Category Archives: Inequality Research

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

Calling Out Older People’s Privilege

Leave a Reply

As I watch the South Australian state election campaign unfold, I feel compelled to call out older people’s privilege – that is, the leveraging of public funds based on a false sense of need and/or entitlement. Spoiler alert: my argument is that poverty and need among older people is about class, not age – but more about that later!

Unequal sign, with wording "inequality - not what you think"

Old Does Not Equal Poor

I will look at a few examples of older people’s privilege below, but first, the basic data that might question the allocation of concessions and government support based on age (or even receipt of the age pension), rather than on need. I have previously posted that 72% of age pensioners own their own home, and around two-thirds of those homeowner pensioners have more than $100,000 in additional financial assets. Not rich, but not poor either – and those statistics say nothing of the self-funded retirees who are too rich for the age pension.

Overall, the data is clear that there is a correlation of age and wealth as many people with steady jobs accumulate wealth over a life-time. ACOSS/UNSW research shows that the average older household was 25% wealthier than the average middle-aged household, and almost four times as wealthy as the average young household.

And yet, we have policies and proposals to provide government support for older people or households that are not available to other, poorer households.

To be clear, an age pensioner who is renting a house and has only limited savings is going to be facing serious financial pressures and in need of support. But that poverty is about not having property and capital income, rather than about their age. And even then, that pension income remains significantly above the income-support and concessions provided to unemployed people of working age.

Yet lurking at the back of our minds and at the front of the social security system is a very old idea of the deserving and undeserving poor. Older people are assumed to have worked throughout their life and deserve support in their retirement, while working-age people who are unemployed or experiencing waged poverty are somehow to blame for their poverty.

This may change as compulsory superannuation opens the possibility of age pensioners being seen as undeserving losers who did not work hard enough to provide for themselves in retirement, but fortunately, we are not there (yet). And in the meantime, despite the obnoxiousness of the deserving/undeserving politics, older people’s privilege is alive and well. Here are a few examples that have come across my desk in the last few weeks.

Examples of Older People’s Privilege

Stamp Duty Promises

At this state election, the Liberal party is promising to provide a $15,000 stamp duty discount for over 55’s looking to downsize, while Labor is promising to abolish stamp duty for over 60s downsizing to a new-built dwelling. Ignoring the impact on government revenue and the fact that stamp duty overall is an unfair and inefficient tax, these tax concessions (almost by definition) go to people who can most afford to pay stamp duty – because they are getting a significant cash bonus from their downsizing. Young families going from their first home unit to a suburban house are likely to need stamp duty discounts more than an aging downsizer, but older people’s privilege prioritises the needs of age over financial position (albeit dressed up as a wider market intervention to free up supply).

The Older People’s Privilege Lobby

The election campaign platform of the state’s leading advocate body for older people, COTA (the Council on the Ageing) proposes free ambulance services for full age pensioners. Nobody wants a cost barrier to people calling an ambulance when they need it, but this proposal is a claim to older people’s privilege (mediated slightly by limiting it to full age pensioners). Age pensioners already get a 50% discount on ambulance cover insurance and ambulance fees, while people who are unemployed (with no children) and those in waged poverty receive no such discounts.

While COTA is simply lobbying for its constituency, a better public policy would be free ambulance services for all South Australians (a universal service). Short of that, the priority should surely be to extend the existing concession to all unemployed people and those in waged poverty, not an extension of an existing subsidy to aged pensioners.

Similarly, COTA’s call for the removal of the work limit for Seniors Card eligibility would see an older full-time worker in a well-paid job receive free public transport and other discounts, based simply on their age. [1] No consideration of income or need, just a claim for older people’s privilege.

Disclaimer – I am particularly grumpy here because these proposals undermine work that SACOSS did to remove unfairness and poverty premiums in the state concessions system – work that was adopted by the SA government in the 2024-25 state budget.

Propaganda in Other Interests

The final example is not an election policy, but a recent Renew Economy article expressing outrage at a proposal to charge all customers a fixed network cost, rather than the usage-based charges currently in place. While fixed charges are usually regressive, this is less the case than the current system where those with money and housing tenure can reduce their chargeable usage (e.g. by installing solar panels and batteries) and avoid paying any network costs.

