تغییر سیستم های مدیریت حسابداری و عملکرد دپارتمانی: نفوذ اطلاعات مدیریتی و عدم قطعیت وظیفه

Despite the proliferation of new management accounting techniques amidst pressures of organizational and global change, the issue of changes in firm-wide management accounting and control systems (MACSs) has largely been ignored in the research literature. This study explores the indirect effect of MACSs change on departmental performance for a cross-sectional sample of 232 medium-sized Singaporean firms. It is hypothesized that MACSs change affects performance but not directly. Instead, this relationship is mediated by managerial-relevant information (MRI) that is impacted by MACSs change, which, in turn, enhances performance. Task uncertainty is expected to moderate the intervening linkages; specifically, the latter are anticipated to strengthen under conditions of more task variability and task difficulty and, thus, augment the indirect effect of MACSs change on performance. The results offer support for the positive indirect effect of improving departmental performance from more MRI, triggered by MACSs change. Although not large, the indirect effect is strengthened when task variability and task difficulty are high. Overall, the findings are consistent with the stated purposes of management accounting that are embedded in normative definitions, and which are relied upon to motivate the framework for analysis.

The issue of accounting systems change has interested management accounting scholars for decades. Dynamic business environments, characterized by unrelenting technological and organizational change from heightened globalization and increasing competition, continue to perpetuate this focus. An expanding contemporary literature that features new planning systems, advanced costing techniques (Cooper and Kaplan, 1992), lateral control systems, balanced scorecard mechanisms (Kaplan, 1990, 1998), and performance management frameworks (Otley, 1999) is ample testimony to the changing landscape of management accounting systems in particular. In addition, escalating globalization has spurred renewed research interest in the cross-cultural design of accounting control systems (Chow et al., 1999; Harrison and McKinnon, 1999), including proposed theories of convergent versus divergent accounting practices in industrialized nations (Granlund and Lukka, 1998). Despite these prolific practice and research advances, most of our cumulative knowledge from them concerns the design merits and implementation issues associated with individual control system components (e.g. ABC) and new management accounting techniques or methods examined individually (e.g. see Otley, 1999). Atkinson et al. (1997, p. 87) contend that better knowledge about change in management accounting systems might be ‘prerequisite to studying the process of change’ for management accountants in general. From a macro organizational perspective, the literature on the possible forces influencing management accounting change is surprisingly silent. Recent comprehensive literature reviews on the current state of research in management accounting indicate no attention devoted to this issue (e.g. see Shields, 1997; Foster and Young, 1997). From a slightly altered perspective, and somewhat remarkable as well, research on whether this wave of new management accounting applications benefits organizations ‘is strikingly absent from the literature’ (Foster and Young, p. 73). Lately, though, the pioneering work of Libby andWaterhouse (1996) along with the subsequent study by Williams and Seaman (2001) have generated evidence on the extent and determinants of management accounting and control systems (MACSs) change1 at the organizational level. These studies utilize contingency theory to argue that MACSs change is dependent on a firm’s context and structure, both of which serve to impel and retard change. However, embracing the challenging issue of whether organizations indeed benefit from changes in MACSs requires the latter to be conceptualized differently. Moreover, it necessitates a methodological shift that elevates MACSs change to the status of an independent variable rather than being a dependent variable as mentioned in the contingency framework above. The central purpose of this paper, therefore, is to develop a testable framework that addresses the relationship between MACSs change and performance at the departmental level, and then to examine it empirically.

The purpose of this study was to investigate the perceived performance benefits from changes in the set of management accounting systems via the intervening effects of MRI that are supplied to managers through these changes. The results appear to support exactly the claimed premise of normative definitions of management accounting whereby increases in MACSs change provide value-added information for managerial decision making and control activity, which then facilitates the achievement of operating departments’ objectives. As predicted in hypothesis H1, the findings did not reveal a significant direct relationship between MACSs change and departmental performance. This is consistent with the weight of evidence from most individual level studies, especially, for example, those involving participative budgeting and performance effects (Shields and Young, 1993). Moreover, there were statistically significant positive linkages between MACSs change and MRI, and MRI and departmental performance (i.e. direct effects), thus supporting the expected intervening influence of MRI on performance from an increase in MACSs change. In aggregate, these relationships produced a positive, but rather small (i.e. 0.053), indirect effect between MACSs change and performance for the total sample (see equation (5) in Table 4). Introducing task uncertainty into the model yielded a positive association between higher MACSs change and more task variability, as well as more task difficulty, thus supporting hypotheses H2 and H5, respectively. The correlation coefficients showed that the direct effect was statistically stronger for task variability as opposed to task difficulty (i.e. 0.24 versus 0.13 from Table 2). Moreover, the direct effects of MACSs change on MRI under conditions of high variability and high difficulty were stronger, as anticipated in hypotheses H3 and H6, respectively. The correlation coefficients of 0.230 and 0.215 shown in Table 3 are both significant at p = 0.02 while the correlation coefficients for low task variability and low task difficulty are not significant. These direct effects are not surprising in the sense that formal management accounting systems can be more objectively quantified to handle the information-sharing function of activities that are identifiable and programmable, even in the face of new technological advancements, such as CAD/CAM and activity-based costing. Changing or adding new management accounting systems under increasing task variability may be a more efficient response for the firm than under increasing task difficulty, consistent with Galbraith’s (1977) arguments. But our results also complement findings reported by Brownell and Dunk (1991) in that the broad focus of management accounting system design under increasing task difficulty is to facilitate the more arduous specification of input/output relationships. Moreover, finding strong direct effects of MACSs change on MRI for departments facing high task variability and high task difficulty is consistent with theoretical arguments that these two uncertainty constructs operate at the work-unit level rather than the individual level (Van de Ven and Delbecq, 1974; Brownell and Dunk, 1991). However, definitive support for hypotheses H3 and H6 can only be claimed if the path coefficients themselves are significantly different between the low and high levels of task variability and task difficulty considered separately. In other words, are the slope coefficients from equation (1) the same or different in each case? Support for a difference in the path coefficient p21 was found (via the Chow test) only between the low and high task difficulty sub-samples but not for the task variability sub-samples. This suggests that providing MRI via MACSs change is more advantageous under conditions of high task difficulty, thus reinforcing the support for hypothesis H6. Despite these results though, no support was found for hypothesis H4, or H7, which predicted that the link between MRI and departmental performance would be stronger under both high task variability and high task difficulty, respectively. All linkages from MRI to departmental performance were highly significant, but no differences for the path coefficient p32 emerged (via the Chow test) between the low and high sub-samples for either task uncertainty variable. Once again, these results suggest that the initial benefit of changes in MACSs in our sample is to produce and communicate relevant information to managers. As anticipated by all normative accounts, it appears that once managers have the appropriate information they utilize it to enhance the operational objectives of their departments without regard to the nature of task uncertainty.