Fifty years ago it was relatively easy to predict how somebody was likely to vote simply by knowing their social class background. While a substantial minority of working class voters supported parties of the right, very few middle class voters supported parties of the left. If you knew someone was middle class, therefore, you could assume with a reasonable level of confidence that they were also politically conservative.

This pattern was found in most western countries, including Australia, although it was probably at its sharpest in Britain. At the UK General Election of 1964, a tiny 2 percent of voters with no ‘working class characteristics’ voted Labour. If someone worked in a white-collar job, did not belong to a trade union, owned their own home, had remained at school past the minimum leaving age, and had grown up in a middle class home with parents who voted Conservative, the odds were 50 to 1 against them voting Labour.[1] Rarely has social science enjoyed such powerful powers of prediction.

A lot has changed in the last fifty years. In Australia as well as Britain, the working class has shrunk and the middle class has expanded. Women have entered the labour force in large numbers, public sector employment has swollen and there has been an explosion in the number of university graduates. Politically, old party loyalties have frayed, support has grown for smaller parties, and the electorate has become more volatile and less predictable. People’s socio-economic characteristics are now much less significant than they used to be as indicators of how they are likely to vote, for not only do many working class people vote for parties of the right, but large swathes of the middle class now vote Labor. [2] Nowadays, if you want to know how somebody votes, sociologists will be little more use to you than tossing a coin.

The weakening of the association between class and voting has been of particular concern for parties of the left, for they have traditionally pitched their electoral appeal in class terms. In Australia (as in Britain) the labour movement created the Labor Party, and the unions still play a key role in the party’s organisation and policy-making. If many manual workers and trade union members are now supporting other political parties while large numbers of middle class professionals are supporting Labor, this inevitably poses the question of what the ALP stands for today, and who its ‘natural supporters’ might now be.

Labor and graduate voters
At the 2004 federal election, trade union members were still much more likely to vote Labor than to support the Coalition (61% supported the ALP on a two-party preferred basis), but trade union votes can no longer provide the ALP with a large electoral base. One of the major social changes of recent times has been the decline of trade unionism, and by 2004, trade union members constituted only a minority of working class voters (only one-third of intermediate and lower-grade manual workers belonged to unions).[3] Support for Labor among manual working class voters as a whole was no higher than support for the Coalition at the 2004 election. Indeed, if lower grade white collar occupations are included as part of the ‘working class’, then the Coalition parties actually achieved a small lead over Labor among working class voters (Table 1).[4]

If Labor is losing its traditional class identity, where is its new social base to be found? Recently, Labor’s Shadow Finance Minister, Lindsay Tanner, commissioned research from the federal parliamentary library which seemed to point to a possible answer. Tanner discovered that Labor’s support has been growing among the more highly-educated sections of the electorate—those with university degrees.[5]

Although Labor support overall fell at the 2004 federal election, Tanner found the Party’s support in constituencies with high numbers of university graduates rose. Labor increased its vote in 10 of the 15 ‘most educated’ constituencies (those with the highest proportions of graduates), and it achieved a positive swing in half of the top 50 educated seats. In contrast, it gained support in only 5 of the 50 ‘least educated’ constituencies, and it lost support in 14 out of the bottom 15.

Interpreting these patterns, Tanner notes: ‘If income is fading as a predictor of voting behaviour, what is emerging in its place? While the picture is still very murky, I believe that across the Western world there are strong signs that education level is growing rapidly as a factor influencing voter choice.’[6]

If this is true, it not only provides some solace for Labor politicians, who can congratulate themselves on being the ‘thinking person’s party,’ but it also bodes well for the long-term future of the Party, for graduate numbers are still on the rise, and traditional blue collar employment is still in decline. The Liberals may have clung to power for ten years by appealing to the working class vote, but Howard’s ‘battlers’ are doomed to dwindle in numbers in the years ahead while graduate voting strength is set to expand.

Before Labor breaks open the champagne, however, we should inspect those graduate Labor voters a little more carefully. Is their education really the key to understanding how and why they vote as they do?

It is certainly plausible to suggest that prolonged exposure to the predominantly left-wing culture of our universities could influence the voting patterns of graduates, and evidence from the Australian Election Study confirms that voters with a university education voted disproportionately for Labor in the 2004 federal election. Indeed, Table 2 shows that the only two educational groups where the ALP won a majority of the votes in 2004 were postgraduates and graduates. While voters with no post-school qualifications divided 55/45 in favour of the Coalition, those with postgraduate qualifications divided 54/46 in favour of Labor. Old political stereotypes, it seems, are getting turned on their heads.

But the patterns in Table 2[7] could be concealing more than they reveal. The association between education and voting might turn out to be spurious if some ‘third variable’ were found to be acting on both. Before accepting Tanner’s thesis we should investigate the possibility that it is not a university degree that draws people to Labor, but is rather some other feature of their lives which is itself associated with a higher level of education.

