Anthony Albanese (Image: AAP/Glenn Hunt)

Labor’s poor performance in the recent byelection in the Upper Hunter, held by the National Party since 1931, has prompted a new round of soul-searching about the party’s failure to maintain the support of its traditional “base”.

Implicitly or explicitly, the base is assumed to be typified by male manual workers, particularly those in rural and regional areas like Upper Hunter, or in industrial cities like Whyalla.

In historical terms, this makes sense. The Labor party was founded after the defeat of the shearers’ strike in 1891, and the party long drew much of its support from workers like shearers, canecutters and miners, as well as from urban factory workers and railway workers. There is plenty of nostalgic appeal in recalling the struggles through the 19th and 20th centuries from which today’s Labor party emerged.

But nostalgia is not a reliable basis for political strategy, particularly not for progressive political strategy. Radical changes in the structure of the labour force, which were accelerated by the reforms of the Hawke-Keating era, mean that it is no longer possible to win elections with a program appealing primarily to blue-collar wage workers.

Many of the occupations and industries that formerly supplied Labor’s core support have disappeared, or been largely eliminated through automation. Canecutters are a distant memory. Wool remains an important industry, but a recent report found a total of 2874 shearers in the entire country.

The mining industry has grown strongly, but mining as an occupation has not. The federal government’s Job Outlook reports that there are currently 58,400 people employed as drillers, shot firers and miners — about 0.5% of the workforce. The mining sector employs many more people, directly and indirectly (perhaps as much as 5% of the workforce), but this number includes lots of white-collar workers, as well as transport workers and construction trades.

The end of industry protection eliminated the huge factories, employing thousands of workers, that most closely approximated our standard conception of the working class. A recent list of the top 100 manufacturers in Australia found only a handful with more than 10,000 employees, and most of these were global companies reporting their entire workforce. No more than 20 manufacturing companies have more employees than the 7000 at the University of Queensland, where I work.

Labour market reforms have also replaced wage employment with (notionally) independent contracting. This has further reduced the number of blue-collar workers. The results can be seen in ABS statistics: employees in the traditional blue-collar occupations (technicians and trades workers, labourers, machinery operators and drivers) now account for about 23% of all workers. That compares to 28% for service workers, 22% for professionals and 9% for managers. Contractors and owners-operators (16%) make up the rest.

These national trends are mirrored at the local level, represented by electorates like Grey in SA, centred on Whyalla. As a recent article by David Crowe in the Nine papers observed, Grey was held by Labor for decades, but is now safely in the hands of the conservative party. But this is not (primarily, at least) because Labor lost the working-class voters of Whyalla. In fact, on a two-party preferred basis Labor won nearly every booth in Whyalla.

The problem is that the closure of the Whyalla shipyards and the shrinking of the steel industry produced a sharp reduction in Whyalla’s population. As a result, the electorate of Grey has been expanded steadily, taking in more and more rural voters and producing a safely conservative seat.

Not only does the call for a return to the blue-collar base ignore the demographic realities, it focuses attention on the subset of blue-collar workers least likely to support progressive politics. In Australia and elsewhere, support for the left is stronger among women than men, among young people than among the old, among employees than among contractors and business owners, and among urban rather than rural voters. (The Australian Election Study is a useful starting point.)

The relationship with education and income is trickier, because education is correlated with income. Holding education constant, higher-income voters are more likely to be conservative, while holding income constant, higher education is associated with stronger support for the left. Mostly these effects work in opposite directions, with income predominating (at least until recently). But where they work together, the effects are strong. Voters with low education and high income (many small business owners, for example) are strongly conservative. By contrast, workers in professional occupations with relatively low pay and status support the left.

What does this say about the “aspirational” blue-collar workers represented as the Labor base by Joel Fitzgibbon and others? They are implicitly cast as male breadwinners, typically of middle-age and older, and in regional areas rather than the much-denounced “inner city”. They are either self-employed or work in the private sector. The word “aspirational” is code for high incomes and a focus on less progressive taxes. In every respect, these characteristics are those associated with the conservative parties. Perhaps some of these voters retain a sentimental attachment to Labor, but making them the focus of electoral strategy is a fool’s errand.

Turning the question around, what kind of worker would represent the archetypal member of the Labor base? The analysis above suggests a young woman in a stereotypically female public-sector occupation requiring post-school education, but with an income well below the average for full-time workers. The archetypal Labor voter, if a concrete example is needed, would be a gen Z enrolled nurse working in a major city hospital.

This is not to suggest that Labor should abandon Fitzgibbon’s blue-collar identity politics in favor of some other form of micro-targeting. Labor’s traditional policies of progressive income redistribution and better public service provision, along with protection of the environment, have been highly successful in attracting support at the state level, and have come close to winning federally in the last two elections. There is no point in dumping them in pursuit of a non-existent “base”.