Hallsworth’s (Reference Hallsworth2022, Reference Hallsworth2023) manifesto is both a worthwhile endeavour for behavioural scientists and important for policy practitioners who deploy interventions in behavioural public policy (BPP). It is timely to take stock of BPP, especially its light-touch version, often called nudge. As stated in the manifesto, these policies have become very popular in recent years without there being a great deal of stocktaking about their general nature. At the same time, nudge has attracted criticism. To ensure the research and policy agenda matures, it is important to take on board the limitations of behavioural science and how it has been applied. By adopting new principles and practices, behavioural science can move ahead, building on past successes. That is what the manifesto promises. But how far does it go?

To engage in an effective critique, the underlying social and political theory behind an approach in public policy and public management needs to be acknowledged and incorporated into the argument. Theory is taken to mean that claims about causal relationships between individual decisions and the structures in which they operate make sense in terms of reasons for individual and group action. Then there are a set of expectations about how institutions guide those actions. Coherence gives theories their cutting edge in terms of offering explanations for social and political actions that are based on a few interlocking elements. Their clear internal logic can help them explain a phenomenon, usually with only a few moving parts, offering parsimony even at the cost of some simplification (Hollis, Reference Hollis1977; Ostrom, Reference Ostrom1990). Lack of coherence would imply that they make incompatible claims and, as a result, they do not have traction even if they might wish to take account of complex relationships.

One candidate for a set of explanations, which has particular relevance for innovations in public policy, is cultural theory. It provides a guide or framework to show how individuals operate in different contexts and offers an explanation of social phenomena, with the advantage of being one step removed from the particular cultural models of change themselves. Introduced into public policy by the anthropoglot Mary Douglas (Reference Douglas1986), cultural theory was developed by Christopher Hood in his The Art of the State: Culture, Rhetoric and Public Management (Reference Hood2000), which followed an earlier essay (Hood, Reference Hood1995). Public management, in particular public management reform, is Hood’s core interest and where he sees the traction of cultural theory. Just in case the link between public management reform and BPP is not seen as obvious, the introduction of nudges into choice architectures and other tools of government, with the idea of securing public policy gains, is a form of public management. BPP would not be able to exist without the bureaucratic and political structures that authorise the interventions and then ensure they are implemented and learnt from. Just like any public management innovation, BPP needs to fit within existing structures and practices and may even be part of an exogenous change in how the state is organised and may bring about organisational reform and a difference in how knowledge is used in the policy process. As a tool of public management, BPP needs a theory of change embedded within it, then advocacy to be able to survive within public organisations, assisting the process of diffusion and entrenching support for BPP in the long term. It is this last issue that the manifesto is most concerned about.

Hood deploys a version of the grid–group theory so as to set out different kinds of worlds or ways associated with public management reform. A grid means ‘the extent to which interactions and relationships are defined by general norms and rules accepted in advance’ (Hood, Reference Hood1995, p. 209). It is about the form of the internal organisation of bureaucracy as to whether hierarchy is stressed or more decentralisation and bottom-up governance prevails. Group is about how society interacts, whether it is connected or not; ‘Group means the extent to which organization members are separated from the society at large, or from one another’ (ibid). This is about the connection of bureaucracy to society and implies a link to common projects, such as in a more egalitarian set-up in Hood’s interpretation. Variations from low to high on these dimensions create different typologies of public management, which are displayed in a two-by-two table in Table 1. First consider the high grid and the low group. This is cell A, which Hood calls the fatalist way, where there is lack of co-ordination and planning. There is little attempt to reconcile society to organisational aims. It tends to atomise organisations, but they remain rule-bound. Then there is the high grid and the high group in cell B, the hierarchicalist way, where there is top-down control but with social purposes entrenched within this pattern of management. This approach can be effective but can suffer from the blind spots of hierarchialism. In cell C, the low grid and the low group create the individualist way, which are characterised by markets, decentralised control and a lack of concern for structures and societal outcomes, at least not directly. It can be an innovative form of decision-making, but there is a lack of central control and linkage to societal projects. Finally, the high group and the low grid create cell D, the egalitarian way, where there is strong participation but weak co-ordination and control. These are bottom-up initiatives that flourish and are not controlled in a formal set-up, such as being part of an official hierarchy.

Table 1. Hood’s typology of public management according to cultural theory

Adapted from Hood (Reference Hood1995, p. 90).

