Selecting party leaders,reform processes and methods: Examining the Australian and New Zealand Labour parties

NZ Labour

Expanding the leadership selectorate began to be considered in the early 1990s, when a party activist promoted the idea of adopting an ‘electoral college’ process similar to that used by the UK Labour Party. Tim Barnett, who would go on to serve as an MP and as party general secretary, was first to raise the issue in a formal way through a discussion paper (interviews 1, 2, 4). Up until this time, MPs alone chose the leader for a three-year term at the beginning of the second year of each parliament. Senior party officials and activists from the time recount there was a general uneasiness within the party’s grassroots, and particularly among progressive activists, in relation to the recent selections of David Lange, Geoffrey Palmer and Michael Moore who had taken the party in a liberal economic policy direction (interviews 1, 2, 3, 4, 5, 19). With the selection of Helen Clark in 1993, the issue largely subsided as the new leader was popular with the membership, and once the party came to power in 1999, the issue of internal organizational reform was off the table as the focus shifted to issues of governing and public policy. Coupled with this was the necessity to adapt organizationally to the new mixed-member proportional (MMP) electoral system, making it an inopportune time for other organizational reform (interviews 1, 4, 19).

Towards the end of Clark’s leadership there remained undercurrents of support for greater involvement of the rank and file. Several prominent MPs from the time stated that they thought the issue was likely to come to the fore after Clark’s tenure, and particularly if the party found itself in opposition (interviews 2, 3, 4). To paraphrase one MP, ‘if Helen’s successor is quickly anointed without broad consultation, there is likely to be a backlash among members saying “hey, why didn’t we have any say?”’

Indeed, this is close to what happened. Clark was quickly replaced by Phil Goff following her 2008 resignation. A senior MP at the time recounts how parliamentary colleagues were on the phone on election night and that a consensus for Goff quickly developed (interview 6). Although Goff may well have been the choice of party members, several senior party officials recounted grumbling among their local supporters that he was ‘anointed’ without consultation with the rank and file (interviews 7, 9, 18). When Goff resigned, following a loss in the 2011 election, the party organized a series of meetings around the country at which the leadership candidates presented their case and took questions from party supporters (Trevett, 2011). The idea was that MPs would consult with local activists before making the choice. Parliamentarians suggest that some of their colleagues took this initiative to heart, whereas others largely ignored it (interviews 13, 18). In the end, the candidate selected by MPs, David Shearer, did not appear to be the first choice of members, increasing frustration (interview 19). One MP referred to this 2011 contest as having ‘let the Genie out of the bottle’ in terms of opening up the selection process, a characterization agreed with by others (interviews 8, 10, 13).

The opportunity to change the process presented itself with the election of Moira Coatsworth as party president in 2011. In her maiden speech, Coatsworth talked about the need for comprehensive organizational reform. She became a champion for greater transparency and member participation in party decision making (interviews 12, 16, 20, 23). After the party’s second consecutive electoral defeat, Coatsworth launched an organizational review process in February 2012. Emerging from this was a proposal to enfranchise party members and trade union affiliates in the leadership selection process. After consultation with MPs, this was formalized by the party’s New Zealand Council into the formula described below. The recommendation was officially adopted at NZ Labour’s November 2012 annual conference (NZ Labour, 2012d).

ALP

The ALP experienced a rather different trajectory of reform. Until 2013, leaders were chosen exclusively by the parliamentary party. Between June 2010 and July 2013 the party underwent a period of leadership instability with the sudden removal of the sitting prime minister (PM), Kevin Rudd, who was replaced by Julia Gillard. Gillard called an election shortly after becoming leader at which the party secured victory and formed a government. However, the legitimacy of Gillard’s leadership suffered from public dissatisfaction with regard to the way in which she was installed and the factional interests that were believed to be at play. The party’s popularity declined and after two unsuccessful leadership challenges, Rudd ousted Gillard as leader in June 2013 (Griffiths, 2013).

Immediately upon becoming leader again, Rudd proposed changes to the leadership selection process to allow the membership a vote. In the event of a leadership ‘spill’ – a motion to declare the position of leader vacant – party members and parliamentarians would vote with both groups weighted equally. The rationale, Rudd suggested, was to ensure the continuity of an elected PM and to give more power to ‘everyday’ members of the party at the expense of factions, which were instrumental within the parliamentary party but less institutionalized within the grassroots organization (Rudd, 2013a). These motivations reflected a desire to address the circumstances in which Rudd was ousted, and to avoid further electoral damage to the party caused by ‘persistent and ultimately debilitating leadership instability’ (Garrett and Dick, 2014: 2). Although the idea of rank-and-file involvement in leadership selection was floated prior to the 2011 ALP national conference, it never made it on to the conference agenda (Gauja, 2017: 153; interviews 26, 27, 28).

