Former prime minister Kevin Rudd.

Former prime minister Kevin Rudd.

SOMETIMES the choices politicians have to make are excruciatingly difficult. However in the case of the present leadership bid being waged by the pro-Rudd forces, the choice before members of the federal Labor caucus is really very simple.

In June 2010, the caucus, led by the Right faction's leaders, deposed Kevin Rudd, one of the more popular prime ministers in the postwar history of Australia, and installed Julia Gillard. Almost no one now thinks it was a clever idea. The reasons given for the coup were rather trivial: a mixture of complaints about Rudd's personality and management style and three months of stormy politics following the decisions to postpone the emissions trading scheme and to introduce a new mining tax.

No one disputes that for more than two years the Rudd government had maintained an always comfortable and usually handsome lead over the Coalition. However, the supposed weakness of Rudd's performance in opinion polls in those final three months was a crucial ingredient in the coup leaders' argument for his removal. At the time and since, this weakness has been greatly exaggerated. On the eve of the coup, according to one poll, Labor led the Coalition, 52 per cent to 48.

After two months as Prime Minister, Julia Gillard scraped through an election with the numbers to form a minority government. In its first year or so Gillard performed reasonably well in the opinion polls, although not nearly as well as had Rudd. In July 2011, however, support for her government collapsed suddenly and dramatically. During the past year federal Labor has averaged about 30 per cent of the primary vote in both Nielsen and Newspoll and less than 45 per cent of the two-party preferred vote. If an election had been held in any week of the past year Labor would have suffered not a decisive but a catastrophic defeat.

The unprecedented nature of the Gillard government's unpopularity is easily demonstrated. Newspoll has been measuring federal voting intentions each fortnight since 1985. In the quarter century to 2011 there was one instance where Labor's primary vote was lower than 33 per cent. In the past 12 months there have been 20. On 11 occasions, federal Labor's primary vote has fallen below 30 per cent, something that had never happened.

The unpopularity of the Gillard prime ministership can be demonstrated in a different way. On several occasions since the coup, Nielsen has assessed people's preference for either Gillard or Rudd as prime minister. In every poll, there has been an almost eerily similar result. Rudd is about twice as popular as Gillard.

Last year, Rudd began campaigning behind the scenes to regain the prime ministership. In February this year, Gillard and her supporters decided to kill off the challenge by forcing the matter to a ballot. Their method was to destroy Rudd's reputation by mounting a series of personal attacks, in essence concerning his character.

At this point Rudd made a tactical error. He could have denied that he was challenging Gillard and forced her to sack him as foreign minister, thus entrenching his popularity with the people as a two-time martyr. Instead, he chose to contest the leadership. Rudd lost by a substantial although not humiliating margin: 31 to 71. He promised he would not mount another challenge. He did not, however, promise that another challenge would not be mounted by others on his behalf.

Gillard's attempt to end the threat of Rudd once and for all, through forcing a leadership ballot, failed for two main reasons. Gillard and her supporters seem to have convinced themselves that Rudd's white-anting was a significant factor in her government's unpopularity. But the polls for Gillard did not improve after Rudd's vanquishing. Gillard also seems to have believed that the character attack on Rudd by several senior ministers would destroy his public reputation. The polls for Rudd were, however, entirely unaffected.

At present the numbers for Rudd in caucus are growing. The gap between 71 and 31 sounds formidable. In fact, it means that only 20 caucus members have to change their minds for Rudd to return as prime minister. As he promised, Rudd will not formally mount a challenge. Rather, he will wait until his supporters can assure him that if a new leadership ballot is called he will have the numbers. His supporters are confident they will eventually be able to find the numbers that are needed. They are more worried that he will pull out of the contest, if the process of garnering the necessary votes takes too long.

In essence, the argument for the return of Rudd, which I support, goes like this. It is as certain as anything in politics can ever be that if Gillard leads Labor to the next election, Labor will suffer perhaps the worst defeat in its history, worse even than 1931 or 1975.

If the defeat is sufficiently disastrous, the Coalition may gain control of the Senate. All the important legislation the Gillard government has passed or promised - the carbon tax, the mining tax, the early stages of a national disability scheme - will be repealed or fail to be followed through. At best, Labor will face the prospect of 10 years or so of opposition. At worst, some Rudd supporters fear that the fortunes of the Labor Party may never revive. Rudd's return is best seen, then, as a salvage operation. What he should do when returned to power is a question for another occasion.

No one should pretend that the transition from Gillard to Rudd will be straightforward. The fate of the present Prime Minister is unclear. Although Rudd will behave generously and will eat some humble pie, certain senior ministers will either decide to resign, or be forced to do so. If there are any resignations from the Parliament or defections among the independents, Rudd might lose the numbers needed for government following a byelection. If his government lasts until an election next year, last February's character attacks by his cabinet colleagues will provide wonderful ammunition for the Abbott Coalition.

Nonetheless, the political situation facing caucus members is straightforward. With Gillard as Prime Minister, Labor will almost certainly be crushed at the next election. With Rudd, the party has, at least, a reasonable chance of minimising the electoral wreckage and even a remote chance of victory.

Before Parliament resumes next month, members of the Labor caucus have in essence to choose between two loyalties: to the Prime Minister, or to their party and their country. The stakes are high, but the choice, in my opinion, ought to be simple.

Robert Manne is professor of politics and convener of the Ideas & Society Program at La Trobe University.