Category Archives: Inequality Research

Highlighting income and wealth inequality data, issues with inequality data, or research (mine or others’) on income and wealth inequality.

Inequality Alarm Bells for South Australia

The data presented in my earlier “Snapshot of Inequality” shows that levels of inequality in South Australia are lower than the national average in a number of areas, but there are alarm bells ringing in the background – and they are getting louder.

The First Warning

At first glance it looks like good news for croweaters. Income distribution between households is slightly more even in South Australia with proportionately fewer households on very high incomes. Compensation for workers in South Australia is higher as a proportion of the state economy than the equivalent national figure (the labour share). The “official” gender wage gap is around half the national average, and households in regional South Australia earned closer to the average Adelaide household than the regional/capital city divide nationally.

However, the apparently good South Australian outcomes come against a background of South Australia having lower average incomes and wealth than the national averages. For instance, the gender wage gap is lower in South Australia – but so are women’s wages with average full-time ordinary time earnings for women $68 per week lower than the national average for women. Similarly, the average income of households in regional SA is closer to the Adelaide average, but $217 a week lower than the national average for regional households.

Wellbeing is about quantum as well as relativity to others (equality). However, the picture here is also complicated by differences in housing costs, service provision and the need for a fuller accounting of income (alluded to in a previous post, but beyond the scope of this work).

The Alarm Bells for South Australia

In the context of this study, what is ringing the alarm bells loudest for South Australia is not the current gap to the national averages, but the general decline in the position of South Australia relative to the rest of the country over the last 30 years. The South Australian share of the national economy (GSP as a proportion of GDP) has fallen from 7.71% in 1990 to 5.7% in 2021. The data in the graph below shows the impact of this relative decline on inequality statistics. Cue the alarm bells: South Australia’s share of national household income, SA labour’s share of the national economy and even the SA female share of the national wage pool, all declined over the period.

Cue the alarm bells: time series (2001-2021) showing downward sloping lines (i.e. declines) of SA shares of national population, household income, labour share, and female wage share.

It is important to note that this is not, or not necessarily, about average incomes declining relative to the rest of country. Equivalised household disposable income in SA was 92% of the national average in 2001, and the same in 2020. Similarly, female full-time total earnings in South Australia were 96% of the national equivalent in May 2001 and the same in May 2022.

Population Impacts

A key factor driving the data, but which is not evident in the average income data is the decline in South Australia’s population share. This decline is evident in the top line of the graph. The drop was less than 1 percentage point (from 7.8% in 2001 to 6.9% in 2020), but this is still significant.

Had SA retained its population share (that is, grown at the same rate as the rest of the country) over the period, there would be around 220,000 or 13% more people in the state. On current averages of household size and income, those population differences equate to around $9.7bn of household income per year extra that would have been in the SA economy (not counting any multipliers arising from extra population).

Clearly the relative loss of population has a major impact on the SA economy. Combined with the changes in the various average incomes, it gives the downward trajectories in most of the data (even where the income share actually increased in relation to the SA economy). Capturing both population and income changes is the whole point of using income shares rather than household averages, and in this case it highlights several concerns for South Australia.

Firstly, with a declining share of national household income, the gap between South Australia and much of the rest of the country is growing. SA is becoming poorer as a community, even if that is not reflected in the individual household data. That is the alarm bell ringing in the background. As I noted in relation to regional South Australia (where this process is heightened with a further decline in regional communities), this impacts on the ability of communities to provide infrastructure, growth and the ability to attract and retain people. It becomes a self-reinforcing cycle, and the alarm bells become a symphony.

Secondly, as the data above shows, the relative decline of South Australia as a whole means that inequalities within South Australia are also made worse – not internally, but in relation to the rest of the country. Put another way, the more egalitarian income spread and distributions within South Australia may ameliorate, but do not overcome the growing inequality between South Australia and the average of the rest of the country (hence the downward sloping graphs).

As a whole, the South Australian data provides an example of how different inequalities interact with eachother. In this case the geographic inequalities undermine greater gender and class equalities (or alternatively, greater gender and class equality ameliorate increasing geographic inequalities).

