Conclusions and further directions for research
In this paper, we have combined the literature on enterprise risk management with research on boundary objects in order to examine how heterogeneous tools, processes and networks of organisational actors can give rise to a seemingly coherent and stable working ensemble, which provides a shared context for risk identification and management in two large organisations. Drawing on the comparative analysis of the two cases, we shed light on what we call the dynamics of (dis)integrated risk management to stress how the ideals of integrated risk management, mobilised by practitioners in the two case-study organisations, are both subject to self-undermining pressures. By uncovering these dynamics, we extend previous work that is critical of standardised forms of enterprise risk management (Hopwood, 2009; Miller et al., 2008; Power, 2007), as well as more recent work on ‘risk talk’ (Mikes, 2016) and ‘dual risk management’ processes (Kaplan & Mikes, 2016). We also contribute to recent studies in the accounting literature that use the notion of infrastructure to explain how vague policy objects become tractable across different organisational sites and institutional arenas (Power, 2015). This paper offers a theoretically-informed analysis of the context-specific development of enterprise risk management in two large organisations. In doing so, this paper may prompt further academic research in at least two areas. Firstly, further work might explore the conditions under which ‘risk talk’ can be sustained over time as an unobtrusive form of risk management that is seamlessly linked to the work of business managers. Compared to previous work (Mikes, 2016), we document potential sources of instability for ‘risk talk’. The commonalities and points of contrast between this study and previous research could help future work to produce cumulative and generalisable insights into the functioning of interactive-rich forms of risk management in different settings. Secondly, further work could be done to examine the dynamics of infrastructure formation. While recent accounting research seems to emphasise the enabling, generative role of infrastructure (Kornberger et al., 2017; Power, 2015), this study suggests that infrastructure is more ambivalent, serving as both an engine and a barrier to integrated control practices. Future work might further explore this apparent tension in the functioning of infrastructure.