Hood sees these ways as in competition with each other, offering alternative paths to public management reform. Each way generates benefits that may, at first, seem alluring, but from which internal contradictions can emerge that come from imposing its vision and purpose too strongly. Managers in cell B can get things done and organised with social purpose, but too much control can lead to a loss of intelligent information gathering and bottom-up resistance to higher-level commands. There is an endless cycle of reform as policy-makers and reformers move from one way to another seeking to overcome its limitations with the advantages of a new approach that, in turn, has its own internal contradictions. The natural reaction to this scheme might be that policy-makers should aim for a middle position that is the best of all worlds, getting the most from each model, such as with the modernisation/convergence of public management, popular ideas in the 1980s and 1990s. Here Hood is at his most trenchant when offering a prediction and recommendation: he says there must be no ‘management for all seasons’ (Hood, Reference Hood1991, Reference Hood2000, p. 20) as this is not a stable proposition. It faces its own internal contradictions that are even worse than the ways of the four cells, which at least have the benefit of initial internal coherence. This expectation applies to any framework that offers to reconcile all the criticisms of other approaches and imposes a bland ‘all things to all persons’ approach. It is probably obvious where this argument is going with respect to the manifesto, but before setting that out it is useful to see how typology fits the extant paradigms of BPP. This is attempted in Table 2.

Table 2. Models of the implementation of BPP according to cultural theory

Cell A, the fatalist way, represents the state of behavioural science before the nudge revolution ushered in by Nudge (Thaler and Sunstein, Reference Thaler and Sunstein2009), and reflects individual and piecemeal initiatives carried out by local agencies and disaggregated central organisations. Such an approach was picked up in a report for the British Academy, which observed how BPPs were adopted patchily in various places, especially locally and in particular sectors, but without much overall impact and buy-in from agencies (John and Richardson, Reference John and Richardson2012). Cell B is the classic bureaucratic framework of hierarchical control harnessing social purpose, which is the essence of BPP, and can incorporate behavioural interventions, such as budges, not orientated to human autonomy (Oliver, Reference Oliver2013). Here all the tools of government become informational so within reach of BPP (John, Reference John2013). Nudge helps the tools get more traction. In spite of BPP’s commitment to light-touch interventions, its practice is essentially pragmatic: design policies for human welfare, evaluate them with randomised controlled trials (RCTs), and then redesign these tools using a ‘what works’ approach. The claim to increasing human autonomy in the libertarian paternalist framework (Sunstein and Thaler, Reference Sunstein and Thaler2003; Thaler and Sunstein, Reference Thaler and Sunstein2003) is often regarded as wide of the mark in achieving autonomist goals. Unobtrusive defaults and prompts can be just as freedom-limiting as any law or financial instrument (Sugden, Reference Sugden2009, Reference Sugden2018; Anderson, Reference Anderson2010) and, in fact, may even be covert and thereby more manipulative than the standard techniques of command and control (Banerjee and John, Reference Banerjee, John, Sunstein and Reisch2023).

Cell C relaxes both the grid and the group, which has the advantage of allowing innovations bloom, within a more decentralised framework and willingness to adopt new ideas. Contracting out and quasi-markets often feature in this approach. It describes the early phase of the UK Behavioural Insights Team (BIT), for example, which followed an entrepreneurial governance model within the British state (John, Reference John2014), naturally leading to the spinout into the non-profit and contracting organisation it is today. The entrepreneurial twist to BPP accounts for the private sector’s interest in nudges and the diffusion pattern to its spread (John, Reference John2019).

Cell D relaxes grid but introduces group more strongly, so that management practices reflect decentralisation and citizen buy-in for social purposes. This more radical approach to BPP stresses agency, bottom-up decision-making and citizen-driven approaches, which are characteristic of deliberation and citizen assemblies (John et al., Reference John, Cotterill, Moseley, Richardson, Smith, Stoker and Wales2019). Additionally, it includes light-touch forms of engagement alongside nudge, such as nudge+ (Banerjee and John, Reference Banerjee and John2021, Reference Banerjee, John, Sunstein and Reisch2023), with its addition of a think to the nudge. This approach seeks to overcome the ethical limitations of libertarian paternalism and its connection to top-down paternalism.