There are two possibilities worth exploring. One relates to public sector employment. If graduates are disproportionately employed in government jobs, and if people who work for the government are inclined to the left, then it could be employment sector, rather than a university education, that is the real influence on Labor voting.

The second possibility is that Labor voting is only associated with graduates in certain kinds of occupations. It may be, for example, that graduates who go into management, business or IT are no more inclined to vote Labor than anybody else, and that left-voting is restricted to those in jobs like teaching or social work.[8]

As it turns out, both of these ‘hidden factors’ play a part in explaining the voting patterns of graduates. Indeed, when their effects are taken into account, the apparent association between education and Labor voting all but disappears.

The importance of being a public sector graduate
Public sector employees tend to be more highly qualified than those in the private sector. Only 20% of voters who were employed in the private sector had university degrees in 2004, compared with 38% of those working in the public sector. Conversely, nearly two in every five private sector workers had gained no qualifications since leaving school, but this was true of only one in five of those employed by the government.[9]

Not only are government employees more highly qualified on average than private sector workers; they also tend to vote Labor in greater numbers than private sector employees do. At the 2004 federal election, past or present private sector employees voted (on a two-party preferred basis) 59% in favour of the Coalition and only 41% for the ALP. This pattern was reversed among public sector workers, 46% of whom supported the Coalition while 54% voted Labor.[10]

This immediately throws some doubt on Lindsay Tanner’s claim that it is education that is driving contemporary voting patterns, for when we divide the electorate according to the source of their employment, the effect of education on voting disappears entirely for private sector workers. They tend to support the Coalition irrespective of their level of qualification—in 2004, private sector graduates voted almost exactly the same way as those with no qualifications at all (see the upper half of Table 3).[11]

Among people working in the public sector, however, a university degree does seem to have some electoral significance. The lower half of Table 3 shows that public sector graduates split almost 2:1 in favour of Labor at the 2004 election while less well qualified government employees voted in roughly even numbers for ALP and the Coalition.

It seems from these results that Labor voting is associated with a higher education, but only for people working in the public sector. To understand why this should be, we need to explore the association between public sector employment and support for the ALP.

Is there a public sector voting bloc?
In Britain during the Thatcher years, some political scientists speculated that social class was diminishing in importance as an influence on voting and was being replaced by a new fault line based on what they called ‘sectoral cleavages.’ [12]The core cleavage they identified was that between those who relied on state employment for their livelihoods (who tended to vote Labour), and those who worked in the private sector and who resented high taxation (who were often attracted to the Conservatives). [13]

This theory of sectoral cleavages looks plausible, for some sections of the population clearly do have a vested interest in high state spending, and it is possible that they vote accordingly.[14] In addition to the large number of workers employed by the government, there is also a growing army of people who are mainly or wholly dependent on welfare payments provided by the government. Added together they represent a substantial chunk of the electorate.

In Britain a recent Spectator article estimated that, ‘More than half the electorate are today, in one way or another, in the pay of the government.’ They include 15% of voters employed by government, 11% living on welfare, and 26% receiving a government age pension. The author concluded from these figures that no party can now win an election in Britain on a platform of tax cuts and smaller government. [15]

The same kind of calculation can be made in Australia. About 14% of electors in 2004 were employed full-time or part-time by the government, another 13% were people of working age who relied almost wholly on welfare payments for their income, and another 23% were retired, of whom 12% were receiving a full age pension and 7% a part age pension.[16] This adds up to 39% of electors relying on the government for virtually all of their income, with another 7% consisting of retired people depending on a part pension. In addition, there are also millions of working-age voters receiving top-up payments such as Family Tax Benefit and Child Care Benefits.

On the face of it, this looks like a formidable voting bloc. But to conclude from these figures (as the Spectator article does) that ‘no party can win power on a platform of radical cutbacks in government’ is to make the huge assumption that all these people vote according to their economic self-interest as state dependents. The reality is that they do not.

Public sector employees and the Labor Party
It is difficult to gauge from the 2004 Election Study data whether Australians receiving welfare payments and pensions are more inclined to support one party rather than another, for while people on benefits do seem to lean towards Labor, the numbers are too small and/or unreliable to draw any very firm conclusions.[17]

We do, however, know from Table 3 that public sector employees are more inclined to vote Labor than those in the private sector, although this only holds for those with university degrees or equivalent qualifications, and the effect is quite weak. The question is why they might think it is in their interests to vote Labor?

Public sector employees certainly benefit when governments raise taxes and expand their budgets. Their promotion prospects grow when government departments expand in size, and their wages and salaries increase when more money flows to the Treasury. All of this is achieved at the expense of taxpayers in the private sector who have to give up more of their income to the government. It is therefore in the interests of public sector employees to support parties which are likely to increase public spending. But does this mean Labor?

