رابطه بین تنوع محصول، استفاده از فناوری های پیشرفته تولید و پذیرش هزینه یابی بر مبنای فعالیت

It is widely recognized that in order to be optimal, costing practices should be closely related to the fundamental work processes in a firm. That is, more complex work processes require more complex cost systems (in terms of the applied overhead absorption procedures) to capture them. Accordingly, from a rational economic theoretical perspective, firms’ adoption of activity-based costing (ABC) should be mainly driven by technological factors, of which especially product diversity is strongly suggested to be a major determinant of the demand for ABC systems (Abernethy et al., 2001 and Al-Omiri and Drury, 2007). Product diversity refers to conditions in which products place different demands on a firm’s activities or activities place different demands on its resources, and is generally considered to be the most important cause of distorted product costs by traditional (non-activity-based) cost systems (e.g., Cooper, 1988 and Kaplan and Cooper, 1998). Since its introduction in the literature in the late 1980s, several streams of empirical research on ABC have developed, of which research on the determinants of ABC adoption is a major one (e.g., Gosselin, 1997 and Krumwiede, 1998). The results of this research, in which the influence of many potential determinants has been studied, are generally inconclusive. This has led researchers into arguing that more research in this area is necessary (e.g., Abernethy et al., 2001 and Drury and Tayles, 2005), since the source of the inconclusive results may not only be substantial, but may also be methodological (Al-Omiri & Drury, 2007). Most studies to date have, for example, used inconsistent definitions of and measurement instruments for both ABC adoption and its (proposed) determinants, and some have used rather simple, bivariate statistics. More importantly, no study has yet explored effects other than linear, additive ones for any of the potential determinants of ABC adoption, although the results of qualitative field research suggest such effects may exist in practice (e.g., Abernethy et al., 2001). Despite its strong theoretical foundation, similar to the other studied (potential) determinants, the results of empirical research on the relationship between product diversity and ABC adoption have also generally been inconclusive, suggesting that there either may be no strong relationship, or that methodological issues may have prevented researchers from consistently finding it. The purpose of this paper is to re-examine the relationship between product diversity and ABC adoption, using a dataset of survey responses from 191 Dutch, medium-sized manufacturing firms. Improving upon the measurement of product diversity and distinguishing between ABC adoption (defined as either using or implementing ABC) and use (defined as using ABC), it examines whether the relationship is curvilinear (inverted U-shaped) and/or moderated by usage of advanced manufacturing technologies (AMT). This paper contributes to the literature on the determinants of ABC adoption in at least three ways. First, previous survey research has been criticized for adopting a rather simplistic approach to the notion of product diversity (Abernethy et al., 2001). In these studies, it has either been measured as the number of products (or product variants) produced in a firm, or by using a composite multi-item scale with respect to the complexity (diversity) of manufacturing and costing in a firm. Both types of measures seem to insufficiently capture the exact nature of product diversity. This study therefore develops and uses a new, more comprehensive measure of product diversity, that essentially combines the main underlying elements of both previously used types of measures. Second, previous survey research (e.g., Al-Omiri and Drury, 2007 and Krumwiede, 1998) has implicitly assumed that the relationship between product diversity and ABC adoption is linear; the assumption has been that increased product diversity leads to corresponding increases in firms’ likelihood of adopting ABC. Positing a linear relationship ignores the possibility that there may be nonlinear effects of product diversity on firms’ likelihood of adopting ABC. The results of this study show that, consistent with the underlying theory, product diversity, on average, is positively related to both ABC adoption and ABC use, but also that these relationships are inverted U-shaped; i.e., that they are positive up to a point and then begin to decline. From a practical perspective, this latter finding means that firms are more likely to adopt and use ABC at moderate levels of product diversity than at high levels of product diversity. Third, previous survey research (e.g., Al-Omiri and Drury, 2007 and Krumwiede, 1998) has also implicitly assumed that the relationship between product diversity and ABC adoption is additive; the assumption has been that product diversity does not have a joint (or interactive) effect with other contextual factors on firms’ likelihood of adopting ABC. In their field study, however, Abernethy et al. (2001) find evidence that the relationship between product diversity and ABC adoption may be moderated by usage of AMT, and argue that this moderation effect requires extensive testing. This study provides such testing and shows that the relationship with ABC use (but not with ABC adoption) is negatively moderated by usage of AMT. From a practical perspective, this finding means that the effect of product diversity on firms’ likelihood of using ABC is stronger when the level of usage of AMT is relatively low than when this level is relatively high. Overall, the results of this study indicate that the inconclusive results of previous survey research on the relationship between product diversity and ABC adoption at least partially seem to have been caused by methodological issues. The remainder of this paper is structured as follows. Section 2 reviews the literature and develops the hypotheses. Section 3 describes the research methods used. Section 4 presents and discusses the results. Section 5 summarizes and concludes.

