Groupthink, Polythink and the Challenges of Decision-Making in Cabinet Government

The Chilcot report set out in detail its ﬁnding that the Blair Government had been prone to groupthink in its decision-making processes when leading Britain into the Iraq War. Subsequent British prime ministers have been in no hurry to change their style of governing in ways that might broaden decision-making circles and introduce the ‘challenge’ that Chilcot said had been lacking. This article draws on the literature on the psychology of group decision-making to examine the extent to which groupthink remains embedded in the processes of cabinet government in the UK. The article argues that the strongest driver of groupthink is the psychological disposition towards conﬂict of individual prime ministers. Drawing on interviews with ministers, civil servants and special advisers we suggest that the political authority of the prime minister interacts with their psychological predis-position towards debate to encourage groupthink, polythink or a more positive style of ‘vigilant’ decision-making.

The premiership of Tony Blair marked a turning point in British politics. The size of his landslide victory in 1997 brought a change not just in the fortunes of the Labour Party, but in the processes of government decision-making. At first sight the clear, united, centralised leadership model offered by New Labour looked impressive. For Labour, the apparent reduction of conflict was presented as part of the change to a more modern form of government. 1 But, as was subsequently outlined in painstaking detail by the Chilcot Report into decision-making 1 In reality there were also, of course, many substantive continuities from the Thatcher and Major years in underlying administrative processes, including the Next Steps programme, contracting-out and civil service modernisation. on the Iraq War, there was a flaw built into this new mode of government. 2 It tended towards 'groupthink', characterised by centralisation of decision-making in the setting of 'sofa government' and the controlled exclusion of dissenting views. Despite the shortcomings identified by Chilcot and others, this new model has shown itself to be remarkably resilient. Subsequent prime ministers have been in no hurry to diminish centralised control by broadening decision-making circles (see, e.g. Bennister and Heffernan, 2012 on Cameron).
How groupthink operates in decision-making at the centre of British government remains under-analysed. The term 'groupthink' has transcended the academy to become part of popular jargon, but there have been relatively few attempts to examine the psychological dynamics of collective decision-making in Westminster parliamentary democracies. Groupthink has become a shorthand way of criticising an apparently over-compliant cabinet, which implies that what is really being attacked is the over-mighty behaviour of individual prime ministers. In particular, this article focuses in on the relationship between the psychological make-up of individual prime ministers and the decision-making regime that is built around them. We ask two inter-related research questions: (1) What factors contribute to groupthink in cabinet government settings?
(2) How do these factors interact with the psychology of individual prime ministers?
We suggest that the structures and practices of modern cabinet government have increased the likelihood of groupthink occurring. This is especially so within Westminster systems of government, where majoritarian traditions prevent the institutionalisation of wider negotiation sites that are an embedded part of the coalition-based parliamentary systems in much of Europe. 3 This article focuses on the UK as an exemplifying case, where the combination of increased centralisation of power and a presidential focus on the prime minister has coincided with decreased opportunities for 'challenge' within the cabinet process. 4 This makes the UK an ideal site for researching how the traditions of cabinet decision-making interact with the personal preferences of prime ministers perceived as having the power to enforce their own choice of decision-making structures.
Crucially, we argue that the overriding precondition towards groupthink is a psychological one. The personal attitude of the prime minister towards conflict governs the shape of decision-making processes. It interacts with cabinet structures in ways that can discourage the kind of 'culture of challenge' envisaged by Chilcot. This occurs because, in the malleable traditions of cabinet government, the structures of decision-making are geared to reflect the individual preferences of the prime minister (see Hennessy, 2001;Weller, 2018). It is the prime minister who ultimately determines which cabinet ministers they meet with and when, and how much power to give to special advisers. While there are restraints imposed by tradition and continuity, it is ultimately the prime minister who determines the size of both formal and informal decision-making groups, and the processes that they will utilise. In order to understand the structural underpinnings of groupthink we must examine the processes of cabinet decision-making through the psychological disposition of prime ministers themselves (see Kaarbo, 2018).
To unpack this argument, we begin by positioning the discussion of cabinet decision-making within the wider psychological literature on small group interaction, in particular the work of Irving Janis. Based on this material, and through abductive interaction with our empirical data, we construct a typology of group decision-making in cabinet contexts. The typology makes clear the connections between the processes of government, the political authority of the prime minister and their degree of psychological comfort with challenge within the cabinet process. The empirical section sets out the views of ministers, civil servants and special advisers on the nature of cabinet decision-making over the last four decades. The discussion section draws out these insights further, mapping them against the ideal types of groupthink, vigilant decision-making and polythink.