For reasons outlined in previous posts, I question the data behind the article’s assertion that the fixed charge proposal takes from the poor and gives to the rich. However, in the context of this post, what is interesting is that the poster child of this injustice is a:

“pensioner in a small cottage home, who makes sure they wear their winter woollies to keep their heating bill down, and always takes short showers and who used their small surplus cash to buy solar to contain their power bill.”

This consumer here is not described simply as a person who owns their own home and has capital invested in solar panels and batteries, they are a (deserving) age pensioner. Yet the distributional impacts of energy network tariffs are mostly about technology and housing tenure, not age – but the age pensioner makes an appearance nonetheless as older people’s privilege is weaponised.

Class, not Age

The reality of all these examples (and a thousand others) is that those with a need for government support need that support because they lack income and wealth, not because of their age. Of course, some older people have low incomes because they are marginalised in the labour market by age-discrimination, but in terms of government supports, this could be equally or better dealt with by reference to employment status than age. And as the stats at the top of this article suggest, many other older households have comfortable incomes and wealth.

Age is simply not a good determinant of wealth or of need. As Piketty made famous, the fundamental driver of inequality is capital accumulation. This is true in the big picture of Piketty, but also in the modest capital of owning a home (or solar power).

When we invoke older people’s privilege to continue to provide (or lobby for) supports based on age rather than need, we are making poorly targeted policy, masking the fundamental inequality of class processes, and perpetuating an antiquated view of the deserving and undeserving poor.

Let’s talk about class (income and wealth) and need, not age.

[1]              The SA Senior’s card is currently available to all residents over the age of 60 who work less than 20 hours per week. There are no income limits.

Revisiting Issues of Affordability, Income and Inequality

This post updates and collects in one place my previous writings about how policy arguments around inequality which are based solely on income data (e.g. income percentiles) fundamentally misunderstand and misrepresent inequality.

The basic argument is that standard income deciles/percentiles are misleading because they create a picture of a continuous income distribution spectrum, rather than differential flows of income to certain parts of the economy. More specifically, they ignore fundamental differences in income arising from housing tenure/ownership, income and capital gains on wealth, and social transfers in kind.

In short, such analysis reflects a 1980s world – before housing costs ate household budgets and superannuation turned wage earners into stock holders.

Housing

Housing tenure matters because it creates differences in effective household income (i.e. actual purchasing power) and living standards. In an earlier post I compared the effective income of a renter and homeowner with identical annual salaries. The homeowner (without a mortgage) ended up nearly $30,000 a year better off than the renter on the same $100,000 p.a. income. This result was driven by:

  • differences in housing costs (imputing rental income to homeowners for the value of housing services received)
  • income from investing the cash that would otherwise have gone to rents, and
  • tax advantages that go with that investment.

This comparison did not take account of capital gains which could heighten the gap, and the renter/homeowner difference is probably worse now with rent prices going up faster than income (so proportionately higher imputed income for homeowners) and a booming housing market seeing higher capital gains.

This is all pretty obvious, and echoes why poverty studies tend to focus on “after-housing” income. However, it does suggest that plotting an income spectrum just based on cash incomes is very misleading when it comes to understanding difference in purchasing power and standards of living. At a minimum, we need to be basing analysis on the intersection of income and housing tenure.

Wealth

While housing is the primary form of wealth for most Australian households, the issues above are magnified when all forms of wealth are taken into account. Capital gains and tax advantages are increased, while wealth also creates additional ability to invest in money-saving technologies (e.g. energy efficient devices) which in turn increases future purchasing power without a change in income. And there are financial, health and psychological benefits of having savings/wealth to fall back on in emergencies.

But what is important here is that we can’t simply assume that income and wealth go hand-in-hand. The last ABS data (before the national statistician created an inequality data black-hole), shows that just under a third (32%) of low-income households also had low wealth, but 23% had moderate wealth (probably owning their own home), while 11% of low-income households had high wealth (See the graph below).

