ELECTING AND EJECTING THE PARTY LEADER

The focus of the literature on leadership successions tends to be on single-country cases and is dominated by three countries with similar polities – Britain, Australia and Canada, whose ‘governing principles derived from similar constitutional assumptions’ (Weller Reference Weller2012: 152). There are notable exceptions, many stressing the importance of contingent contextual factors and comparative analysis (such as Bynander and ‘t Hart Reference Bynander and ‘t Hart2007, Reference Bynander and ‘t Hart2008; Cross and Blais Reference Cross and Blais2012; Davis Reference Davis1998; ‘t Hart and Uhr Reference ‘t Hart and Uhr2011; Kenig Reference Kenig2009; Walter Reference Walter2013; Weller Reference Weller1994).

Within the single-country-based literature there is considerably more material assessing leadership successions within British political parties, providing a rich analytical basis for our two-country comparison. Scholars have developed their work around the following themes. First, there is work that appraises individual contests to examine the quality of candidates and the significance of the campaigning period (for example, Alderman Reference Alderman1996, Reference Alderman1998; Alderman and Carter Reference Alderman and Carter1991, Reference Alderman and Carter1993, Reference Alderman and Carter1995; Denham and O’Hara Reference Denham and O’Hara2008; Dorey and Denham Reference Dorey and Denham2006, Reference Dorey and Denham2011; Heppell Reference Heppell2008, Reference Heppell2010). Second, there is work that seeks to examine the variables that may have influenced voting behaviour at the level of the parliamentary party (for example, Cowley and Bailey Reference Cowley and Bailey2000; Cowley and Garry Reference Cowley and Garry1998; Heppell and Hill Reference Heppell and Hill2008, Reference Heppell and Hill2009, Reference Heppell and Hill2010). Third, there has been a focus on the significance of amending the party leadership selection rules, or the continuing merits or otherwise of existing procedures (e.g. Quinn Reference Quinn2004, Reference Quinn2005, Reference Quinn2012; Stark Reference Stark1996). Fourth, there has been work that analyses the ease, or otherwise, with which incumbents can be removed from the party leadership (for example, Alderman and Smith Reference Alderman and Smith1990; McAnulla Reference McAnulla2010). What the above appraisal demonstrates is that the dominant focus has been on elections, whether for vacancies (the majority of cases) or challenges which have been successful. The academic analysis of when party leaders survive, despite considerable questioning of their leadership, is thus the under-developed aspect of leadership selection studies.

THE NON-REMOVAL OF GORDON BROWN

What made Brown vulnerable to eviction? Some of his critics, such as MPs Charles Clarke and Frank Field, would claim that Brown lacked legitimacy flowing from the way in which he had acquired the Labour Party leadership in May 2007 (Quinn Reference Quinn2012: 88). Brown had been elected unopposed as his only possible rival for the leadership, John McDonnell, failed to secure the requisite number of parliamentary backers to initiate the electoral college. However, as Brown defeated McDonnell by 313 to 29 nominations (McDonnell needed 44 to pass the nomination threshold or 12.5 per cent of the Parliamentary Labour Party), the argument that he lacked legitimacy at parliamentary party level is erroneous. His legitimacy only came into question when his authority was undermined by his poor performance as prime minister (Foley Reference Foley2009).

Brown’s deficiencies provoked limited sympathy within his own party. This was partly due to the reputation that he had acquired during the years waiting for Tony Blair to step down. His fears that an alternative to himself as leader would emerge had contributed to him developing a reputation as a ‘scheming fixer’ and ‘petty infighter’ (Hughes Reference Hughes2010: 3). Labour had experienced an upsurge in poll ratings in the first few months of Brown’s prime ministerial tenure and there appeared an opportunity to enhance his leadership legitimacy. With this chance to secure his own mandate to lead, expectations that Brown might call a snap general election were allowed to develop (Rawnsley Reference Rawnsley2010: 496–515). When a Conservative recovery in the opinion polls emerged, Brown chose not to dissolve Parliament, but tried to claim that his decision was not influenced by the opinion polls. Brown’s credibility never recovered from the election that never was in the autumn of 2007 (Foley Reference Foley2009: 500).

The cumulative impact of these factors contributed to a series of plots being initiated to remove Brown from the leadership, and as the speculation about possible challenges intensified, so did Brown’s ‘hyper-sensitivity to potential rivals’ (Kenny Reference Kenny2009a: 503). The speculation disfigured Labour Party politics between mid-2008 and around January 2010. There were three substantive plots against him, which were interpreted by political commentators as preludes to the removal of Brown. These occurred in July 2008; June 2009 and January 2010. The first rumoured plot involved the foreign secretary, David Miliband, whose advisers decided to make a ‘significant intervention’ (Hasan and MacIntyre Reference Hasan and MacIntyre2011: 139). Miliband’s Guardian article ostensibly did nothing more than outline what Labour needed to do to reconnect with the electorate, but his failure to mention Brown by name was interpreted as indicating his own willingness to be an alternative leader of the party (Hasan and MacIntyre Reference Hasan and MacIntyre2011: 140).