Applying psychological theories of decision-making to the cabinet process
A great deal of attention has been paid in recent decades to examining how prime ministers in numerous countries have centralised power around themselves at the expense of both the parliament and the rest of the executive. In the British case, such analysis has focussed specifically on the growth of Number 10 Downing Street. There is strong evidence that Number 10 has indeed gained power at the expense of departmental ministers (Foley, 2000;Poguntke and Webb, 2005), although intense debate continues about the true extent of that power shift (see Burch and Holliday, 2004). The trend towards a stronger centre has coincided with a more choreographed style of decision-making, tilting the psychology of cabinet government away from open debate and towards the avoidance of conflict. While no prime minister could ever hope to prevent all conflict, one feature of the choreographing is that prime ministers who seek to prevent conflict will shift dissenters out of larger meetings where they might achieve backing into smaller meetings where prime ministers will be able to suppress dissent more effectively.
Psychologist Irving Janis (1972) broadly identifies three styles of group decision-making: groupthink, polythink and 'vigilant'. While these are ideal types, they reflect three different underlying approaches towards group discussion. Janis argued that groupthink results when overly cohesive groups become prone to suppression of dissent, resulting in lack of consideration of alternative courses of action. If a leader is too dominant, other members of the group may get into the habit of suppressing their own views in order to conform to what they believe he or she wishes. The 'groupthink' that results manifests as a failure to weigh up decisions from all angles. This can, in turn, lead to perceptions of policy failure (see King and Crewe, 2013, pp. 255-267;Barr and Mintz, 2018).
Groupthink has been shown to emerge particularly problematically in crisis situations and foreign policy contexts, where the presentation of a united front is seen as important ('t Hart et al., 1997;Hermann, 2001;Verbeek, 2017). Especially during crises, voters reward clear messaging and unambiguous decision-making as being signs of strong and decisive leadership (Boin et al., 2009). Perceptions of decision-making deadlock can be disastrous (see Kowert, 2002). However, 't Hart argues that governments adopting a 'command centre' view can in reality mask an array of sub-optimal psychological machinations. 'The mere fact that with groupthink there is at least a unified, predictable policy emerging from the often conflictual and erratic corridors of power is less important than the high likelihood that this policy is seriously flawed in its assumptions and reasoning' ('t Hart, 1998, p. 310). This exposes the dilemma that contributes to the persistence of groupthink in cabinet decision-making: the apparent stability and unity of purpose that accompanies groupthink brings political dividends, while simultaneously leading to potentially sub-optimal policy outcomes. It can leave prime ministers having to decide between good politics and good policy.
Groupthink in effect anchors one end of a spectrum when considering group decision-making. The opposite extreme to groupthink is 'polythink', where the number of different views is so diverse that it leads to 'intragroup conflict, a disjointed decision-making process, and decision paralysis and inaction as each group member pushes for his or her preferred policy action' (Mintz and Wayne, 2016, p. 4). Polythink offers less room for dysfunction to hide. Unfortunately, any benefits this might offer in terms of transparency are offset by perceptions of chaotic decision-making. To voters, this looks like a cabinet that does not have its act together. Rather than the premature imposition of a group consensus on individuals too ready to fall into line, polythink leaves groups desperately divided between individuals and factions unwilling to move towards agreement at all (Mintz and Wayne, 2016). In a cabinet situation, this manifests as a tug of war in individual directions, rather than a debate leading to a shared position. The much-publicised travails of the British cabinet in its struggles to deliver Brexit under Theresa May provide an excellent case study of what polythink can look like in practice.
Like all spectrums, there remains space between polythink and groupthink for a middle ground. In particular, Janis identified an ideal type of decision-making which he characterised as 'vigilant decision-making ' (1989, ch. 5). Vigilant decision-making relies considerably on group leaders making a conscious effort to include people with different views (see Kelman et al., 2017, pp. 246-247;Barr and Mintz, 2018). They set a tone where members feel that they can express their viewpoints openly. The group reaches what is likely to be a more effective decision 'after adequate search, appraisal and planning manifest in absence of defects like selective bias, [and] omission of alternatives' (Janis, 1989, p. 91). Vigilant decision-making involves a fine balance; enough freedom to express divergent views but also enough convergence for a decision to be reached and polythink avoided. There is a constant interplay between collegiality, leadership and the structural processes of cabinet decision-making (Vercesi, 2012).
To operationalise the insights from the groupthink literature and apply them to the analysis of British cabinet government, we propose a typology ( Figure 1) that seeks to capture the interaction between the leadership style of a prime minister and the government's decision-making process. It suggests that the prime minister's authority within their own government interacts with their psychological preferences and leadership style to lead to particular types of decision-making. The tendency within modern government to strengthen the centre at the expense

Groupthink in Cabinet Government
of the periphery, with a concomitant concentration of decision-making advice in smaller groups, builds in a tendency towards groupthink. Should the prime minister lose their authority for any reason, leading to a weaker centre of government, this groupthink can very quickly deteriorate into the equally undesirable polythink.
