Uncertainty: A Diagrammatic Treatment Sheila Dow
Economics Special Issue on Radical Uncertainty and Its Implications for Economics
In what follows we use the term ‘uncertainty’ in the Keynesian sense to signify low confidence in expectations, regardless of probability (quantifiable or not). This definition potentially includes uncertainty about quantitative probabilities, or ambiguity, which has been an increasing focus of the mainstream analysis of uncertainty in the wake of the crisis. But ambiguity is a special case of uncertainty, to be distinguished from fundamental uncertainty (or ‘radical’ uncertainty), which is the outcome of an absence of quantifiable cardinal probabilities.
A non-dualistic understanding of uncertainty
Integral to Keynes’s [descriptive] theory of decision-making under uncertainty is the role of psychology (Dow 2011). Further he argued that it is conventional to give more credence to extrapolation from past trends than we know is reliable. It may therefore be conventional, under fundamental uncertainty, to act as if there was a preferred model and a relevant set of information, with uncertainty relevant only in the sense of ambiguity. But since the underlying subject matter does not yield correct models or notions of completeness of information, this convention is vulnerable to being confounded by actual developments.
The scope for instability implied by this epistemology is moderated, not only by conventions, but also by institutions which perform a cognitive role (Dequech 2000). Indeed, considered over the evolution of society, institutions have evolved precisely to provide a grounding for decision-making under uncertainty. Uncertainty is thus endogenous (Dow 2014).
Uncertainty is characterised as varying in degree, from mere risk to fundamental uncertainty. It is multidimensional. (Mathematically, one can say that a probability P(A|B) depends on the context, so one can have P(A|B)=p being a definite probability, uncertainty about the value of p or its stability, or uncertainty about the context or its stability.)
Risk, uncertainty and the crisis
The crisis was … seen as … the inevitable outcome of the ever-increasing fragility of the financial system in the preceding period of apparent stability. … Nevertheless the timing of the crisis and the particular event which would reveal the extent to which pricing had been based on over-confidence were both subject to fundamental uncertainty.
But, while the mainstream approach presumes a return to [‘normal’ risk], the Keynesian approach cannot presume a return to [a similar situation]. … Fundamental uncertainty remains relatively high.
The absence of more radical solutions follows from the mainstream understanding of stability as the norm ensured by freely operating markets and the identification of increased uncertainty with [exogenous] shocks.
But from a Keynesian perspective transparency may in fact be highly damaging. … A change in regulation is required to allow practices which sustain confidence in expectations.
From a Keynes/Minsky perspective, instability is the norm and the incidence of crisis cannot be prevented, because of the nature of financial markets. Given fundamental uncertainty, there are no true prices to act as benchmarks, to which markets can return after a crisis. Yet conventional judgements build up, often in defiance of reason and evidence, which fuel instability. The first focus of policy therefore has to be vigilance in monitoring financial markets for signs both of unreasonable confidence in expectations and increasing fragility (Dow 2014). The focus of policy addressed to moderating any tendency for increasing uncertainty would involve promoting stability through appropriate design of practices, conventions and institutions.
It is urgent that policy address the potential for uncertainty to aggravate continuing instability. But until the different ways of analysing uncertainty are acknowledged and understood, the policy discourse will be mired in confusion.
The last paragraph is critical, but I have some quibbles about some of the rest. In my understanding stability is the norm, if only for the reason that it what most people expect and that expectation tends to promote stability of a kind. But from time to time there will be critical instabilities (such as those identified by Minsky). Looking forward, there will be many possible future stabilities, sometimes including a continuation of the existing one. What Dow describes is that we tend to see the current apparent stability and to rely on it more and more, until we have an unforeseen crisis leading to psychological uncertainty and (hopefully) a new ‘world order’. But it is not clear that a new, stable, order must always emerge. Perhaps we could enter a sustained period of instability. It seems to me that, in a sense, the problem that Dow describes is one of increasing apparent stability and order, so the solution may be a sustained period of moderate disorder, while having enough regularity / order / stability for societies (including businesses) to thrive. Where business is supposed to thrive on creativity, it seems to me that each epoch should encourage creativity very widely, challenging the status quo. Thus the stability will have radical ‘shocks’ that are exogenous to its current mainstream understanding, but fostered by its general nature. It seems to me that transparency is only highly damaging if one is relying on ‘closing down’ debate, which would be even more damaging in the longer term.