Prime minister Kevin Rudd’s proposal to prevent the ALP caucus from deposing a sitting prime minister has raised the hoary issue of the Westminster system.
The dominant view among pundits is that under the Westminster system, voters elect members of parliament - at least in the lower houses of the institutions using the system - and in practice the party with a majority of members on the floor of that lower house (usually known as the House of Representatives or Commons) choose the prime minister.
Yet there are key differences between the Westminster system as developed and practised in London, and the version that, while closely related, has evolved in Australia. The Australian Senate is elected by the people. The governor-general is, in theory, Australia’s head of state, rather than the monarch of the United Kingdom of Great Britain and Northern Ireland, as is the case in the state that gifted Australia its political architecture.
More important is the fact that Australia is a federation with a written constitution, while the UK has no written constitution and is a quasi-unitary state. The importance of the constitution is qualified by the fact that it prescribes very little about the way the country should be governed. It does not mention prime ministers or political parties, even though both were well-established features of the Westminster system when it was written.
So, the relationship between the prime minister, the majority party in the lower house (if there is one) and the parliament as a whole is one of convention, and convention can evolve over time. The way the Australian system has worked in practice - at least since the creation of the Liberal Party by Robert Menzies - enables voters to choose between candidates based on perceptions of the party leader and the party label. Well-liked local candidates may gain a handful of percentage points in the ballot, but this is of trivial importance overall. Since party allegiances are relatively stable, the decisive factor is normally the choice of leader.
The complicated part of the story is the way in which leaders are changed. Roughly speaking, the rules that have evolved are as follows: leaders defeated at an election normally resign, and can expect to be deposed if they do not; the opposition can change leaders whenever the majority of the parliamentary party judges it to be desirable; incumbent prime ministers can be deposed only in the face of imminent electoral disaster; and defeated leaders can return if they are willing to wait long enough for their successors to fail.
There have been only a handful of exceptions to the first rule in recent decades. Labor was more loyal to defeated leaders in the past, giving Ben Chifley, H.V. Evatt and Arthur Calwell three tries each. Whitlam was given a second go after 1975 but resigned immediately after his loss in 1977. Beazley had a near-win in 1998, and was also given a second go in 2001.
The second rule is uncontroversial, though of course the actual choices may be debated. Instances of the fourth include Howard in 1985, Peacock in 1989, Howard again in 1995, Beazley in 2005 and now Rudd.
The real difficulty, illustrated in the case of Rudd and deposed prime minister Julia Gillard, arises with the replacement of sitting prime ministers. On this issue, there is a clear conflict between the majority of voters - who assume that they are supposed to choose the prime minister - and the advocates of the Westminster system, who say that this right belongs to the parliamentary caucus of the party with a lower house majority. As noted, the Constitution is silent on this matter, so the question can only be resolved by looking at actual experience.
The two previous cases in the post-war era were those of John Gorton and Bob Hawke. Gorton only held office because the seemingly obvious choice, William McMahon, had been vetoed by Country Party leader John McEwen. As soon as McEwen retired, Gorton was dumped. This anomalous case tells us little, except perhaps that McEwen was a good judge of character.
Hawke’s defeat by Keating tends to support the caucus supremacy view. However, Hawke’s defeat took place in dire electoral circumstances, comparable to the recent switch from Gillard to Rudd. The deep recession, and Hawke’s inability to respond to the Fightback! package proposed by opposition leader John Hewson created a general view that Labor was doomed without a change.
Arguably, the procedures now proposed by Rudd might have been invoked in this case. At the state level, the case is similar. The dumping of premiers Morris Iemma and Nathan Rees by NSW Labor and Ted Baillieu by the Victorian Liberals took place when polls indicated electoral defeat looming, although the NSW moves proved unavailing and the case in Victoria has yet to be tested.
In all of this, then, the Rudd dumping in 2010 stands as an anomaly. Despite a couple of bad polls, Rudd was certainly not facing inevitable electoral disaster. As has been made abundantly clear since 2010, the main reasons for his removal were that a number of his colleagues - including some of the most senior - strongly disliked him, and that he had incurred the displeasure of factional heavyweights who commanded the majority of votes in caucus.
Although Labor scraped back into minority government in 2010, it became clear over time that these grounds for removing a sitting prime minister were not regarded as sufficient by the majority of voters. Despite some substantial achievements, Gillard was never regarded as legitimate by many voters, and every misstep she made was evaluated in this light. The interpretation of the Westminster system favoured by Canberra insiders has been weighed by the Australian public and found wanting.
The changes proposed by Rudd represent, in essence, a codification of the existing situation rather than a radical change. It was the coup against Rudd in 2010 that was a breach of the rules as understood by their ultimate arbiters: the Australian public.