Category Archives: By-Elections

  • -
Dunkley Election - VIC Demographic Analysis, Profiling & Mapping

Dunkley By-Election

Category:By-Elections,Dunkley By-Election

I’ve written a profile of the by-election on Saturday in Dunkley for the Australian Financial Review, putting it in the context of our 2022 National Election Personal Vote Profiles and our profiles in Aston and Fadden.   

See AFR story here: 🔗 https://www.afr.com/politics/federal/dunkley-exposes-further-shift-away-from-major-parties-20240303-p5f9dw

The three individual seat profiles use ADS modelling of personal votes to  estimate the impact of the retirement or death of an MP with a large personal vote on the subsequent by-election result.

The personal vote for sitting members tends to be higher for provincial city or rural seats, as they are demographically more stable and voters are more likely to live and work in the same seat. These voters are easier for sitting members to locate at work or home and they tend to live in the same seat for a few elections and so are more likely to have a personal knowledge of their MP either socially through family sporting events, or officially, if they seek assistance with a federal Government matter, to do with transfer payments for example. The personal vote therefore builds up over time for these MPs.

To illustrate this point ADS modelling of the 2022 General Election showed a personal vote of more than ten percent for three female ALP MPs in Richmond, (Justine Elliot), Eden-Monaro (Kristy McBain) and Corangamite (Libby Coker) and without this personal vote, Labor would not have won their seats in 2022. So the personal vote is crucial when it comes to winning marginal seats.

Because the average personal vote of sitting Government members is typically about three to five percent, this figure is often confused by commentators with an average by-election swing of three to five percent against the Government if the Government MP leaves office unexpectedly in a by-election.

Unfortunately this simplistic use of swings comes unstuck when an Opposition MP with a big personal vote, such as we saw in Aston, retires and the resultant re-allocation of his personal vote back to the Labor tally produces a swing towards the Government. Oops.

Knowing this personal figure for the seat as a whole, we then profile the range of by-election swings for individual seats across booth catchments and profile these against key demographics across the same booth catchments.

This is a tricky exercise, especially with increasing numbers of pre-poll votes in a limited number of booths and a small number of matched pairs of booths. So caution is exercised.

But when we are looking at a strong inferential relationship significant to more than 99 percent confidence levels between votes and big demographics like Female Professionals and Male Construction workers in Dunkley, it is possible to make some comments, especially when it is consistent with longer term trends at general elections and demographic break downs of regular aggregated polling data.

Our Senior Demographic Mapper Dr Jeanine McMullan used our polling booth catchments to show some of the spatial links between Dunkley vote swings and Dunkley demographics at the by-election. Using the various layers, the reader can form their own impressions from the relationship between the swing to Labor and Professional Women and between the swing to the Liberals and Men working in Construction.

The link for the map is here:

🔗https://egs-au.maps.arcgis.com/apps/instant/exhibit/index.html?appid=8e5d29bb7cb54b8a8c217d25cebf459d

 


  • -
Eden Monaro By Election snapshot of results

National implications of Eden-Monaro by-election

Category:By-Elections Tags : 

The Eden-Monaro by-election was held on July 4 and won by Labor’s Kristy McBain, despite a small 2PP swing to the Liberal candidate Fiona Kotvojs.

Notwithstanding the small net swing, the range of swings for and against the ALP were 25 percent, which indicated that there was a considerable range of political views about each candidate. Modelling of the booth returns showed strong demographic drivers underlying the swings.

As Eden-Monaro is an excellent representative sample of Australia, demographically, spatially and politically, we profiled the 2PP results by pre-poll and election day booth catchments and projected these onto all Australian federal seats.

We stress that the by-election results are only a snapshot of how individual candidates performed at a certain time and in a specific set of circumstances and these circumstances will have changed significantly between July 4 and the next election.

Click to view demographics for Eden-Monaro by 4th July,2020

Eden Monaro By Election snapshot of results

The booths swinging to Labor tended to contain higher percentages of lower-income families and retirees, often employed part time in tourism and hospitality jobs which had been heavily impacted by the Covid-related lockdowns impacting this sector.

Those booths swinging to the Coalition tended to contain higher income families in secure white-collar jobs, least impacted by Covid jobs lockdowns.

For seat by seat projections of the Eden-Monaro swings onto all current federal seats, see the attached map.

 

Comments from John Black, founder of ADS and Education Geographics and map from Dr Jeanine McMullan, CEO of Health Geographics.