The typology in Figure 1 illustrates that the prime minister's personal preferences on cabinet processes interact with the politics to encourage particular decision-making outcomes. When prime ministers are in dominant positions within their own government and party, their strength makes either groupthink or vigilant decision-making possible. Which will prevail depends on the leader's willingness to allow room for challenge and the conflict it creates. Where a leader is predisposed to avoid conflict, and/or deliberately limits decision-making to small, 'trusted' groups, groupthink is more likely to occur as tight decisionmaking structures keep 'challenge' within a small, inner circle. When prime ministers are in a less dominant position within their own party, groupthink becomes less likely. But what replaces it is a form of ineffective decision-making as more decisions are debated by more voices more often. When that larger group of voices loses discipline and the leader is forced into relationships of open distrust with colleagues, polythink results. We analyse these categories more fully in the empirical section that follows.

Methods
Cabinet as an institution relies for its survival not on its formal, legislated structure-because none such actually exists-but rather on the continued willingness of Cabinet ministers, advisers and officials to buy into a set of traditions about what it means to be bound by collective decisions (see Rhodes et al., 2009, pp. 78-112). This article is based on interviews conducted from an interpretivist, 'actor-centred' perspective (see Rhodes, 2017) and analysed using thematic analysis (Braun and Clarke, 2013). We conducted interviews with key actors in the cabinet system in order to explore inductively their thoughts on collective government. This enabled us to look past the formal position someone holds to examine more deeply how their beliefs on what cabinet government is help to actually shape how it works. Following Rhodes, 'an interpretive approach argues it is not possible to identify people's beliefs by appealing to the allegedly objective social facts about them. Instead, we must explore the beliefs and meanings through which they construct their world' (Rhodes, 2011, pp. 2-3). Placing the interpretivist lens at the heart of the research design enabled us to see cabinet government through the eyes of its participants.
We sought the views of a range of actors. We collected these views from three separate sources. First, we draw on bespoke semi-structured interviews (n 14) with a range of cabinet ministers, special advisers and senior civil servants who served under prime ministers Thatcher, Major, Blair, Brown or Cameron. We chose these three categories of interviewee-rather than simply cabinet ministers alone-to reflect the fact that modern cabinet processes rely heavily on all three groups of people. Where respondents accepted the invitation to participate, they were interviewed by both authors in face-to-face settings, with interviews recorded to ensure accurate transcription. Interviews were conducted on the promise of anonymity, and we have therefore anonymised specific quotes and attributed them only to the category of person as a minister, adviser or civil servant.
Secondly, these interviews are supplemented by a range of in-depth interviews now on the public record, including the set of BBC interviews (n 5) with individual cabinet secretaries and a publicly available dataset of interviews with former Cabinet ministers (n 45) compiled by the Institute for Government. Thirdly, we triangulated these sources with memoirs, speeches and press reports for the period under-study to create a diverse picture combining how the actors see themselves with how they have been perceived by others and in contemporary scholarly works assessing this period in British Government.
We focus here on the extent to which the sources show whether prime ministers encouraged divergent views, both through their personal style and the particular processes of decision-making they put in place. We present our findings analytically, based on the factors that most align with the characteristics of groupthink, polythink and vigilant decision-making. These include the process, personnel and politics involved, alongside the dynamics of group size.
As the literature on core executive studies has now argued for nearly three decades, power resides not in someone's position title but in their capacity to shape events. This waxes and wanes over time. So too do the attitudes of different prime ministers and their cabinets to the business of collective decision-making. No single prime minister will necessarily stay within the 'vigilant decision-making' quadrant throughout their time in office. Temporal factors, the flow of events and electoral pressures may see them retreat into less consultative modes of governing at different points.
The focus of our empirical research was on the cabinet system as a whole, including both formal and informal meeting sites. This includes full cabinet and cabinet committees, but also the informal meetings and networks that affect decision-making in formal fora. In this article, we use the term 'cabinet process' to include all of these interactions, each of which have an impact on the nature of cabinet deliberation and decision-making. We structure our empirical material below by firstly setting out the ways in which structural decisions on the size, composition and processes of group-decision making can contribute towards groupthink. We then analyse how the psychological preferences of individual prime ministers interact with these structures to either encourage or discourage particular styles of decision-making.

Structural factors-group size, composition and processes
The literature on the psychology of decision-making in groups indicates that group size is one of the most important variables. The sweet spot where groups have a wide enough membership to have a diversity of views represented, but not so wide as to become unwieldy, is somewhere between six and ten (Kelman, Sanders and Pandit 2017, p. 253). This correlates strongly with the size of 'war cabinets' and some cabinet committees but juxtaposes awkwardly against the much more unwieldy cabinets of peace time. As one senior civil servant told us: First thing you have to remember is there's never been such a thing as cabinet government as a collective unit. The [full] cabinet is too big. No serious organisation runs on the basis of 20 people around the table. (Our interview, 2018) As a result, every prime minister, whatever their psychological preference, is initially faced with the fact that the full cabinet is simply too large to be effective as a debating body. What is more, the UK cabinet has in fact grown rather than shrunk in recent decades. Thatcher, Major and Blair each had about 23 cabinet Ministers. Brown started the 'two tier' process of 'associated members' on top of the core 23, who were invited to attend when relevant; there were nine associated members by the end of his time in office. Cameron was forced to expand further in the coalition years, with up to 30 in total entitled to attend.