Unlike the reform process in NZ Labour, the ALP reforms were ushered through the parliamentary caucus after discussion with the leadership and full ministry teams (Rudd, 2013a). According to Rudd, a change to the caucus rules would be sufficient for changing the process and the proposal need not go to a meeting of the National Conference (Rudd, 2013b). The reforms were officially adopted in a special meeting of the parliamentary party on 22 July 2013. The timing of the reform announcement (less than three months out from a general election) was effective in minimizing dissent from caucus colleagues. At the 2015 National Conference, the process was partially codified in the ALP’s National Constitution. What was enshrined was the requirement, in cases in which the parliamentary party decides that the leadership selection will be open to the members, for members’ votes to be weighted equally with those of parliamentarians (Article 27). The constitution states that the rules for the conduct of the election are matters for the National Executive and parliamentary party to determine, maintaining a significant degree of flexibility to change the process.

Comparing reform trajectories

There are several points to be drawn from this overview. First, organizational reform is more likely to occur when a party is in opposition and after an electoral setback. This is a pattern observed by Cross and Blais (2012a) and Harmel and Janda (1994) in many parties. The common explanations are that there is a shift in the balance of power from the parliamentary to the organizational party and that parties removed from government, typically, find themselves in need of organizational rejuvenation, have the time and capacity to focus on organizational matters and see the expansion of the leadership selectorate as a way of attracting and energizing local supporters.

The NZ Labour parliamentary party was dispirited in the aftermath of the 2011 defeat, having won a historic low of just 27% of the party vote. At the same time, the organizational party was led by a new leader committed to reform. There was also widespread sentiment among the leadership that it had not successfully made the organizational adaptations necessary under MMP. The result was little to no resistance from MPs when the party launched its comprehensive reform project. Parliamentary leader Shearer, along with Coatsworth, was co-chair of the project, and a team of MPs was included in the project’s leadership. Although the literature often refers to the importance of party leaders in pushing for reform (Gauja, 2017: 149), this case illustrates leadership from the organizational party.

The experience of the ALP was quite different. Kefford (2016: 518) has argued that ‘none of the central conditions Cross and Blais . . . identified fit neatly with the ALP case or the process of reform’. Although the party was not in opposition, it is clear that electoral considerations were still a primary motivation. The changes sought to achieve stability, mitigating the negative electoral impact successive leadership challenges had created for the party. Perhaps the issue on which the ALP differed most from NZ Labour is that change did not take place with a corresponding shift in power to the extra-parliamentary organization. As we explain in the following section, this trajectory of reform influenced the range of associated decisions. As an elite-dominated process, issues such as the role of trade union affiliates were never addressed.

Defining the selectorate

Once a decision is made to expand the selectorate, the two principal sub-decisions are who to include and how to allocate vote shares. The most commonly used process is the delegated conference (Pilet and Cross, 2014). Parties adopting this method often automatically appoint their parliamentarians and other party leaders as delegates to ensure there is balance between rank-and-file and elite-based representation. Although variants of this method were at the time in use by the Green Party and the Maori Party in New Zealand, there is no indication that it was given serious consideration by Labour. Similarly, this method was not considered in Australia. This choice is consistent with the pattern of individualized rather than assembly-based voting that has characterized party membership reforms in recent years (Gauja, 2017; Scarrow et al., 2017).

The obvious starting point in considering who should be enfranchised is the formal party membership (Table 1). In the New Zealand case, some consideration was also given to including registered party supporters (interviews 17, 23). Parties in many countries are struggling with declining membership and some suggest this stems from an unwillingness to formally join (Gauja, 2015; Scarrow, 2015). The Canadian Liberals, for example, allowed ‘supporters’ to vote in their 2013 leadership contest as have the Italian Partito Democratico and the UK Labour Party in their own contests. In the end, NZ Labour decided not to include supporters. The inclusion of supporters was not considered in Australia – the dominant rhetoric surrounding reform was centred on increasing the rewards of membership (Rudd, 2013b). Parties including their ‘members’ must also determine how to define membership and this is taken up in the next section.