 


  • -
The Eden-Monaro by-election will be held on July 4.

Eden-Monaro by-election 2020

Category:By-Elections Tags : 

By JOHN BLACK, Founder of ADS

The Eden-Monaro by-election will be held on July 4. With a week to go, Labor’s Kristy McBain is the bookies’ favourite on $1.70 to win, with published robo-polls giving her 53 percent of the two-party preferred (2PP) vote.

Despite trailing in the robo-polls and with Sportsbet paying $2.25, I think the Liberal candidate Fiona Kotvojs is pretty good value for money if you like a flutter. There are a few reasons for this, but first some background.

Eden-Monaro is pretty close to being Australia’s version of Magic Town, named after the 1947 movie featuring a fictitious version of one of the original United States polling companies. The pollster – played by James Stewart – is obsessed with finding a small town which is a perfect demographic and political sample of the USA. This meant it could be milked cheaply for perfect poll results on everything from toothpaste preferences to voting intention.

With similar demographics to Australia, Eden-Monaro tends to vote like Australia and return the same result: and this means the party winning Eden-Monaro tends to also win a majority of seats and form the Government.

Mike Kelly first won the seat for Labor when Labor’s Kevin Rudd won the 2007 election. He won it again in 2010 when Labor also held Government, and he lost it in 2013, when Labor lost Government. So far, so good for Eden-Monaro’s reputation as Australia’s political Magic Town.

But then the popular Dr Kelly stood again for Eden-Monaro in 2016 and won, despite Labor losing the national election narrowly to the Coalition. He eroded the Eden-Monaro’s Magic Town reputation further in 2019, by holding the seat on 50.9 percent 2PP vote, despite Labor’s failure to regain Government.

When we look into the demographics of Eden-Monaro, we see it’s very similar to Australia, in terms of its socio-economic indicators which cover job status, education status and income. It also contains some urban overspill from Canberra, some farming and fishing communities and a slice of tree-changing and sea changing retirement communities.

Click to view demographics for Eden-Monaro

The Eden-Monaro by-election will be held on July 4.

The differences between Eden-Monaro and Australia seem minor but explain why the seat went against the trends in 2016 and 2019, sticking to Mike Kelly as a well-regarded political identity.

Compared to Australia the seat, on average contains more farmers and farm workers and a disproportionate number of retirees. Our demographic profiling of elections back to 1966, shows farm workers and retirees have a weak class vote and are more inclined to cast a large personal vote for any effective MP seen to be doing a good job.

We did several national demographic models of the 2019 election and on the one we chose, Mike Kelly polled 6.6 percent more than the model predicted.

With a standard error of estimate of 4.5 percent, this was a strong performance and history tells us that when a popular sitting member like this retires, their personal vote tends to be redistributed back to the opposing party, especially when that party is nominating the same candidate.

In the case of Eden-Monaro, that party is the Liberal Party and their candidate is again, Fiona Kotvojs. It’s important to bear in mind we model 2PP votes, so one candidate’s overperformance, is also their opponent’s underperformance.

So, it’s statistically safe to say Kotvojs would have won Eden-Monaro in 2019 with up to 55 percent of the 2PP vote against a less effective sitting MP than Mike Kelly. Hence my tip earlier on the value for money bet.

But we’re now in 2020 not 2019. Relevant factors to consider since 2019 include.

Favouring Labor:

Anthony Albanese is now the Labor Leader, not Bill Shorten. Both men were on similar levels of satisfaction – 41 percent – at similar stages of the election process, but Shorten was on a dissatisfaction score of 49 percent before the last election, compared to Albanese’s 38 percent.

Scott Morrison performed below par during and after the bushfires which devastated the east coast of the electorate over Christmas and this region was just getting back on its feet when it got whacked again with job losses from the Covid lockdown.

Labor now appears to be running an effective postal vote campaign in 2020 compared to 2019, when it ran no campaign at all. This is a sleeper for Labor.

The National Party and the Shooters are reportedly up to some preference allocation tricks in Eden-Monaro to disadvantage the Liberals. Even if true, I suspect these factors can be disregarded. Voters have more important things to worry about at the moment.

Favouring the Coalition:

Scott Morrison has performed above par during and after the Covid-Pandemic and lifted his personal pre-2019 election satisfaction score from 46 percent to 66 percent and his Better PM rating from 47 percent against Bill Shorten to 56 percent against Anthony Albanese. Tick one for the Liberals. But this is a pretty big tick, especially in a crisis such as we’re now facing, when voters tend to rally behind the Government of the day.