The second substantive plot presented a more serious threat to Brown. The dual impact of the fallout of the expenses scandal and poor local and European parliamentary election results led to a series of ministerial resignations. The scandal over MPs’ expenses claims forced two cabinet ministers – Jacqui Smith and Hazel Blears – to resign, and Caroline Flint also resigned, claiming that Brown ran a two-tier administration that marginalized female ministers. Of greater significance, however, was the resignation from cabinet of James Purnell. His resignation letter, which claimed that Brown continuing as prime minister made a Conservative victory ‘more, not less likely’, ended with the request that Brown ‘stand aside to give our party a fighting chance of winning’ (Quinn Reference Quinn2012: 89). Press speculation now focused on whether other cabinet ministers would resign, thus assuming that Brown would depart voluntarily if they did, without a formal challenge being needed. However, with senior cabinet figures such as Harriet Harman, Jack Straw and, critically, Peter Mandelson rallying behind Brown, the most likely beneficiary of this plot, David Miliband, decided not to resign, and thus Brown survived (Quinn Reference Quinn2012: 89).

One final attempted putsch was initiated in January 2010 and was coordinated by former cabinet ministers Geoff Hoon and Patricia Hewitt. A letter was circulated amongst Labour MPs demanding a vote of confidence in Brown. This request contravened the constitution of the party, which had no provision for confidence motions regarding the party leader. The intention was to attempt to circumvent the procedures by initiating a ‘frontbench coup’. Cabinet members dissatisfied with Brown remained out of sight and did not answer their telephones in the aftermath of the circulation of the letter. Their reluctance to endorse him publicly was expected to force Brown into resigning, given that the cabinet were no longer backing him (Quinn Reference Quinn2012: 90). However, as the hours passed and Harman and Straw backed off from asking for Brown’s resignation (which was supposedly part of the plan), so cabinet ministers gradually began to offer their support. David Miliband, who had been ‘waiting to see what happened’, eventually stood outside the Foreign Office and announced he was ‘getting on with his job’ (Hasan and MacIntyre Reference Hasan and MacIntyre2011: 169).

THE REMOVAL OF KEVIN RUDD

Kevin Rudd’s removal can be seen as a consequence of deep personal flaws and institutional pressure. One simple answer as to why the Australian Labor Party ousted Rudd is that it had the institutional capability to remove the leader and was willing to use this collective caucus power. Since 1945 in Australia there have been several challenges to the sitting prime minister in the party room. Liberal Prime Minister John Gorton was challenged shortly after winning the 1969 election and again in 1971, when famously a tied vote saw him casting the deciding vote against himself. Andrew Peacock failed to unseat Liberal Prime Minister Malcolm Fraser in 1981. Under Labor, Bob Hawke managed to see off Paul Keating’s first party room challenge in 1991, but after a destabilizing six-month backbench campaign from Keating, he lost the prime ministership. Evaluations of the first short-lived Rudd period of office and its dramatic end divide between the personal (Jackman Reference Jackman2008; Marr Reference Marr2010; Megalogonis Reference Megalogonis2010, Reference Megalogonis2012; Stuart Reference Stuart2010) and weightier structural analyses (Evans Reference Evans2010; Rhodes and Tiernan Reference Rhodes and Tiernan2013; Walter Reference Walter2011).

Rudd had arrived in Canberra in 1998, determined to make his way to the top of the Labor Party. His political experience was honed as Chief of Staff under Queensland Premier Wayne Goss. He had a singular desire to lead the federal party and ‘He didn’t hide his ambitions’ (Marr Reference Marr2010: 40). Rudd, though, had to bide his time as Labor went through a succession of leaders in opposition. While the Liberal National Coalition of John Howard remained in power, Labor tried first Kim Beazley, then Simon Crean, Mark Latham and Beazley once again. It was Crean who made Rudd foreign affairs spokesman and he subsequently established a more public persona, becoming a serious challenger for the leadership by 2003. Although his popularity grew with the public, he struggled to muster enough support in caucus (a recurring refrain in the Rudd story) to formally stand in 2003 and 2005. Hanging on to his foreign affairs brief, he was well placed to capitalize on the Australian Wheat Board scandal that broke in 2006 over kickbacks paid to the Iraqi regime. Rudd mounted a most effective attack on the government, simultaneously presenting Howard as a manipulator of the truth and himself as a professional, diligent politician acting with integrity. Rudd stepped up to challenge Beazley and became Labor leader on 4 December 2006, winning the backing of the Labor caucus by 49 to 39 votes, but only when Julia Gillard and her supporters accepted that Rudd, though less popular in caucus, was polling better as the alternative to Beazley.

Rudd initially took to the leadership with gusto. As John Howard’s own leadership creaked and strained, the focus shifted from Labor’s leadership troubles onto the government. In particular, Howard’s own position as leader in advance of the 2007 election came under much greater scrutiny. When battle was joined in earnest in October 2007, Rudd ran an overtly ‘presidentialized’ campaign based on ‘new leadership’ for Australia (van Onselen Reference van Onselen2010; Wanna Reference Wanna2008). A slick campaign largely based around the Rudd brand as Kevin07, saw Labor gain a 16-seat majority (Jackman Reference Jackman2008; Stuart Reference Stuart2010). Although not an overwhelming electoral endorsement, with Howard dramatically losing his seat, it was a victory that Labor could savour after so long in the wilderness. The euphoria of a return of a Labor government, led seemingly by an intellectual man of conviction, proved short-lived. Rudd, the Mandarin-speaking former diplomat, was ushered in as a ‘man for all seasons’, a ‘geek’ who would appeal to a broad range of Australians. He presented a cautious and conservative form of leadership, reaching out to the median voter in a highly personalized campaign. His early activism as prime minister, signing the Kyoto Protocol on his first day in office and delivering an apology to the stolen generations on the first day of parliament, appeared a promising start. By early 2010 it had all turned sour.