Policy Response

South Australia is not alone in this predicament. The same is probably true of other smaller jurisdictions (although I have not done the numbers), and it is certainly true of regional South Australia where the decline of local communities is truly alarming. Clearly a policy response is needed: the current policy settings are not working as South Australia and the regions are being left behind. However, the interactions of different forms of inequality make the policy response complicated.

I have previously noted the importance of government redistribution through mechanisms like the sharing of the GST pool. The data here clearly shows the ongoing need for a sharing from the high-income states to the rest of the country. I have also noted that traditional responses in economies at the periphery (be it regional areas, struggling Australian states or developing countries) has been to attempt to intensify the use of natural resources. However, academic literature suggests that this “extractivism” offers no guarantee of genuine development and may exacerbate inequality and environmental degradation.

In South Australia’s recent history, the Rann/Weatherill Labor governments, like the Liberal state government before them, were desperate to develop new mining operations in the regions and defence technologies in Adelaide. The Marshall government offered a high-tech space future, and the current government spruiks a hydrogen future. Whatever we make of these ideas, the obvious point is that they did not stop the relative decline of South Australia and regional South Australia in particular, and there was little consideration of the multiple layers of inequality in those policies. A far more comprehensive approach is needed.

There are no easy answers here, and there are contradictions in dealing with ininequality. For instance, as my post “the super-rich don’t live here” indicated, the biggest areas of geographic inequality between South Australia and the rest of the country, and between regional SA and Adelaide, exist at the higher end of the income spectrum. This creates a contradiction because policies aimed at creating high paying jobs in particular areas would be a step towards reducing inequality between geographic areas, but would increase inequality within those same areas (as there would be a greater spread of incomes). For instance, policies which pay highly skilled workers loadings for working in regional or remote areas may increase incomes at higher levels (and may be necessary to ensure service provision), but they will increase the income gap between those people/households and most other households in that area. Similarly, regional development based on extractive industries may increase regional incomes and geographic equality, but is also likely to favour male workers and exacerbate the gender wage gap.

Conclusion

The contradictions of development and inequality are difficult to navigate, and the focus on multiple dimensions of inequality (in this case, how the inequality between South Australia and the nation as a whole intersects with inequalities within South Australia) complicates the picture of inequality. However, such a focus is important.

The examples above and the use of data on gender and class inequality alongside household income remind us that inequality is built in to economic activity. In turn, this suggests that addressing inequality can’t simply be a distributional after-thought where we drive policy for narrow economic goals, then catalogue how the results are distributed and perhaps seek to address inequalities at that point through welfare provision.

At best, with that approach we will always be playing catch-up – looking for payments and services to paper over fundamental cracks. More likely, we will simply lose. The macro-dynamics will overwhelm any welfare provision.

I will say more about that in a future post, but from the inequality data presented here and in previous posts, it is clear that greater economic intervention is essential. If left to themselves, market forces and demographic patterns will continue to channel money and population to existing growth areas, and continue the decline of South Australia. And if intervention and governance is done without a focus on inequality, or with too narrow a focus only on some inequalities, we will see people or groups within South Australia (and elsewhere) left behind.

The alarm bells are ringing – we need to pay attention.

Snapshot of Inequality – South Australia and National

This snapshot of inequality summarises my series of recent posts on this subject. The series has been a journey. I didn’t completely know where the data would take me and I am now asking different questions than when I started. However, frustrated by the mono-dimensional analyses that often dominate discussion of inequality (including in my own work) I was keen to explore the multi-layered nature of the beast.

Inequalities in income and wealth distribution between households, across states and regions, and between structurally differentiated social groups all matter, so I was keen to analyse the data on all of these – even if in an iterative fashion. I also wanted to look particularly at South Australia, partly for local relevance and partly because state-level data is often not factored in to the analysis of inequality.

Snapshot of Current Inequality

Overall the data examined in this series (almost all it sourced from the Australian Bureau of Statistics) showed significant levels of inequality across the country, as summarised in the table below.