The big group relatively unimpacted by the Covid jobs Lockdown is of course public servants, and Eden-Monaro contains a lot of these, due to some fast-growing, high-SES, urban overspill south of Queanbeyan and this included some swings to the Coalition in 2019. I suspect this is a sleeper for the Coalition.

This leaves us to consider the Robo-polls now putting Labor’s Kristy McBain on 53 percent of the 2PP vote in Eden-Monaro. If you were one of the candidates, you’d rather be McBain on 53 percent, than Kotvojs on 47 percent.

But other pre-election Robo-polls overestimated Labor’s 2019 vote by three percent at the last election by oversampling IT-savvy, highly-paid professionals and younger highly-mobile, agnostic, Tertiary students, with strong dependence on social media, who love to fill in online media surveys and polls. These were the groups swinging to Labor in 2019. These are the groups now most impacted by the Covid Lockdowns.

The robo-polls in 2019 also under-sampled working families with two reasonably secure jobs, a mortgage to pay off and dependent school aged kids to look after, and they also under-sampled church goers, especially those on the fringes of our major cities. These are the groups which swung to Scott Morrison and the Coalition and they are the group least impacted by the current Covid Lockdowns.

The big unknown here are how the public votes during the Covid pandemic and associated jobs lockdown and none of us will know this for sure until we get the results on Saturday night.

But on the evidence, Fiona Kotvojs is good value at $2.25 for a win. Very good value.

MAP NOTES:

We can tell from the online interactive map that the areas hit hardest in Eden-Monaro by the jobs’ lockdown in March/April, were those showing the strongest signs of recovery in May. You can track the impact of the changes by clicking on the stages from 1 to 5.

We also include map layers showing swings and votes at the 2019 election. The votes provided are based on modelling of all the votes cast in Eden-Monaro, not just the booth votes which are becoming increasingly irrelevant, due to major increases in pre-poll voting.

 

Comments from John Black, founder of ADS & Education Geographics and map from Dr Jeanine McMullan, CEO of Health Geographics.


  • -
Longman By-Election - 2018 Qld

LONGMAN BY-ELECTION 2018

Category:By-Elections,National 2019 Tags : 

The ADS team has prepared an exhaustive profile of by-election swings in Bennelong, New England, Longman and Mayo.

A full report of the swings and their implications for the 2019 election appears in the Weekend Australian.

Below is a link to the maps of the swings and the associated demographic drivers for Longman.

Click away and compare the swings to and from the major parties with key demographics in our statistical profile.

Click on the Map to view

Longman By-Election - 2018 Qld


  • -
Malcolm Turnbull campaigns in Caboolture, in the Queensland seat of Longman

DIVINING VOTES ALL ABOUT SWINGS AND ROUNDABOUTS AT BY-ELECTIONS

Category:By-Elections,National 2019 Tags : 

Provided your humble correspondent is not eaten by a grizzly bear next week while trout-hunting in the wilds of northwest Canada, he will be back in time to help our team sift through the demographics of Super Saturday.

We will be looking for demographically driven swings across the booths in Longman and Braddon that are consistent with the swings we saw in New Eng­land and Bennelong; and tangible links between these patterns of swing and the policy offerings of the government and the opposition, especially as they relate to the hip-pocket nerve.

I’m looking particularly at swings by voters concerned with imputation tax increases for retirees and ­income tax increases for aspirational voters.

If we can see patterns among bigger demographic groups living in marginal seats then we can draw some inferences as to their impact at the next federal election. The other seats will provide a bit of a sideshow to the main event in Longman and Braddon.

South Australia’s Mayo, for ­example, is a seat a popular, local Liberal should have won back easily from former Nick Xenophon protege Rebekha Sharkie, who won Mayo from the less-than-popular local Liberal MP Jamie Briggs. However, polls show about 60 per cent of Mayo’s new generation of prosperous and professional commuters support Sharkie over the Downer dynasty’s Georgina Downer. After generations of political mulishness that has splintered the ­Coalition vote in South Australia — dating back to the original Liberal Movement — the Adelaide political establishment, like the old French aristocracy, learns nothing and forgets nothing.

