09
1
0.764
0.90
PEU-ECO
-0.09
0.31
0.09
1
0.757
0.90
COMPT
0.10
0.24
0.18
1
0.665
1.11
STRA-A
1.71
0.93
3.38
1
0.066
5.56
STRA-D
0.44
1.12
0.15
1
0.696
1.55
STRA-P
0.49
1.03
0.22
1
0.636
1.63
SIZE
0.86
0.32
7.28
1
0.007
2.37
Test
Wald Chi Square
df
p
Omnibus test of model coefficients
28.19
11
0.003
Goodness-of-fit test (Hasmer & Lemeshow)
4.74
8
0.676
Model summary: -2 log likelihood=66.35, Cox and Snell R2 = 0.14, Nagelkerke R2 =0.35
Table 9: The Roles of Predictors on ABC Implementation Stages
Panel
Independent variables (Predictors)
IT
DIVER
OVER
PEU-IND
PEU-FIN
PEU-ECO
COMPET
STRA-A
A
N.S
N.S
N.S
N.S
-
N.S
N.S
+
B
-
+
+
-
-
-
+
+
C
N.S
+
N.S
-
N.S
N.S
N.S
+
* N.S: It is not significant.