 AustraliaSouth Australia
Household Income (between states) (2019-20)Australian average (mean gross) income $2329 per week, but state averages range from $1,736 (Tas) to $2,422 (NSW)Average household income $1,989 p.w. = 85.4% of national average, although difference mainly at top end. SA receiving 6.3% of all household income (which is below its population share).
Household Income (within states) (2019-20)  Bottom 40% of households received 13.4% of the total income. High-income households (90th percentile) received 9 times the income of low-income households (10th percentile).Inequality broadly reflects national patterns, but relatively lower incomes at the top end of the income spectrum
Household Income (Regional Areas) (2019-20)Regional Australia accounted for 31% of the population, but received only 27% of income.Regional SA share even smaller with 19.9% of population, but just 17.7% of income.
Household Wealth (2019-20)Distribution more unequal than distribution of income: highest wealth quintile held 62.2% of all household wealth, while the bottom 40% of households held only 6.1% of wealth.Distribution data not available, but wealth holdings in SA have a different structure (relatively less wealth in home ownership, more in financial assets).
Labour Share of the Economy (2021)“Compensation of employees” at historically low levels (47.7% of GDP).Labour share slightly higher (49.3% of GSP)
Gendered Wage Patterns (2022)Gender pay gap of 14.1% in full-time ordinary-time earnings. Bigger gaps when all earnings and all employees included.At 7.4%, f/t ordinary-time earnings gap is around half the national average, but on lower earnings and relatively lower male earnings. Difference between national and SA figures narrows when all earnings and employees included.

Changes Over Time

The above snapshot of inequality is precisely that – just a snapshot at the current point in time. Arguably, a more important story is evident when changes over recent decades are considered.

That story is not straight-forward and has been explored more fully in the earlier posts. However, the short version is that, at the national level:

  • income inequality between households has increased slightly,
  • inequality between cities and the regions, between capital and labour, and inequalities in household wealth have all increased more markedly,
  • gender wage inequality was the only measure where inequality decreased.

These trends are evident in the graphs below which trace changes in the share of the various pools of total income (or wealth, or production), alongside the same data for South Australia (i.e. share of SA total income/product). The time periods vary depending on data availability.

Household Income and Wealth

As can be seen, nationally, the share of total household income of the lowest two (equivalised) income quintiles has been relatively stable, peaking in 1996-97 at 21.4% and falling to 20% in 2019-20. However, while this fall in income share appears small, every 0.1% change represents over $22.6m (in 2019-20) going from the lower to higher income quintiles. Even more significantly, the share of national wealth held by the poorest two wealth quintiles fell more markedly, although the data is more limited, and is not available for South Australia.

Time series snapshot of inequality: share of household income and wealth captured by the bottom two (equivalised)  income quintiles, 1994-2020.

Households in Regional Areas

The share of total income received by households outside Australian capital cities fell from 30.7% of the national total in 2000-01 to 27% in 2019-20. (Note: not all years are included in this data). The South Australian data is more volatile, and shows a lower share overall (with proportionately fewer households outside the capital), but the trend is similar.

Time series snapshot of Inequality: Share of national and SA income captured by households in respective regional areas, 2000-2020.

Class

Nationally, labour’s share of the economy fell from 48.8% in June 1994 to 47.7% in June 2021, with each 0.1% change in this data set representing $2bn (in 2021) lost from labour payments. However, the South Australian data here is different, falling more swiftly from a higher share of the economy to a low point in June 2004, then recovering to 49.3% of Gross State Product in 2021 – a larger share of the economy than the labour share nationally.

Time series snapshot of inequality: labour compensation's share of the economy, SA and Australia, 1990-2020.

Gender

As noted above, the gender wage share is the only indicator to see a reduction of inequality with women increasing their share of the national wage pool from 33.1% in November 1994 to 39% in May this year.

Time series snapshot of inequality: female share of total wage pool, SA and national, 10094-2022.

Caveats and Conclusions: Why the Numbers Matter

This snapshot of inequality focuses on shares of total pools of income/wealth, rather than the more traditional but disparate average income figures. My approach enables some consistency of analysis across the different data sets, but I make no claim of a causal relationship or that the inequalities are comparable in nature. Clearly, inequalities can contribute to each other and the data sets overlap, but they are analytically separate and the different trajectories show why it is flawed to simply focus on one dimension (usually household income) when considering questions like whether inequality is increasing or not.

In a future post examining the South Australian data in more detail I will explore more specific interactions between different axes of inequality, but the point of this snapshot of inequality is simply to summarise the data and note the importance of considering multiple forms of inequality alongside each other, rather than the usual mono-dimensional focuses.

