The near conclusion of the long running saga called the Coalition’s Demographic Train Wreck.

After spending the last year and a half ferreting out every bit of 2007 election polling I could get my hands on, I think we might have – finally – a good set of estimates for the two party preferred vote at the 2007 election by age cohort. Most pollsters have broad brush age cohorts that are over 15 years wide, but none alone have sample sizes large enough to justify breaking down the age results into smaller chunks. But now, we can look a little deeper.

Most of the data for this didn’t come from just the usual suspects (although the usual suspects were certainly helpful), but from a mix of major pollster data, unpublished polling and exit polling that was undertaken by a surprisingly large number of groups either very, very close to election day, or in the case of exit polling – after election day itself. So thank you to all who willingly (or often grudgingly!) contributed to this – you know who you are.

See how power works in this country.

News done fearlessly. Join us for just $99.


The most difficult part was sorting the 50-60’s from the 60’s and over, and those three cohorts contain the largest uncertainty. We can’t really give a margin of error for it because of the mathematics used to derive it ( in most cases I didn’t have access to respondent level data, but had age cohorts bundled into 5 and ten year groups)– but it’s fair to say that the over 60’s may be a point undercooked for Labor while the 50-60’s may be a point overcooked. But the results below are, by far, the highest probability result. Another reason I’m confident in the figures is that if you weight these results by the age breakdown of the electoral roll, the ALP two party preferred comes out at a national 52.8% – which is only 0.1% off the actual election result.

The most interesting thing about the result is the way the two party preferred numbers work in terms of the actual size of the voting populations in each age cohort. What really stands out is the way the Coalition avoided suffering a complete wipe out in 2007 singularly because of the level of support they enjoy in Generation Blue – that group of people born before 1947. The chart below has the ALP two party preferred vote by age cohort, as well as the proportion of the electoral roll that each of those cohorts make up.


You know you’re a political tragic if you looked at that chart and went “Faark me dead!”. The consequences of it are pretty profound. To tweak the gravitas of it even further – the average ALP two party preferred vote among the under 60’s was an astonishing 56%. Sit on that thought for a moment and ponder.

Generation Blue are an interesting lot – their relative love affair with the conservative side of politics isn’t merely something that has come with age, but appears to have been a long term pattern of electoral behaviour.

Using public Newspoll data going back to 1987, much of it unearthed from the dungeons of the Newspoll database by Ian Watson, if we look at the size of the Coalition primary vote by periods of birth, we can see that those born before 1947 have had a large partisan lean toward the Coalition:


Those 2007 figures for the cohorts born after 1947 are a few points skewed to Labor, as it looks like some of the data was apparently taken before The Narrowing happened (and remember, don’t let that blip in 1998 fool you, that was the One Nation effect where preferences in that age cohort flowed back very strongly to the Coalition). That said, using non-Newspoll figures, Gen Blue had a Coalition primary vote at the last election in the low to mid 50’s. It’s also worth noting that the old story about people getting more conservative in their vote as they get older doesn’t hold true. Rather than a gradual increase, there seems to be a small “one off” structural jump that occurs in a person’s late 20’s early 30’s and from there the vote merely wanders around.

As a specific cohort, the pre-1947 group have been between 6 and 16 points more favourable to the Coalition primary vote than the average of all other demographics that have come after them, and Generation Blue has done this for at least the last 8 elections.

In 2007, the youngest members of Generation Blue had just turned 60, and combined they made up over 26% of the electoral roll. We can actually see their composition by looking at a basic age distribution chart using the 2006 Census data:


Unfortunately for Generation Blue – and the Coalition, it must be said – these folks are now running into the pointy end of human mortality. To highlight how quickly the political demographics can change, we can chart Australian mortality rates using the latest ABS Life Tables:


On the left is the probability that a person of a given age will die in the next 12 months. As you can see, after 60 the mortality rates start increasing, after 75 they climb dramatically.