Column graph showing percent of households in lowest, middle and high income brackets that have low, moderate and high wealth.
Source: ABS, Household Income and Wealth

A concrete example of this wealth-income divergence emerges from the government’s data on age pensioners. The data for the September Quarter 2025 shows that 72% of pensioners own their own home, and around two-thirds of those homeowner pensioners have more than $100,000 in financial assets beyond their home. These pensioners have low-moderate incomes (otherwise they would not be eligible for the pension), but substantial enough capital to be protected against poverty and to have a better standard of living than many renters on higher incomes.

In short, low income does not necessarily mean low wealth or low purchasing power, and an income spectrum based solely on income figures misleads as to who is likely to be struggling.

Social Transfers in Kind

The final piece of the puzzle would be the inclusion of social transfers in kind – that is, the receipt of public services such as education, public health care, child care subsidies as well as a range of rebates and concessions. Many of these transfers go disproportionately to those on lower incomes, which then increases their effective consumption and standard of living. In turn, this decreases inequality – which was the finding of a leading Australian scholar in this field, Yuvisthi Naidoo, whose work I have summarised here.

However, the analysis is more complicated. A very useful recent briefing paper from the e61 Institute shows that while social transfers in kind are generally progressive (i.e. disproportionately benefit those on lowest incomes), there are significant differences between different transfers. The graph below from their report shows the distribution of transfers across both income and wealth quintiles. We can see, for instance, that pharmaceutical concessions are one of the more progressive transfers when plotted against income, with about two-thirds going to those in the lowest two income quintiles. However, those pharmaceutical benefits are far less progressive when plotted by wealth – in part because older people have more needs and eligibility, and are also likely to have accumulated more wealth (mostly in the form of home ownership).

Bar graph showing the percent of each of 15 different government transfers going to each income quintile. Social/public housing is the most progressive, while community health services and private health insurance rebate are the least progressive.
Source: e61 Micronote: Welfare for the Well Off?

It is worth tracking the comparison of progressivity in this graph for each transfer, and there is further discussion below on energy concessions, but the main point here is simply that inequality looks different when wealth and social transfers in kind are considered.

Why Does It Matter?

Overall, all this matters because it means that the level of inequality we see in standard income spectrums may be misleading, but also because actual households will be in different places on the income spectrum when extended incomes are taken into account. Naidoo’s research showed that nearly a half of all older people (65+) were in the lowest standard (money) income quintile, but the inclusion of imputed rent and social transfers in kind reduced that to 22.5% (because older Australians are disproportionately more likely to own homes and benefit from health services). By the time imputed wealth annuities were included in the analysis, only 17% of older people were in the lowest income quintile. At the other end of the spectrum, accounting for extended income meant that 26% of older households were in the highest quintile, up from 7.1% when based on standard income alone. (Naidoo, Appendix Tables C8-11) .

These issues have very direct implications for policy fairness. My attention was recently drawn to this in relation to energy affordability, where there is a legitimate concern to alleviate and avoid energy costs for low-income households. Obviously we don’t want households to go without power, or be bankrupted by power bills, but targeting energy assistance to those on low incomes may be poor targeting. Worse, a focus simply on income might mean imposing more network and other costs on those least able to pay.

Consider the pensioner households noted above. The nearly one-half of age pensioners who own their own home and have more than $100,000 in financial assets can easily afford solar power and energy saving technologies (if they have not got them already). They are far less likely to be facing energy hardship than renters on the age pension without access to the same technologies (who, incidentally, would be seen to have a higher income due to receipt of Commonwealth Rent Assistance). This is important because both pensioner households would receive the same energy concessions (at least in states like SA where the concession is a flat rate) because concession eligibility is based on income rather than ability to pay.

Further, those homeowner pensioners are also far less likely to be in energy hardship than renter families in waged poverty, yet the age pensioner will get an energy concession while those in waged poverty may not qualify. This is a different type of income fetishism (based on income type rather than quantum), but again we see income as an unreliable indicator of affordability and need for support.

More broadly, we can see in the e61 graph above that energy concessions are more progressive by income than by wealth. Nearly half of all concessions go to those in the lowest income quintiles, but only around a quarter go to those in lowest wealth quintile.

