Radical Uncertainty

John Kay and Mervyn King, an economist and a former Bank of England governor, argue that economics and finance have systematically misapplied probability to situations where it does not belong. Their distinction is between resolvable uncertainty, where the possible outcomes are known and probabilities can be attached, and radical uncertainty, where the future contains possibilities nobody has enumerated.

Their historical argument is that Frank Knight and John Maynard Keynes both understood this distinction in the 1920s, and that the profession lost it when Savage's subjective expected utility framework made it possible to attach probabilities to anything. Once every uncertainty could be treated as a probability distribution, the distinction disappeared from economics, and models built on it were applied to questions they could not answer. They treat the risk models used before 2008 as the clearest case: value at risk calculations produced precise numbers about events whose distributions were unknown and unknowable.

Their alternative is narrative reasoning. They argue that people facing radical uncertainty use reference narratives, meaning coherent accounts of what is going on and what is likely to follow, and that the right question is not what is the probability but what is going on here. They defend heuristics and rules of thumb as rational responses to non-stationary environments rather than as biases, which puts them at odds with the behavioural economics tradition. Critics respond that they conflate different targets, that they underrate what probabilistic models achieve within their proper scope, and that reference narratives are not a method anyone can apply consistently. The critique is stronger than the alternative.