Did WorkChoices impact on the NSW election results?

Ben Spies-Butcher, Macquarie University
Shaun Wilson, Macquarie University

The 2007 NSW election—like the Queensland and Victorian elections before it—produced little overall swing against the incumbent Labor government. All state elections since the Coalition’s federal triumph in October 2004, and in the era of WorkChoices, have produced Labor wins. While the small swing against Victorian Labor seemed to fit the experience of the Bracks government—competent and scandal free—the same could not be said for the small swing against a scandal-prone Beattie government (now in its fourth term), and Morris Iemma’s even more scandal-prone government in New South Wales attempting the same.

Labor has won all state elections since the Coalition’s federal triumph in October 2004.

Why then, have there been small swings against incumbents with troubled records when voters had the opportunity to protest without changing governments? Could a counter-protest—one against WorkChoices—be part of the answer?

With the federal battle drawing nearer, conservative opinion quickly mobilised to explain New South Wales, pointing to incumbency as still key to the results. Initially, the Prime Minister said WorkChoices did not make a difference, citing the anti-Labor swing in Menai in Sydney’s south (The West Australian). But leading commentators reluctantly conceded some role. The Australian’s Paul Kelly (2007) wrote ‘These laws were a factor in the state election, but how much is a matter of intense dispute between the parties’. Gerard Henderson (2007) mirrored Kelly’s assessment in The Sydney Morning Herald:

There is some, just some, evidence to support Iemma’s thesis. A Herald / AC Nielsen poll, published last Friday, found that 18 per cent of those surveyed described industrial relations as the most important issue in the NSW election campaign. An exit poll, conducted for Unions NSW on election day in nine Labor-held marginal seats, found that industrial relations was a decisive or significant factor in the decision of more than 60 per cent of electors who voted Labor.

But these reluctant concessions conceal a greater political amnesia: the last two changes of government at state level both claimed IR-reforming conservatives: Jeff Kennett’s government lost in 1999, which had attracted huge labour protests, and Richard Court’s government was defeated in early 2001, after re-awakening Western Australia’s labour movement. Both governments had received strong endorsement from the media—Kennett was the country’s leading reformer for a while.

Australia’s electoral system punishes the Labor side of politics.

Howard’s public recognition of the problem came later. As Labor struggled with aggressive media over its alternative to WorkChoices, Howard saw opportunity to address ‘perceptions’ of the new industrial relations system, introducing a new test that gives conditional and limited protection to those signing new AWAs. It was a small but important concession—one that Howard said proved he was a listener and which made possible a new round of pro-WorkChoices advertising paid for by taxpayers. Still, it is a gamble. After previous efforts to promote WorkChoices failed, Howard again risks putting the unpopular laws at the centre of political debate, reminding voters of their problems. Political scientists call this increasing the ‘salience’ of an issue. Politicians usually try to keep unpopular policy out of voters’ minds. Perhaps the Government is aware of this risk but believes it now has no choice. Howard may have calculated that to win, he must somehow turn the tide on WorkChoices and the question remains—WorkChoices may be unpopular but will it be decisive in the context of so many other electoral issues?


Conservatives still push the argument that incumbency will preserve Howard’s eleven-and-a-half-year-old government later this year. They may be right. Australia’s electoral system punishes the Labor side of politics. In the post-war period, Labor has frequently won more votes than the Coalition—and in 1954, a majority of primary votes—but not claimed government.

Still, elements of ‘the incumbent will prevail’ story struggle with unusual facts that alarm supporters of WorkChoices and the government. Federal Labor has led or tied the government (in two-party preferred terms) in 80 per cent of Newspoll results since the July 2005 when WorkChoices became a public issue. Since July 2006, Labor has led or tied in 19 out of the last 20 Newspolls. We must return to 2000–2001 to find similar leads for Labor. Before that, the longest period of opposition dominance in opinion polls took place in 1995 and 1996, when the Howard‘s opposition led (or tied) Labor on primary votes for 32 consecutive polls before taking office.

This time, however, Labor’s primary vote over the last six months has exceeded its earlier peak in 2000–2001, and Coalition’s vote is lower. The ALP has probably never before stayed above 55 per cent of the primary vote for a period of eight two-weekly opinion polls, as it has recently achieved. Moreover, this higher Labor vote appears to be a trend that no state opposition has come even close to achieving over the same period. Even if the Labor’s vote declines from here—in the same way it did before both of the last two elections—Kevin Rudd would still lead the ALP to a comfortable victory on votes.

