Applying Power’s (2003b) notion of calculative cultures sheds some light on the underlyingcurrents that resulted in these particular risk management and control patterns. Calculative idealism and calculative pragmatism describe two extreme attitudes of users towards the products of highly quantified models, which are shrouded in analytical mystique. In the context of risk management, calculative idealists believe in the power of risk models and given sufficient data quality, they regard the outputs of the models as economic representations of risk, suitable for limit setting and control.
Calculative pragmatists, on the other hand, typically regard numbers as only attention-directing devices with no intrinsic claims to represent reality. Risk models are useful only when they reveal trends in risk for management purposes and help to steer behaviour in theright direction. In order to make risk numbers a basis for control, calculative pragmatists areready to tamper the numbers produced by models with judgment and intuition.
Calculative idealists and calculative pragmatists are ideal types, in practice one finds peopleholding a mixture of these views (eg, short-term pragmatists being long-term idealists etc), aswell as moving from one attitude towards the other (eg, as they rise in the organizational hierarchy or change occupational groups). However, in order to determine the remit of risk management and the corresponding risk management methodologies, the senior risk officers interviewed had to decide to what extent they regarded risk numbers as representing economicreality.
The interview material suggests that BWT’s senior risk officers regarded the risk numbers asattention-directing devices with little intrinsic claims to represent reality. They saw their contribution in helping to steer line management behaviour in the right direction. Further, in a control setting where conflicting risk and return objectives wrestled with one another, seniorrisk officers sought to direct the attention of decision makers to issues that warranted priorityat any given time.
At Fraser it appeared that risk officers displayed a great deal of calculative idealism in thatthey aimed to represent the cost of true economic capital based on high quality data and they worried constantly about the ‘robust’ and ‘hard’ nature of their risk analysis. This calculative idealism was challenged in the process of allocating the Group’s capital to business units. It took a second team of risk capital controllers to realise that calculative idealism had to be combined with political shrewdness. Fraser’s own organisational culture imposed a constant consensus-seeking behaviour on decision-makers. Hence in the process of risk capital allocations risk people’s calculative idealism was toned down to the extent that the next economic capital team was willing to compromise their preferred technique for the sake of reaching a compromise between the competing risk-return interests within the group. Nevertheless, by involving ‘experts’, the language of these negotiations remained technical . Different Economic Capital allocation methodologies wrestled with each other, representing the different risk-return interests of business units – until a final compromise was achieved. The quasi-technical process of decision making by compromise (Burchell et al 1980) gave risk capital allocations sufficient credibility to make it integral to strategic planning and performance management at Fraser.
The existence of different calculative cultures, based on the evidence presented here, is detectable in the attitude of senior risk officers to the outputs of risk management models in use. Apparently, senior risk officers develop personal philosophies about the ‘manageability’ of risks by quantitative models. While there appears to be much consensus on the manageability of certain risks (eg, market risks), the issue of non-quantifiable risks and that of internal capital allocation are contestable. BWT’s senior risk officers who had doubts about the use of quantitative models in these contested locales chose to define their area of competence broadly (encompassing risks outside the quantifiable risk framework). Conversely, senior risk officers at Fraser who had more confidence in the reliability of the risk models were able to make them work in the contested locale of capital allocations and performance measurement. However, by doing so, they confined their area of influence to thatof measurable risks. This underlines the role of managerial discretion (at least in part) in the selection and use of ERM systems.