In Australia in recent years, it has been John Howard’s Liberal government that has blown record revenues in a splurge of spending from which public sector employees have benefited.[18] The number of Commonwealth public servants has grown by 20% since 1999 (an annual rate of staff increase of 3.5%), and the number of senior executive service grade staff has grown even faster, by 30%. Public sector wages have also risen faster than private sector wages; Commonwealth wages in certified agreements have risen 4.5 percentage points higher than private sector wages in the last five years.[19] And to add insult to injury, the government has now established a ‘Future Fund’ using taxpayers’ money to pay for its unfunded employees’ superannuation pay-outs.[20]

Government employees have therefore fared well under the Coalition. Nevertheless, Labor is traditionally the party associated with support for high levels of government expenditure on public services, while the Liberals are popularly associated with low taxes. Almost irrespective of what the parties actually do, these stereotypes remain fixed in the public’s mind.[21] It follows that, if you want the government to push up spending, you are most likely to vote Labor, which is exactly what public sector employees tend to do.

Public sector workers are not shy about defending and promoting their sectional interests—46% of them belong to a trade union compared with 17% of private sector workers.[22] Nor are they reticent about endorsing policies which are likely to benefit them, for they support higher taxation in greater numbers than private sector workers do. They are less likely to see tax as an important election issue, are less likely to agree that high taxes destroy work incentives, and are less likely to think that taxes could be reduced by cutting wasteful expenditure.[23] Asked to choose between reducing taxes or spending more on social services, only 30% of public sector workers favoured tax cuts while 46% went for higher spending. Among private sector workers, 38% opted for tax cuts with 33% wanting more spending.[24]

Although statistically significant, however, the differences between public and private sector workers in the way they vote and their attitudes to taxation do not appear particularly strong. One reason for this is that employment sector only seems to make a difference for those in middle class jobs. In the administrative, professional and managerial grades, there is a clear distinction between public and private sector workers in their preferences as between lower taxes and higher spending, but in lower grades this distinction all but disappears (Table 4).[25]

The same is true when it comes to voting. Middle class public sector employees are significantly more likely to support the ALP than are their private sector contemporaries, but among intermediate and manual grade occupations, the distinction between private and public sector employment has little electoral impact.[26]

It seems, therefore, that sectoral cleavages divide the middle class but have little impact among the working class. This casts some doubt on the argument that public sector employees are voting Labor and supporting higher taxation out of self-interest, for if this were true, it should presumably hold right across the public sector, not just in the middle class. It also suggests that the large size of the public sector payroll does not necessarily represent a barrier to any government looking to cut taxes and roll back expenditure, for outside the middle classes, employment sector does not seem to influence the way people vote.

The Opinionators and the Labor Party
Lindsay Tanner’s claim that there is an important link between level of education and Labor voting has been dented, for we have seen that this association only holds for public sector employees. In the private sector, university graduates are no more likely to vote Labor than are unqualified school leavers.

The Tanner thesis takes a further knock when we add a second factor to the analysis—the nature of the jobs in which Labor-voting graduates are concentrated. Inspecting the occupational data more closely, it becomes clear that it is not because they have a degree, but because they are a teacher, a social worker or a graphic designer, that certain graduates are today supporting Labor in such large numbers.

Labor’s graduate support is concentrated among that section of the educated middle class that could be called ‘the opinionators’. These are the people whose job it is to develop, process, transmit and interpret ideas, values and opinions in our society. They include academics (who drive research agendas and provide ‘expertise’); journalists (who report current affairs and set developments in a context we can understand); artists and TV programme makers (who help define and shape our values through written texts and visual images); political advisers and community activists (who develop, analyse and influence public policies); and teachers (whose job it is to transmit our culture to the next generation). Taken together, these people constitute no more than 10% of the employed population, but they occupy pivotal ideological positions in our society.

It is not their socio-economic status that makes the opinionators such a politically distinctive group. A university teacher may have a similar level of income and education to a solicitor or a vet, but he or she is much less likely to vote for the Coalition. Giving every occupation a score between 0 and 100 to indicate its ‘socio-economic status’ (the level of income and education typically associated with work at this level),[27] the mean (average) occupational status score of Coalition voters in 2004 was 38.0 (roughly the equivalent of an hotel manager or an electrical engineer), while that of ALP voters was only fractionally lower at 35.9 (roughly the score of a library technician or a fitness instructor). The difference in these mean scores is not statistically significant.[28] In other words, knowing somebody’s occupational status tells you nothing about how they were likely to vote.[29]

Labor’s relatively strong standing among certain sections of the graduate population reflects the specific kinds of jobs these graduates are doing. Opinionators represent a distinct stratum of the middle class made up of professionals working in education, the arts and ‘social’ fields like counseling and community work. At the 2004 election, these people were strongly supportive of the ALP and the Greens, and given Green Party preferences, this translated into a strong voting bloc for Labor.[30] At an election where Labor won only 45% of the vote on a two-party preferred basis, 64% of opinionators voted Labor. Elsewhere in the middle class, just 40% of scientific, technical, business and health professionals supported Labor, and only 30% of managers and administrators did so. The next highest level of Labor support registered by any occupational stratum was 54% among unskilled manual workers.