This brings about a quite rapid political realignment that we are just starting to see. In 2006, Generation Blue made up 24.5% of the Australian population aged 20 or over. In 2011 that proportion will reduce to 19.5% from mortality alone. By 2016 it’s down to 14% and by 2021 it will be sitting around 9.5% – and that’s using the assumption of zero immigration. So the real figures will actually manifest as a sharper decline because of the age distribution of Australia immigration that will further water down their overall population proportion.

But Possum, what does all this mean for the Coalition vote”, I hear you ask – and what a mighty fine question that is! 😀

To highlight the impact of Gen Blue mortality on the Coalition two party preferred, let’s take a squiz at a couple of hypothetical situations and model them out.

Firstly, let us assume that in 2006 there was an election (coinciding with the census data) and Gen Blue voted 57/43 like they did in the actual 2007 election. Let us also assume that everybody else in Australia outside of Gen Blue voted 50/50. Let us also say that for hypothetical elections in 2011, 2016 and 2021 – Gen Blue continues to vote 57/43 while the rest of the population votes 50/50. This what what the results look like:


As we can see, natural attrition from the mortality rate alone knocks around half a point of two party preferred vote off the Coalitions end result every 5 years. Yet, the rest of the country doesn’t actually vote 50/50. So let’s run two more models where firstly, the rest of the country votes 53/47 to the ALP (which is close to the 1998 and 2001 results), and another where they vote 56/44 to the ALP as they did in the 2007 election.


By 2011, if the country votes exactly the same way they did in 2007, the election result will be, at best, 53.5/46.5 to the ALP. “At best” for the Coalition because, remember, this ignores Australian migration where the vote of migrants heavily favours the ALP.

Taking migration into account, we should expect that in the 2010 election, the ALP should generically have around a 0.5% two party preferred head start on their 2007 election result.

By 2016 the ALP will have a generic head start of around 1.3% in their two party preferred – but taking into account migration projections, that will actually be closer to 2 percentage points in real terms.

This creates a great deal of political grief for the Coalition – for without reforming their political ideology to be more reflective of the views and values of people currently under 60 in Australia, their only real chance to stay in the game in the short term is to start stuffing around with the electoral roll, reducing the number of younger voters turning out to the ballot box – something we witnessed in the final term of the Howard government. Yet they are no longer in a position to do that, suffering as they are from the legislative impotency of the Opposition benches.

But that style of franchise bastardry does little more than play around the fringes, the reality of demographic change is close to swamping the efficacy of any further attempts in the future to manipulate the roll.

To prevent the Coalition from facing a generation or more in the political wilderness, not only do the turgid old dinosaurs like the Bronwyn Bishops, the Kevin Andrews and the Tony Abbotts of the world need to be removed from the public face of the Liberal Party, but the views they represent – views which have only minority support among those under 60 – need to be isolated from the Liberal Party mainstream platforms.

The problem with having a political party where the membership is mostly over 60, is that the political party itself starts to represent little more than the views of those over 60. That’s all good and well while there are plenty of them around – but when their numbers start to fade, and when the views they pursue are alien to the large majority of those outside of that cohort – the road to irrelevance is pretty much charted.

For decades the Liberal Party has ridden on the coat tails of the ballot box premium that Generation Blue has provided them, allowing them to effectively ignore the fact that on most issues they have often been on the wrong side of public opinion as seen through the eyes of those born after World War 2. Without pivoting towards some form of modernity, and doing it pretty quickly, they will be up shit creek without an electoral paddle in a sinking demographic canoe – to really strangle a metaphor to death.

The problem, of course, becomes one of trying to break the institutional grip that Labor has on younger demographics. They have to hope that Labor doesn’t posses a voter premium among Boomers and Gen X and Y that the Liberals did with Gen Blue – for if Labor does hold an unbreakable premium there of 6 to 8 two party preferred points, then the next Liberal Party Prime Minister is probably still in high school.

Unfortunately for the Liberal Party – Labor probably does.

See how power works in this country.

Independence, to us, means everyone’s right to tell the truth beyond just ourselves. If you value independent journalism now is the time to join us. Save $100 when you join us now.

Peter Fray
Peter Fray
SAVE 50%