Denis Villeneuve’s Dune: Part II, the sequel to his 2021 Dune film, released to box-office triumph and critical acclaim. The films are based on the popular 1965 science fiction novel Dune by Frank Herbert, which tells the story of a young prince, Paul Atreides, and his battle to control the universe’s all-powerful spice resources. While Dune may seem merely a fictional space epic, it contains fascinating parallels with a RAND approach to policy analysis.
In Herbert’s universe, spice acts as a hallucinogenic that unlocks the power of the mind. It facilities interstellar trade and travel by expanding human consciousness, enabling navigation through the vast uncharted regions of space. This makes spice immeasurably valuable. The galactic emperor sends young Paul Atreides and his family to govern the hostile sand planet Dune, home to the empire’s sole spice reserves. However, shortly upon arriving on Dune, the Atreides family is ambushed by political rivals: Paul’s father is assassinated, the Atreides army decimated, and 15-year-old Paul is cast into the desert to die an exile.
The desert of Dune is entirely inhospitable, filled with radical warriors, intolerable heat, and colossal, subterranean sand worms. Paul seems destined to die. Yet the desert has other plans. As Paul wanders through the desolate wilderness, the planet’s spice unlocks his mind: he is struck with visions of the future, often blurry and contradictory. In some visions, he dies in the desert. In others, he retakes control of Dune and avenges his father. Armed with his visions, Paul plots his re-ascendancy. He turns his competing visions of his future into strategic assets. By considering all prospective futures, Paul charts a prophetic-like path through an otherwise perilous environment; a path to lead to him safely out of the desert and back to the throne.
Years after Herbert’s authorship, but well before Villeneuve’s movies, RAND scholars Steven Popper, Rob Lempert, and Steven Bankes formalized a Dune-like approach to policy analysis. Traditional approaches to policy analysis attempt to statistically predict the future and then prescribe appropriate policy remedies. But this “predict then act” paradigm becomes increasingly problematic when faced with deep uncertainty. Deep uncertainty is characterized by wickedly complex environments, fluctuating or unknown variables, and models that rely heavily on fundamental assumptions. Under these conditions, predictions can be inaccurate, misleading, or at worst, flat wrong. To solve this dilemma, RAND researchers pioneered a new approach to decisionmaking under deep uncertainty. They called it Robust Decision Making (RDM).
RDM assumes the future is unpredictable and must be met with “robust” strategies, impervious to a wide range of possible conditions. RDM works for policy analysts in much the same the way spice worked Paul: it envisions a vast range of possible futures (tens, hundreds, or even thousands) and identifies a policy strategy that performs satisfactorily across many of these futures. Such a strategy is robust in the face of deep uncertainty. Where conventional approaches attempted to predict the future conditions, this methodology would illuminate successful policy strategies regardless of external circumstances. Inspired by Dune, RAND economist Steven Popper named one of the early RDM computer modules “Spice,” a method that would make it seem as if policymakers could prophetically traverse through seemingly unnavigable situations.
The RDM methodology is straightforward. First, frame the objectives, inputs, outputs, causal relationships, and relevant policy levers in a model. Then, simulate a multitude of outcomes by varying the model’s assumptions: what happens if the inputs change dramatically? What happens if the causal relationships are different than assumed? What if you enact one policy, but refrain from others? The final step of RDM is to identify strategies that are systematically successful across these futures, regardless of external conditions and model assumptions. RAND has applied the RDM methodology to a diverse range of policy issues, including climate change, transportation, and natural resource planning. It even has applications for long-run policymaking (see Shaping the Next Hundred Years).
In the same way that Paul Atreides sifted through his visions of the future in search of answers, RDM sifts through model assumptions and inputs to determine the most robust policy approach.
In the same way that Paul Atreides sifted through his visions of the future in search of answers, RDM sifts through model assumptions and inputs to determine the most robust policy approach. RDM prioritizes many “what could be” scenarios over a single (likely erroneous) “what will be” prediction. It circumvents traditional analytical challenges by varying foundational premises and simulating an array of possible futures. Instead of asking, “What will the future look like?” RDM asks “What should we do, regardless of what the future looks like?” It is not about making better predictions; it is about making better decisions. The result is policy that is effective in both good and bad times. In a world of increasing geopolitical, technological, and ecological volatility, RDM is more necessary than ever.
Dune is science fiction. But it reminds us that even in the real world the future is always uncertain. Conditions of deep uncertainty require novel approaches. Just as spice allowed Herbert’s characters to traverse the vastness of space and the dangers of the Dune, RDM empowers the world to steer safely through unprecedented times. It provides a map to navigate the unknowable future, a pathway through the desert of uncertainty.
