Agency,institutional stretch and structural adjustment

Introduction

The capacity for political leaders to engage in institutional stretch has been well established in Westminster systems (Bennister, 2007; Foley, 2000; Heffernan, 2005: 69). But how much impact can party leaders have on structural reform? Can they really act as change agents or are they defined by the structural conditions they face? This article will explore these questions by examining the way two recent leaders of the Australian Labor Party (ALP), Julia Gillard and Kevin Rudd, engaged in institutional stretch and structural reform. To achieve this aim, this article will utilise a contextual analysis to examine the forces and conditions which gave rise to one of the most remarkable periods in the history of the party. This article will also demonstrate that when it comes to the ALP, ambiguity, opportunism and calls for unity have been central to the capacity of party leaders to stretch their institutional authority and to drive structural reform in the party. Whilst these three justifications have been employed at opportune moments by party leaders, this alone fails to explain the events of this period. In fact as will be shown, politically strong and politically weak leaders alike were able to exercise agency to alter intra-party norms, move beyond their prescribed authority and drive structural reforms.

As recently as two decades ago, political leadership was a largely neglected area of the literature. Fortunately this has been reversed of late thanks to new frameworks and innovative approaches which provide scholars with the tools to conceptualise and identify better changes in both the way leaders exert power and in the institutional environment they work within. In particular, the rise of approaches grounded in comparative politics has stretched tired old debates (Rhodes, 2013: 318). Contextual analysis or the ‘interactionist approach’ as Elgie (1995: 8) has referred to it, has become an important part of this rejuvenation in the study of political leaders. This approach positions context and the environment in which political leaders operate as incredibly important, while also acknowledging that individual agents are important and can shape their institutional environment (Elgie, 1995: 5-8). More recently, this has been updated further. Walter (2013: 53) demonstrates, after exploring the performance of four prime ministers during the global financial crisis (GFC), that outcomes are a combination of both circumstance and character. While ’t Hart (2014: 211) contends it is not only ‘when and how context shapes leadership possibilities’ that is important but also when and how ‘leaders are able mould the context in which they operate’. This approach to studying agents, who in this instance are party leaders, to shape their environment can also be seen as having ties to historical institutionalism (HI). For instance, in the view of Bell (2011, 2012) and Bell and Feng (2014), by avoiding ‘sticky versions of HI’ and instead by using an ‘agents in context’ approach to historical institutionalism, both exogenous and endogenous changes can be successfully explored. The approach of this article is to explore the context in which the ALP as an institution evolved, while recognising the importance of party leaders as central agents in these institutions. By linking the study of political leadership more explicitly to contemporary debates about institutions and agency, the goal is to move beyond what Rhodes (2013: 318) has referred to as the ‘hoary old chestnuts’; broadening and deepening the theoretical basis of the field. This approach is therefore informed by the work of scholars who have used a nominally HI approach to examine agents working within various institutional arrangements. Hargrove’s (1998) ‘skill in context’ approach which was designed to examine the US presidency, but has been applied comparatively, is one example of this. According to Bell et al. (1999: 529), ‘the model permits the comparison and assessment of personal political skill in relation to contextual factors’. A second example is the highly influential work on US presidents from Skowronek (1993), in which this author argues that patterns exist when the US presidency is examined which are related to the context presidents inherit from their predecessors. This process tracing of the importance of political context has also been transplanted to Westminster systems. Laing and McCaffrie (2013: 98–99) in examining three Australian prime ministers contend that Skowronek’s theory can be valuable in Westminster systems as well, but most of all, that context should not be underestimated (see also Helms, 2012: 8; Bennister, 2012: 8-12).