In arguing for a broader focus on the multiple forms or layers of inequality, I am not calling for endless sets of data or the infinite division of society until we are left only with individuals (as in the neoliberal dream). Rather, my point is that statistics (and all research data) is a reflection of the questions we ask and the theoretical understandings underpinning the research. Our national economic statistics are a product of the neoclassical and Keynesian theories that gave rise to them (that was Chapter 1 of my PhD). So too, the data we use to describe inequality reflects particular theoretical standpoints.

More than that, the data can limit the way we see society and the policies we might pursue to address both inequality and political economy more generally. For instance, data on the distribution of income between households tends towards a tax-and-transfer redistribution (after the fact) to support households in the lowest income brackets. By contrast, labour share data begets industrial policies, while regional data inevitably leads to development debates.

In saying that, I am mindful that my analysis is not comprehensive. I am sure that people with better statistical programs and skills could provide more nuanced numbers, and not all structures of inequality have been examined. Most notably, there is no consideration of structures of racial inequality, although with relevant census data to be released later this year I may be able to add to the analysis later. Perhaps more importantly, I am also acutely aware that race, gender and class inequality is ultimately not reducible to numbers – or even to the economic.

All that said, the economic aspects of inequality are important, and the numbers do provide useful points of reference. At its most basic, it seems to me to be important to have some sense of the scale of inequality, whether things are getting better or worse, and in what areas.

South Australia’s Small Gender Pay Gap: Is it Good News?

The gender pay gap in South Australia is around half the national figure – but is this really good news?

This is the sixth post in the series on inequality in South Australia, with the early posts dealing with inequality between households and the later posts looking at structural inequalities. This post continues the focus on structural inequality, and looks at gender inequality – or at least one aspect of gendered economic inequality.

Gender Pay Gaps – Official

The “official” gender pay gap figures used by the government’s Workplace Gender Equity Agency show that the gender pay gap is significantly narrower in South Australia than in the country as a whole. At 7.4% the gender pay gap in South Australia is around half the national average, and is the lowest in the country. However, there is devil in the detail of this official figure.

While the gender pay gap is lower in South Australia, this gap is on lower wages overall and does not necessarily reflect women doing relatively better. ABS Labour Force data shows that the $1541 per week full-time ordinary time earnings for women in South Australia was 95.8% of the equivalent figure at the national level. For men in SA, full-time ordinary time earnings were 89% of the national figure. In theory, this difference could mean that SA has proportionately more women in higher paying jobs, but more likely it reflects labour market segmentation where proportionately more women are in jobs with wages set nationally (e.g. minimum wages or modern awards), while men are overly represented in non-award industries where wages may differ more across the country.

This suggests that much of the difference in the national and SA gender pay gaps is not about women’s pay, but rather because men’s wages in SA are disproportionately lower than the national average. Put another way, rather than the labour market in SA bringing women’s wages up closer to men’s, it is an “equalling down” based on relatively lower men’s wages.

This creates particular challenges for the left and the union movement: how to increase wages in South Australia to closer to national averages without also increasing the gender wage gap? Obviously, a focus on gender segmentation and increasing wages in highly feminized industries like childcare is a start, but it should also be a factor in changes to enterprise bargaining currently under consideration.

Bigger Gender Pay Gaps Beyond the Headline Figure

The “official gender pay gap” based on full time ordinary time figures is only one measure of gendered pay inequality. The table below shows a range of key gender pay gap and wage share data for Australia and South Australia and is necessary because the full-time ordinary time data is limited. It ignores (and arguably institutionalises) gender differences in access to overtime and bonuses.

The first line in the table is the official full-time ordinary time data, but as can be seen in the second line of the table, when overtime, bonuses and other extras are taken into account in total full-time weekly earnings the gender pay gap is much larger. It is still less of a gap in South Australia, but the difference in the national and SA figures is narrower.

However, even these full-time figures ignore the over-representation of women in part-time and casual work. As evident in the third line in the table, the gender pay gap jumps markedly when the average weekly earnings of all employees are included. The gap between South Australian and national figures is much lower here, presumably because of a higher proportion of women in part-time jobs in South Australia. This is confirmed by the gender wage share data in the bottom line of the table, where the difference between the national and SA figures is lowest.

