Nuclear risks are rising globally. The Russian invasion of Ukraine in February 2022 provides a stark reminder of how quickly tensions between governments can turn into war. Advances in psychology and neuroscience now allow for a much better understanding of human decision-making, especially under high-stress conditions. Policymakers in some areas – such as healthcare, social security or taxation – have already begun to incorporate these learnings into how decisions are made, how leaders are trained and how policy changes are communicated to the public. But nuclear weapons policy has so far remained fairly insulated from this discussion. As part of Chatham House’s programme of work on complexity in nuclear decision-making, this research paper explores the implications of behavioural insights for nuclear policy and decision-making.
Shortcomings in human decision-making across fields from personal finance to health are well documented. Humans are easily swayed by short-term gains over long-term ones; they tend to take the easiest option available and are inclined to follow social norms. Policymakers as individuals, and institutions that bring together individual preferences and biases, unsurprisingly also demonstrate these behaviours.
Humans are easily swayed by short-term gains over long-term ones; they tend to take the easiest option available and are inclined to follow social norms.
The field of behavioural insights attempts to address some of these shortcomings. It has focused on interventions targeting general populations, such as in climate change, economics and public health, as well as on improving decision-making environments to achieve better outcomes. For instance, the introduction of default enrolment into workplace pension schemes in the UK drastically increased pension savings among those not previously saving for retirement. There is a growing interest in understanding how behavioural insights can be applied by policymakers and their institutions to help develop solutions to the behavioural shortcomings that might impede effective decision-making in other areas. Indeed, the role of institutions as ‘behaviour change machines’ – designed to shape the behaviour of those working within, or interacting with, them to achieve institutional goals – is growing as practitioners search for the most effective policy solutions.
In nuclear weapons policy, the concept of the ‘rational actor’ continues to hold a great deal of power. The security of the international system in the nuclear age has rested on theories of nuclear deterrence developed in the early Cold War that are believed to work if leaders behave rationally. The rational actor model assumes that decision-makers will make choices after careful deliberation of their full suite of options and then choose the pathway that best aligns with their objectives. But critiques of rational actor theories have pointed out that incomplete information environments and psychological factors pose challenges to the assumption of the rational and clear-headed leader who operates with complete information and is capable of making calm decisions in a crisis.
There are numerous factors which make good decision-making in nuclear weapons policy particularly difficult. Nuclear weapons policy features, on the one hand, decades-long struggles to ratify treaties and ensure compliance and, on the other, the most acute time pressures surrounding crisis decision-making with civilization-defining consequences. Neither of these extremes are optimal for human decision-making. The stakes are extremely high and the opportunity for learning and practice low. Most challenging of all, nuclear weapons policy necessarily operates internationally, requiring policymakers to communicate across cultural and linguistic barriers to reach a common understanding and get close to a shared form of rationalism.
Despite these immense challenges, this paper is intended to provide a constructive basis for discussion about where positive changes in nuclear decision-making at all levels may be possible. As has been demonstrated in other policy areas, an empirical approach to human behaviour can be used to better understand failings and to find ways to work around them. Each of the paper’s four sections covers facets of nuclear policy where both the room for growth and the appetite to do better have been identified. One feature of behavioural insights is that it points to a wider mindset shift about how to approach decision-making across all levels of policy, whether in the initial stages of policymaking or in implementation. Each section concludes with recommendations that could be implemented across different levels of the nuclear policy apparatus, as well as ideas and case studies of how such changes could be made.
Methodology
This project used a qualitative research methodology, consulting with 35 former officials and experts from 13 different countries covering a wide range of different positions on nuclear policy. The group was consulted in three ways: in a group setting, testing and discussing ideas and case studies; through in-depth research interviews with a selected sub-group of former officials from different nuclear weapons states, delving into detailed questions on nuclear decision-making; and by asking a separate sub-group of former officials for detailed feedback on an earlier draft of this paper.
To narrow down the wide range of potential nuclear policy areas to the four areas discussed in this paper, the project team convened an initial scoping roundtable in May 2022 that brought together policymakers and experts from 14 different countries representing a wide range of policy positions on nuclear policy. The roundtable discussion answered initial questions about the concept of behavioural insights and explored initial areas of nuclear decision-making where they might be applied. Following the roundtable, the project team conducted additional desk research and expert consultations to investigate the most promising areas to explore in greater detail. The project team conducted eight in-depth consultations ranging in length between one and two hours. The eight people consulted were former officials from nuclear-armed states who had spent their careers in the nuclear decision-making architecture and had held high-level decision-making positions. The consultations provided insight into several decades of nuclear decision-making, and allowed the project team to identify areas of agreement or disagreement, as well as areas in which the experts thought there was room for improvement in decision-making structures or mechanisms. This series of consultations was supplemented by a feedback workshop with a larger group of experts – comprising 10 former officials and experts, representing eight different countries – to review an early draft of this paper.