With modern government often characterised as too fast-paced for traditional processes, smaller groups seem to offer more nimble options for making faster decisions. 5 The unwieldy size of full cabinet has led prime ministers and their advisers to have the substantive discussions elsewhere. One adviser from the Blair era noted that issues should be 'thrashed out' outside cabinet processes because you could not do it 'in a useful way' round the cabinet table. 'The human dynamics of discussion in large groups' mean 'you will not have an honest discussion' (Our interview, 2018). In this adviser's view, full cabinet is 'not a debating chamber. . .it's a device that allows you to keep collective responsibility, and for people, when they really feel strongly on something politically, to make a statement, to do something'. 5 The bypassing of full cabinet is reflected in the length of its meetings; Blair's were much shorter than Thatcher's, Brown deliberately held meetings longer than Blair's at the start of his premiership to signal a more open approach, but attempted to shorten them again as time went on (Mandelson, 2010, pp. 442-443). Theresa May would have liked to keep them short but was sometimes forced to hold protracted 'away-days'. Among our interviewees, multiple advisers and ministers felt that the cabinet process had in reality become a 'show'. They believed that decision-making now happens largely in small informal groups and bi-lateral meetings. However, officials and some ministers, from both the Labour and Conservative parties, thought the system needed to be reformed to reboot the formal process. They believed a key benefit of the formal process was that, by having fixed membership, it made it more likely that divergent views would be expressed. They offered two proposals, both attempting to bring the group size closer to the six-ten optimum: some kind of super cabinet, or better use of formal cabinet committees.
One former senior civil servant said, 'it may be too late for this', but he would propose a 'smaller cabinet' of about six for 'strategic leadership or government'. Each of the six would co-ordinate a number of departments. The cabinet office would expand to brief the six. This civil servant believed the 'older system', even with increased use of the formal cabinet committees, could no longer be done 'sensibly' in part because of lack of trust within the full cabinet. There is at least one historical precedent for a peace-time super cabinet. A cabinet of 'overlords' similar to the World War II war cabinet was briefly resurrected by Churchill in 1951. Hennessy and Welsh (1998, p. 69) cite the then Cabinet Secretary, Norman Brook, as objecting to the plan partly because it was 'inconsistent' with the principle of individual ministerial accountability and partly because it was contrary to Cabinet traditions that all members were equal.
The other oft-cited option involves increasing the use of formal cabinet committees, which was one of Chilcot's suggestions. These committees are more workable in size than the full cabinet but retain the formal authority of the cabinet structure to add weight to their deliberations. As one civil servant put it, 'what doesn't have to be a myth is the idea that a smaller group of ministers make the decisions on issues under the ultimate authority of the prime minister' (Our interview, 2018). The utility of cabinet committees as sites where vigilant decision-making can flourish depends not only on whether the prime minister regards them as important but also on who does the chairing. Theresa May for instance, chaired many committees herself, with the capacity to thereby manage the debate within them and increase her 'positional power' (Allen and Siklodi, 2016;Dunleavy, 2018). In contrast, under the Blair and Brown governments the then Deputy Prime Minister John Prescott frequently took the chair of cabinet committees, allowing the airing of divergent views (Hennessy, 2005;Clarke, 2015, ch. 21). Under Cameron, Deputy Prime Minister Nick Clegg reputedly performed a similar role as chair of the Home Affairs cabinet committee. In his memoirs, Clegg recalls this committee fondly: 'There were challenging, collegiate debates between ministers on both sides of the coalition on everything from the future of social care to early years provision-debates that were not possible in the more perfunctory format of the weekly meetings of the cabinet' (Clegg, 2016, p. 78).
It is worth noting that Clegg (2016, p. 78) felt that the quality of debate in the committee diminished over time as partisan rancour intervened more regularly. This suggests that alongside group size, group composition effects whether groupthink, polythink or a more vigilant style of decision-making prevails. Nowhere has this impact become more readily apparent than through the role now played by special advisers in both formal and informal decision-making settings. Since 1979, there has been a steady increase in the number of special advisers and the roles that they play in British governments, with a particular growth in numbers since the arrival of the Blair Government in 1997. This has coincided with a seeming decrease in the breadth of input and access given to cabinet ministers, and indeed to the senior Civil Service. In other words, the composition of the personnel in the room when decisions are made has changed. This is visible at each level of the process, whether it is staffers being allowed into full meetings of the cabinet or only very small groups of people being invited onto the metaphorical sofa for the key discussions.
