Book Reviews

by Amalia Frommelt

Radical Uncertainty: Decision-Making Beyond the Numbers provides individuals a narrative-oriented framework for addressing risks in a radically uncertain environment. Artificial intelligence has several attractive applications in today’s complex decision-making environment, but John Kay and Mervyn King warn against the abandonment of human judgement. Ultimately, human and non-human approaches to risk management will likely carry biases and repercussions of their own. (Photo credit: Kyle Glenn on Unsplash)

John Kay and Mervyn King advocate for embracing uncertainty and human judgment over artificial intelligence in Radical Uncertainty: Decision-Making Beyond the Numbers.  The authors, both British economists and policymaking elites, are well-equipped to provide insight on policy making under conditions of uncertainty; King served as Governor of the Bank of England for a decade, and Kay served as a member of Scotland’s Council of Economic Advisors. They argue that “a single approach to rational choice under uncertainty has dominated economics,” and that continuing this tendency will be problematic for policymakers (Kay and King, 64).

Kay and King’s introduction makes three propositions. Firstly, economics and other social fields aren’t governed by unchanging scientific laws; they are instead “non-stationary.” Secondly, it is normal and rational for a human to not “optimize.” And thirdly, “humans are social animals and communication plays an important role in decision-making” (Kay and King, 16-17).

Central to the book’s argument is the distinction between uncertainty that can be described probabilistically versus the uncertainty “which surrounds every unique project or event” (Kay and King, 14-15). Probabilistic uncertainty can be thought of as a puzzle which has “well defined-rules and a single solution,” whereas the latter approach to uncertainty can be thought of as a mystery. which offers “no objectively correct solution” (Kay and King, 20-23). Kay and King acknowledge that while puzzles may be more fun, our everyday lives increasingly deliver ill-defined mysteries with unknowable outcomes. 

The authors sample various anecdotes and case studies in the first section of the book, which are primarily underpinned by Merton’s concept of “reflexivity” as a distinctive property of the social system. Kay and King indicate that “reflexivity undermines stationarity,” given that “the system itself is influenced by our beliefs about it” and serves as a justification for their first proposition regarding the changing nature of social systems (Kay and King, 36). The book’s second section provides another comprehensive recitation of historical and contemporary cases where radical uncertainty went unacknowledged, despite probabilities being objectively unhelpful. For example, President Barack Obama considered his decision to approve the raid by U.S. Navy SEALS on Osama bin Laden’s compound to be a coin-flip; he reflected on this as a situation where he received  “probabilities that disguised uncertainty as opposed to actually providing you with useful information” (Kay and King, 9). 

Radical Uncertainty is a product of a complex, late-modern world. In her book, Risk, sociologist Deborah Lupton argues that today, “change is rapid and intense and the number of choices to be made have proliferated,” which ultimately “renders choice-making very difficult and fraught with uncertainty” (Lupton, 143-171). Correspondingly, Kay and King assert that since reflexivity undermines stationarity, behavioral models should be phased out by risk-based approaches in decision-making, given that humans often deviate from expected outcomes provided by behavioral models.

However, they fail to identify the consequences of incorporating risk-based approaches. Lupton’s discussion on risk discourses helps to fill many of the gaps left open by Kay and King, who fail to identify the implications of applying risk-based approaches in decision-making. Risk-based policy gives individuals the impression that risk can be eliminated through risk-averse tendencies; however, those tendencies might be less effective in reducing risk and more effective in providing a false sense of security. For example, some individuals have fallen prone to the ritual of handwashing as an attempt to exert all of their agency to prevent sickness, when in reality, a multitude of factors are required to ensure one’s health besides obsessive hand-washing. 

Risk discourses depict “humans as far more able to exert control over their environment,” so in a late-modern society that is underpinned by reflexivity and constantly confronted with novel risks, advantageous policymakers are able to exploit those fears rather than create policy that addresses root issues (Lupton, 143-171). Proposals for a wall to be built on the U.S. border with Mexico infamously reveal the implications of subjective risk-based policy. Donald Trump’s campaign inflated perception of risk at the border to justify installing a border wall, which had no proven ability to reduce the flow of migration (Bier). 

Kay and King’s partiality towards reflexivity and uncertainty is commendable but naive as they fail to address the potential fallout of adhering to risk-based approaches; Lupton would warn that individuals “develop and exercise reflexivity in response to expert knowledge” rather than in response to their own perceptions and experiences with risk (Lupton, 143-171). 

Kay and King’s second proposition claims that humans deviate from optimization and outcomes predicted by probabilistic methods like Bayesian rationality. The authors argue that humans are well postured to deal with radical uncertainty because it is a byproduct of human evolution, creating evolutionary rationality—the basis of human judgments and actions (Kay and King, 154-176). As coined by influential German historian Leopold von Ranke, “evolutionary rationality” provides humans the ability to describe “wie es eigentlich gewesen ist,” or “how it really was.” 

Kay and King’s third proposition identifies humans as “social animals” who require forms of communication to make decisions, such as narratives. Narratives can provide a conceptual framework that goes beyond rational probabilistic calculations, whether that be in the field of economics or foreign policy. Radical Uncertainty provides a conceptual framework for applying these narratives and advises economists to diversify the data that they use to make decisions. Ultimately, Kay and King fail to explain why decision-makers continue to adhere to rigid economic models in the first place. While the authors succeed in asserting that even the most rational set of models can be subjective, they also fail to identify the impact that narratives have on unquantifiable pathologies of risk that distort perceptions and drive individuals to act obsessively or habitually in an attempt to eliminate a risk that can not be eliminated (Rothstein et al.). 

The book’s argument is premised on the rationality of the individual, but given the authors’ backgrounds, it would be more appropriate to discuss the rationality of policymakers who create risks to justify policy, rather than create policy to address risks. Ulrich Beck’s book, Risk Society, asserts that modern society’s infatuation with identifying, assessing, and preventing risk ultimately creates more risks to address in the future (Beck). In the same way that an over-dependence on rational behavioral models can be dangerous in decision-making, prioritizing the creation of policy based solely on risk can become subjective and distort our perceptions at a societal level. For example, while the War on Terror was launched to address the risk of terrorism and to protect Americans post-9/11, the policies allowed the U.S. government to commit many human rights and civil liberties violations, in turn creating new risks for American citizens who became subject to racial profiling and surveillance (Costs of War). 

Kay and King’s suggested framework for decision-making takes a step away from quantitative methods but still carries the underlying biases, motives, and assumptions associated with risk management. This oversight could be because Kay and King are elites themselves; as key players in British policy-making, they fail to recognize their role in contributing to our “risk society” as the individuals who prescribe risk-based policy to their constituents, furthering society’s unrealistic infatuation with eliminating risk. 

Radical Uncertainty: Decision-Making Beyond the Numbers provides a nuanced and robust history of the over-quantification of decision-making. The authors isolate the notion of uncertainty and encourage it within the decision-making process. As a result, they fail to provide a deeper understanding of the other drivers that contribute to the difficulty of decision-making, particularly risk. Kay and King’s suggestion to live with uncertainty is a helpful point of departure, but individuals and decision-makers alike should remain wary of suggestions provided by elites whose work continues to contribute to our risk society.

*********