تمرکز زدایی تصمیم گیری در استراتژی های مدولار: غلبه بر موانع در سیستم های حسابداری ناکارآمد

Research on product modularity is dominated by an operations management (OM) perspective, through which numerous models to predict optimal modularization strategies have been developed. However, we argue that implementation of these predictions is hampered by prevailing project accounting systems which distorts the economic effects of modularization at the level of the individual product. This has the implication that decisions on modularization can only be made by top management if decision authority and relevant information are to be aligned. To overcome this problem, we suggest a solution that aligns the descriptions of the economic consequences of modularization at the project and portfolio level which makes it possible to decentralize decision making while making sure that local goals are congruent with the global ones in order to avoid suboptimal behaviour.

Rapid product development is key to sustainable competitive advantage, both in terms of satisfying the seemingly ever increasing customer demand for variety and to keep competitors off the market. Product development is, however, a rather resource demanding activity, and speeding up this process will therefore typically incur more costs. In addition, even though empirical results are somewhat ambiguous (cf. Anderson, 1995), it is generally assumed that increased product variety has a negative impact on company costs (e.g. Miller and Vollman, 1985, Banker et al., 1995 and Kaplan and Cooper, 1998). With increased product portfolio heterogeneity the company will typically have to source, manufacture, and sell in smaller batches and the support functions of the company will typically have to be expanded to accommodate the increased internal demand for activities such as planning, change over, quality testing, etc. Fisher et al. (1999) suggest two general approaches to handle increased costs caused by increased variety. One approach is process oriented and the other product oriented. Whereas the process approach focuses on introducing flexible manufacturing equipment, the product approach is concerned with designing products that make it possible to offer variety at low costs. Product modularity, i.e. using the same components (modules) in different products and thereby allowing for variety at low costs (Ramdas, 2003), is one example of the latter approach. The concept of modularity varies across literature (Fixson and Clark, 2002, Fixson, 2005 and Salvador, 2007), however, we use it equivalently to component sharing (Ramdas et al., 2003) and component commonality (Labro, 2004). Product modularity is not a new phenomenon (see Starr, 1965 and Starr, 2010). However, it seems to attract growing attention. Recent years have seen a rising number of publications on this (e.g. Kim and Chhajed, 2000, Baldwin and Clark, 2000, Nobelius and Sundgren, 2002, Salvador et al., 2002, Doran, 2003, Zhou and Grubbström, 2004, Antonio et al., 2007, Doran et al., 2007, Brun and Zorzini, 2009, Brun et al., 2009, Jans et al., 2008 and Gomes and Joglekar, 2008) and it is claimed that companies across industries increase their use of modularity (e.g. Sanchez, 1999 and Kim and Chhajed, 2001). Indeed, according to Heikkilä et al. (2002) modularity has become one of the most discussed issues in product development, technology, and business management. As argued by Labro (2004), modularity or commonality has primarily been researched from an operations management (OM) perspective, whereas there are only few contributions from management accounting researchers. This is somewhat surprising considering the apparent relevance for this domain, given the fact that surveying the relevant literature, it is evident that product modularity is supported by an economic rationale. OM research has, for example, shown that modularizing companies may enjoy increased economies of scale (Krishnan and Gupta, 2001), reduced demand for support activities (Thonemann and Brandeau, 2000) and reduced investment in tools (Fisher et al., 1999). Other more soft approaches have identified reduced assembly time due to learning curve effects (Nobelius and Sundgren, 2002), and increased market share (Nobeoka and Cusumano, 1997 and Sanderson and Uzumeri, 1995). The arguments in this paper are developed from a management accounting perspective, but based on insights from OM research. Ramdas et al. (2003) present three organizational structures for approaching modularity: the coordinated projects approach, the project-by-project approach and an intermediate partially coordinated approach, i.e. a mix of central and local decisions. They develop a model that predicts the coordinated projects approach to offer the most cost-efficient path. Ramdas et al. (2003) suggest that model implementation may be challenged by designers who perceive the use of such a model as a distraction from more important development work. In this paper we add another challenge. We argue that the implementation or continued support of the optimal strategies predicted by the OM models will encounter difficulties in practise—not because they are formal or because they must necessarily rely on a set of assumptions, but because companies’ accounting systems will act as barriers to implementation. We base our discussion on Ramdas et al. (2003), but we believe that our claim is generally valid across OM models—irrespective of the correctness of the model because if the accounting systems used in practise do not provide information that supports the OM model prediction, implementation will be impeded. This is the case in many companies where the accounting system is set up to account for the individual projects’ performance. Unit costs of the common components probably differ from the cost of the unique component being substituted—to the benefit of some projects and detriment of others—and development cost of the common component is split between participating projects in a heuristic manner. Consequently, the economic consequences of modularization are not described adequately in the individual projects’ accounts and decisions to modularize or not should be based on the aggregate consequences across all participating projects only. This has the implication that decisions must be made by top management if decision authority and decision information are to be aligned. In this situation all projects should be coordinated centrally, corresponding to Ramdas et al.’s coordinated project approach. This is probably infeasible, since “[a]ttention is the chief bottleneck in organizational activity, and the bottleneck becomes narrower and narrower as we move to the top of the organizations” ( Simon, 1976:294). The objective is therefore to change the project accounting system so that it does not act as an obstacle to decentralization and in such a way that it encourages project managers to make local decisions in congruence with global goals. We develop the boundaries of such a system and within this a rule-based profit and cost allocation approach that makes this possible. The remainder of the paper is organized as follows: Section 2 presents a review of OM literature on economic decision making for modularity. In Section 3 we develop a changed project accounting system that makes decentralization of modularization decisions possible and illustrate the idea by an empirical example in Section 4. In Section 5 we discuss some practical challenges and Section 6 concludes the paper.