Under Thatcher and Major, the only people actually in the cabinet room for meetings of full cabinet were the ministers themselves and a cabinet secretary. Blair started the practice of inviting advisers, which has continued. They are only supposed to 'observe' but have the potential to act as an inhibiting presence. One Minister was astounded to enter their first full cabinet meeting under Cameron in 2010 to find an array of unexpected people present: All around the wall, around the entire cabinet room, [were] people, some of who I knew, some of whom I didn't. I recognised ex-Murdoch journalists, all sorts of people sitting around . . . and they could. . .hear if any cabinet minister was coming up with some stray and unexpected line. (Minister, our interview, 2018) The presence of advisers, even if they remain silent, contributes to the sense that full cabinet is too unwieldy to make effective decisions. It also contributes to the likelihood of leaks. The increased leakiness of cabinet was a recurrent theme amongst our interviewees. People in full cabinet could not be trusted to ensure that divergences of opinion would be kept in confidence. This creates extra incentives for prime ministers to retreat to a smaller circle of trust for their discussions. Leaks have always been a challenge within cabinet settings, but the increase in special advisers and the centralisation of government in Number 10 has made it easier for prime ministers to simply circumvent their cabinet colleagues altogether in order to minimise unwanted leaks.
The cabinet is a uniquely political setting in which people simultaneously position themselves to take internal and external credit. Letting slip in public that they fought against an unpopular measure in Cabinet offers ministers a chance to burnish their own public reputation at the expense of their collective colleagues.
At moments of perceived prime ministerial weakness, this temptation becomes simply too hard to resist. When the tradition of cabinet confidentiality is ignored it not only undermines cabinet solidarity and stops cabinet being a 'safe place' (O'Donnell, 2018), it also discourages prime ministers from building a culture of challenge into government decision-making.
2.2 When psychology meets process: the impact of Prime Ministerial preferences regarding conflict 2.2.1 Thatcher and Major Some prime ministers enjoy conflict and argument while others shy away from it where they can. Margaret Thatcher was in the former camp. One of her Ministers noted having many 'endless arguments, rows' (our interview, 2018) with her. She liked and respected those with strong views who were prepared to express them. As former Cabinet Secretary Lord Wilson (2018) has recalled: 'she wanted your raw advice as a sort of untutored official'. Lord Armstrong, her cabinet secretary from 1979 to 1987, said 'Oh she liked a good argument and she didn't like somebody who. . .gave in as it were' (Armstrong, 2018).
And sometimes she could be persuaded; one Minister boasted to us that after a series of rowing one-to-ones 'she. . .had been persuaded of my method of doing it' (our interview, 2018). Even in full cabinet 'she didn't always get her way and she wasn't always in the majority. . .she relied on persuasion' (Our interview, 2018). While she was often bombastic and less inclined to listen as time went on, Lord Wilson recalled that 'she wasn't always like that and there were occasions when she was quite thoughtful' in the full cabinet or committee (2018). She had a large number of bilateral meetings but civil servants have attested she was usually prepared for policies to be discussed in the formal committees. One Minister described it as 'classic cabinet government' and 'a very collective process' (Our interview, 2018). He found hearing his ministerial colleagues' views helpful: We did actually have discussions. I found, when I was a cabinet minister, it was really quite valuable. Because when you're working out what you were going to do, . . .the comments from your colleagues could be some of the most useful. . .and give you better understanding of how you were going to have to sell it in the political world if you were given the go ahead. (Our interview, 2018) Lord Butler, having been her Principal Private Secretary 1982-1985, noted that by the time he came back to Downing Street as Cabinet Secretary in 1988, Thatcher was listening less (Jago, 2017, p. 291). The result was that she also encouraged effective challenge less, making a form of groupthink more likely. Ministers became more reluctant to express their views because there was little chance of persuading her. This could have policy consequences, such as her strong support for the poll tax against opposition from some in her cabinet. Because Thatcher listened less as time went on, by the time John Major became Prime Minister in 1990, there was a desire for a return to a more collegiate cabinet process. As a former cabinet colleague noted, 'he wanted much more cool and collective and. . .consensus discussion' (Our interview, 2018, and see Kavanagh and Seldon, 1999, p. 224). Cabinet Secretary Lord Butler also describes a change under Major: 'I think he wanted to re-establish cabinet government, to be less dominant' (Butler, 2018). As Major himself recalls, Margaret had often introduced subjects in cabinet by setting out her favoured solution: shameless but effective. I, by contrast, preferred to let my views be known in private, see potential disasters ahead of the meeting, encourage discussion, and sum up after it. . .Margaret had been at her happiest confronting political dragons; I chose consensus in policy making. (Major, 1999, p. 437) Major was 'an excellent negotiator, a good chairman, skilled in teasing out a position acceptable to people of different viewpoints' (Kavanagh and Seldon, 1999, pp. 207, 224;see too Hennessy, 2001, p. 439). This more open psychological disposition however, as suggested in the typology in Figure 1, could not guarantee vigilant decision-making when the prime minister's own political authority became too weak to sustain it.