This paper uses a dataset of survey responses from 191 Dutch, medium-sized manufacturing firms to re-examine the relationship between product diversity and ABC adoption. Improving upon the measurement of product diversity and distinguishing between ABC adoption and use, it examines whether the relationship is curvilinear (inverted U-shaped) and/or moderated by usage of AMT. The results show that product diversity, on average, is positively related to both ABC adoption and ABC use, but also that these relationships are indeed inverted U-shaped; i.e., that they are positive up to a point and then begin to decline. From a practical perspective, this latter finding means that firms are more likely to adopt and use ABC at moderate levels of product diversity than at high levels of product diversity. The results also show, however, that the rela- tionship with ABC use (but not with ABC adoption) is negatively moderated by usage of AMT, which from a practical perspective means that the effect of product diversity on firms’ likelihood of using ABC is stronger when the level of usage of AMT is relatively low than when this level is relatively high. Overall, these results indicate that the inconclusive results of previous survey research on the relationship between product diversity and ABC adoption at least partially seemto have been caused by methodological issues. Some of the empirical results merit further discussion. First, prior studies have used many inconsistent definitions of and measurement instruments for both ABC adoption and its determinants, which is likely to at least be partially responsible for the inconclusive results of these studies to date. This study adds to this inconsistency as it develops and uses a newmeasure of product diversity, which is argued to be more comprehensive than existing measures and to better capture the exact nature of product diversity. The usage of this newmeasure may have contributed to finding results that differ fromthose of some of the prior studies, for example on the positive, but curvilinear relationships between product diversity and ABC adoption and ABC use. Future research could further develop and validate the measurement instrument used in this study to measure product diversity. Second, this study finds that the relationship between product diversity and ABC use (but not ABC adoption) is negatively moderated by usage of AMT, supporting qualitative findings of Abernethy et al. (2001). Their findings also suggest, however, that the joint effect of product diversity and usage of AMT on cost system design choices may be mediated by the cost structure (i.e., the proportion of batch- and/or product-related costs) of firms, which was not tested in this study. Future research could examine this mediation effect, as well as the joint (or interaction) effects of other contextual factors on firms’ ABC adoption choices. Third but related to the former two points, this study distinguishes between ABC adoption and use, and examines the relationship of product diversity (and a number of other technological factors) with both of these stages of the ABC imple- mentation process. Overall, the associations of product diversity and usage of AMT with (and therefore their effects on) ABC use appear to be somewhat stronger than those with (on) ABC adoption. Future research could further examine these differences in findings for ABC adoption and ABC use, and could also explore to what extent the effects found in this study also apply to other stages in the ABC implementation process. As with any study, the findings of this study are subject to a number of potential limitations. Because cross-sectional research can establish associations, but not causality, the direction of effects cannot be established with certainty. Also, there may be omitted variables which may bias the results. Another issue that may potentially influence the findings is measurement error. The dependent variable in this paper, ABC adoption, is a typical example of a practice-defined variable (cf. Luft & Shields, 2003), and such variables are more likely to suffer from measurement error than theory-defined variables. Research developing a better measurement instrument for this variable could significantly contribute to the literature.16 Finally, there is the issue of generalizability. The response rate in this study is rather low and, consequently, the sample size rather small. Although comparisons with the sampling database show that the sample is representative in terms of industry, it may be biased with respect to other (unknown) variables. Despite these potential limitations, this study presents a step further in our understanding of the determinants of ABC adoption in practice.