We have argued that OM models on modularity (like Ramdas et al. (2003)) will encounter difficulties in getting implemented in practise. Dysfunctional accounting systems will act as barriers in that they at the level of the individual project distort the projects’ contribution to product portfolio as a whole when common modules are at play. Typically cost-efficient modules through prevailing heuristic cost allocations benefit some projects but hamper others that actually contribute to the modules’ cost efficiency. This has the consequence that relevant information is only found in the aggregate and at the organizational top. This prevents decentralized decision-making and project account responsibility. To solve this we identified the boundaries within which allocations rationally can take place so that profitability at the project and portfolio level is aligned. Within these boundaries we developed a rule-based allocation procedure that encourages project managers, on their own initiative, to cooperate in an attempt to exploit profitability of modularization. We also found that to maintain responsibility accounting and the benefits of accountability and motivation inherent in this management technique, it can be necessary to adjust the cost allocations between cost-hierarchy levels to accommodate additional financial performance measures used. Our rule-based allocation model has some limitations. First, we have not explicitly taken into account situations where the modularization effort is conducted across company jurisdictions, i.e. cooperation in modularization between independent companies in networks or supply chains. However, in some contractual situations it becomes simpler, for example, in situations where the supplier quotes different unit prices of the common module as a function of the degree of functionality and volume. This simplification entails that no allocation of development cost is involved then; and this can be dealt within the current procedure. The optimal degree of modularization will be different to in-house modularization though, since some degree of profit is allocated the supplier. In other contractual situations, for example, where more suppliers and various supplier activities are paid for separately (e.g. development cost) more research is needed. Another limitation is that our procedure is static, in the sense that we look at simultaneous product development projects. Lack of dynamism is also referred to by Fisher et al. (1999:312 ff.) stating that the “challenge of component sharing is increased as the decision is viewed dynamically. In most industrial situations, there already exists a portfolio of products and the managerial problem is to decide which components to reuse, which components to replace, and which new components to develop. This problem is complex and deserves further research attention”. Our model is not sensitive to decisions concerning reuse and replacement of existing components/modules, though—past development costs of these are ‘sunk’ and thus irrelevant to decisions and allocations. However, our model is sensitive to module needs of future products beyond the company’s product portfolio planning horizon. One way forward might be to incorporate real options theory since the flexibility achieved by modularization has a value which can be estimated using this theory (e.g. Baldwin and Clark, 2000 and Gaynor and Bradner, 2001). We acknowledge the logic and complexity captured in the OM models on component/module sharing and the insights they bring to management’s decision-making processes. At the same time, though, we propose further research into the organizational intricacies in which these models are to be implemented. Such research will have to bridge theoretical domains, as we have tried to bridge OM and management accounting in this paper. We believe that other domains, like organization, finance and marketing, may also benefit from bridging with OM—and there is certainly room for more bridging from the management accounting domain too.