OPEN IN VIEWER

Table 1.

Critical questions for leadership selection reforms.

Question	Range of options	NZ Labour decision	ALP decision
Who should vote in leadership elections?	• Membership • Supporters • Parliamentary party • Affiliated groups (including labour unions) • Local elected officials	Members, parliamentary party, labour unions	Members, parliamentary party
How should vote shares be allocated?	• Split equally • Weighted to increase influence of certain groups	40% members; 40% parliamentarians; 20% labour unions	50:50
How should membership be defined?	• Existing dues-paying members • Dues-paying members with required minimum tenure • New members joining during contest	Dues-paying members at time contest is called	Dues-paying members at time contest is called
What electoral system should be used?	• Preferential • Runoff • Majority requirement • Quota	Majority, preferential	Majority
How long should the campaign be?	• 14 days to 6 months or longer	21 days	32 days
Who has authority to remove the leader?	• Parliamentary party • Membership (full vote or conference delegates) • Mixed group including parliamentary party and membership	Parliamentary party	Parliamentary party

NZ Labour: New Zealand Labour; ALP: Australian Labor Party.

A subsequent question is whether the vote should be limited to members or include other groups. Although many parties have moved to selection through a simple vote of their rank and file (30% of the Pilet and Cross cases), others have chosen to adopt what is commonly called an ‘electoral college’, in which vote shares are distributed among party members, MPs, trade unionists and, in some cases, local elected officials. This method was championed by UK Labour when expanding their selectorate in 1981 (Stark, 1996). Although the percentage of the vote awarded to each group shifted over time, it was shared between members, parliamentarians and trade unions. The method was subsequently adopted by other parties including the Canadian New Democrats, as well as Fianna Fáil and Fine Gael in Ireland who split the vote three ways between parliamentarians, party members and local elected officials. The New Democrats split it two ways between trade unions and members. Both the New Democrats and UK Labour subsequently moved to simple membership votes.

Restricting the vote to members received little support in the reform process in both New Zealand and Australia. Instead, both parties seem to have been strongly influenced by UK Labour, which was still using an electoral college, and quickly settled upon sharing the vote between the parliamentary party and the membership. In the New Zealand case, several leading figures suggest there was strong sentiment that MPs should retain significant influence in the process as they know the requirements of a party leader well and have first-hand knowledge of the skills of their colleagues. It was also suggested that removing any privileged role for MPs would have caused a considerable backlash and placed the reform project in jeopardy (interviews 13, 15, 17, 18). In the Australian case, the process was driven by the parliamentary leadership and subject only to their approval, and this likely explains the ease with which this decision was made.

NZ Labour ultimately decided on a three-way sharing of the vote between MPs, party members and trade union affiliates. Interviewees suggest that some of the party’s supporters, who more closely aligned with its ‘social issue’ sectors, such as the environment, gay rights, gender and youth groups, were sceptical about including the trade unions (interviews 15, 16, 17, 23). At least one regional party organization submitted a remit to the conference calling for the vote to be restricted to party members and MPs. Party leaders suggest that it was important to include labour affiliates as the party is dependent upon their organizational support in election campaigns and could not afford to alienate them (interviews 9, 21, 22). Here too, reference was made to the UK Labour Party. In the end, the ‘compromise’, was to allocate unions a vote share significantly less than that reserved for party members and parliamentarians.

By contrast, unions were not given a vote share in the ALP. There was some criticism from the union lobby when the decision was announced, as the reforms missed a significant number of union members (around a million) who could have been enfranchised, and it was said that the party risked drifting too far to the left if controlled by the rank and file (Financial Review, 10 July 2013, p. 1). However, apart from a few days of media coverage, the unions’ concerns did not impact significantly on the trajectory of reform. Rudd’s position on the issue was that ‘what we are seeking to do here with this reform is say to all our friends and supporters in the trade union movement, we want you active in the branches of the Australian Labor Party’ (Rudd, 2013b).

The New Zealand model allocates 40% of the vote to both MPs and the rank and file, with the remaining 20% to the trade unions. Some in the parliamentary party argued for a majority share, whereas others countered that reform was only meaningful if party members had, at minimum, a vote allocation equal to MPs (interviews 13, 15). The experience in other parties varies significantly, from parliamentarians having 65% of the vote total in Ireland’s Fine Gael to them having no votes at all in those parties with pure membership votes.