Leadership, agency and power

Political leaders have institutional power as a result of the position they hold, but the capacity of agents to exert power, or what Dowding (2013: 58) refers to as ‘power to’, is also shaped by the contextual and situational. Mahoney and Thelen (2010: 4) argue that ‘institutional change often occurs precisely when problems of rule interpretation and enforcement open up space for actors to implement existing rules in new ways’. This argument about the problems of rule interpretation is especially acute for parties such as the ALP which have undergone clear changes in direction in the last four decades. The literature on the evolution of political parties in liberal democracies is long and distinguished. Since the 1960s scholars have examined the way that parties have transformed from the classic mass party model to catch-all, electoral-professional and cartel parties (Kirchheimer, 1966; Panebianco, 1988; Katz and Mair, 1995). These changes have had a dramatic effect on intra-party power and the relationship that party leaders have with their parties. Likewise some scholars contend that the personalisation of politics in parliamentary democracies has had an enormous impact on the types of campaigns parties run and the way party leaders interact with their party (McAllister, 2007: 582; Heffernan, 2007). ¹ While Foley (2000) referred to the autonomy leaders could possess as spatial leadership, according to Karvonen (2010: 4) personalisation refers to a change in the prominence of individual actors over collective groupings such as parties. While the extent that personalisation impacts on voting intentions has been hotly debated for the best part of a decade, the evidence indicates that leaders have at the very least a small effect (Karvonen, 2010; Bittner, 2011; Webb et al., 2012).

In contrast to these debates internationally, the Australian literature has been emphatically supportive of the personalisation thesis. Bennister (2012: 163–166), Walter and Strangio (2007: 62) and Young (2003: 109) have all identified personalisation in the Australian polity. Australia’s majoritarian system and the norms that encompass the polity provide party leaders with advantages which contribute to a particularly leader-centric style of politics. Publicly funded elections, compulsory voting, tight election cycles and one of the most concentrated levels of media ownership for a liberal democracy in the world all contribute to this. It is evident since the late 1960s in Australia that both major parties, the Liberal Party and the Australian Labor Party have shifted towards an electoral–professional model (Brett, 2013: 177; Ward, 2006; Warhurst and Parkin, 2000). While the Liberal Party has historically given the party leader substantial authority, the ALP in contrast was once the very model of a mass party. However this is now in the past. According to Walter and Strangio (2007: 11), ‘Once sceptical of leaders, the Labor Party shows signs of having become equally, if not more, addicted to the messiah complex than the Liberal Party’. Arguably, changes to the media and the three-year election cycles contribute to this and the ‘the permanent campaign’ which has become a defining feature (Rhodes et al., 2009: 96). Ultimately, the decline of the collectivist nature of many parties (especially centre-left parties) has contributed to rising ambiguity over who has and who should have power to decide policies and procedures. It is here that opportunistic ALP party leaders have sought to mould the institution around them, leaving an institutional legacy for those that follow.

This article will explore the case study of the ALP between December 2006 and October 2013. It will examine the way pre-selections were dealt with in the first Kevin Rudd leadership period of the ALP (2006–2010), that of his successor, Julia Gillard (2010–2013), as well as in Rudd’s second period in charge (2013). This case demonstrates the extent of the changes in the ALP, with party leaders in this period engaging in institutional stretch to exert authority over the pre-selection of candidates for Australian federal elections. This is far beyond the prescribed authority they hold. Furthermore, this article will demonstrate how Kevin Rudd, during his second period as party leader, exercised agency at an opportune moment to reform the party structurally. The transformation of the process to select the federal parliamentary leader of the ALP, poses certain conceptual challenges to the dominant literature on expansion of the leader selectorate (Cross and Blais, 2012). The importance of this case is not only that the ALP has been resistant to structural changes, but also that Australian political parties more broadly have been slow to reform in ways that have occurred in other Anglo parliamentary systems. ² Moreover, the pace of change and the absence of the variables which have defined transitions internationally indicate that this is a vital case to examine and to compare to the international literature. Therefore, not only is the case deviant, but the process which led to reform will also be shown to be deviant. This article will conclude by placing the ALP case within the international literature and by highlighting the way that the literature on party leaders can be placed within the broader literature on agency.