Following the research and consultation phase, the list of topics for inclusion was narrowed down to four:
- Overconfidence. Several of those consulted cited examples from their own careers of when overconfidence had posed a problem in certain areas of nuclear policy. Examples mentioned included absolutist beliefs in both the strength of deterrence and the ability to control escalation in the event of conflict. The relevant section includes ways to improve how decision-makers assess their state’s policies and capabilities to avoid overconfidence.
- Miscommunication. Everyone consulted expressed concern regarding the ability of policymakers to communicate clearly the intent of nuclear decisions. Clear signalling and communication are perennial security policy challenges, especially during times of heightened international tension. The section discussing miscommunication introduces some best practice recommendations, as well as new behavioural insights to help improve the clarity of communication.
- Errors. The US experts in the sample group were able to point to US government learning from past errors as one way to improve nuclear policy implementation. However, not all nuclear weapons states have been equally transparent about how they have addressed similar challenges. Experts from other states pointed out that it was important to be specific to the political and strategic context of each state, as those states differ in the extent to which they have implemented some of the recommendations in this section. This section shares best practice for dealing with nuclear safety and security, and discusses the value of transparency.
- Public and political salience. UK and US experts both highlighted the relationship between public and political salience, pointing to cases of treaty ratification in which public mobilization helped sway elected officials’ positions, or in which protests played a role in influencing policy. This point is particularly important at a time when public awareness of nuclear weapons risks was rising again, due to Russia’s nuclear threats against Ukraine. The relevant section includes possible strategies for engaging both the public and politicians on nuclear issues.
In addition to the individual and group consultations outlined above, the project team used extensive desk research and case studies to identify areas where behavioural insights could be applied. As behavioural insights are intended to make the choice architecture of decision-makers more transparent, the tools suggested in this paper could potentially be applied to other cases.
Limitations
Discussing the methodology also requires transparency over limitations. One of the project’s main limitations is the small number of well-documented, unclassified cases in which policymakers’ struggles with nuclear decision-making can be studied closely. Thanks to a range of declassified documents on both the Soviet and US sides, the Cuban missile crisis of 1962 has become the go-to case for studying various aspects of nuclear crisis decision-making. The thoroughness of the documentation available, the duration of the crisis and its varied components make it a useful case study.
However, it is important to acknowledge that lessons from the Cuban missile crisis cannot be overgeneralized. The episode must be seen in context: it was one of the first nuclear crises, taking place when the Soviet Union and the US were still defining many aspects of their relations and the implications of their nuclear-armed status. It was also specific to the political moment, marked partly by the make-up of both governments and partly by the interpersonal relationship between Soviet premier Nikita Khrushchev and US president John F. Kennedy. The leaders’ personalities, previous interactions and assumptions about each other played a large role in how the negotiations unfolded. The states’ military capabilities were also a significant factor: both states have since pursued parity policies and expanded their nuclear arsenals in ways that affect their deterrence relationship. Other nuclear crises would therefore look different in terms of advisers’ concerns about second-strike capabilities.
It must be noted that the dynamics between countries and different regional contexts are complex and culture-specific, and the particularities of each nuclear weapons state must be considered. The decision-making of the Cuban missile crisis cannot therefore simply be expanded into a ‘how to’ guide to preventing future nuclear crises.
Several of the recommendations in this paper assume that those making certain nuclear decisions can be trained to improve their resilience to stress or, simply, to make better decisions. Some former officials consulted for this project recounted the difficulties they had encountered when trying to train high-level officials, who sometimes had the attitude that they did not require additional training. The former officials consulted made the point that it would be easier to implement the training requirements at lower levels of government, and to put in place incentives to engage senior decision-makers. Such incentives could include access to the highest-level decision-maker or time spent with their peers that they would not want to miss out on. These insights point to the role of culture in implementing recommendations: if a certain institutional culture considers training or feedback to be a weakness rather than a strength, or believes training to be unnecessary beyond a certain level of seniority, that culture makes change difficult to achieve. Further research could compare institutional cultures across different nuclear weapons states as potential catalysts or inhibitors for change.