A range of factors coalesced to turn the opinion polls against the Rudd-led government. The opposition (not immune itself to leadership turnover) managed to unite behind the right-wing Tony Abbott, after Brendon Nelson and Malcolm Turnbull had brief stints as Liberal Party leaders, and started to make a dent in Rudd’s personal popularity rating. The Rudd government was hit by a succession of policy failures, U-turns and over-promises (many derived from the grand 2020 summit championed and led by Rudd in April 2008). The most damaging policy reversal proved to be the decision to dump Labor’s commitment to the launch of an Emissions Trading Scheme (ETS) by 2011. As several have reflected (Evans Reference Evans2010; van Onselen Reference van Onselen2010), Rudd’s problems really began to kick in from this point. Having made climate change the centrepiece of his 2007 campaign and initial premiership as ‘the great moral issue of our time’, to then drop the flagship commitment dealt a severe personal and collective blow to Rudd and Labor. In addition to the debacle of the Emissions Trading Scheme, Evans (Reference Evans2010: 272) lists six other policy failures (refugees, home insulation, primary schools building project, child care, Northern Territory intervention, mining super tax) that ‘alienated the electorate and, most significantly, radicalised opposition within his own party’.

Rudd also faced a growing backlash against his style of governing (Strangio Reference Strangio2013; Tiernan Reference Tiernan2008). The critique of Rudd’s style of management centres on the impact of a dysfunctional leader on the various aspects of prime ministerial power. Rudd failed to work with and through his cabinet (Evans Reference Evans2010; Marr Reference Marr2010). Decisions were concentrated in a ‘gang of four’ of Rudd, Treasurer Wayne Swann, Lindsay Tanner and Julia Gillard. Despite early policy successes the signs were there that Rudd might be storing up trouble. He ran his affairs at a breakneck pace; insider comments such as ‘it is impossible to exaggerate the degree of personal intervention by the prime minister. It’s his personality’, became commonplace (Tiernan Reference Tiernan2008). His insistence on personal control across the full spectrum of the government’s policy and media management put a huge strain on his staff and inevitably led to a backlash against such micromanagement.

A constant link is made between the policy failures, dysfunctional governing style and Rudd’s apparently flawed personality and fitness to lead. The impact of his personality was such that policies were never followed through to conclusion, left in limbo, U-turns performed and government mired in indecision and failure. The slide in the polls created alarm, but it was Rudd who had brought this on himself. Yet it is impossible to assess the removal of Rudd without analysing the factional cleavages that any leader of the Australian Labor Party must take heed of in order to survive (Davis Reference Davis2011; Gauja Reference Gauja2011; Leigh Reference Leigh2000; Warhurst and Parkin Reference Warhurst and Parkin2000). The ability to hold the support of the parliamentary party is crucial to remaining in power (Bennister Reference Bennister2007, Reference Bennister2008, Reference Bennister2012), even more so with a parliamentary caucus of 115 in the Australian Labor Party.

In early May 2010, with an election due later in the year, the consequences of the downturn in the Labor government’s fortunes hit party strategists. The party’s primary vote had slumped to 35 per cent and Rudd’s own approval rating fell into negative territory for the first time (Stuart Reference Stuart2010). Rudd had generated this autonomous relationship with the public to reach beyond the party factions, but it left him vulnerable when the polls fell in such a dramatic fashion (even allowing for an electoral cycle depressing support for the incumbent government in the winter months). Australian Labor Party ‘machine politics’ kicked in and the powerbrokers began to think the unthinkable: replacing a sitting prime minister in his first term of office (Kent Reference Kent2010: 313). The window of opportunity meant that any challenge had to be concluded successfully by 24 June, the last parliamentary sitting day of the term. MPs would break for winter and caucus would be dispersed after this date. Any earlier move and the new prime minister would be swiftly exposed to the opposition. As Julia Gillard was the only credible alternative to Rudd, it was hoped she would benefit from a political honeymoon over the winter before the polls picked up in spring, leading to the sunny uplands of the impending general election.