Blair and Brown
Some argue that Tony Blair's determination to avoid challenge within the cabinet process stemmed from his observation that factionalism undermined the Labour cabinets and shadow cabinets of the 1970s and 1980s, when the meetings were often acrimonious and smacked of dysfunction to the electorate (our 2018 interview with Minister in Blair's cabinets; Blair in Applebaum, 2001, interview). Our interviewee also stressed Blair's desire to co-ordinate as a response to the structural pressures of modern government (see also Burch and Holliday, 2004). However, there is evidence that a third, and, we argue, significant reason that Tony Blair sought to avoid open conflict in decision-making is that it made him personally uncomfortable. One person did constantly challenge him, his chancellor Gordon Brown, but Blair took steps to ensure that any direct face to face conflicts were conducted when they were on their own or with only a couple of witnesses, such as Deputy Prime Minister John Prescott. In his memoirs Prescott is one of many to describe Blair's aversion to open conflict, always preferring to 'persuade' potential challengers through one-to-one discussion.
He hated to say 'No, we can't do that,' or 'I'm ordering you to do this.' He preferred to talk you round slowly. He'd do it so well you'd find you were beginning to agree with him. (Prescott, 2008, p. 311) One senior unnamed Whitehall figure told Hennessy: It's not the fear of the press going on about splits that stops the Cabinet from discussing things, it's because the PM doesn't like argument. Cabinet these days is just a series of self-congratulatory remarks. (Hennessy, 2001, p. 525) Blair's aversion to challenge led him to shun groups as large as cabinet committees, preferring bi-laterals or regular 'stock-takes' with individual ministers.
One key adviser attempted to justify Blair's predilection for one-to-ones by arguing that even though cabinet committees were smaller than full cabinet they did not generate 'informed discussion' because of their fixed composition (Our interview, 2018). Then, after the invasion of Iraq and criticisms in the Butler report of Blair's avoidance of the formal cabinet process, his advisers were forced to persuade him to use cabinet committees more. But, Seldon and Snowdon argue that he never felt comfortable with them.
What's happened to my stock-takes?' he exclaimed after one of the first new committee meetings. . .As Blair said, 'These [cabinet committees] don't really function for me, and they don't enable me to have the sort of discussions I want to have. (Seldon et al., 2008, pp. 328-329) A key civil servant interviewed for this project confirmed that conclusion: 'Blair liked smaller groups and had to be cajoled by one of his cabinet secretaries. . .to use cabinet [processes] more' (Our interview, 2018).
Patrick Diamond, who served as a special adviser under both Blair and Brown in the Policy Unit, confirms that picture of 'most government business' being conducted bilaterally, which 'also remained the case for much of Gordon Brown's premiership' (Diamond, 2011, p. 153). A Cabinet minister who had several run-ins with Brown over specific policy issues told us Brown was 'very rarely ready to have a direct conversation, and when he did have a direct conversation, he would invariably do it completely privately, just the two of you, with nobody else' (Our interview, 2018, see also Turnbull, 2018). Brown continued to ask advisers, as Blair had done, to prepare the ground before cabinet meetings by identifying ministers likely to disagree: And advisers, you know, you had to kind of talk to people before they went in. . .the Prime Minister would often meet people before Cabinet. 'Can you come to Gordon's office?' People would know they were being talked to before they went in (Brown adviser, our interview, 2018).
Therefore, despite his different personality, Brown also avoided situations where there might be open challenge in a group.

Cameron David Cameron's decision-making arrangements incorporated
at least two factors mitigating against wider challenge and debate. First, the people he chose to be in his inner circle included a few key ministers, and a group of special advisers who came from similar social backgrounds and had known each other and the PM for many years, dubbed the 'Cameroons' or 'Notting Hill set'. Cameron's inner circle was far more harmonious than those of previous prime ministers, and stayed that way throughout. One member of it accepted the truth of the characterisation in an interview with us: We all marched together to do one thing. I think that's fair to say. A lot was made of the fact that we were friends, all from the same background, which is actually overplayed. Of course, like all things, there's a truth. . . (Cameron adviser, our interview, 2018) Secondly, Cameron did not challenge the Blair/Brown tradition of suppressing conflict in the wider executive. Cameron was an admirer of the Blair model of centralisation as a way of making the executive work more smoothly. He reversed his initial decision to scale down the Number 10 policy unit, and he kept the core structure and practices developed by Blair at Number 10. A senior official present as Cameron started his premiership told us 'I suspect some of the traditions of how it went were very much borne out of how the Cabinet had been before' (Our interview, 2018).
