The Uncertain Meanings of Risk. Calculative Practices of Rating Agencies As Ordering Devices of Financial Action

Based on Frank Knight’s distinction between uncertainty and risk, the idea that economic action is pursued under fundamental uncertainty has been at the center of economic sociology – serving not only as a basic micro-sociological assumption, but also “as the vantage-point” (Beckert 1996:817) for the sub-discipline. But while much work has been done in economic sociology to understand the institutions, networks, and cultural artifacts that help dealing with uncertainty, the concept of risk itself has often been underappreciated. Risk is often regarded as a fiction of neo-classical economics, one that overestimates the extent of the calculative abilities of market actors and the calculability of future events as such. But declaring risk a fiction neglects that this concept is a central category of knowledge in financial markets. Financial actors themselves do not start with the assumption of fundamental uncertainty and are therefore not concerned with such things as “social mechanisms” that reduce uncertainty. Instead, they work very hard to increase the calculability of future events. In fact, part of the financial logic – and a fundamental precondition of financial markets – is precisely the actors’ assumption that the future can be brought to the present by probabilistically calculating it. For financial actors, (financial) calculation has thus more than just a legitimizing or “appeasing” function to “conceal” their inability to predict an inherently uncertain future. It is the central ordering device that guides financial action.

Therefore, even if we as social scientists believe (rightly or wrongly) that risk and its calculability are a fiction, we need to engage in studying the concepts of risk and the practices of calculation that market actors use to “transform” (perceived) uncertainty into (perceived) risk. Risk (and not uncertainty) has become a “powerful organizing category” (Power forthcoming) of financial markets. A disciplinary division putting risk in the realm of economics and uncertainty in the one of sociology might reproduce the “Parsonian” division between both fields. Just as the determined and enthusiastic agenda of economic sociology has been to reclaim the economy as their territory of study, the concept of risk in financial markets should be reclaimed by sociologists as well. And instead of just challenging the assumptions and calculative practices of financial actors, sociology should treat them as central socio-cognitive phenomena that constitute financial markets and therefore study them empirically.

This paper studies calculative practices of risk assessment as social phenomena that shape the current conceptions of risk of financial market actors. My theoretical argument is that calculative practices are inherently intertwined with the knowledge they produce. Through the decisions of which aspects of a phenomenon are measured and which ones are left out, what is compared, and through the specific calculative manipulations, the object that is calculated is created in practice (Kalthoff 2005). It ceases to be an abstract idea and becomes specific and concrete. Therefore, in order to understand the de facto meaning of a concept in practice, it is necessary to look at the calculative practices that construct it. Calculative practices of risk assessments therefore are the concrete manifestations of ideas of risk in financial markets, which fundamentally shape financial (and more generally: social) action.

To better understand the different conceptions of risk and the corresponding risk calculations prevailing in financial markets, I turn to credit rating agencies (CRAs), which attempt to assess credit risks. By studying CRAs’ methodological publications and interviewing rating analysts at Moody’s and Standard&Poor’s, I identify two fundamentally different methodological approaches for producing ratings, which I call the “diagnostic” and the “technical” approach. While the former is merely descriptive and does not claim predictability, the latter is completely based on probabilistic statistics, with a claim to predictive capabilities. These two epistemic cultures fundamentally shape the respective conceptions of credit risk both that co-exist and compete within CRAs, and which I describe in an ideal-typical sense (challenging typical sociological dichotomies, such as those by Luhmann (2005), Giddens (1990), or Castel (1991)): The first one is a “risk-as-danger” conception of risk. Risk is seen as opposed to security, and as an only partially calculable and predictable problem or hazard that should be avoided or minimized. This conception of risk corresponds to the everyday use of risk (Lupton 2013). The second one is a “risk-as-opportunity” conception of risk. This conception implies that risk should not be avoided but managed and exploited. It sees risk as something calculable, controllable and manageable and as an opportunity to gain profits. This latter conception of risk corresponds to the general financial logic (Wigan 2009): The main concern in this perspective is thus not to reduce risk, but to manage it. This means that the main challenge is to measure risk, to be able to calculate the “right”, “correct” or “fair” compensation for investing in a future endeavor. The measurement of risk therefore has the purpose of calculating the “risk-adjusted yield” of an investment.

This research shows that different conceptions of risk are not only possible, but can also co-exist within a single financial organization. The calculative practices of CRAs shape the conceptions of risk, depending on how it is calculated. The assessment of risk is not a simple and undisputed or purely ‘technical’ process. The practices of assessing risk are necessarily inscribed with specific ideas, epistemological assumptions, and different normative implications. Calculative practices of risk assessments are therefore the concrete manifestations of ideas of risk in financial markets, which fundamentally shape financial (and more generally: social) action.