The effects of uncertainty on the roles of controllers and budgets: an exploratory study

Abstract

This paper explores how contextual uncertainty and the use of the budgetary system explain cross-sectional variation in the organisational role of business unit controllers. We argue that there are complementarities between the role of the budgetary control system (i.e. coercive vs. enabling) and the role of the controller (i.e. corporate policeman vs. business partner). Thus, we explore both the direct effect of uncertainty on the role of the controller and the indirect effect through the role of the budgetary control system. Using survey data from 134 business unit controllers, we find that uncertainty provides a partial explanation of the variation in the role of budgetary control systems and in the role of controllers. In particular, our data suggest alignment between the coercive (enabling) use of the budgetary control system and the role of controllers acting as corporate policemen (business partners). These findings add to our understanding of the functioning of business unit controllers within their organisational context.

Keywords:

• budgeting
• controllers
• management accounting
• uncertainty

Acknowledgements

We appreciate the help and/or comments from the editor, two anonymous reviewers, Paolo Perego, Sergeja Slapnicar, Marcel van Rinsum and participants at the 2011 EAA annual meeting and a workshop at the University of Ljubljana.

Notes

Throughout this paper, we use the job titles ‘management accountant’ and ‘controller’ interchangeably.

Two other styles of use described by Burchell et al. (1980) and Hopwood (1980) are use as an ammunition machine and use as a rationalisation machine. If the system is employed as an ammunition machine, organisational participants use it to manipulate information cues that enter decision-making processes to promote their own particular positions. If it is employed as a rationalisation machine, managers use accounting figures ex post to legitimise and justify their actions and decisions (Burchell et al. 1980). These styles of use may be considered non-functionalistic in the sense that they reflect accounting's role in the organisational politics surrounding decision-making processes instead of a potentially value-adding activity.

In addition, the original classification of Hopwood (1972) has been criticised for its lack of analytical sharpness. Recently Hartmann et al. (2010) showed that the evaluation styles may be multidimensional and comprise both general leadership characteristics and the choice of (non-)financial performance metrics.

While we assume that the emphasis placed on alternative controller roles represents a design choice made by the organisation, we also acknowledge that in some organisations individual controllers might have some discretion in choosing how to enact their alternative roles.

Chapman (1998) did not explicitly differentiate between types of uncertainty, as uncertainty was related to the strategic choice of the firm. The firm showing the highest performance had chosen an entrepreneurial strategy that led to both high task and environmental uncertainty. In this firm, the interaction between controllers and line managers appeared to be the highest of the four firms investigated.

This project also provided the data for Maas and Matějka (2009). The items used to measure the controllers' business partner role in our study were also used in Maas and Matějka (2009). Environmental uncertainty, task uncertainty and interdependence served as control variables in that study, which focused on the controller's engagement in data misreporting.

The ‘Controllers Instituut’ (CI). See www.ci.nl for a more detailed description, in English, of the CI and its activities.

We drop controllers who work for multiple units because they might play different roles in different units. Given the items that we use to measure our variables, this would make it difficult to interpret their responses.

The same test using all 186 questionnaires received also indicated that there was no non-response bias.

The instrument used by Chapman and Kihn (2009) only measures enabling control and does not address coercive control. Their instrument comprises 20 items that reflect various managerial behaviours that the budgeting system evokes, in four categories. Note, however, that their hypotheses refer to the global construct of enabling control rather than its four underlying dimensions. The global construct is the focal point in their analysis.

We use Smart PLS 2.0 (Ringle et al. 2005).

The descriptive statistics refer to construct scores that were calculated as the mean of their indicator scores and not to the scores obtained by PLS. PLS calculates construct scores as a weighted average of standardised item scores and, as such, the PLS means are not very informative as descriptive representations of the sample data.

The single exception is the AVE for the corporate policeman role, which is 0.30.

We thank one of the reviewers for this suggestion.