Defining membership

Parties take very different approaches when they invite the ‘membership’ to participate. At one extreme, some require active membership of a year or more (the Irish Labour Party for example). Parties who make this decision wish to reward their committed members and encourage other supporters to maintain an ongoing activism. Limiting the franchise to long-term, active members also reduces the possibility of new entrants influencing the process who have little or no ongoing commitment to the party. At the other extreme, parties such as those in Canada have very open processes allowing virtually anyone to join after the leadership election is called. These contests often result in dramatic expansion of the membership (by as much as 500%, Cross, 2015). Supporters point to this growth as a positive way of rejuvenating the party, whereas opponents suggest it gives the choice over to party outsiders and could be used as a mechanism to ‘water down’ the voices of committed activists.

UK Labour offers an informative example here. In its early electoral college votes, the franchise was limited to those who were party members at the outset of the contest. The rationale was that voice in the leadership contest should be limited to those who were active in the party. Mobilization was not a motivation. This began to change with the contest that selected Ed Miliband in 2010. In this case, a two-month window was provided for new members to join and be eligible to vote. There was, however, very little indication that candidates used this opportunity to mobilize potential supporters. This changed dramatically with the selection of Jeremy Corbyn. In 2014, the party adopted a one-member-one-vote system, and redefined the selectorate to include members, affiliated union supporters and registered supporters. A group of over 552,000 supporters signed up to vote in the 2015 contest. These comprised 292,000 members (up from 190,000 in 2013), 148,000 affiliated supporters and 112,000 registered supporters. As in 2010, votes were cast over two months, a time period that caused considerable controversy and anxiety as allegations of infiltration were made to the effect that supporters were being actively recruited to support Corbyn. This generous window was removed at the next leadership contest in 2016.

NZ Labour chose something of a middle ground. Eligibility to vote was limited to those who joined the party no later than 24 hours after the announcement of the leadership election. This rule provides almost no opportunity for recruitment of new supporters. However, an exception was made for those who had belonged at some point during the past three years; they were allowed approximately two weeks to renew their membership. Some were uncomfortable with these rules and suggested that by being so restrictive the party forfeited the opportunity to capitalize on the enthusiasm the contest generated (interviews 15, 16, 22, 23). Even without offering voting rights, the party membership increased by nearly 20% during the 2013 contest. This may have been a result of requiring membership for attending the leadership candidate hustings. Several in the party leadership countered that with a rather small membership, the party risked having the process taken over by outside groups who might easily mobilize enough people to determine the outcome if given the opportunity (interviews 19, 22).

There was relatively little debate within the ALP as to who should be counted as a party member. This may have been due to the fact that only the basic principles of reform were announced by Rudd, and the details of the contest had to be worked out by the National Executive once the contest was imminent. On 16 September – four days before nominations for the contest closed – the National Executive decided that everyone who was a dues-paying member of the party as of 7 September would be able to vote. Previously, it was expected that only people who had been members for at least two years would be enfranchised (ABC News, 16 September 2013; Australian Labor Party, 2013). The reason for this extension was so volunteers who had joined to get involved in the election campaign could have their say (Jenny McAllister, ABC News, 16 September 2013). Although party officials and parliamentarians cited an increase in membership numbers (of around 4000) during the contest (Australian Associated Press, 2013), it was not specifically designed as a tool for recruitment.

The electoral system

Some parties in jurisdictions with geographically defined electoral systems weight votes for purposes of providing an equal share to each electorate. This was the case with regard to UK Labour when it first adopted reforms (Punnett, 1992; Stark, 1996) and remains the case for the Canadian Conservatives and Liberals. ³ A rationale for this approach is that it rewards candidates who have geographically dispersed support and mirrors the way general elections are won. There appears to have been no consideration given to weighting votes by electorate by either NZ Labour or the ALP. In the case of NZ Labour, this decision may have reflected the primacy of the national party vote in determining electoral success, rather than the outcomes in individual electorates. In the Australian case, the decision reflects both the speed and the process by which the reforms were implemented, which hindered the ability to debate detail, and the fact that agreement on the obvious way of weighting in the Australian context – on a state-by-state basis reflecting the federated nature of the party – would be hard to achieve.

In terms of the electoral ballot and count, both parties decided upon a majority requirement and a preferential ballot. The vast majority of parties – three-quarters in the Pilet and Cross study (2014: 230) – require leaders to win majority support regardless of who chooses. This includes all the principal Westminster parties (Cross and Blais, 2012a: 68). The rationale is that the leader has widespread support and is not the favourite of a single faction. Requiring majority support often also means that candidates have to seek second and third preference votes, requiring appeals to those who may not be their natural supporters and making significant intra-party squabbling less likely.