The ALP and the power of its leaders

Australia’s major political parties are extremely strong and highly disciplined (Jaensch, 2006: 24). They have a unique combination of advantages which empower them and contribute to their continued dominance of the Australian political scene. Compulsory voting, publicly funded elections and preferential voting in the lower house are part of this. While some evidence exists of voter disillusion with the major parties, demonstrated by growing levels of informal voting (Green, 2013), Australia still has comparatively high levels of party identification. In 2006, Jaensch (2006: 24) argued that the ‘parties and the party system are more than ever, the key to understanding the polity’ in Australia. There is little to challenge this view nearly a decade on. The ALP is Australia’s oldest continuous political party and one of the world’s oldest ‘Labour’ parties (Dyrenfurth and Bongiorno, 2011: 13). According to Rhodes et al., (2009: 105), its traditions are ‘determinedly collegial’. In other words, the party has a strong historical sense of collectivism. The key documents of the party spell out clearly defined roles and procedures for the intra-party institutions. The parliamentary caucus is theoretically bound by decisions of the national conference as they were viewed historically as the delegates of the broader movement, which in turn is legislated as the chief decision-making body of the party. Elected at each conference is the National Executive, which is viewed as the chief administrative body of the party. The party also has strong factions, some of which are related to its continued relationship with the union movement with others being state-based groupings. The state branches still play an important role in assisting the national office and in organising their own campaigns during their respective state elections. Leaders of the party are therefore required to interact with a complex web of parliamentary and organisational wing institutions. Successfully navigating the Byzantine structures of the party has historically been a challenging job for even the most persuasive and respected of leaders. In contrast to similar parties such as New Labour in the UK, the ALP has, since the late 1960s, largely resisted major structural changes to the intra-party institutions and key processes.

Warhurst and Parkin (2000: 31), Jaensch (2006: 24) and Walter and Strangio (2006: 56) all agree that changes to the ALP really began from the time Gough Whitlam was elected leader in 1967. Before this time, the politicians had little representation within the intra-party institutions, and their primary source of power was derived from their election to parliament. These institutions – such as the National Conference and the National Executive – were dominated by non-parliamentary actors. By the time Whitlam became party leader in 1967, he and others were adamant that the party required structural changes and in 1969 the party resolved to reform the intra-party institutions. These reforms included dramatic changes to the structure of the National Conference, providing representation for the parliamentarians, and granting the parliamentary leadership full voting rights on the National Executive (Oakes, 1973: 154–155; Freudenberg, 2009: 94). ³ While there had been broad support for these changes and the structural conditions were in place for such a change, these reforms were not easily achieved. Whitlam and others had to negotiate, cajole and twist arms, the process starting even before Whitlam became leader in 1967. This last instance of major structural reform contrasts greatly with the pace and processes of change in 2013. ⁴ While differences between these periods should be expected, the contrast highlights changes in intra-party norms and the personalised nature of party leadership in the contemporary ALP. Arguably it was from the Whitlam reforms that party leaders had an increased potential to act opportunistically, using ambiguity over rules and calls for unity to involve themselves in processes that were beyond their prescribed authority.

Institutional stretch and leadership instability

By the end of 2006, the ALP had been routinely frustrated by the Liberal Party and its coalition partners, The Nationals, for a decade. However, from the time Kevin Rudd became leader of the ALP in December 2006, the party remained well in front in the polls and in November 2007, the party won a resounding election victory. Despite this, the ALP’s election victory masked deeper problems within the party. Towards the end of 2006, it became apparent that three contenders would vie for the parliamentary leadership. This included the incumbent, Kim Beazley, Rudd and Julia Gillard. While Rudd was more popular in the electorate than his two main rivals according to Newspoll (2006), inside the party Rudd was far less popular. It became widely known in Australia, that Rudd had at best the same level of support as Gillard, which individually was still less than that of Beazley in the lead up to the caucus ballot (see Hartcher, 2009: 158; Dyrenfurth and Bongiorno, 2011: 186; Jackman, 2008: 56). Ultimately, Rudd and Gillard decided to run on a joint ticket, with Rudd as the leader and Gillard his deputy and, on 4 December 2006, Rudd won the caucus ballot 49 votes to 39.