Labor’s primary vote over the last six months has exceeded its earlier peak in 2000–2001.

Proponents of the incumbency argument still claim these leads are meaningless. But that cannot account for the different fortunes of oppositions at state and federal level. If their argument was true, we would expect to find state oppositions with huge leads over their Labor predecessors, only to find them diminish to nothing as elections near. That has not happened: no state opposition has experienced the turnaround in fortunes that the federal opposition has.

To hold, the incumbency story must account for the strong signs of a trend to Labor in federal polls, but no sign of a trend to the Coalition in the states—despite ample electoral opportunities. A more complicated version might claim that, nowadays, voters have an inbuilt bias to Labor at the state level, and the same for the Coalition federally. This would accord with the ‘electoral logic’ of fiscal federalism: swinging voters want both good services and tax cuts and thus vote Labor at state elections (where responsibilities to deliver services are greatest) and Coalition federally (where tax cuts are most possible). Accordingly, the pro-spending mood of the Australian electorate—evident since the early 2000s—would benefit state Labor governments, but not necessarily damage the conservatives federally.

This story may help explain long term dynamics, but does the return of Labor governments indicate an underlying anti-Coalition mood that lies beyond an incumbency argument?


The alternative argument—one advanced by the NSW Premier in his victory speech—is that the electorate shows signs of rejecting WorkChoices. Was Morris Iemma offering his union allies a convenient gloss on the result (in return for their support) or could his statement be supported by analysis of the NSW results? Let’s consider these results in more detail. We start by stating that we cannot conclude with certainty that WorkChoices was the underlying federal issue in New South Wales. Rather, we claim there is a plausible interpretation of the results supporting this argument.

The NSW election result was influenced by a complex range of factors.

The NSW election result was influenced by a complex range of factors. Urban commuters were said to be furious with the NSW government failures on public transport. Voters in the Hunter were said to be furious with scandals involving local politicians. The Coalition hoped that its heartland would return strongly after 2003. In the end, the election was comfortably won in Sydney, where no seats changed hands between Labor and the Coalition. The Opposition managed to re-gain Manly and Pittwater, but failed to secure the seats in Sydney’s Anglo southern fringe that vote Liberal federally.

Before establishing whether industrial relations made a difference in New South Wales, we must first consider public opinion about the laws. Are they actually unpopular? The evidence seems pretty clear. Opinion polls and survey data has shown that WorkChoices has the support of a small minority of voters, perhaps around 20 per cent (see van Wanrooy 2007). The most recent surveys show that WorkChoices worries low to middle income voters and younger voters most. There are also good a priori reasons for suspecting that the material effects will be greater for these groups, both of whom are more dependent on earnings from work.

However, we argue that WorkChoices is unpopular at two levels. First, the laws are unpopular, as we say, because voters fear their personal impact. In the absence of government data about the real effects of the laws, the task of assessing the reality of these fears is made harder. Still, documents leaked to The Sydney Morning Herald on 17 April suggest that many AWAs breach even the Government’s pared-back standards of labour contracts (Davis 2007). Second, WorkChoices is unpopular because it represents a symbolic threat: Australians don’t want the industrial relations system too closely aligned with employer interests, and don’t think the new laws are good for jobs or the economy.

Newspoll data confirms these two threats. Men, younger voters, and low income earners are all more likely to say that WorkChoices will make them personally worse off. The results are particularly strong for young voters—44 per cent claim this. Overall, 33 per cent think WorkChoices has been bad for them personally—a large minority. It is admittedly lower than the 58 per cent who claimed something similar for the GST in a May 2005 poll (see Newspoll 2005). However, bear in mind that the GST was an overnight change to the tax system with universal impact. WorkChoices, by contrast, has a more gradual impact by its nature. For that reason, we might expect the unpopularity of the laws to increase as its impact on industrial relations affects more workers. The government anticipated hostility to WorkChoices would diminish as voters came to see change as more gradual than claimed by the ACTU. But Newspoll evidence shows no decline in the number of voters adversely affected, and a growing number of voters who think the laws are bad for jobs and the economy.

