Risky Mark Latham is just the ticket for the battling Opposition team

The most common adjective used to describe newly elected Labor leader Mark Latham over recent weeks has been “risky”. “Fasten your seatbelts – we’re in for a bumpy ride”, one Labor politician was heard to remark. But as any sports fan knows, when your team is behind and the clock is running out, it is time to play “catch-up football” – switching to a high risk, entertaining style of play.

Historically, opposition parties that are neck-and-neck in the polls tend to choose leaders who are electorally safer – such as Bob Hawke, Kim Beazley or John Howard. Oppositions that know they are facing a huge battle at the coming election are more likely to opt for riskier leaders – such as Gough Whitlam, John Hewson, or Mark Latham.

How likely is Latham to win the next election? John Howard’s response has been to play down Latham’s success in the polls as simply reflecting a normal “bounce”. To parse out honeymoon effects from real effects, we need a more reliable indicator than the polls – one that helps reveal Latham’s true potential in the 2004 election.

In an article published in the Australian Journal of Political Science last year, we argued that betting markets provide precisely such an indicator. In the 2001 federal election, the daily odds posted by Centrebet, Australia’s largest private bookmaker, predicted voting patterns far better than any pollster. By forcing punters to put their money where their mouth is, the betting market aggregates information across the nation’s political pundits, insiders and informed gamblers.

According to Centrebet’s posted odds just before Crean stepped down, Labor’s chances had drifted down to 28 per cent. This figure remained unchanged when Crean first announced his resignation, indicating that the ABC slogan (“Anybody But Crean”) was not supported by the betting market.

Since Latham’s elevation to the leadership, there has been over $15,000 bet on Labor, with the latest odds of a Latham Prime Ministership improving to 33 per cent (about the same as John Howard’s odds six months prior to the 2001 election). Centrebet’s Gerard Daffy describes Latham as “fairly impressive to date”.

We can also use these betting odds to assess what would have happened under Beazley. If the pre-ballot probability of a Labor win (28 per cent) reflected an average of the election-winning abilities of candidates Beazley and Latham, we can infer from the fact that with Latham as leader Labor is a 33 per cent chance, Beazley would have been about a 23 per cent shot. According to the punters, the Labor caucus backed the more likely winner.

As for the current opinion polls, our research suggests that long-range polls tend to be wildly inaccurate, and are generally best ignored. While these polls have historically predicted vote shares with an average error of around 5 percentage points, simply guessing that previous election’s results will recur is about as accurate, and guessing that the election will be close actually yields more accurate predictions. So if you want to predict the outcome of the next federal election, remember, the current favourite in the election forecasting stakes isn’t the polls – it’s the bookies.