Shortly after Rudd became leader, he attempted to demonstrate why things would be different under his leadership. In particular, he claimed that he would not follow a rule established in 1905 that said caucus alone had the right to pick the front bench (Hartcher, 2009: 146). Despite this being reported as a key change in the way the party functioned, substantial evidence emerged that Rudd was still negotiating with the factions (Hartcher, 2009: 145–146 and Stuart, 2010: 69). ⁵ In April 2007 the party conducted its national conference well aware that an election would in all likelihood be called in the second half of the year. At this conference, a controversial decision was made relating to the pre-selection of candidates for the forthcoming federal election. Traditionally, the pre-selection of ALP candidates for Australian federal elections was entirely decentralised. According to Marsh (2006: 4), ‘Local committees whose primary attachment was to a district, not to a central organisation, selected candidates’. However, the conference resolution advocated by Rudd and the parliamentary leadership made specific reference to the limited time the party had before the election. So, in order to maximise the chance at success, the conference resolution noted that it gave the National Executive ‘specific authority to pre-select candidates in the House of Representatives for the 2007 federal election’ (ALP, 2007). The resolution also noted that this would only apply to 2007 federal pre-selections and that this process would be applied exclusively to seats in the state of New South Wales (NSW). The resolution did however note that the National Executive would only intervene where ‘application is sought by the Administrative Committee of another State or Territory Branch, and only in those seats where the National Executive Committee has unanimously agreed upon a process to select a candidate’ (ALP, 2007). After the conference, then National Secretary, Tim Gartrell, defended the position by arguing ‘Sometimes the rank and file gets it wrong; sometimes the executive gets it right’ (cited in Farr, 2007: 13). In reality, the application of the policy was much wider than initially envisaged with at least 10 seats in the state of New South Wales alone having their candidates chosen by the National Executive (Coorey and Humphries, 2007; Kefford, 2013: 140). Moreover, while sections of the party were unhappy with this decision, the leadership called for unity because it was an election year and they were in front in the polls.

In the first few months of 2010, as the federal election drew closer, it became apparent that intervention by the National Executive into pre-selections would not remain restricted to 2007. The ALP, having governed since November 2007, while remaining in a strong position in the polls (Newspoll, 2010a), still needed to begin the process of pre-selecting candidates for the election. By March 2010 the evidence suggested that the National Executive had maintained, if not increased their willingness to exert authority over the process (ALP, 2010; Bowe, 2010; Maddison, 2010). The evidence also suggests that the National Executive were not using the motion passed at the 2007 National Conference to justify their actions. Instead, they were using the plenary powers under Section 7 (f) in the ALP Constitution (Cavalier, 2012; ALP, 2010). These powers, which had been used extremely rarely until 2007, were now being used extensively to decide candidates for pre-selection in NSW as well as in Queensland and internal opposition started to swell (ALP, 2010; Maddison, 2010; Faulkner, 2011: 8). Rodney Cavalier (2010a: 185), a former NSW Labor Minister, argued that by using the plenary powers that the leadership group had ‘presumed to determine pre-selections for any seat that Labor has a chance of winning wherever the local membership might select a candidate unacceptable to that group’. He added of Rudd, that ‘No previous Labor Leader has so casually but consistently used the National Executive to impose his will in candidate selection’ (Cavalier, 2010b). Furthermore, in 2010 two specific examples of Rudd personally intervening in pre-selection battles emerged. In April 2010 it became widely known that Rudd had personally insisted that the National Executive overturn the decision of the Tasmanian state office to select Kevin Harkins, a union official, for the top spot on the state’s Senate ticket (ABC, 2010; Denholm, 2012). ⁶ Furthermore, evidence emerged in the Australian Capital Territory (ACT) that Rudd had exerted pressure on a leading candidate for the seat of Canberra to withdraw from the process (Giesecke, 2012: 16). One ALP member, who wrote an account of the pre-selection process in the ACT, argued that Rudd had become ‘Lord High Executioner’ and that ‘previously the Parliamentary leader’s writ only extended to the parliamentary party. Issues such as preselections were left to State/Territory branches and the National Executive’ (Giesecke, 2012: 16).