There are lots more intricate issues around who bears (and should bear) the necessary costs of the energy transition and how network costs are paid for (apportioned between customers). But what is clear is that a distributional analysis of energy costs based on a simple income spectrum would be misleading in terms of both ability to pay (income) and access to energy-saving technology (cost).

The Way Forward

Energy is just one area where there is a need for a far more sophisticated analysis of income inequality. We need an analysis of affordability for a range of essential expenditures that takes account of housing tenure, but also extended incomes and the real ability to pay for essential consumption.

Ultimately what I would like to see is, firstly, for the ABS to get themselves resourced and organised to do another Household Expenditure Survey (the last one was 2025-16!), and then to be able to analyse those expenditures based on an extended income spectrum combining wealth, housing and income. Only then will we really know which expenditures are genuinely regressive (have disproportionately highest impact on those with the least ability to pay) and where and how to target support.

In the meantime, caution and an analysis based on housing tenure is advised.

Gender Equality @ Work – South Australian Data Questions

I have previously highlighted the fact that the gender wage gap in South Australia was (in 2022) significantly lower than the national gap, primarily because of the relatively lower average male pay in SA. I have not tracked the gaps since they, but was made curious again by the recent launch of the Gender Equality @ Work Index Report and a talk I heard from one of its lead authors, Professor Rae Cooper.

The Gender Equality @ Work Index

The Index is much broader than simply the gender wage gap and covers 7 dimensions of workplace gender inequality:

  • Participation – labour force participation and labour utilisation rates
  • Pay – hourly and average weekly pay gap
  • Hours worked – paid and unpaid
  • Stratification – vertical distribution of men and women across workplace levels (including senior executive representation)
  • Segmentation – horizontal concentration of men and women in different industries and jobs
  • Job security & access to leave entitlements
  • Safety – injury rates and sexual harassment at work

I like a good index (hence SACOSS’ almost famous Vegemite Index!), and the Gender Equality @ Work Index is particularly good because of its unique combination of those 7 dimensions (underpinned by 16 different data sets). The result is an index that brings together fairly mainstream but disparate data sources into a snapshot of gender equality at work, but one which also tracks changes over time in key factors driving the broader outcomes.

For instance, as the summary graph below shows, overall gender equality at work has improved over the last 10 years, but progress has been uneven. The biggest improvement has been in stratification (more women entering management and senior positions), while safety has gone backwards and, despite some improvement, segmentation remains a significant problem with the lowest scores in the index.

Gender Equality @ Work Index 2014-2024 Line graph showing trajectory of each of the 7 elements of the index.
Source: Gender Equality @ Work Index Report. Note: the index scores each dimension out of 100, with 100 representing equality and numbers below 100 showing the extend of the gap between men and women.

For me, the inclusion of safety was one of the most important things in the index, while the relative quantification of gender segmentation was particularly useful. I have long heard that Australia had one of the most gender-segmented workforces in the OECD, but the index clearly shows the importance of this in driving overall inequality.

Beyond that, I will let the Index speak for itself (it is an easy read!), although for reasons outlined here, I would have liked to see the gender wage share included in the pay measure. However, indexes are predominantly about changes over time rather than the nuances of their construction, so it is no big issue. And the data did make me go back to the South Australian data, because the Index is only presented at the national level and I wondered if there might be geographical differences.

South Australian Data

I only looked at SA data for a couple of parts of the Gender Equality @ Work Index, participation and pay, and I used slightly different and more limited ABS data sets than those used in the index. In my data:

  • the participation figures are simply the ABS labour market participation rate (without the underutilisation calculation in the Index), and
  • the pay rate is average weekly earnings – total earnings (a different indicator, and without the separate accounting for differences in hourly rates in the Index).

These differences don’t matter much as I was not trying to replicate the Index, but rather to query whether there might be differences in state and national level data.

The two graphs below show the comparison of the national and South Australian data for labour force participation and pay respectively. To match the Index, the gender gap is expressed as a ratio of female to male participation/remuneration, and while there is much greater volatility in the South Australian numbers, there are some clear differences in the national and South Australian trajectories.