The two West Australian by-elections of Fremantle and Perth will be interesting as a guide to how many intending Liberal voters, when denied a candidate, will vote for Labor over the Greens. Way too many to cancel the loss of Mayo, would be my guess.

Let’s look at what we know.

Last year’s by-elections in New England and Bennelong showed an average swing to the government of 1 per cent and a range of swings across the booths of about 12 per cent. The biggest swings against the government were in urban Bennelong booths dominated by progressive Left Sydney voters who hated having to vote yes in the same-sex marriage plebiscite and by conservatives who hated losing.

However, urban middle-class mainstream voters could not see what the fuss was about and quietly saved Liberal John Alexander.

In New England, Nationals flag-bearer Barnaby Joyce had his vote boosted by the big group of Howard battlers who had drifted back to Labor since 2007. We’re talking here about welfare recipients, tradies and hospitality workers living in rented accom­­mo­­­­dation in country towns where they can find affordable housing.

Battlers are also pretty thick on the ground in Longman in Queensland, a state where, in 2004, about one in seven electors voted for Labor premier Peter Beattie at the state election and then for Coalition prime minister John Howard eight months later at the federal election. This splitting of votes at state and federal elections is a characteristic of the Howard battler, a demographic that can be sentimentally supportive of favourite leaders but ruthless towards parties they regard as taking them for granted — especially with their Senate votes.

The trick for politicians is to match the sentimental rhetoric with what these voters see as their economic self-interest.

In Longman, they voted for Beattie because they saw themselves as Labor supporters and they voted for Howard because he stopped the boats, looked after their pensions and made the economy run on time. They had no problem holding what many commentators would regard as contradictory positions.

Plenty of Howard battlers are found in Braddon, where one in seven locals split their primary vote in the last state and federal elections. In mid-2016 the federal primary Liberal vote was 41.5 per cent for MP Brett Whiteley, but the primary vote for state Liberal candidates in March this year was 56.1 per cent.

In Longman and Braddon, polls are showing an average swing towards the Coalition of about 2 per cent, meaning both seats could go either way next Saturday. This is broadly consistent with the 1 per cent average swing to the government in New Eng­land and Bennelong, perhaps even a slight improvement for the Coalition.

Given some of the economic difficulties facing the federal government, compounded by its gaffes, even a small swing to the Coalition on its 2016 figures in these two seats would be exceptional, especially considering that Labor’s new MPs in Long­man and Braddon would have seen their vote rise by a couple of per cent since 2016 because of the personal vote benefits of ­incumbency.

Any opposition should comfortably win by-elections in seats it already holds. It should come close to winning more marginal seats like Bennelong.

Hopefully the demographic range of swings across the booths in Braddon and Longman will shed some light on why the opposition is underperforming and tell us what this could mean at the federal election due before the middle of May next year.

I’ll be paying particular attention to the range of swings across booths dominated by the different income groups to see which tax policies look like winning the most votes — with the opposition favouring those earning below $90,000 a year and the Coalition favouring those earning above that amount, particularly up to $200,000.

Like my Canadian trout taking a breather in a deep pool on their upstream spawning run, taxpayers tend to concentrate in income tax ranges just below a big jump in their marginal tax rate, so we’ll check the reaction from working voters who aspire to earn more.

We will see if we can discern any impact from the reductions in dividend imputation for retirees, although this one could be messy in practice.

With this sort of research, we go where the evidence leads us and we could see, for example, some impact from the campaign run by Catholic education against some of the federal funding ­reforms, which would be easy enough to measure given the ­detail in our education database.

We may see some increased support for the government’s quiet cuts in immigration, which would show as an increased vote for the Coalition among the huge mainstream groups of English-speaking and Australian-born. It’s pretty hard to lose an Australian election when you’re getting a swing towards you from Australian-born voters.

My working hypothesis is that Labor is making gains among younger, professional voters, ­especially those benefiting from the opposition’s big spending promises on new jobs in health and education.

We find these voters in the high-priced houses of the inner cities, the city seaside suburbs or bigger blocks with a view in the outer suburbs. But the Liberals should be able to withstand this sort of movement in what are typically their more comfortable urban seats, unless they repeat the mistakes of Mayo.

Labor is also making gains in Sydney and Melbourne among some conservative, non-English speaking migrant groups living in safe Labor seats, who respond well to big-spending promises as they are direct beneficiaries. However, these gains in votes by Labor often don’t bring commensurate gains in seats.