By June of 2010, however, the fortunes of the ALP and Rudd started to deteriorate. Polling numbers for the party and Rudd personally were beginning to follow a downwards trend, though most polls still had the ALP in front (Newspoll, 2010b). In particular, perceived policy failures and difficulties Rudd’s colleagues were having with him were cited as central reasons for this. While it seemed unlikely that the party would change leaders, given that this was still a first term government, Rudd’s parliamentary colleagues informed him on 23 June that he no longer had their support. The following day, Julia Gillard, his deputy, was elected unopposed. While the evidence from this period indicates that at times Rudd worked with the National Executive – and as a result the factions – he also intervened personally in at least three pre-selections. While Rudd working with the National Executive and the factions to decide pre-selections could be viewed as a centralisation of the process, his personal intervention into some pre-selection battles is evidence of institutional stretch. If the interventionism into pre-selection had then ceased, it could have been viewed as contextual or conditional due to the actors involved. This includes Rudd, who had a history of attempting to centralise decision-making (Kefford, 2013: 138). That Rudd’s successor continued to intervene in pre-selections, however, reveals much about changes to the way party leaders of the ALP could make calls to an individual rather than a collective mandate to expand their sphere of influence. This highlights the personalised nature of party leadership in the contemporary ALP.

In January 2013, Gillard demonstrated this when she chose to exercise what she described as a ‘Captain’s pick’ to select Nova Perris for pre-selection for the top spot on the Northern Territory Senate ticket for the party. This was despite the current Senator, Trish Crossin, wanting to contest the position again (Cullen, 2013; Ireland and Harrison, 2013). Moreover, in the months leading up to Gillard’s decision to promote Perris over Crossin, Gillard was politically and electorally weak. The polls were very unfavourable with most showing the ALP well behind on a two-party preferred basis when the decision was made. Gillard was also unpopular on a personal level with a long term negative satisfaction rating (see Newspoll, 2013; Essential Research, 2013; Galaxy Research, 2013). In addition, Gillard was beset by accusations about the functioning of the minority government she was leading, including the agreement the ALP had made with the Greens. This example serves to highlight some important points. First, both strong and weak leaders were able to appeal to this individual rather than collective mandate. Second and inter-related, the personalisation of party leadership in the ALP is apparent in two leaders of different styles and with a different history in the party. While Rudd was an ‘outsider’ in many ways, Gillard was very much a traditional ALP ‘insider’. Third, that while Rudd had primarily worked through the National Executive ‘court’ with factional and competing interests, his personal interventions had been to veto candidates, not to actually individually choose candidates. Gillard, while still needing the National Executive to agree to the selection of Perris, stretched the authority of the leader even further than Rudd had up to this point. ⁷

Structural adjustment, opportunism and the institutional stretch continues

The changes in leadership were not over, though. In June 2013 the parliamentary party changed leaders again. On 27 June Rudd was re-elected by the caucus (57 votes to 45) after a sustained and relentless campaign which included a prior challenge in February 2012 which Gillard comprehensively won (71 votes to 31). Shortly after returning to the leadership, Rudd once again demonstrated his opportunistic nature, arguing that he wanted to make two changes to party processes. The first change was to the way the party selected the leader of the federal parliamentary party. The reforms which the caucus endorsed were almost identical to what Rudd had proposed and included: (1) grassroots members of the party being allocated 50% of the vote in choosing a parliamentary leader, with the remaining 50% allocated to the federal caucus; (2) that if a leader wins an election they remain leader for the duration of the term; the exception to this is if the leader resigns, requests a leadership election, or if 75% of caucus members sign a petition calling for a leadership ballot ‘on the grounds that the current leader has brought the party into disrepute’; and (3) candidates for the leadership must be nominated by 20% of caucus and ballots to decide the leader will automatically occur following an election loss (Murphy, 2013; Kenny and Aston, 2013).The reforms were unanimously endorsed by caucus, despite reports of deep opposition to the changes because a federal election was only months away. ⁸