His advisers saw their role as being primarily to shadow departments assiduously to ensure they were not diverging from Number 10. Two of Cameron's team of staff described how they continued the previous practice of shadowing departmental ministers, identifying who had gone 'off piste' (Cameron adviser, our interview, 2018). They engaged in systematic choreographing: 'So you'd make sure that you'd squared all the relevant people and. . . you wouldn't leave it to chance not to know where people stood on stuff' (Cameron adviser, our interview, 2018). One Cameron adviser described continuing this practice of 'the rolling of the pitch' coming into cabinet meetings, whether committee or full.
Cameron's continuation of Blair/Brown traditions then played out in the use of cabinet meetings as choreographed endpoints of verification rather than a site for debate. In 2010-2015, there was arguably even greater need for choreographing than usual as two parties were involved. One senior civil servant commented to us that the Cameron cabinet appeared to be 'processing issues' that had been decided elsewhere: 'You know I can remember only a handful [of instances], as I say, that you thought 'oh well that was a defining moment really' (Civil servant, our interview, 2018). He echoed a view among other civil servants and ministers with long experience that: Quite a lot of it was I wouldn't say going through the motions, but not far off that really -if I'm being very honest . . .there was no sense of a developed debate. . .they rarely challenged the fundamental premise of the paper or the issue, in a way that might engage in a kind of a dialogue. (Civil servant, our interview, 2018) One example was decision-making about whether to hold a referendum on membership of the EU. Having discussed the issue with Number 10 staffers in November 2012, Cameron held individual meetings with key cabinet ministers: May, Hammond, Duncan Smith and Ken Clarke (Seldon and Snowdon, 2015, p. 263). In an interview with us, one adviser suggested that because it was party policy not government policy, it did not have to be ratified in cabinet. In their view, there was adequate discussion in the form of one-to-one consultation with senior cabinet members: We had the speech in Number 10 and everyone came and read and talked us through.. . .So, it was widely discussed before the speech. That speech was not a surprise to the senior party. (Cameron adviser, our interview, 2018) However, one Cabinet Minister told us he thought the decision should have gone to political cabinet at least, rather than being discussed one to one in Number 10.
He [Cameron] consulted some cabinet members. Mainly his entourage the Notting Hill set. . . .They completely sold the pass. . .I don't recall it ever coming up on the [political cabinet] agenda. . .That's an example of cabinet government collapsed. (Minister, our interview, 2018) The extent to which the role of Number 10 in suppressing internal conflict has become an embedded tradition of British cabinet government is exemplified by the remarkable similarity of the phrases used in interviews by advisers to three different prime ministers: Nevertheless, there is some evidence that Cameron was less personally averse to conflict than Blair or Brown. He recognised the dangers of groupthink inherent in having such a tight inner circle. One senior adviser noted that they did 'really, really try to reach beyond' the horizons of their close-knit social group. The adviser said there were some 'outsiders', including Steve Hilton who 'didn't come from a similar background' although 'he did end up being Oxbridge', and Andy Coulson who was 'quite a different animal'. However, for various reasons Hilton and Coulson did not last long. To inject more divergent views, Cameron's team enlisted Ameet Gill, even though he 'was never really a Conservative' (Adviser, our interview, 2018). After the 2015 election, they were conscious that the group was still close and no longer had the challenge of the Liberal Democrats; groupthink might become even more of a problem. They hoped 'there would be someone who'd emerge whose view was a little bit questioning, who we bought in. . .because . . .it's always good to have someone to challenge' (Adviser, our interview, 2018).
Perhaps more significantly, Cameron set up an ad hoc but regularly meeting group which provided a forum for challenge by his Coalition partners. Known as 'the Quad', the core of the group consisted of Cameron himself and Chancellor of the Exchequer George Osborne with Liberal Democrat leaders Nick Clegg and Danny Alexander. While the group was small, there are at least two aspects which counteracted any tendency towards clique-style decisionmaking. First, the Cabinet Secretary and other relevant ministers sometimes attended, so the Quad was not necessarily as closed in its membership as appearances would suggest. In fact, the effective group size often came close to the optimal six to ten sizes for vigilant decision-making. Secondly, Cameron did not abandon holding the meetings when they started to become heated, which our interviews indicated they often were. He was prepared to play a robust role in them but accepted some 'challenge' over a five-year period.

Discussion and conclusion
We have positioned these events as the re-making of traditions around the nature of cabinet decision-making, particularly from the Blair premiership onwards. We argue that the suppression of challenge in the cabinet process has accelerated in the last 30 years. While some of the factors driving that change were already in train before 1997, Blair's personal antipathy to conflict was important. He and his advisers conflated the need to suppress and manage conflict with broader needs to modernise and centralise decision-making, to the extent that they undermined the capacity to grow a 'culture of challenge' within cabinet processes. Although a different personality, Brown also disliked open conflict and continued much of the Blairite tradition. Cameron and key advisers were aware that, within their inner circle, views might converge more than was healthy, but this did not result in actual change to cabinet processes to ensure that a culture of challenge was actively supported.