Gender Pay Gaps, May 2022

 AustraliaSA
F/T Ordinary Time Earnings14.1%7.4%
F/T Total Earnings16.5%9.6%
Average Earnings29.7%27.7%
Women’s Participation Rate62.2%58.4%
Gender Wage Share39.0%40.0%
Source: Calculations from ABS Average Weekly Earnings and Labour Force

Gender Wage Share

I have set out elsewhere a rationale for considering, alongside the traditional average pay gap data, the gender wage share – that is, the female share of the total wage pool. The importance of that wage share data can be seen here.

Both nationally and in South Australia, the female half of the population takes home around 40% of the total wage pool. In actual dollars, the May 2022 figures show that men as a whole in South Australia take home $221m more than women each week, for an annual return to gender of $11.5bn. To put that in the context of the local economy, the total sales revenue of South Australia’s biggest company, SANTOS was around $6.9bn in 2021, around 60% of the gender wage share differential.

As I argued in an earlier post about the corresponding national figure of a $200bn annual wage differential, the aggregate gender wage share difference is a significant economic flow that not just reflects but contributes to the reproduction of that inequality.

Changes Over Time

The 2022 data above shows, in short, that while the official gender pay gap is much lower in South Australia, the benefit of this is diluted by lesser participation in paid work by women (both relatively lower participation in the workforce, and relatively fewer hours by those engaged in paid work). As we will see below, these patterns are not new.

The following graph shows the full-time (all earnings) gender wage gap in Australia and South Australia over the last 25 years. While the data on South Australia is more volatile, the gender pay gap in South Australia has generally been below the national figure. In both data sets, the gender pay gap shrinks in the late 1990s, but grows again from 2000 nationally and from 2004 in South Australia through to 2014. The gap then shrinks over the next few years before flattening out in the last few years.

Time series of gender pay gap in full-time total earnings in South Australia and Australia.

However, the gender wage share data tells a slightly different story. There is little difference in the time-series data on South Australian and Australian women’s share of their respective total wage pools. In both data sets there is a small but sustained increase in women’s share (apart from a small dip after the Global Financial Crisis in 2008) (see the black SA line in the graph below). In South Australia women’s share of the total wage pool increased from 35.1% in November 1994 to 40% in May this year. The national figures showed an increase from 33.1% to 39% over the same period.

Crucially though, this story is set against the backdrop noted in a number of my previous posts of a general decline in South Australia’s economy relative to the national economy. This was evident in SA’s share of household income and the labour share of the economy. The graph below shows that this decline also impacts on the gender wage share. While the top line shows that women have been increasing their share of the South Australian wages pool, with that pool in relative decline, South Australian women’s share of the national wages pool has actually shrunk from 2.6% in 1994 to 2.41% this year.

Time series (1994-2022) of female share of total wages pool in South Australia, and South Australian female wages as proportion of Australian wage pool.

While this decline in South Australian women’s share of the national wage pool is alarming, it is made even more problematic because, as noted in the previous post, the wage pool itself is declining as a proportion of the economy.

Summing Up

The official gender pay gap data shows South Australia doing relatively well with the smallest gap in the country, but a closer look at the data tells a less rosy story. The smaller gender wage gap is based on lower women’s participation rates and relatively lower male wages in South Australia. It is not a story of greater female agency in the labour market. The national/SA differences almost disappear when the share of the total wage pool is considered.

And regardless of the comparative story, the gender wage inequalities in South Australia remain significant – with a 27.7% gap in average earnings leading to an aggregate $11.5bn annual wage gap. Further, while women’s share of the South Australian wage pool has increased over the last 25 years, these gains are undermined by the shrinking of that wage pool relative to the national wage pool – a national pool which, with historically low labour shares of GDP, is itself shrinking relative to the economy as a whole.

Beneath the headline gender pay gap data, for South Australian women workers there is a particularly problematic intersection of gender, geographic and class inequalities.

The Labour Share: No Paradise for Workers

This is the fifth post in the series on inequality in South Australia. While previous posts used ABS data to identify and track inequality between households, this post shifts the focus to structural inequality, and questions of class in particular – as evidenced by the changing labour share of the economy. Future posts will look at other structural inequalities based on gender and race.