Arguments that party reforms were vital can be traced as far back as 1979 (ALP, 1979). These recommendations are identifiable in the writings of party insiders (see Button 2002; Latham, 2005), and in many of the internal party reviews of election results (see ALP, 2002; ALP, 2010). However, most of the reforms proposed had far more to do with encouraging grassroots democracy rather than dealing with the way the parliamentary leader was selected. In spite of this, intra-party inertia has ground many proposed reforms to a halt. ⁹ So, when Rudd proposed his dramatic changes to the way the federal parliamentary leader was selected, some commentators were understandably sceptical of Rudd’s motivations. The rationale for the changes, including giving the rank and file an equal say as that of the caucus in deciding the leader, were spelt out in unambiguous terms by Rudd. On 22 July, after the changes gained the support of the caucus, Rudd argued that the changes meant that ‘the prime minister the Australian people vote for is the prime minister the Australian people get’ (Rudd, cited in Ireland, 2013).

The second change Rudd was publicly promoting was a return to ‘local democratic ballots’ (cited in Ireland, 2013). Commenting specifically on the removal of Crossin for Perris, Rudd (cited in Harrison, 2013) claimed ‘The bottom line is this: I don't like the way in which it was handled, that's the truth’. In spite of the rhetoric Rudd was using in public, in private he was once again working with the National Executive to impose candidates for a number of electorates across the country. This included seats that had already conducted local pre-selection ballots. For example, in the seat of Forde, in the state of Queensland, Labor’s pre-selection process was complete before Rudd announced on 4 August that the election would be held on 7 September. On 8 August, however, it was announced that the democratically elected candidate had been replaced by former Queensland state premier Peter Beattie. The evidence suggests that the decision to parachute Beattie into the seat, while ratified by the National Executive, was done so at the behest of Rudd (Griffiths, 2013; Marszalek et al., 2013). This was not the only case where intervention from the parliamentary leadership and the National Executive were apparent. According to numerous reports, Rudd and his supporters had also unsuccessfully been attempting to parachute a Department of Foreign Affairs official into the safe Labor seat of Lalor in Victoria, because Julia Gillard had decided to retire from politics and the seat she had held since 1998 (Crook, 2013; Millar and Willingham, 2013). Moreover, two party officials asserted that Rudd had initially planned to intervene in five additional seats in Queensland, replacing candidates who had also already been pre-selected (Party Official A, 2013; Party Official B, 2013).

Speaking before the 2010 federal election, Lindsay Tanner (2010), then Minister for Finance and Deregulation, argued that,

the leader has a certain amount of what you might call reserve power in regard to pre-selection which can enable him or her to determine the outcome of an individual pre-selection on the odd occasion…and in my view can only be exercised pretty rarely. If we ended up with a position where we had a leader who was seeking to directly interfere in pre-selection contests across the board and to determine outcomes across the board that would create a lot of controversy and political upheaval.