Many AWAs breach even the Government’s pared-back standards of labour contracts.

This leads us to consider the second and larger threat of WorkChoices: voter assessment of its broader impact on jobs and the economy, and its symbolic threat. A majority of women and voters of all income groups, as well as most voters from the core 18–49 working age group now agree the laws are bad for the economy (see Table 1). In fact, perceptions of the economic impact of the laws—as measured by Newspoll—are actually deteriorating. The poll reveals layers of hostility. Some voters are yet to register any personal impact, but concede the laws are bad at a general level for jobs and the economy. Other voters worry about their future prospects while still others are untroubled by any personal impact but view the laws as breaching widely respected balance in the workplace. The larger symbolic impact is registered in the persistent majority opposition to the laws.

Table 1: Low income and younger voters register the
personal impact of WorkChoices, per cent, 2007
worse off
Bad for
Male 35 48
Female 31 53
Household Income:    
Less than $30,000 na 51
Between $30,000 and $70,000 36 52
Above $70,000 32 56
18–34 44 55
35–49 28 51
50+ 24 47
Total 33 51

Source: Newspoll March 2007; n= 695 (for personally worse off column)
and n=1205 for the economy.

The best way to find out how much WorkChoices swung NSW voters would be to ask them. To our knowledge, no detailed survey research is publicly available. Without this, we have few reliable ways of probing for deeper trends. But we can undertake a more modest strategy, using NSW Electoral Commission data to find out which factors predicted seat-by-seat swings. We have developed a model to test whether electorates with more voters that fit an ‘anti-WorkChoices profile’ behaved differently to other electorates without that profile. Our model uses results from 82 seats for which there are results that measure the swing away from Labor (a further eleven seats produced two-party contests between the Coalition and another non-Labor candidate). We use a mix of other variables to find out those factors that best predict the size of the swing away from Labor. (The full model and an accompanying explanation are available in an appendix.)

A majority of voters now agrees that the industrial laws are bad for the economy.

Our results undercover fascinating trends. It appears voters in the outer-metro areas of Sydney were in the mood to punish the Government. Outer-metro seats were more likely to register substantial swings than either voters in the inner city, or in rural and regional New South Wales. Layered on top of the outer-metro swing, was the ‘Hunter’ effect. Controlling for other factors, Labor appear to cop an additional 5 per cent swing in this area. The absence of sitting Labor members appears to have mattered too. According to our model, Labor faced an additional swing of around 2 per cent where the incumbent did not recontest. (We also tested whether Labor was helped when the Coalition was also fielding a new candidate in earlier versions of the model, but found that they were not.)

But did the two factors used to capture the effect of WorkChoices—seats with more working-age voters and low-to-middle income earners—make a difference? The answer is yes—and in both cases. These electorates registered significantly smaller swings against Labor than others. Put plainly, seats with younger voters and more working families—the voter blocs most worried about WorkChoices—stayed with Labor.


Post-election, Sydney’s electoral map highlights these trends: north of the Harbour is Coalition blue, the rest of the city is Labor red. Clearly, a return to traditional voting patterns has emerged with the west and south staying Labor, and the north strengthening for the Coalition. Commentators, including Antony Green, would say that this is just a return to ‘electoral normality’ after Carr’s unusually strong win in 2003.

Voters in the outer-metro areas of Sydney were in the mood to punish the Government.

This argument would be correct, but miss important novelties. That Labor electorates—traditionally representing younger and lower income voters—remained loyal to Labor is significant. Hit by commuter and service failures, these electorates had no less reason to swing against Labor. And they certainly have in the past, ending the tenure of the Keating government in 1996 and leaving federal Labor further behind in 2004. But, this time, they did not. Perhaps policies that disproportionately impact on working families explain why these voters did not desert Labor in New South Wales. Most representative of these policies is WorkChoices.

Consistent with their tendency to volatility, outer suburban Sydney did swing against Labor—compared to other areas. However, the final outcomes in many of those seats were countered by other electoral forces. Excluding the north, outer suburban Sydney has higher numbers of the two groups of voters who say they are affected by WorkChoices. Our interpretation is that their higher numbers in outer metro seats helped contain the swing against Labor, providing insights into likely trends at the upcoming federal election.

Will outer suburban volatility rebound later this year?

