Labor is set to have itself a nervy little Christmas. It's not too late to make 2021 sing

The most recent Newspoll showed the Coalition government ahead of the Labor opposition by 51-49% on a two-party-preferred basis. It's a relatively small margin, but Labor MPs have indelibly etched on their minds that the final Newspoll before the 2019 election had Labor ahead on 51.5%, yet it won only 48.5% of the vote and suffered a shock loss on election day itself. This was within Newspoll's margin of error, but that was little comfort for those banking on a win because of the headline number.

Newspoll's design was tweaked in the wake of the 2019 election but two-party-preferred estimates remain vulnerable to discretionary polling design decisions and pollster "groupthink" - that is, the herd behaviour that leads to models generating results clustered around pollsters' intuitive assessment of how the major parties are travelling.

If the 51.5% forecast for Labor at the 2019 election turned out to be 48.5% on election day, how much worse than Newspoll's 49% last week might Labor's vote be if an election was held this weekend?

In How To Win an Election, I argue one of the big lessons from the 2019 election was to focus on that part of polls least subject to discretionary design decisions and pollster groupthink: the primary vote.

Since Federation, Labor has only once formed a majority government with less than 40% of the primary vote. That was in 1990, when the Hawke government won with 39.4%. In 2010, Julia Gillard patched together a minority government with just 38%. The latest Newspoll has Labor's primary vote on 36%.

Hence the very nervy Christmas Labor's marginal seat holders face this year. Right now, they're giving speeches about JobSeeker. Privately, they're wondering whether next year they'll be on it.

Labor seems stuck. Opposition Leader Anthony Albanese is well liked and enjoys widespread loyalty. The pandemic has made Albanese's job extra difficult this year.

However, the government has presented a rich target on a range of issues from robodebt to the repatriation of stranded Australians overseas, without the opposition landing a convincing blow. The feeling has taken hold in Canberra, and generally, that Labor is drifting inexorably to another loss.

Albanese's stolid leadership is the central concern. Prime Minister Scott Morrison is easily besting him with the flimsiest of performances, and despite a pressing array of problems besetting the government.

Morrison's behaviour suggests the LNP considers Albanese a political asset worth preserving. The government relentlessly attacked Bill Shorten when he was opposition leader, yet treats Albanese as Labor did Malcolm Turnbull in the previous parliament, and for the same reason: because they see him as eminently beatable compared to the alternatives.

The conversation being had around many end-of-year work functions goes like this. "Do you think Albo can win?" "No." "Who do you think would win?" The names that come up are Tanya Plibersek, Jim Chalmers and, occasionally, Chris Bowen and Richard Marles.

When you ask who would be the best deputy to their first choice, it's Plibersek or Chalmers. Overwhelmingly, Plibersek-Chalmers or Chalmers-Plibersek is seen as the election-winning team.

Wistful musings then follow along the lines of "what happened to Albo?" One puzzled person commented to me that "he went from a 'people eating out of the palm of his hand' kind of guy you'd want at your barbecue to boring dude you avoid at office parties quick smart". When your own supporters say things like this, it's hard to see how you could win swinging voters over no matter how sharp your political cunning, the thing Albanese considers his competitive advantage.

These are the kinds of conversations Labor's marginal seat holders are going to have at barbecues all summer as they contemplate a very uncertain 2021. Do they continue loyally marching into the valley of political death behind a leader no-one thinks can win, or do they save their political hides, and the interests of Labor voters, by installing someone who can?

The "how" is easier than many realise. By a simple majority vote, caucus can amend in any way it likes the rule change Kevin Rudd instigated to shore up his position as party leader in 2013. That amendment can instigate a fresh leadership contest, either through a straight caucus vote as occurred for a century before the Rudd rule change, or with rank-and-file party member involvement in the way that has occurred since 2013. Albanese could emerge victorious and re-energised, or Labor could have a new leader who can do better than the uninspiring effort it is weighed down by now.

After all, the Liberal Party does it all the time - and keeps winning elections. In its previous term in office, Liberal MPs replaced Turnbull with Morrison and won the 2019 election. The term before that they replaced Tony Abbott with Turnbull and won the 2016 election. The Liberals' all-time record was replacing John Hewson with Alexander Downer, and then Downer with John Howard, in one term of opposition, setting up Howard's 1996 election win.

Having the fittest leader is one of the keys to how to win an election. Labor still has time, if it gets a move on, to ensure it does.