In both participation and pay, the national data shows a steady improvement in the female-male ratios over the last ten years. That is, there has been a relatively steady increase in equality in these two dimensions of gendered workplace equality. However, the South Australian data show that momentum towards greater equality here has largely stalled since COVID (2020).

Line graph showing differences in female and male labour force participation rates in SA and nationally from 2014 to 2025.
Line graph comparing gender wage gap (expressed as a ratio of female to male average weekly earnings) in SA and nationally from 2014-2024.

As evident in the first graph above, the participation ratio in South Australia was higher than the national level for much of the pre-COVID period (due partly to declines in the male participation rate in SA from 2014 to early 2017, and again from early 2019 until the COVID recovery). Significantly though, female participation in the SA labour force has not grown much from pre-COVID numbers, and included a large drop in the first half of 2024, hence the ratio in the post-COVID years is mostly lower than the national average.

Similarly, as evident in the second graph, with the exception of the 2024-25 financial year, the pre-COVID improvements in the gender pay ratio in South Australia stalled after COVID. Between November 2020 and May 2024, average weekly earnings for women in South Australia increased by 16.9% and by 16.5% for men (thus basically maintaining the gender pay gap). By comparison, the national figures were a 18.7% earnings increase for women and 12.9% for men, narrowing the gap and increasing the female-male wage ratio.

Conclusion

I don’t have the expertise or resources to replicate the whole Gender Equality @ Work Index for South Australia, but it would seem that there are some potentially significant differences at the state level. This is not to deny the validity or usefulness of the national index and data. However, if the drivers of inequality play out even a little differently in some states, potentially because of different legislation, policy or industry structures, then the Index creators may want to consider state-based indexes as a future expansion of the project.

Of course, that would add complexity in both data analysis and presentation, and one of the strengths of the Index is that it presents key data on complex issues in an accessible way. So there would be pluses and minuses to incorporating geographic differences in the analysis, but even without state data, the Gender Equality @ Work Index is useful as both an educational tool and a policy starting point in addressing workplace inequality.

Compare the Pair: Income Inequality on the Same Income

Can a homeowner really be $30,000 a year better off than a renter with the same income?

We are all familiar with superannuation ads asking us to “compare the pair”: two otherwise similar workers in different superannuation funds getting very different financial outcomes. Today I want to do a similar exercise for the extended incomes of two people on identical wages, with the only substantive different in their lives being that one owns their own home, while the other one rents.

The different financial outcomes that result go beyond just the amount of rent that one of the pair pays. Nor are they based on lifestyle or consumption differences, clever investment strategies, complex tax minimisation planning (beyond a basic voluntary superannuation contribution), or any other scheme beloved of financial planners. As such, and with everything else about their lives the same, the compare the pair exercise enables us to draw some interesting conclusions about the role of housing and taxation in inequality.

Inequality - not what you think

Compare the Pair

Renter-Greg and homeowner-occupier-Greg both work as professionals in suburban Adelaide and earned $100,000 last financial year. They both live alone, but are neighbours in the same set of home units. Renter-Greg pays $350 a week in rent, while home-owner-Greg is an owner-occupier who paid off their mortgage six years ago. Ever since paying off the mortgage, homeowner-Greg has been making voluntary superannuation contributions equivalent to renter-Greg’s weekly rent – leaving both with the same weekly consumption expenditure (neither saves or invests any further or has any other source of income).

You get the picture: a standard simplified model where all other factors are equal so any differences in financial outcomes are only the result of different housing costs and the ability of homeowner-Greg to put money into superannuation. Of course, the ability to buy a home and invest in super may be conditioned by all sorts of social factors, but the model could also simply be viewed as a comparison of potential financial outcomes of the same person with different housing tenure.

As the table below shows, the bottom line is that homeowner-Greg is nearly $30,000 a year better off than renter-Greg, despite them having the same employment income.

TABLE 1: Comparing the Pair, 2023-23

Compare the Pair: Table showing renter and homeowner comparison, with homeowner $3k better off after tax, then earning $18,200 in imputed rental income, and $8,565 in return on extra superannuation contributions.

(a) Tax is calculated using the ATO simple tax calculator, with homeowner tax based on income tax on $81,800 salary and 15% of $18,200 voluntary super contribution.