When analysing the differences between the NSW result in the 2001 federal election and the 1999 NSW election, Antony Green (2002) examined the growing differences between the votes of the two major parties at state and federal levels. While Labor achieved only an extra 1.6 per cent of the two-party preferred vote in the 1995 NSW election compared to the 1996 federal election, that gap grew to 8.1 per cent between the 1999 NSW election and the 2001 federal poll. The gap has continued to grow: the NSW election in 2003 recorded a swing to Labor, but NSW voters swung further to the Liberals in the federal poll of 2004.

Transposing federal results onto the state electoral boundaries, Green found that the gap in two-party support for Labor between the 2001 Federal result and the 1999 NSW result was concentrated in a string of outer metropolitan seats, such as in Penrith (22.0 per cent gap), Campbelltown (19.1 per cent), Riverstone (17.6 per cent), Mulgoa (17.3 per cent), Londonderry (16.6 per cent), Heathcote (16.6 per cent) and Macquarie Fields (16.4 per cent). They all recorded more than twice the average statewide gap between Labor’s state and federal two-party preferred vote. Labor held all of these seats at state level, but half would have had notional Coalition majorities on the federal results.

WorkChoices may symbolise threats to working families.

These seats recorded large swings to the Coalition in the 1996 federal election, but also recorded swings (often from already high bases) to Labor at the 1999 NSW election. Clearly, outer-metropolitan Sydney seems to have the largest proportion of voters who change political allegiances between elections. Having a larger proportion of swinging voters, these seats also appear most likely to lead the overall electoral trend. This appears consistent with our analysis. The Hunter notwithstanding, outer-metropolitan Sydney was the region that most likely to follow the general electoral trend away from Labor.

It seems clear that many outer-metropolitan swinging voters voted against the Iemma government. Their decision had to do with the performance of NSW Labor, and won’t add to any swing against the Howard government. But there are good reasons to believe this volatility can change direction, and in the upcoming federal election, substantially affect Coalition majorities.

If our analysis is broadly right—and the impact of WorkChoices contained the swing against an unpopular incumbent—then the industrial laws may amplify the swing against the Coalition later this year. Rather than the two effects cancelling each other out—as they did in New South Wales—they could add to each other, producing larger swings in these seats than that recorded overall. The message from our analysis is that these swings are most likely in seats with younger populations and poorer working families.

Critics might argue that what we infer about younger voters and working families wrongly attributes their current mood to WorkChoices. Housing shortages, rising inflation, interest rates and higher petrol prices are all likely to influence the same group of voters who tell pollsters their worries about industrial relations. But here lies a possible, important connection. WorkChoices may symbolise threats to working families—not only in the workplace, but in their larger position in the economy as living standards in cities—as they are buffeted by rising costs and competition over everything from schools to houses.

The coming battle of interests and symbols

The Coalition is advantaged by symbolic issues that appeal to conflicted conservatives.

In attempting to explain their eleven year loyalty to the conservatives, outer-suburban voters have been variously described as ‘aspirational voters’ and ‘Howard’s battlers’. We think a good general description applying to this group—one found in all affluent democracies—is offered by American political scientist, James Stimson. He calls them conflicted conservatives. Describing their impact on American politics, Stimson says:

The conflicted conservatives are an interesting group. Large enough to swing all elections one way or the other, their votes are potentially available to both parties. They want more liberal policies and respond to specific Democratic appeals to do more and spend more on various domestic priorities. They think of themselves as conservative and respond to Republican identification with conservatism. Which has the stronger appeal, liberal policies or conservative symbols is a close call and so varies with the times. Where demand for liberal policies is at a low ebb, symbols prevail and Republicans win. When that demand is strong – think 1960 or 1992 – then the policies carry the day and Democrats win. That makes these voters hugely important in determining outcomes (2005, p. 93–94).

The specifics are to do with America, but the message applies more broadly. Outer metropolitan Australia responds to the same politics. In the last election, the Howard government was able to combine both conservative symbols and material interests—an attractive combination of Australian nationalism, employment growth, and rising asset prices. This combination left Labor with little to marshal a campaign around. This time is different. While the government will claim that it still offers superior economic management, higher interest rates and inflation—and on Newspoll evidence, now WorkChoices—qualify this.