Structural Inequality

The shift in focus is important because while the household data reveals much about patterns of inequality, it also has many limitations. Household income and wealth data is based on an assumption that households share resources. This assumption is probably unwarranted for many households, including the 13,000 “group households” in South Australia (272,000 nationally) where some expenses may be shared, but income is likely not shared. Further, viewing the household as the basic economic unit hides dynamics of potential inequality within households – most obviously gender and age dynamics, which then have ramifications in society more widely.

Further, the household data produced by the ABS, and used in many studies of inequality (for instance, the Productivity Commission report and the ACOSS/UNSW research) presents a mono-dimensional stratification of income distribution: a continuous spectrum from lowest to highest income/wealth on which all household/individuals are located. The measures of inequality are then based on the relationship of arbitrary points along this continuum such as income quintiles, top-bottom deciles. They do not tell us much about what is driving this stratification.

While it is possible to look at different demographic characteristics of households at various points on this spectrum, it is still largely only a scoring of an end point of distribution – not the mechanism of unequal distribution itself. By contrast, structural inequality is not simply a different location on a spectrum, it is a systemic and often conflictual relation, buttressed by a range of institutional arrangements where the economic inequality drives or at least contributes to the reproduction of that inequality.

Class – the Labour Share of GDP

This description of structural inequality and some of terms above are loose and contested, and there are mountains of academic writings on the subject. I have previously discussed Erik Olin Wright’s attempt to draw together classical Marxist, Weberian and Durkheimian approaches to class. All these approaches offer different insights, and all are added to and cross-cut by analysis of other structural inequalities. However, with no claim to being comprehensive or determinative, in this post I want to keep it fairly simple with just one measure of one structural inequality: class, as measured by the relative financial returns going to capital and labour.

The usual measure of these class-based economic flows is the share of national income going to labour (“compensation of employees” in ABS-speak), or alternatively, the labour share of the whole economic pie (that is, the share of Gross Domestic Product going to labour). I am sure there is a PhD somewhere critiquing the notion that “GDP = the economy”, but nonetheless, the labour share of GDP is a convenient measure as there is a robust ABS data set and changes in the labour share reflect changes in class power within the economy.

The Labour Share Data – National and South Australian

As is well-known, labour’s share of GDP in Australia has been falling in recent times. In 2018, the Journal of Australian Political Economy dedicated an excellent issue to highlighting the issue and examining the causes, and the labour share remains a cause of concern for the labour movement and many on the left.

Rather than add to the debate about causes or solutions, in this post, I simply want to look at the figures and compare the situation in South Australia with the rest of the country. As can be seen in the table below, despite lower average wages and lower workforce participation rates, the most recent ABS data (June 2021) shows that labour share of the South Australian economy was slightly higher than the labour share nationally.

AustraliaSouth Australia
Average Total Weekly Earnings$1,797.10$1,626.00
Participation Rate66.2%62.7%
Labour Share47.7%49.3%

While this suggests a slightly better distribution of income to workers in South Australia (i.e. a more equitable outcome in class terms), as the graph below shows, this is not historically consistent. The labour share in South Australia is more volatile and slumped below the national average in the decade from 1996 to 2006.

Time series showing labour share of economy in SA and nationally from 1990 to 2021. Long term decline in national series, while SA slumped but regained ground from 2010.

It should also be noted that the national data here does not show the current “historically low” labour share discussed in the media. That discussion is rightly based on seasonally-adjusted figures, but these are not published at the state level so I have used the ABS “original” data series to ensure a like-for-like comparison. That said, the seasonally adjusted figures show an even lower labour share nationally (46.1%), so either way, it would appear that SA labour’s share of the state economy is higher than labour’s share at the national level.

While this may appear to be good news for South Australian workers, this is a higher share of a decreasing part of the national economy. In 1990 the SA economy (Gross State Product) accounted for 7.7% of the national economy (GDP). In 2021, SA’s share was 5.7%. As the graph below shows, South Australian labour’s share of the national economy (in orange plotted on the right axis) has declined over the last 30 years, even while largely retaining its share of the state economy (black line plotted on the left axis).