The Coalition seems to be still going well among middle-class, Australian-born families in mainstream urban areas and among Howard battlers in the middle to outer suburbs and in some pro­vincial city-rural seats. This is not the picture of a comfortable Labor majority, either in votes or in seats.

Roll on, Super Saturday, and I’ll report back if the bears don’t get me first.


  • -
New England and Bennelong By Elections in NSW

NEW ENGLAND AND BENNELONG BY-ELECTIONS

Category:By-Elections,National 2019,NSW Tags : 

See for yourself in our Esri online map how local demographics fed into the results in New England and Bennelong by-elections and the Same Sex Marriage plebiscite.

Click on the link to the map below. It will open with a default map of the Bennelong estimated 2PP swing to the Liberals, with the dark blue streets swinging slightly to the Liberals and the lighter areas swinging strongly to Labor. Now click on the Layer icon  at top right of your screen to open the Layer list. You can see the legend in the map by clicking on the small arrow to the left of the layer called Bennelong Liberal-Swing . You can click on the layer Bennelong Polling Booths  to show each booth and click on the booth icon itself to see the results.

Save this map of the swing by taking a screen snip or leave it open in another screen and then click off the swing and open the other Bennelong layers. You can open more than one at a time and see the impact of the various demographics on the swing. So you can open Layers for creative arts, Green votes and Agnostics, to see the combined impact of all three. This isn’t rigorous statistical modelling, but it gives you an idea of how it works.

The strongest predictor of the swing was not the Bennelong Chinese born males (or females), but Bennelong Males with No Religion. These Agnostics at the National level were the strongest supporters of the Yes vote in the Same Sex Marriage Plebiscite and when we fed this variable into the modelling, ethnicity did not contribute any additional explaining power. Chinese born persons in Bennelong tend to be Agnostics, but it was their lack of religion, rather than their ethnicity which tended to accompany the biggest political swings to Labor in Bennelong.  This shows us we should look past obvious ethnic stereotypes if we’re trying to explain voting behaviour.

When you have finished with Bennelong, you can click on the Bookmark icon  at top left and select New England which opens a default map showing the estimated 2PP swing to the Nationals, with the darker green areas registering the largest pro-National swing. Once again, save a copy of this map and compare it to those larger local demographics most strongly linked to the by-election swing to the Nationals: 2016 ALP voters, Tradies, or those demographics dominating the booths with smallest by-election swings to the Nationals: in this case men working in the Agricultural industry.  We reversed the direction of the legend in this last one so the areas with the darker green colouring contain the fewest farmers and farm workers and the biggest pro-National swings. This confirms that the smallest swings to a party often occur in the areas of strongest support for the same party.

Click on the map or link below to view.

New England and Bennelong By-Elections - See for yourself in our Esri online map how local demographics fed into the results in New England and Bennelong by-elections and the Same Sex Marriage plebiscite. See for yourself in our Esri online map how the local demographics fed into the results in New England and Bennelong.

https://egs-au.maps.arcgis.com/apps/webappviewer/index.html?id=cd901e880d5e44aabcae6a86cd8b0edd

 

 


  • 0

2008 GIPPSLAND BY BY-ELECTION

Category:By-Elections Tags : 

2008 Gippsland by By-Election

The first thing that stands out with the booth profile of the Gippsland by-election results is the strong link between Labor’s 2007 vote and the 2008 swing to the National Party candidate Darren Chester.

This was an election at which Gippsland voters, who had voted Kevin in 07, decided Kevin wasn’t so great in 2008.

When we fed our 600 variable database into the 87 booths in Gippsland and benchmarked them against the 2007 Labor 2PP votes and the June 28 2PP by-election swing to Labor in every booth, we first checked the relationship between the 07-08 pro Labor swing and the 07 Labor vote and the correlations were nudging up to minus 0.5, well into significance territory.

 

Read Full Report – 2008 Gippsland by By-Election

 


  • 0

Mar 2001 – Ryan By-Election Analysis

Category:By-Elections

ALP Two Party Preferred Swings, Demographic Correlations for Ryan 1998 Federal Election and 2001 By-election.

 

Download file icon pdf Ryan By-Election Analysis March 2001

 

 


  • 0

Jul 2001 – Aston By-Election Analysis

Category:By-Elections

Demographic Correlations for ALP Swing in Aston By-Election, July 2001

Download PDF - Aston by-election July 2001 ASTON BY-ELECTION ANALYSIS