Two different possibilities exist for accomplishing this: runoffs between the top candidates and preferential voting. When MPs made the choice, both parties used runoffs. NZ Labour explicitly rejected this method and instead adopted preferential balloting using the alternative vote method. This decision was largely made for the practical reason of requiring only one round of voting with a larger, more dispersed electorate (interviews 17, 20).

This is a common change made by parties when expanding their selectorate. It was a reform adopted by UK Labour in 1993 and also in later leadership votes by the Canadian parties. The principal reason advanced for the change is administrative efficiency. When MPs make their selection they are typically gathered in one room for the vote and subsequent balloting can be held immediately, with no additional cost or delay. This is not the case when the selectorate is expanded to include party and affiliate members.

There are some potentially important consequences with regard to the choice of electoral system. Preferential voting makes it less likely for long-shot candidates to emerge victorious as there is no chance for a trailing candidate to build momentum as a compromise choice. There is also no opportunity for those supporting one of the leading candidates to change their votes between counts. The result is that winners in preferential balloting have almost always finished first or second in the first ballot. This is not the case with runoffs. With regard to both selection in the parliamentary party room and among broader selectorates, multi-ballot contests often result in third or fourth place candidates (or even candidates not in the first ballot) emerging as victors (See Cross and Blais, 2012a: 72–73).

One of the more contentious issues in the New Zealand case related to how the 20% of the vote assigned to trade union affiliates would be cast. Those opposing participation of the affiliates feared they would vote as a bloc, thereby exercising disproportionate influence in the contest. They argued that the affiliates should be required to conduct a ballot of their membership and cast their votes in a proportionate way reflecting members’ preferences (interviews 16, 19, 21, 22, 23). The affiliates countered that this would be too administratively difficult. In a compromise, an agreement was reached whereby each affiliate would have the option of balloting its members or conducting a vote among those attending a meeting of the most inclusive and participatory governing body of the affiliate. These issues were absent from the Australian debate as the detailed rules for the contest were only announced after it was clear there would be two candidates, ensuring a majority winner in the first round.

Length of the campaign

When party officials in both countries were interviewed before the adoption of the new processes, they pointed to the three-year electoral cycle as a reason why they could not adopt more inclusive leadership selection (interviews 1, 2, 3, 24, 26, 28, 29). Their argument was that these processes take months to complete and that parties facing elections every 30 months or so cannot afford the competitive disadvantage of being without a leader for any significant period. As ALP Senator Stephen Conroy argued, ‘We’ve got no leader, no frontbench, no shadow spokespersons who are able to lead the debate for us’ (Ricci, 2013). The advantage of selection solely by the parliamentary party is that it can be completed within a matter of days. In adopting new rules, NZ Labour was particularly attuned to this concern, including a provision that: ‘The system will be capable of delivering a result within 21 days of being triggered’ (NZ Labour, 2013 [revised 2014]: 3.5.1).

Contests in other parties using broader selectorates typically take considerably longer. This likely does reflect the 4–5 year electoral cycles in countries such as the UK and Canada but also the fact that parties in these countries use the first weeks and months of the contests to recruit new members. The 2010 UK Labour formal campaign period lasted almost 20 weeks, whereas the 2013 Canadian Liberal campaign lasted five months.

NZ Labour’s (2013) contest effectively began on 26 August. The winner was announced 21 days later. This was remarkably quick and efficient. The 2014 contest lasted somewhat longer with Cunliffe resigning on 27 September and the new leader named on 18 November. The latter contest began just days after a general election and there was less rush to expedite the process given that the next election was not on the horizon, whereas the 2013 contest was held less than a year before the next campaign would begin.

In proposing the ALP reforms, Rudd was adamant that the process should take no longer than 30 days (Rudd, 2013a). However, the 2013 ALP membership vote had to be extended for a further two days late in the campaign, due to delays in the receipt of ballot papers in some states and remote areas (Australian Broadcasting Corporation, 2013). This raises the important issue of the method of casting votes. With regard to both NZ Labour and the ALP, party members voted by post, but this may prove to be an outdated method as citizens undertake more activities online. ⁴ Both of these contests illustrate that more inclusive leadership selection processes need not drag on for long periods of time if a party does not view the contest as a way of growing its membership.