After Rudd’s return to the leadership in 2013, he attempted publicly to capitalise on the unpopularity of the structures of party, while privately attempting to exert authority over the process. The evidence therefore suggests that the opportunism displayed by both leaders, combined with ambiguity over who holds the final authority over pre-selections contributed in some way to a stretching of party leader authority. However, this is not the whole story. Evidence of the personalisation of the party leadership is also apparent. The initial push in 2007 could be logically explained as contextual in that the ALP had its best chance of electoral victory since before losing office in 1996. ¹⁰ Also, in the lead up to the 2010 election, a possible explanation could be that, after winning office, the party was prepared to grant an electorally popular prime minister some leeway. Where political context has less explanatory power is in what followed. While Gillard was a sitting prime minister, the context she found herself in was far different to that of Rudd. Therefore, the evidence suggests that while Rudd used context to stretch his sphere of influence, Gillard relied on the institutional power of the prime ministership to impose her will. The capacity of Rudd to drive unexpected structural reform after returning to the leadership also appears heavily reliant on context, which in this instance was potential electoral defeat. Nevertheless, in each of these moments the calls to a personalised over a collective mandate are identifiable. The personalised nature of the ALP leadership led not only to institutional stretch, it also led to structural adjustment. Considering how comparatively slow Australian parties have been in changing this process and the inertia within the ALP to reform, this highlights this case even further (see Cross and Blais, 2012).

Both the process which led to a democratisation of party leader elections and the ALP case poses challenges to the current party reform literature which need to be reconciled. In particular, the conditions in which the ALP moved to broaden the party leader selectorate challenges the four central hypotheses in Cross and Blais’s seminal work (2012). These are:

parties will only expand their leadership selectorate after an electoral setback…parties are more prone to broaden their selectorate when in opposition…new parties will more easily adopt rules that allow a greater role for rank-and-file members…parties will be more inclined to move to an expanded selectorate when one of their competitors has already moved in that direction. (Cross and Blais, 2012: 130)

Conclusions

This article has examined the case study of the ALP in the period December 2006 to October 2013. The ALP case during this period is an important one to examine, for a number of reasons. First, it highlights the impact of personalisation on Australian politics and indicates just how much the ALP has changed. This is shown by the contrast to the difficulty Gough Whitlam had during his time as party leader. In particular, it shows the way party leaders are able to stretch, twist and shape their institutional environments more than previously. The personalisation identifiable in this instance is what Balmas et al. (2012: 37) argue is centralised personalisation, which is where ‘power flows upwards from the group to a single leader’. Arguably, it is the idiosyncratic structural conditions in Australia which have contributed to personalisation.

Second, this article has examined the context in which the party underwent dramatic structural reforms. The change in the way the party selects the federal parliamentary leader was swift and unconventional. In fact, compared to changes in the other Anglo parliamentary democracies, the ALP represents a deviant case of institutional change. Gerring (2007: 89) argues that deviant cases ‘deviate from some cross-case relationship’. None of the central conditions Cross and Blais (2012: 130) identified fit neatly with the ALP case or the process to reform. This is perhaps not surprising, however, because Australian parties have for a long time been seen as outliers to the other Anglo democratic parties. In spite of this, it is worth remembering what Lijphart (1971: 686) has argued about deviant cases: ‘Deviant cases weaken a probabilistic hypothesis, but they can only invalidate it if they turn up in sufficient numbers to make the hypothesized relationship disappear altogether’. Therefore, the theory about reforming the leader selectorate needs to be reconciled with this new evidence from the ALP case. Moreover, whether a contagion effect does occur in Australia will need to be charted and included in any revision of the current theory.

Helms (2012: 8) has argued that, ‘…how much a leader mattered for the outcome of a given situation cannot really be assessed strictly empirically’. This is certainly the case in the events examined in this article. While evidence of party leader agency is readily available, causation is still part of a broader ‘black box’ problem. At a minimum, however, we can say that the evidence indicates that both Rudd and Gillard as party leaders exercised their agency to stretch their institutional authority and, in Rudd’s case, structurally altered the leader selectorate of the party.

That both strong and weak leaders (politically and electorally) were able to shape party norms was vitally important in demonstrating the personalisation effect on party leadership . By using a contextual analysis, this article has been able to chart the way institutional authority can endogenously and incrementally change in a political party such as the ALP.