The plotters, none of whom was a household name, quickly became dubbed the faceless men in the subsequent accounts of the ejection. Such labelling further damaged the Australian Labor Party, harking back to the notorious case of the 36 ‘faceless men’ deciding Australian Labor Party policy in 1963 while the leader waited outside the conference hall (Warhurst and Parkin Reference Warhurst and Parkin2000). The key individuals were MP Bill Shorten, parliamentary secretary for disability and children’s services, but more importantly former secretary of the Australian Workers Union (AWU); Paul Howes, national secretary of the Australian Workers Union; and Mark Arbib, senator and employment minister (Wanna Reference Wanna2010). In addition, two senators, David Feeney, assistant national secretary of the Australian Labor Party, and Don Farrell were involved. These individuals cut across the left and right factions: although the motive was revitalizing the party fortunes, some had personal grievances with Rudd (according to Stuart Reference Stuart2010: 270). Shorten and Howes brought the influence of the wider Labor movement via the Australian Workers Union. Howes was the first to call for Rudd to go, but he subsequently played down his actual influence on the Australian Labor Party in his diaries (Howes Reference Howes2010). Arbib, who had previously delivered the numbers for Rudd against Beazley, organized the powerful New South Wales (NSW) right faction. Feeney did the same job in Victoria. Although Gillard was nominally aligned with the left, gaining her seat in Lalor with support from the Victorian left faction, Kent (Reference Kent2010) suggests this was more ‘organisational than ideological’. Gillard gained her political experience as Chief of Staff to John Brumby when he was opposition leader in Victoria and she was a supporter of Mark Latham’s doomed Australian Labor Party leadership. Both men were significant figures on the Australian Labor Party’s right. Gillard, it seems, was beyond factional alignment, a pragmatic and opportunistic politician, rather than ideologically driven. Therefore the fact that these individuals were all from the right faction should not be so surprising. Rudd, who had owed his position, at least initially, to some of the powerbrokers on the right (although the complexities of the December 2006 leadership contest that put Rudd in place even though Gillard had the numbers meant there was not a clear left–right factional split in caucus), saw his support disappear. The reality of the situation became apparent on the eve of the challenge as he commented, ‘It has become apparent to me in the course of the last period of time … that a number of factional leaders in the Labor party no longer support my leadership’ (Hartcher Reference Hartcher2010).

Gillard rebuffed approaches, but as it became obvious that Rudd was isolated and Gillard had the numbers she decided to stand (Gauja Reference Gauja2011: 11). Surprisingly few MPs and journalists were aware of how developed the coup had become, with many senior players still in the dark (Stuart Reference Stuart2010). Although vowing to fight on, after taking soundings Rudd realized he had little support in the caucus; one conservative estimate gave him no more that 20 of the 118 caucus votes in the party room and these were largely sympathy votes (Stuart Reference Stuart2010: 278). Rudd withdrew from the contest, allowing Gillard to be elected unopposed at the caucus meeting the following morning.

In contrast to Rudd’s lack of solid numbers in caucus, Gillard had realized early on that maintaining a factional powerbase (appealing to both left and right factions) was crucial in the ‘gang warfare’ of Australian Labor Party politics (Davis Reference Davis2011; Kent Reference Kent2010). Rudd failed to heed the lesson of Labor leaders; he led no faction within the party and owed his position to a broader constituency based on the opinion polls. He had not shaped the party around himself as Gough Whitlam had done. He was not steeped in the Labor Party as a favourite son as Bob Hawke had been and he had not battled to the top after a long internal struggle as had Paul Keating (Marr Reference Marr2010). The lack of a stable and loyal powerbase in the party (as shown by his inability to develop sufficient numbers of supporters within the caucus), when added to his centralizing tendencies, combined to create a vulnerability to the Rudd leadership (Stuart Reference Stuart2010).

THE REMOVAL OF JULIA GILLARD

When the dust settled on Gillard’s three years as Labor leader and prime minister, the extent of the oppositional forces she had to battle with became apparent. From inside the party Rudd led a campaign to destabilize her and retained a degree of caucus support and wider electoral support, the Liberal-led opposition had regrouped around Abbott’s leadership and the forces of the press circled around Australia’s first female prime minister (Bennister Reference Bennister2013). After ousting Rudd in June 2010, Gillard called an early election for 21 August in an attempt to establish a stronger mandate, mindful of the manner of her taking office. The resultant minority Labor government supported by independents provided her with another monumental challenge.

Gillard had more than managed to keep Labor afloat in parliament by astutely negotiating deals from her position as leader of a minority government, but failed to register with the electorate. Again the short electoral cycle fuelled leadership speculation as Rudd waited in the wings. Gillard had tried to accommodate him as foreign affairs minister when she named her first ministry after the 2010 election. Rudd, bitter at his removal, spent the time undermining Gillard. He stood down abruptly on 22 February 2012 on a trip to Washington, prompting Gillard to face him down five days later in a caucus vote. She won 71 to 31, but despite Rudd’s assurances that he would not challenge again, his backbench presence continued to haunt Gillard. Her personal leadership ratings and those of the government continued to flat-line. In almost a mirror image of the Gillard challenge in 2010, Rudd seized the final opportunity before the parliamentary recess on 26 June 2013. Conscious of the likely drubbing in the election, 26 Labor MPs switched sides to give Rudd a 57 to 45 advantage (of all the ‘spills’, this proved to be the highest number of caucus supporters Rudd could muster). The crucial factor proved the transfer of Bill Shorten, factional leader and former head of the Australian Workers Union, to Rudd. Shorten had been a prominent Gillard supporter and put his own future leadership ambitions at risk by jumping across to the Rudd camp.

There are multiple explanations and commentary on the reasons for the failure of Gillard’s tenure as prime minister. Prominent factors include an inability to shake off the impression of ruthlessness in the way she obtained the premiership and her policy reversal after ruling out a Carbon Tax in 2010, then proceeding to introduce it. Many (e.g. Brett Reference Brett2013; Johnson Reference Johnson2013) also note her skilful management of a minority government and the piloting of several important pieces of legislation through parliament, in particular the Gonski education reforms and national disability insurance. The mechanics of the party meant Gillard was always in a perilous position, having to battle on several fronts at once. As Strangio (Reference Strangio2013) noted, her premiership was under a constant shadow: ‘Her entire prime ministership has dangled under a Sword of Damocles of leadership speculation. More has been said and written about when she will be dumped from office than about what she has done in office.’