Compared with other occupational groups, opinionators are hugely over-represented in the public sector (54%, compared with 27% overall), but this does not explain why they tend to vote Labor. Rather, it is their presence in the public sector which generates the statistical association between the public sector middle class and Labor voting. Similarly, opinionators are also very highly educated (more than 3 in 10 of them has a postgraduate degree, and nearly two-thirds have first degrees), but again, this does not explain their voting patterns. More exactly, it is because so many of them are so well qualified that Tanner was able to find an overall link between Labor voting and a university education. Table 5[31] confirms that Labor swept the board among opinionators, irrespective of whether or not they had letters behind their names (and among other occupations, the Coalition secured a narrow majority of votes cast at all levels of education).

Whichever sector of the economy they work in, and no matter what their level of qualification, opinionators are therefore much more likely than any other occupational grouping to support the ALP.

If Lindsay Tanner is looking for a new social base for the ALP, he will not find it among graduates as a whole, nor in the middle class as a whole, nor even in the public sector workforce as a whole.[32] He will, however, strike gold among that specific stratum of the educated middle class employed in the artistic, social and educational professions. He probably already knows this, for many members of this stratum sit beside him on the Opposition benches in Canberra.[33]

The rise of the Opinionators
The opinionators stand out from all other voters, not only in the way they vote, but also in their wider political beliefs, values and attitudes. They are not only a distinct voting bloc; they are also a distinct ideological bloc, quite sharply divided on many issues from their fellow Australians.

Given their strong support for the Greens, it is not surprising that their attitudes on environmental issues tend to be more radical than those of the rest of the population. Nearly half of them claimed to belong to, or to have considered joining, an environmental group, but this was true of less than one-quarter of other voters.[34] Asked to assess the urgency of a series of environmental concerns, opinionators consistently thought matters were more urgent than other voters did, particularly on high-visibility election issues like logging, or on touchstone issues like GM crops.[35]

Opinionators are also more heavily unionised than other voters—42% of them belong to a trade union, compared with 25% of all voters—although this is partly explained by their heavy concentration in the public sector where rates of unionisation are much higher.[36] They are also more supportive of trade unions. Compared with other voters, they are much less likely to agree that ‘trade unions have too much power’ and are less inclined to support stricter laws to regulate trade unions (see Table 6).[37]

Table 6 also shows that, compared with other voters, opinionators are more supportive of Aboriginal rights, more tolerant about immigration and asylum seekers, more permissive on law and order issues and on questions of child rearing, and more censorious about the Iraq War and the ‘war on terror’.

They are also much more passionate than other voters about the desirability of paying higher taxes to support higher levels of public expenditure. Asked to choose between lower taxes and more spending on social services, only one opinionator in ten strongly favours reducing taxes, compared with almost a quarter of other voters. Conversely, almost four out of ten strongly favour higher spending, an option supported strongly by fewer than two in ten other voters.[38]

At a time when there seem to be few clear political or ideological divisions left between different occupational and social class groups, the opinionators stand out sharply as a distinct section of the Australian electorate. With the rest of the country inclining somewhat to the right, they stand prominently as the major bulwark of the left. This is not without significance for the so-called ‘culture wars’, for the people whose job it is to develop and interpret our culture is more divorced from how the majority of Australians are thinking than any other single group in the population.

Conclusion
Lindsay Tanner thinks that graduates may be Labor’s new core voters. He is wrong. The graduates inclined to support Labor are those working in a specific slice of the arts, education and social professions. In the private sector, a university degree has no impact on voting patterns, and even in the public sector, its influence is limited to those I have identified as the opinionators.

The opinionators are a clearly distinct political stratum who stand out from every other section of the electorate. Not only do they vote for Labor (and the Greens) in much larger numbers than any other occupational group, but their political values and opinions are generally at variance with the way other groups are thinking. On nearly all the big issues of contemporary Australian political debate, they are out on a limb, whether it be support for Aboriginal land rights and asylum seekers, or opposition to tax cuts.

In terms of sheer voting strength, opinionators will never form a large enough group in the electorate to give Labor power. Indeed, given their relatively radical views, they may represent more of a hindrance than a help for the Party’s political ambitions, for given the chance they are likely to drag it further away from mainstream opinion and values. In terms of their wider ideological importance, however, the opinionators occupy many of the key positions within our core educational and cultural institutions. Their political significance should not be measured in votes.