Time series graph repeating SA labour share of GDP, but also showing long even decline in SA labour's share of national economy (GDP)

For South Australian labour there is then a double-challenge – the class challenge of retaining and increasing its share of income, and the development challenge of growing the economic pie overall. Of course this struggle is not unique, but it is a particular challenge given the arguments in previous posts about being at the economic periphery.

Caveats and Conclusion

It is worth emphasising again that the labour share is a class distribution, an economic process rather than a positioning of people or households. Many people and households (particularly higher-income households) have multiple sources of income including government transfers, bank interest, investment income, imputed rents and capital gains. These are different economic flows (class processes) and those households’ labour income does not necessarily determine their total income, lifestyle options or position in society.

That said, labour income is the major income source for the majority of Australian households. Accordingly, changes in the labour share of income are a key contributor to the household income spectrum, while also being important in their own right as a reflection of structural inequality.

In this context, my interim conclusion (to be added to in future posts) is simple: with a declining share of the national economy, it is no paradise for workers[1] – either in South Australia or nationally.


[1]  With apologies to Ken Buckley and Ted Wheelwright for stealing their classic title.

Household Wealth Inequality

This is the fourth post in the series on inequality in South Australia. The previous posts focused on different aspects of inequality in the distribution of income. This post looks at the distribution household wealth, again based primarily on the 2019-20 ABS Housing Income and Wealth data released this year.

Wealth Distribution Between Households

It is well-known that the distribution of wealth is far more unequal than the distribution of income. This is a function of both the cumulative impact of income differentials over years, as well as the ability of wealth to beget wealth (through capital gains and higher average returns on capital).

Nationally, the households in the highest wealth quintile hold 62.2% of all household wealth in Australia, while the bottom 40% of households account for on 6.1% of wealth. The P90/P10 wealth ratio sees the high-wealth households (90th percentile) owning some 52 times the wealth of the low-wealth households in the 10th percentile. (By contrast, households at the 90th percentile earn around 9 times the income of the low-income (10th percentile) households.

There is clearly a massive inequality in wealth distribution, but it is important to note that these high wealth households are not necessarily high-income households. Most obviously wealth accumulates over a life-time and at some point people retire which may leave them with some wealth but low incomes. The ABS data suggests that the wealth-income correlation is not straightforward. For instance, 10% of high-wealth households had low-incomes, while only one-third of low-income households also had low wealth. However, Piketty’s work highlights the correlations at the very top of the income/wealth spectrums where income from capital (wealth) becomes the dominant income stream.

Unfortunately, the ABS does not publish wealth ratios or wealth-by-quintile of data on a state basis, so there is limited data on wealth inequality within South Australia. However, there is some useful data on wealth distribution between states.

SA share of total household wealth

As with previous posts, I start with aggregate wealth rather than the ABS household averages because the aggregates adjust for differences in population. The graph below shows the various state and territory share of total wealth held by households in 2019-20, but plotted against the share of total income and population. It can be seen that NSW and Victoria have a greater share of wealth than of total current income, while Queensland and WA a higher share of income than wealth. South Australia’s share of the total wealth held by Australian households is quite close to its share of income (wealth 6.5% of total, income 6.3%), but both are below the state’s population share.

Column graph showing state/territory shares of national household income, population and net household wealth.
(Note: NT data does not include remote areas)

While this graph is a snapshot of 2019-20, the long-term trends are not easy to discern. My previous post showed that South Australia’s share of national household income has declined over the last two decades, driven in large part by a declining share of the population. However, the columns in the graph below, which show the state’s share of total household wealth, do not show a similar pattern of decline – or any pattern really. This could be a trick of data unreliability (wealth data is less reliable than income data), or of different structures of household wealth (see below) or different levels of indebtedness (which would impact on net wealth – although the data is not available over the long term).

Time series showing SA declining population and income share (as % of national aggregate), but no real pattern in share of  wealth bouncing between 5.6% (2016) and 7.2% (2006), and overall approximately the same in 2020 as in 2004.
(Note: data not available for 2000-01 and 2007-08)

I have included this time series data for completeness, but I draw no conclusions from it. Nor can I draw any conclusions on the distribution of wealth between households in Adelaide and the rest of the state. Even though the ABS publishes it, the regional wealth data has such high margins of error I would not rely on it. I think the primary take-out of the data above is that South Australia’s share of national household wealth is lower than its share of population. Or put another way, wealth (like income) is disproportionately held elsewhere – and with only around 6% of both wealth and income, this makes it difficult for South Australia to have an impact on Australia’s political economy.