Closer analysis of how caucus turned against Gillard showed that it was not just Shorten’s actions, but the loss of the New South Wales right faction that had previously turned against Rudd to install Gillard in 2010. Bob Carr, the former New South Wales premier who had been made foreign affairs minister under Gillard, succeeding Rudd, abandoned her soon after professing loyalty. New South Wales MPs deserted her almost en masse, fearing they would be wiped out in the election. Polling in Paul Keating’s former constituency of Blaxland in New South Wales suggested that even a 12 per cent lead was under threat (Heath and Swan Reference Heath and Swan2013). Panicked MPs grabbed the chance to jump ship and return to the disliked, but potentially damage-limiting, Rudd. In the event, Rudd’s second go at the premiership lasted only a few months and amounted to a survival exercise. His return may have prevented an Australian Labor Party meltdown, but the loss of 20 seats did not amount to a triumphant return. Rudd stepped down from the leadership and, although he initially remained in parliament, formally announced his retirement as MP for Griffith on 22 November 2013.

UNDERSTANDING EASE OF REMOVAL: PARTY INSTITUTIONS, CULTURE AND CIRCUMSTANCES

The explanation for the removals of Rudd and Gillard and the non-removal of Brown can be understood by comparing the respective party institutions, party cultures and political circumstances. That is to say, it is more than just the institutional arrangements that define the leadership succession as Weller has implied. Rather, procedures operate in conjunction with the cultural norms which shape the respective parties, as well as circumstances that determine how easy it is to remove incumbents.

In the case of Brown it is clear that the electoral college constitutes a protective shield for incumbent Labour party leaders, and that protection is even more pronounced when in office as the principle of initiating a contest requires the approval of party conference (Quinn Reference Quinn2005: 809). Thus Weller is justified in noting that incumbents benefit from strong procedural obstacles that enhance their security of tenure as there are considerable disincentives for challengers (Quinn Reference Quinn2012: 82–94).

In order for a challenger to proceed to the electoral college they need to secure the backing of 20 per cent (or 71) of Labour MPs, which is a high threshold for a challenger to pass (Dorey and Denham Reference Dorey and Denham2011: 289). What also impedes challengers is the way in which the electoral college works. Its procedural inflexibility means that a prospective challenger has a series of costs that may put them off initiating their challenge. These costs can be defined as decision costs, financial costs and disunity costs. Decision costs mean shifting the focus away from policy implementation and effective governing, alongside critiquing the opposition, as the party turns in on itself. Financial costs reflect the varying burdens that can be imposed upon the party depending on the selection procedures that it utilizes. Disunity costs reflect the risks associated with rival candidates condemning their respective policy positions. With the rules permitting prolonged periods for electioneering, and the time for formal balloting to be conducted, the consequence is political paralysis for the party if the candidates attempt to use the electoral college (as intended) whilst in office (Quinn Reference Quinn2005: 795–6).

Given those constraints, potential challengers (and their supporters) had to ask themselves the following question: Is it viable for a governing party, on the brink of facing the electorate in the midst of an economic recession, to set aside around four months for an expensive leadership challenge, which by its very nature will set Labour elites and factional blocks against each other? (Quinn Reference Quinn2005: 799–801). With the costs of mobilizing a challenge so high, the Parliamentary Labour Party became trapped in a ‘vicious circle’. The credibility of the government would be undermined on a frequent basis. Each crisis that undermined the government became linked to debates about Brown’s competence. As a consequence, political commentators would repeatedly refer to plots and rumours to unseat Brown, whereupon the Parliamentary Labour Party would ‘hit the brick wall of the procedures which make it so difficult to unseat the incumbent’ (Heppell Reference Heppell2010: 193).

There were also wider risks for potential challengers such as Miliband. These involved risks for them personally, but also a calculation of the risks for their assumed supporters. This reflects the strategic choices that they would have as prime ministerial aspirants. ‘Serious’ candidates for the succession who wanted to succeed Brown as an alternative leader had to challenge him directly (Weller Reference Weller2012: 153). The in-from-the-start rules (IFTS) meant that they could not use a backbench or alternative frontbench stalking horse candidate to challenge Brown. This option, in which the alternative challenger would wound Brown to such an extent that he stood down, creating a vacancy, whereupon the leading candidate would enter, was not available. The necessity of having to challenge directly incurs increased risk, but not just for the candidates themselves. A potential alternative would presumably have a ‘court’ of parliamentary backers and they also have to calculate the cost–benefit analysis for themselves personally, in terms of their careers. To initiate a contest the challenger would need to acquire the backing of 71 out of 363 Labour MPs. If the challenger resigned from cabinet themselves they would find it difficult to persuade those occupying ministerial office (approaching 100 Labour MPs were ministers in the Brown government) to resign in order to sign nomination papers backing the challenge. Their fear would be that Brown might survive, meaning that their disloyalty to Brown would be punished. Not only would a speedy return to ministerial office be blocked, but they had to factor in the way in which Brown dealt with disloyalty – that is, his reputation of negative briefing against his rivals and critics (Quinn Reference Quinn2012: 91).