Where does the manifesto fit within this scheme? The most plausible answer is that it is a manifestation of a very traditional form of BPP as represented in cell B. To deal with the problem of BPP, a strategy of intensification is proposed whereby more knowledge and science are applied to public issues. Control problems are addressed by better intelligence. Effectively BPP is not doing its job properly by thinking in terms of silos and not seeing the wider pattern of causal interactions. It must ‘see the system’. It needs to do RCTs better, such as by replicating studies and building up scientific knowledge more effectively. RCTs need to be more adaptable and extensive.

Even though manifesto aims to change the ambition and scope of BPP, the form of decision-making appears to remain much the same as the organisation and authorisation of standard BPPs. Hallsworth does not say much about the politics of BPP, such as who decides these policies, but surely that is important in addressing the problems he raises, in particular the lack of understanding of system processes. One way to do that is to relax the grid while keeping or lessening the group so as to release more intelligence in the system. It may be that the problems he identifies are a consequence of the bureaucratic and hierarchical structure in which BPP emerges. By exercising more control, the implementation of the manifesto might exacerbate the contradictions within the hierarchicalist way.

Hallsworth does move a bit into cell D when he writes about the need to acknowledge marginalised voices and communities, but the recommendation is not central to his argument and the wider question of who organises BPP is not addressed. He needs to be clearer in his overall view of BPP, so it is not in his words ‘a view from nowhere’ that is decontextualised from the very communities he wishes to include. The bureaucracies and political principals who sponsor BPP do not figure in the manifesto, so it is not clear how these marginalised voices are going to be incorporated other than from the wishes and orientations of these commissioners who on the whole need to work within the ‘what works’ framework or in response to wider political considerations. There is nothing in the manifesto that signals an institutional change within which BPP operates. Where is the agency that is going to implement the change in perspective? What would motivate decision-makers to change their behaviour as they have their own biases and habits, just like citizens have? Without dealing with these institutional and wider political-organisational considerations, Hallsworth cannot fully move into cell D, which means that the incantation of ‘management for all seasons’ might be invoked as the manifesto is not coherently placed in one cell but is bestride at least two. This form of concept expansion (cf. Sartori, Reference Sartori1970) is also signalled by the text on organisation embedding (p. 315), which goes down the innovation and private sector route of cell C, if only tangentially.

In spite of these cross-overs of cultural categories, the main problem of the manifesto is not so much the critique of management for all seasons, but that it does not depart sufficiently from the constraints of cell B which is hierarchy-mobilised public management orientated for social purposes. The decentralisation and entrepreneurial themes are not strongly articulated. They are mainly add-ons. The implication of the manifesto is that no matter how many of the recommendations are implemented, BPPs can be commissioned in the same way as during the classic period of the upswing of interest in nudge. Only now, they can be done more effectively with attention to the complexity of social systems and by overcoming the limits to knowledge that arise from relying on just a small number of RCTs. For that reason, the manifesto probably does stand up as a coherent piece even if it nods towards decentralisation (cell D) and entrepreneurialism (cell C). As a result, it might still suffer from the tensions within cell B. There are limits to hierarchicalism or top-down governance, which can lose its legitimacy and shut out innovation. Even though the manifesto aims to be radical, it offers a standard approach to BPP.

It could be the case after all that there is nothing to fear from management for all seasons. Embracing contradictory tendencies is just what policy-makers and even researchers have to do. They need to ‘contain multitudes’ in the words of Walt Whitman (Song of Myself, 51). Working through these contradictions could be a source of creativity and innovation. A complete move to bottom-up governance in cell D and the entrepreneurial governance of cell C are probably not possible, given the current set of institutions and political practices. But what can be done is to moderate the tendency to conservatism in cell B. These ideas appear in the manifesto even if they are not developed. Rather than criticise Hallsworth for defending management for all seasons, he does not offer enough of the all-seasonal approach. In fact, the appearance of the work of the author of this comment in all four cells suggests that public management for all seasons can be developed, even if critiqued. The nudge+ framework is just such a hybrid wanting all the benefits of nudge and BPP while at the same time seeking citizen engagement and participation. The same can be said of other agency-enhancing interventions, such as boosts and debiasing (Banerjee et al., Reference Banerjee, Grüne-Yanoff, John and Moseley2024). Thinking through how power is exercised within institutions, and who makes BPP and other policies, should be an important element to any BPP prospectus or manifesto. In that sense, moving towards quadrant D is a good aim for everyone working in the field.