Confidence votes and trigger for a new contest

When parliamentarians select the leader, they almost always retain sole removal authority. This has traditionally been the case with the major New Zealand and Australian parties. This can become more complicated when the selectorate is expanded. Parties in different countries have taken different approaches to this question and it is not uncommon to find asymmetry in the selection and deselection processes. Pilet and Cross (2014: 230) find fewer than 1 in 10 parties grant members the authority to remove the leader either through a vote of conference delegates or a plebiscite of the entire membership.

The UK Conservative party caucus maintains the removal authority even though selection is through a membership runoff vote of the top two candidates. This can create unusual situations such as that of Iain Duncan Smith. Duncan Smith was never the favourite of the parliamentary caucus, winning only 54 of 166 parliamentary ballots. However, by winning the party’s 2001 membership vote he became leader. Only 13 months later he was removed from office by a 90:75 vote of the parliamentary party, actually winning more votes from parliamentarians than he had in being elected leader (Heppell, 2008). Other parties, such as those in Canada, have aligned the removal authority with the selectorate, thus placing it with the membership. UK Labour has a complicated procedure that requires action from both sections of the party. Experience in other countries suggests that when the removal process involves the party membership, it is more difficult to invoke and often involves long and divisive struggles typically played out in front of a public gallery. In contrast, parliamentary party removals are legendary for the speed with which they occur (Cross and Blais, 2011).

NZ Labour opted to leave removal authority with the MPs. Although there was some discussion of possibly providing a trigger for the membership to spill the leader, there was no widespread support for this position and it does not seem to have been seriously pursued. More controversial was the question of when and how MPs would review and potentially remove a leader. When the removal rules are changed during the tenure of an existing leader who wishes to remain in the position, it is not surprising that they are viewed through the prism of what they mean for the incumbent. The case of NZ Labour was no exception as many saw this question as a battle between the Shearer and Cunliffe camps (interview 15).

In its July 2012 Organizational Review Recommendations, the New Zealand Council endorsed proposed rules requiring an automatic vote of MPs within three months of each general election in which the incumbent had to receive simple majority support to maintain the position. Outside of this mandatory vote, the leader could only be removed by agreement of two-thirds of MPs. This was in sharp contrast to the status quo in which the incumbent could be removed at virtually anytime through a simple majority vote. The New Zealand Council defended this position, suggesting it would provide needed stability in leadership and Coatsworth referenced parties in other countries with similar rules.

There was an immediate backlash to this proposal by those who saw it as ‘incumbency protection’. ⁵ In the end, opponents of the proposed rule carried the day and the status quo was, essentially, maintained, allowing MPs to remove the leader at any time with a simple majority vote and adding an automatic post-election vote, in which the incumbent would require 60% support to maintain the position (NZ Labour, 2014: 378, 379). The super-majority requirement to retain the leadership is unparalleled in the Westminster parties.

Because the changes to ALP leadership selection were made in the context of leadership churn, the removal of the party leader was the most controversial aspect of the debate. As in the New Zealand case, no real consideration was given to allowing the extra-parliamentary party any role in the removal process. Previously, the party leader could be removed by a simple majority vote. Rudd’s original proposal for reform retained this power exclusively for the federal caucus. A party leader could only be removed and a new contest called in four circumstances: an automatic spill would occur following a general election if the party did not win government; if a leader resigned voluntarily; at the request of a leader; or if at least 75% of the parliamentary party signed a petition saying that the leader had brought the party into disrepute.

Immediate concerns were voiced that the 75% requirement was too rigid and in practice meant that a leader could only be removed in the most exceptional of circumstances. This would alter the balance of power between the leader and the caucus, reducing accountability between periodic leadership contests. At the special caucus meeting that approved the rule changes, Rudd’s proposal was amended so that only a petition of 75% of the caucus could remove a leader in government, but an opposition leader could be removed with a petition of 60%. However, even in the amended form of the rules, the high threshold for removing the leader was controversial. Speaking out against the changes, Julia Gillard argued that the ‘rules protect an unsupported, poorly performing incumbent rather than ensuring that the best person gets chosen and supported for the best reasons’ (Gillard, 2013).

Despite these criticisms, the amended proposal passed through the caucus relatively smoothly. This is surprising given the amount of power that the parliamentary party ceded, but is explained by the timing and context of the reform process. Considerations with regard to the electoral damage caused by leadership change in recent years were of greater concern to parliamentarians than the prospect of holding future leaders to account.