Risk, fear of defeat and fear of the consequences of defeat undoubtedly deterred possible serious candidates. However, there were particular and distinct circumstances at play in the 2008–10 period that made the removal of Brown harder to engineer than, say, the removal of Blair in the 2005–7 period. Brown’s survival was also a by-product of there being no clear consensus on who was best positioned to defeat Brown, and then win the forthcoming election. Whilst some advocated Alan Johnson as an alternative, and Straw and Harman were said to be positioning themselves, the most openly ambitious cabinet heavyweight was David Miliband. The fact that opinion polling evidence suggested that Labour would be only marginally better positioned for retaining power if he replaced Brown created doubts amongst potential backers. Despite the risks and costs identified above, it might have been worthwhile if there had been clear evidence that David Miliband could retain power for the party, but overwhelming evidence of this was not forthcoming. Miliband had to calculate not only whether he could defeat Brown, but whether he could defeat David Cameron, the leader of the opposition Conservative Party. Regardless, was it best to experience a brief tenure in Downing Street, between defeating Brown and losing to Cameron, potentially tainting the Blairite brand? As Miliband’s younger brother Ed reportedly told him (at the height of the June 2009 speculation), the leadership was his anyway after Brown, whether in office or opposition (Hasan and MacIntyre Reference Hasan and MacIntyre2011: 187).

Before considering how the survival of Brown compares with the removal of Rudd and Gillard, it is worth considering how the position of Miliband in the 2008–10 period compares with that of Brown prior to 2007. The above analysis suggests that the removal of an incumbent Labour Party leader is immensely difficult due to the procedural obstacles that characterize the nomination processes and the functioning of the electoral college. However, McAnulla (Reference McAnulla2010: 593) develops an argument that seems to challenge this assumption, by arguing that Blair had been ‘compelled to leave office earlier than intended following internal party pressure’. The circumstances that Brown faced as the leading alternative to Blair prior to 2007 were different to those faced by Miliband (and Johnson, Straw and Harman) and reflected the importance of circumstances. Brown was unwilling to challenge Blair formally, but Blair had publicly declared in 2004 that he would not seek to lead the Labour Party into a fourth term. He did, however, stipulate his intention to serve a full third term, implying that the leadership transition would be scheduled for around 2009. Whilst considerably later than Brown had wanted (and expected, according to the infamous and unclear Blair–Brown deal of 1994, which he felt stipulated the mid-point of the second term), acknowledging his intention to stand down provided Brown with an opportunity. As the authority of Blair drifted away, Brown applied informal pressure upon him to set a timetable for his departure. The September 2006 coup initiated by supporters of Brown forced Blair into naming his departure date, after a series of coordinated ministerial resignations (and threats of more) unless Blair agreed to depart (Rawnsley Reference Rawnsley2010: 402–3).

It was because Blair had pre-announced his intention to depart (making the question when, not if) which meant that informal pressures made more impact in the 2005–7 period for Brown. The likelihood of it succeeding was aided by the fact that it was early in the parliament, whereas later on the parliament plotters had overcome the fear that selecting two unelected prime ministers within one parliament might not be acceptable to the electorate (Quinn Reference Quinn2012: 90). And therein was the central question about the direction of the post-Blair Labour Party: the debate about whether, how and when to remove Brown stemmed from the circumstances of his accession to the party leadership. As Kenny (Reference Kenny2009b: 666) argues, the nature of the Blair–Brown feud, both ‘towering figures’ with ‘destructive entourages’, would ‘leave precious little space for other possible successors to emerge’. However, the nature of the transition from Blair to Brown suggests that there are more than two ways to see Labour Party leadership disputes. First, there are forced exits triggered through the formal procedures; second, there are attempts to force an exit by informal pressures which are ignored by the incumbent; and third, there are forced exits triggered by internal pressures but activated outside the formal procedures. As Cross and Blais (Reference Cross and Blais2012) suggest, the third scenario of ‘resigned under pressure’ best describes the departure of Blair, whereas the second scenario best describes the experience of Brown.

The end for Rudd’s leadership of the party in 2010 and Gillard’s in 2013 was swift and brutal. As with Bob Hawke before him (in 1991), Rudd had to endure a hastily arranged press conference, gathering his family around him to shed the tears of an ousted prime minister on live television. Gillard gave a dignified and measured performance, but also close to tears hoped she had ‘made it easier for the next female prime minister’. The simplistic analysis is that Rudd and Gillard were removed because the Labor Party could. Australian party politics had managed to resist the trend towards expanding leadership selection beyond the parliamentary party to the membership (Cross and Blais Reference Cross and Blais2010; Kenig Reference Kenig2009). With only the minor parties in Australia expanding the leadership selectorate, there was an absence of any contagion effect. Party officials interviewed by Cross and Blais cite the short three-year electoral cycle as the most common reason for maintaining the status quo. The main parties were reluctant to indulge in extended leadership selection and be ‘leaderless’. Therefore power to select the party leader remained firmly in the hands of the parliamentary caucus, containing elected members from both the House of Representatives and the Senate. The concentration of elite power exacerbated the role of factions within the Australian Labor Party. Former party leader Mark Latham (Reference Latham2005: 186) called Labor a ‘virtual party controlled by a handful of machine men’. The Australian Labor Party’s factions are not only more entrenched than any other Australian party, but arguably more entrenched than any other social democratic party in the Western world (Boucek Reference Boucek2009; Leigh Reference Leigh2000: 427; Warhurst and Parkin Reference Warhurst and Parkin2000). Factions have a formal presence at all levels of party politics; they hold regular meetings, elect office-bearers, produce newsletters and often policy papers (Leigh Reference Leigh2000).