The Blair government, as Chilcot and others went to great lengths to demonstrate, was clearly prone to groupthink. 6 It was governed by a strong leader, with a penchant for short and 'closed' cabinet discussions rather than open debate. This juxtaposes clearly against the Thatcher years. The evidence suggests that at least in the early years she combined her desire to win arguments with a willingness to allow challenge. The result was that ministers felt that they had a chance of swaying things if they were bold enough and persuasive enough to shift the prime minister's position. This presented a combination of a strong form of leadership coupled with genuine opportunities for deliberation. When that culture began to close itself down, and distrust crept in, the quality of decision-making suffered. Thatcher began to consult less, leading to less vigilant decision-making, and then greater and more open rancour as her own authority as prime minister began to fade. This illustrates the importance of prime ministerial authority as a precondition for groupthink as outlined in Figure 1. Once authority begins to fade, the psychological disposition of the leader becomes less of a factor for the simple reason that they can no longer be sure of actually enforcing their preferred style.
The evidence suggests that Thatcher relished challenge to a greater extent than Blair and Brown because she thought she could win the arguments that ensued, whereas Blair and Brown were less disposed to having the arguments in the first place. In terms of the inner circle at Number 10, Cameron faced a greater potential for groupthink than either Blair or Brown because of the close-knit nature of a group of people who had known each other for two decades. However, psychologically, he was perhaps more prepared than his two immediate predecessors to tolerate some challenge in the wider cabinet system, particularly in the Quad during the coalition with the Liberal Democrats.
The signs are that Theresa May continued the Blair/Brown/Cameron tradition in the brief period before the 2017 election when her two gatekeepers, Nick Timothy and Fiona Hill, exercised significant power over cabinet ministers. She continued to try to reduce the chance for challenge after the election too, but her diminished political position simply no longer allowed it. 7 As one civil servant 6 It is worth noting that, even with more centralised decision-making and the tendency towards groupthink, many of the daily demands of UK policymaking still relied heavily on the work of departments. Even with a more centralised structure, Number 10 could not choreograph all day-to-day decisionmaking across Whitehall. 7 Our interviews were conducted with political actors from the pre-May period and therefore future research is needed to accurately assess the May style. Based on evidence to date, a common perception at the time was that her cabinet was conflict ridden to a dysfunctional degree (Singh, 2018;Rentoul, 2019). Seldon and Newell (2020) estimates she had the highest annual rate of ministerial resignations outside reshuffles in modern history. However, there are also some initial analyses suggesting she had greater control, in other words less polythink, than might be imagined, at least until 2018 (Worthy and Bennister, 2020). commented to us (Our interview, 2018), her July 2018 presentation of the Chequers Plan at an away day without consulting key ministers first was testament to a desire to control, while simultaneously demonstrating the weakness of her position. The leakiness of her cabinet, resignations and open divergence from collective positions saw May's cabinet ultimately become dysfunctional and close to a textbook case of polythink. However, we argue that what made May's difficulties more acute was the juxtaposition of her cabinet disagreements against the tradition established in the preceding twenty years that any public disagreement is in itself a sign of prime ministerial weakness.
Examining the study of cabinet government through the lens of group decision-making theory offers some potentially fruitful avenues for further research. It offers a set of questions on when and how particular cabinets might find themselves shifting from function to dysfunction. It also highlights the central importance of the psychological disposition of the prime minister towards conflict. Further research is needed to shed a more detailed light on the variables that govern the interaction of the personal psychology of an individual prime minister with the structural architecture of cabinet government. Further research too is needed on the extent to which the Westminster system of government is more or less prone to encouraging groupthink than other systems.
Part of the solution to issues of groupthink, as suggested by the Chilcot report and many others, may be to strengthen the architecture of decision-making so that it becomes less reactive to the personal style of the prime minister. But, it is important not to over-correct. The realities of politics dictate that prime ministers must be able to shape policy and administration rather than simply becoming prisoners of structure. Equally, mandating certain structures may cause prime ministers to simply find new ways to work around them by having the 'real' discussions elsewhere. Further research is needed to identify the ways that challenge can be built into a system, without generating 'rubber stamps' that simply divert debate to new and ever more secretive forums. As we have argued in this article, polythink and groupthink are merely the extreme ends of a wide spectrum. Defining any 'perfect' combination of process and personality for cabinet decision-making remains a fraught task, but the spectrum from polythink to groupthink helps to provide conceptual grip on the nature of dysfunction and the variables that contribute to it.