The Coalition is advantaged by the symbolic issues appealing to conflicted conservatives. But occasionally, symbols favour Labor. Analysing the 1993 election, Clive Bean (1994) argued that the belief that the Hewson-led Coalition lost the election because of their GST proposal was too simple. Instead, Bean highlighted the role played by four issues—unemployment, taxation, health, and industrial relations. With 11 per cent jobless, the first favoured the Liberals, but the others favoured Labor. For these three, the Coalition proposed significant changes: a new tax in the GST, the replacement of Medicare, and the marginalisation of collective bargaining. Bean argues that all three were important vote-changing issues; indeed, his analysis suggests health was the most powerful.

The new laws threaten both the policy and symbols broadly endorsed by Australians.

These three policy issues go beyond the normal appeal of Labor policy to conflicted conservatives. Instead, they hold a special symbolic value that defines an Australian egalitarianism, which many voters—both Labor and Coalition—hold dear. Before the 1996 elections, John Howard shifted his party on all of them. He ruled out a GST, committed himself to Medicare, and promised ‘no worker would be worse off’. When the Coalition decided to reintroduce the GST, the subsequent election saw a 5 per cent swing to Labor, giving Labor a majority of the two-party preferred vote (though not government). WorkChoices represents a similar threat to Coalition incumbency: it gives Labor such a powerful weapon because these new laws threaten both the policy and symbols broadly endorsed by Australians. Howard must hope the new test on AWAs—and an unrelenting campaign against unions by his government and the commercial media—minimises the damage.

At the same time, the Coalition’s most trusted symbols—populist positions on refugees and border protection—have been diminished. The effective ‘veto’ of a resistant group of backbenchers to anti-refugee politics and a new agenda dominated by climate change means that the Government has lost some of its most reliable appeals.

But most importantly, the Government has compromised its claim to represent middle Australia in its industrial relations policies. These laws bring together interests and symbols. Many voters worry about the impact of the laws and a majority see the laws as bad for the economy and jobs. When Australia’s social contract was in better shape—and most could expect a living wage out of work—economic management meant something similar to a wider group of people. It was about managing the costs of living and keeping people in work. Today, perceptions of economic management depend increasingly on the voter’s position in the economic system. It matters whether you are buying, own, or are renting your home; or whether you own shares or are only precariously attached to the workforce through casual or contract labour. Depending on who you are, WorkChoices may come to symbolise an economy managed for some, but increasingly, not for others.

Click here for the technical appendix to this article.


Bean, C. 1994, ‘The 1993 election and Australian Electoral Studies in the 1990s’, Australian Journal of Political Science, vol. 29 (Special Issue).

Davis, M. 2007, ‘Revealed: How AWAs strip work rights’, Sydney Morning Herald, 17 April [Online], Available: http://www.smh.com.au/news/national/revealed-how-awas-strip-work-rights [2007, May 7].

Electoral Commission of NSW, Electoral District Profiles [Online], Available: http://www.elections.nsw.gov.au/state_government_elections [2007, May 7].

Green, A. 2002, Implications of the 2001 Federal Election for the 2003 New South Wales Election, Background Paper no. 1/02, NSW Parliamentary Library Research Service.

Henderson, G. 2007, ‘Work Choices spin just doesn’t make sense’, The Sydney Morning Herald, March 27 [Online], Available: http://www.smh.com.au/news/opinion/work-choices-spin [2007, May 7].

Kelly, P. 2007, ‘No Howard retreat on WorkChoices’, The Australian, Blog, 26 March [Online], Available:http://blogs.theaustralian.news.com.au/paulkelly/index.php/theaustralian [2007, May 7].

Newspoll 2007, Industrial Relations Changes, 3 April [Online], Available: http://www.newspoll.com.au/ [2007, May 7].

Newspoll 2005, GST Poll, 31 May [Online], Available: http://www.newspoll.com.au/ [2007, May 7].

The West Australian 2007, ‘NSW not about WorkChoices: PM’, 25 March [Online], Available: http://www.thewest.com.au/aapstory.aspx?StoryName=367531 [2007, May 7].

Stimson, J. 2005, Tides of Consent: How Public Opinion Shapes American Politics, Cambridge University Press, Cambridge.

Ben Spies-Butcher is a lecturer in Sociology at Macquarie University.

Shaun Wilson is Lecturer in Sociology in the Division of Society, Culture, Media and Philosophy at Macquarie University.