Yet the machine extends beyond the caucus to extra-parliamentary powerbrokers to the ‘faceless men’ and other key actors. In the wake of Paul Keating’s defeat, the centre left faded (largely through a combination of retirements, defeat and shift of the remaining to become unaligned) and although the right is the dominant faction, the party is largely polarized between two core factional groups of left and right (Gauja Reference Gauja2011). Rudd used the right faction to gain the leadership, but one of his many errors was not cultivating this support in the party room. Factionalism only became entrenched in the Australian Labor Party when a more orchestrated system emerged to allocate party positions in the 1980s. John Howard was adept at working the Liberal party room to neuter his challenger Peter Costello, ruthless in removing his opponents and assiduous in cultivating his supporters (Bennister Reference Bennister2008, Reference Bennister2012). Rudd’s relationship with the powerbrokers in the Labor Party was key to understanding his rise to the top, his demise and then return. With such a small caucus and party room, a prime minister who wishes to survive needs to put in the hard yards in Canberra. Rudd’s efforts to override ‘the machine’ were bound to come unstuck. When elected as prime minister, Rudd broke with tradition (and parliamentary party rules) in announcing that he would not be consulting with factional leaders when appointing his first cabinet (Gauja Reference Gauja2011; Strangio et al. Reference Strangio, ‘t Hart and Walter2013). Initially this was interpreted as an indication of the decline of factional power in the party, but closer investigation shows that it made little difference; faction leaders were consulted and as ever a leader needs to balance state and factional interests (Kefford Reference Kefford2013: 139). Leigh predicted in 2000 that factionalism in the Australian Labor Party, largely driven by ideological bases now no longer as relevant, had reached its zenith. Cavalier (Reference Cavalier2005, Reference Cavalier2012), though, points to factions as ‘executive placement agencies’, having succeeded in capturing the Australian Labor Party organization whereby a small political class dominated decision-making, increasingly disconnected from the Australian public.

The Rudd–Gillard battle demonstrated that factions still maintained a hold on the upper echelons of the Australian Labor Party and the operational mechanisms of an increasingly dysfunctional party. Although policy differences were minimal between the two, the consequences of losing factional support were considerable. Gillard, who owed her position to the factions (coming from the left, but supported in her challenge to Rudd by the right), understood this dynamic better than Rudd. Conscious of how the public tends to be most aware of the Australian Labor Party’s factional politics during such leadership disputes (Economou Reference Economou2010), she realized she needed to legitimize her leadership early on by calling an election. Gillard could not see off Rudd during her three years as prime minister nor counter the notion that she had ruthlessly ousted her predecessor. Much as Brown had posed a constant threat to Blair, so Rudd undermined and destabilized the Gillard minority government. Brown, though, had been inside the tent, Rudd after the 2012 failed challenge was outside the government, acting more like Keating, sitting on the backbench carping and plotting against Hawke.

A simple structural analysis would present the return to Rudd as an acceptance by caucus that they got it wrong in elevating Gillard in 2010 and seized the opportunity to return to the Rudd project. Such analyses expose the dysfunction evident in the Australian Labor Party: ‘There are few checks and balances within the party itself; with whom do these caucus members actually consult before deciding whom they’ll support in a leadership ballot? A few faction bosses, perhaps a union official here and there; but let’s not confuse such methods with democracy’ (Bongiorno Reference Bongiorno2013a). The context of the Gillard premiership does need to be considered. Having seen off the Rudd challenge in February 2012, her position should have been strengthened. However, Gillard lurched from scandal to scandal involving the speaker of the House of Representatives (Peter Slipper), then former MP Craig Thomson’s misuse of union funds (Wanna Reference Wanna2013). Both cases had unfortunate sexual aspects. Gillard came out fighting in Parliament on 9 October 2012 with a stirring attack on Tony Abbott’s misogyny (footage of which went viral). Gillard, as Australia’s first female prime minister, faced a set of unique circumstances that had an impact on her leadership and played a considerable part in the framing of her period of office and ultimate removal. Prior to 2010, Gillard had been subject to sexist attacks regarding her childlessness and marital status, but once she became prime minister the intensity and vitriolic nature of the abuse rose considerably (Sawer Reference Sawer2013). The sexual vilification of Gillard fed into the framing of her as a ‘liar’ (for her broken campaign promises) and ‘traitor’ (for the way she ousted Rudd). She was judged by different standards from her male counterparts (Sawer Reference Sawer2013; Summers Reference Summers2012). Gillard’s admirable effort to counter the media and opposition narrative in Parliament gave her a slight personal rating bounce and bought her some time. She sought to capitalize with a surprise announcement at the end of January that the next election would be on 14 September 2013. In a move designed to end speculation of her leadership and place the party on an early election footing, she had hoped to remove uncertainty from her tenure. It had the opposite effect as it presented Rudd with a timetable to mobilize against her.

