Winning five elections may be an ask too far

The conventional wisdom among many journalists and opinion writers is that mid-term polls don’t matter much in assessing election outcomes (For example, see Dennis Shanahan in The Australian, October 6, 2006) Voters, they assume, behave differently anytime short of a few weeks before an election - mid-term they are not switched on to politics, governments can pork-barrel in the campaign, come the election voters prefer the incumbent in good economic times, and so on.

This wisdom has been very much to the fore in Australia over the past decade, particularly in The Australian newspaper and other Murdoch outlets. I agree that polls have to be read with caution and that what voters tell opinion pollsters can be different to how they behave in the ballot box.

But many commentators have elevated to almost mythical status John Howard’s campaigning ability rationalising that, in spite of mid-term polls, the punters are still more likely to favour him on election day (For example, Paul Kelly in The Australian, October 18, 2006). They certainly had the bookmakers in as, despite the Coalition being behind Labor during most of 2006 on a two-party preferred basis, Howard’s team was consistently long odds on to be returned. (But maybe the bookies are finally getting the message that opinion polls do actually count for something: at the time of writing Centrebet at last has the two parties neck and neck.)

I always treat individual opinion polls with caution and prefer to look for trends over time. In the chart below (using data from Newspoll) I compare the Coalition’s two-party preferred (2pp) performance for the (almost) two years leading up to the October 2004 election, with the same timeframe, to-date, for the current cycle.

I think the comparison valid as most predict an election in October 2007, or a month either side. So it is appropriate to see where we are now at the same stage as 2004. The purpose is to try and see how the “Howard always overcomes bad polls” thesis stacks up.

I could have chosen comparisons with other cycles, but settled on the 2004 election lead-up because I believe the most valid comparisons in politics are recent comparisons - too many social and economic variables change the longer the time frame. For example, the 2001 election was uniquely influenced by the Tampa incident and New York attacks, and in 1998 Howard lost on the 2pp and only scraped in (although Howard winning these two elections has gone along way to building his mythology).

The main point that the chart illustrates is that, although the Coalition holds a similar position at around 45 per cent 2pp in both cycles at about now (February-March), in the 12 months prior it travelled much better in 2003 than 2006. During 2003 there was scarcely a point where the Coalition was below 50 2pp percentage points, while in 2006 there is scarcely a point where it was above.

“So what!” might the orthodox thinkers respond. Howard recovered from this position last time, and he can do it again. Maybe, but I think there are good reasons to think there is more to it than that.

First, the chart indicates that John Howard spent much more political capital during 2006 than he did in 2003. Sure, he may have been at a low point in March 2004, but he had the glory of the previous 12 months to fall back on. Voters were used to him being on top and used to Labor being, frankly, hopeless. They dumped Howard briefly for the exotic Mark Latham, but quickly came to their senses during the campaign. Howard, my argument goes, had good form “in the bank” and was able to draw on it when needed.

Contrast that to the present situation. A sin of many writers has been to discount Labor’s dominant poll position, under Kim Beazley, over Howard during most of 2006. While these commentators were blithely ignoring what their own eyes should have seen, the voters were quietly spending John Howard’s political capital.

That is one factor, I believe, that has helped Kevin Rudd sustain his dominating poll position: Beazley shifted the voters’ mind set over to Labor during 2006, leaving Rudd the easier task to build on the status quo, whereas Latham had to change it. It is in the polls to read, but few see it.

A second reason why 2007 is not 2004 revisited, is that Kevin Rudd is not Mark Latham. Latham just scared the pants off voters. I have written previously in On Line Opinion on what I believe opposition leaders need to do to have their best chance of getting elected (using, in that instance, a West Australian election as a case study), and everything I see in Mark Latham and Kevin Rudd confirms my views.

Voters judge opposition leaders much more harshly and with much less tolerance than the incumbent leader, and so far Rudd has managed to operate largely within this narrow band.

Opposition leaders have to be seen to be safe, capable and with perhaps a few good, but not scary, ideas. They never (or rarely; maybe and only maybe a case for Whitlam) “sweep” or “muscle” their way into office. Boring I know, but boring has worked for Howard up to now and it is working just fine for Rudd.

My final point is simply that winning five elections, if history is any guide, is nearly always an ask too far. Put simply, the “cost of government” just wears governments down.

Yes, I know this is the era of the incumbent, due largely to risk-averse voters not wanting to upset the economic apple-cart, but in the end voters will weary of a government. Are they simply sick of the sight of John Howard? Maybe that is the message in the polls over the last 12 months? Relaxed and comfortable can look like yesterday’s man very quickly. Nothing personal, John, perhaps is their message, but a fresh face on the box would be nice.

My intention in this article is not to predict who will win the 2007 election but rather to challenge the orthodox thinking that we are in some sort of ground-hog cycle where at every election a wily John Howard pulls a victory out of the hat. Opinion polls have their uses and their failings, but in this instance I think they help tell a story: this election is not a foregone conclusion and has not been for some considerable time.