The research is contained in a new report from the Centre for Future Work at the Australia Institute.  It shows that average nominal wages in Australia grew just 1.7% per year between 2012-13 (when the wage slowdown took hold) and 2016-17 (most recent tax data available). That’s below the average national rate of inflation over that period (1.9%), resulting in a decline in the average real wage.

While the wage slowdown was experienced across the country, some regions were particularly hard-hit. Real wage losses were especially large in Queensland and Western Australia. Moreover, the impact was disproportionate in regional communities in both states — located in some of the most fiercely contested electorates in the current federal election campaign. This suggests that public anger over falling real wages could be politically pivotal to the result on May 18.

The new research uses a novel source of data on wages: personal tax returns, summarised in Australian Tax Office reports. The data is produced on a financial-year basis (less frequently than other wage statistics, such as those publishedby the Australian Bureau of Statistics). However, the tax data allow a more precise disaggregation of wage trends by state, community, and electorate.

The 1.7% estimate of annual national wage growth from 2012-13 through 2016-17 based on the ATO data was a full half-point slower than the 2.2% growth reported for the same period in the ABS’s commonly-cited Wage Price Index (issued quarterly by the Bureau). The WPI series makes adjustments for changes in the composition of employment, in order to create a hypothetical fixed “bundle” of jobs. As a result, the impact of changes in job quality (such as the rise of part-time work, casualisation, and ‘gigs’) is not reflected in the WPI results. In the tax data, in contrast, all of these factors affect the evolution of realised average wages and salaries reported per tax-filer. This is thus a more complete and accurate indicator of the evolution of earnings actually received by Australian workers.

Analysis of tax data also confirms that while wage growth in all parts of the country fell to historic lows, two states were especially hard-hit: Queensland and Western Australia. In those states, wages grew more slowly than elsewhere, and real income losses were larger. Real wages fell by over 3% in Queensland, and over 5% in WA, during that four-year period.

Those two states are home to some of the most tightly contested electorates in the current federal election. The tax data allow calculation of wage trends by electorate – a level of detail not possible with other data sources.

The report estimates wage trends for 17 marginal electorates in the two hard-hit states: including all electorates decided by less than a 5% margin in the 2016 election. Real wages fell in every one of the marginal electorates. In 7 seats (6 of which are currently represented by Liberal or LNP members) the cumulative decline in real wages exceeds 5%.

“Perhaps it is not a coincidence that some of the tightest contests in the current federal election are precisely in those communities where real wages have declined the most,” said Dr. Jim Stanford, Economist and Director of the Centre for Future Work, and author of the report. “Public anger over cost-of-living issues is clearly understandable, given this hard evidence that real wages in these communities have fallen substantially.”

“It would be a poetic irony for the whole election to be decided by frustrated voters in these hard-hit marginal electorates, which have been seemingly left behind by economic growth in the rest of the country.”