Rudd moved swiftly on resuming the premiership (he saw the Gillard period as an unfortunate aberration); in contrast to Gillard’s early election announcement in 2010, Rudd announced a package of party reforms. Party change in Australia had always bucked the trend identified in the literature, whereby factional conflict is temporarily suspended when a party achieves vote success (Budge et al. Reference Budge, Ezrow and McDonald2010). The Federal Parliamentary Labor Party standing orders stated that all members of the caucus were eligible to vote in the leadership election, with the winning candidate determined by a run-off ballot (Gauja Reference Gauja2011). Rudd sought to change this on reassuming the leadership. Central to the reform package, and in the hope of capitalizing on his newfound caucus support, he pushed through a change to the leadership selection rules. The proposals to widen the electorate for leadership ballots to the party membership giving equal weight with the caucus required a formal rule change at a special parliamentary caucus meeting on 21 July 2013. In contrast to the electoral college in the British Labour Party, the unions were not part of the new franchise (Gauja Reference Gauja2013). Proposals to reform the election of party leaders were not new; in fact Rudd had floated the idea in 2011. As noted, Australia has appeared out of step with most other liberal democratic parties in concentrating the election in the hands of a small parliamentary caucus. However, the appearance of a sudden democratic conversion should be tempered with a political reality check (Gauja Reference Gauja2013; Manwaring Reference Manwaring2013a; Quiggin Reference Quiggin2013). Rudd drew his strength from the party membership and by presenting reform early in his second go as party leader (with his political capital momentarily high again) forced through the issue to make sure that his personal position was bolstered. Rudd also set a high threshold at 75 per cent of the caucus required to remove an incumbent leader between elections (though this was swiftly reduced to 60 per cent by caucus). Rudd had perhaps hoped to entrench his own position as leader, by effectively creating a fixed-term leadership, protected from intermittent caucus challenges. In the event the Australian Labor Party election defeat in September 2013 saw Rudd step down as leader. Bill Shorten became the first Australian Labor Party leader to benefit from the selection changes, defeating Antony Albanese on 13 October 2013. Shorten won with 64 per cent of the caucus vote and 40 per cent of the party membership vote, giving him an overall weighted vote of 52 per cent (Harrison Reference Harrison2013). Ironically it was the caucus vote that secured the leadership for Shorten. The new rules, though introduced hastily (party members ended up paying for it), appeared to discourage any factional strife and initially at least presented a more united party (Bongiorno Reference Bongiorno2013b; Manwaring Reference Manwaring2013b).

The political culture of party leadership in Australia has always been ‘brutal’. A ‘spill’ could be organized at short notice and defenestration was swift and ruthless. Though more common in opposition than government, it has been a powerful feature of party politics. Davis (Reference Davis1998: 172) observed that ‘beyond doubt, party leadership in Australia operates on a Darwinian scale unmatched elsewhere in the Western parliamentary democracies’. The oligarchic nature of party organization ensures that party leaders need to satisfy, placate, manipulate or cajole their peers to survive in post (Bynander and ‘t Hart Reference Bynander and ‘t Hart2007). As Rudd and Gillard found out, once leadership speculation gets going in Canberra a cocktail of party powerbrokers and political journalists can easily destabilize an incumbent prime minister. A devastating critique of Rudd by journalist David Marr (Reference Marr2010) represented a tipping point in Rudd’s fortunes for the public, meanwhile concerted internal party opposition was mobilizing (Evans Reference Evans2010: 261). Although Rudd’s micromanagement tendencies, short temper and indecision were known, Marr put a damaging perception of Rudd into the public domain. Once such speculation is set in motion it becomes a self-fulfilling prophecy, as leadership consolidation is an elusive commodity in Australian politics (Bynander and ‘t Hart Reference Bynander and ‘t Hart2007). Rudd and Gillard, in common with many of their predecessors, had to deal with interpersonal conflict and rivalry at the heart of the party. Factional politics meant he had no choice but to work with and through his deputy, who clearly coveted his position. Challengers regroup and fight again, as Keating did in 1991, but also leaders can hang around to fight to regain the crown, as Howard did successfully and Peacock unsuccessfully in the Liberal Party. Beazley had two spells as Australian Labor Party leader and so Rudd’s return was not so unusual. The vanquished in Australian politics are reluctant to leave the stage, desperately clinging on to power (Weller Reference Weller2012). Keating summed it up in characteristic forthright style: ‘You know, prime ministers have got Araldite on their pants, most of them. They want to stick to their seat. And you either put a sword through them or let the people do it’ (Bennister Reference Bennister2012: 128; Brett Reference Brett2007: 24). The parliamentary caucus dynamic and machine politics create an Australian leadership setting in which ‘hypocrisy, deceit and plotting are endemic’ (Bynander and ‘t Hart Reference Bynander and ‘t Hart2007). The widening of the leadership selectorate is unlikely to remove this abrasive political culture – after all, factions still control pre-selections and national conference – with such a ruthless ‘coup culture’ ingrained at state as well as federal level (Bryant Reference Bryant2013). Yet the Australian Labor Party may have set in train a party reform that will provide a level of stability all leaders crave.