Structure of Household Wealth

Beyond these aggregate figures, the published data based on average household wealth shows some interesting differences between South Australia and other states in the structure of wealth holdings.

The graph below shows a state-by-state comparison of different components of average household net wealth (that is, the value of household assets minus liabilities). At the highest summary level, it is clear that South Australian households on average have less wealth than Australian households generally (approximately 88% of the national average), and significantly less than states like NSW, Victoria and the ACT. This is seen in the height of the columns on the graph, but it is interesting to note that South Australian households have the fourth highest average wealth. By contrast, SA households have the second lowest average household income. Both WA and Queensland have higher average household incomes, but lower average wealth.

South Australian households fare relatively better in wealth distribution than income distribution, but this is even clearer when we look at the components of this total net wealth. One of the key reasons for SA lagging behind the richer eastern states is the higher value of wealth tied up in owner-occupied dwellings there (the blue at the top of the columns). That is, households in other states where housing prices are significantly higher are likely (on average) to have more of their wealth tied up in their homes. Not counting owner-occupied dwellings, the wealth held by the average South Australian household is just $15,000 less than the national average. The SA average is 98% of the national average.

Column graph showing different components of wealth (real estate, and various financial assets) - net of liabilities. SA has 4th highest level of wealth, but below national average - but much closer if owner-occupied housing assets are excluded.

The amount of wealth held in owner-occupied housing is important because there is considerable debate as to whether owner-occupied housing should actually be considered as wealth. At one level it is clearly an asset which can be sold for money, and (as I have argued in an earlier post) provides housing services which is a form of in-kind income which can be imputed as non-cash household income. On the other hand, as Anwar Shaikh and others have argued, it is not real “wealth” because housing is a necessity and cashing in the asset is not simply transferring a physical asset to a cash asset (with an attached transfer of income from imputed rent to other investment income). Liquidating an owner-occupied house actually leaves the owner worse-off as they will now have to pay rent.

Obviously though, the same is not true for housing properties not occupied by the owner as these are held as investments and can be liquidated in a simple swap to another form of asset.

The wealth tied up in owner-occupied housing is also important because lower housing costs in South Australia (relative to other states) can also translate to SA households holding relatively more financial assets (because their “savings” are not tied up in housing costs to the same extent). This is evident in the graph in the relatively high shareholdings and business assets held by South Australian households (seen in orange), although the ABS warns that there is a high margin of error on these particular figures and some caution is required.

Further though, (despite lower incomes) lower relative housing costs also potentially translate into lower average household debt. In 2019-20, the principal outstanding on loans for owner-occupied dwellings in SA averaged out at $88,500 per household, which was 77% of the national figure, while total household debt as a proportion of household wealth was also lower in South Australia (14.4% of total household assets in SA, 16.4% nationally).

The debt figures are averages of households with and without debt, so those in debt will have much higher amounts owing than these average figures, and the averages will be impacted by the proportions of households with/without debts. However, on its face it does suggest interest recent and predicted interest rate rises will impact slightly less in South Australia than in states (and territories) where average debt is higher.

Conclusion

What all of the above suggests is that, despite the average income of South Australian households perennially lagging behind the national average, SA fares better in the distribution of household wealth. Indeed, when the wealth data is considered without reference to owner-occupied housing (which is not transferable or income-earning in the same way as other assets), average net household wealth in South Australia is pretty close to the national average.

However, at the aggregate level, South Australia’s small share of both income and wealth shows that there are still significant issues around the state being relatively marginal to income distribution and wealth accumulation in the national economy. This has not (yet) flowed down to impact on average household wealth, but the wealth data does little to relieve the concerns highlighted in previous posts about SA being at the economic periphery.

And finally, this focus on geographic inequality should not blind us to the massive inequality of distribution of wealth within the state. There is little reason to assume South Australia is immune from the national pattern where the majority of wealth is held by the richest households and the lowest quintiles hold almost no wealth. The dual challenges remain: to get a bigger share of national economic development for South Australia, while distributing that share more equitably within the state.