(b) Income from voluntary super contributions does not include the voluntary contributions themselves, only the income on the accumulated balance of these contributions, calculated using an industry average rate of return (9.2% for the 2022-23 year, and 5.8% for each preceding year).

Explanation

The first step in comparing the pair is simple enough and is just based on salary income and tax. As can be seen above, homeowner-Greg has a higher disposable income because of the tax concessions on voluntary superannuation contributions.

When we include housing costs, the difference is starker. As noted in a previous post, accounting for housing-costs in income comparisons is common in poverty research (which is usually based on after-housing incomes), while the national accounts also recognise the value of housing ownership by including a value for rent that owner-occupiers are deemed to pay themselves in their measure of the size of the economy. The value of this imputed rent is added to homeowner-Greg’s income (because the value of this free housing service is income-in-kind). The result is a further increase in income inequality, with homeowner-Greg’s after-housing income being 28% higher than renter-Greg’s income.

The next step is to account for the extra income homeowner-Greg receives from being able to make a voluntary superannuation contribution. When this investment income is included, homeowner-Greg’s extended income is nearly $30,000 higher in the year than renter-Greg’s. Again, this is based solely on home ownership and the ability to invest the money saved on housing costs in superannuation.

Alternatives

A significant part of the difference in financial outcomes above comes from the superannuation investment. If homeowner-Greg just put the savings into a standard bank account, they would not get the benefit of the tax concession and would have lower returns on accumulated savings. However, with a relatively modest interest (2% ->3%), they would still get some $2,600 more income than renter-Greg in the 2022-23 financial year.

Of course, homeowner-Greg could also just spend the money not going on rent rather than invest it, which would mean less difference in the long-term, but significantly higher weekly consumption and standard of living than renter-Greg on the same income.

There is one final calculation that could be made to include the value of the capital-gain on the house in homeowner-Greg’s annual income. Home unit prices in Greg’s neighbourhood increased by 8.8% in 2022-23, so homeowner Greg would have “earned” $39,600 in capital gains on his $450,000 unit (average unit price). That would bring the difference between renter-Greg and homeowner-Greg’s income to $68,650, or more than two-thirds of their starting income (gross salary). I have argued previously that capital gains are an important driver of inequality which are overlooked in most data, but there is also an argument to exclude capital gains for owner-occupied housing. This is because unlike pure financial investments, where capital gains are a key return, owner-occupied homes are not simply an investment product. They fulfill a basic need (housing), and while a capital gain may be realised upon sale, most people selling are buying another house in the same inflated market – so they are really simply swapping one house for another, not gaining wealth through the capital gain.

The arguments here are complex, and excluding capital gains on the owner-occupied residence probably underestimates the differences between renters and homeowners. However, the compare-the-pair data above shows that even on the conservative estimate, homeowner-Greg’s extended income is 39% higher than renter-Greg, despite doing a similar job for the same pay.

Conclusion and Implications for Analysis

The most obvious conclusion from this data is that it pays, literally, to be a homeowner. However, it is also important to note that the capital income from accumulated voluntary super payments, and therefore the extended income differences between home-owners and renters, will grow over time as investment income increases with capital gains and further contributions.

The compare the pair exercise also has broader political economy implications. It provides further evidence that the standard statistics on income inequality which deal only with money income hide significant inequalities between people/households who appear to have similar incomes.

Further it supports the argument put by Lisa Adkins and others that capital gains and capital income, rather than employment income (which in the Gregs’ case is identical) are the preeminent drivers of inequality.

It is also evident in the comparison above that the tax system is failing its redistribution function. In this case the tax system not only does not tax aspects of tax capital income, the concessional tax arrangements applying to superannuation promote the inequality between these two people on the same money income.

Implications for Advocacy

Finally, this compare-the-pair exercise raises questions for my own work at SACOSS and the approach of many anti-poverty advocates who have traditionally focused on championing rental affordability and renters’ rights. This focus undoubtedly supports those likely to be most disadvantaged in the housing market. However, the data above shows that the fact of them renting puts renters on the wrong side of increasing inequality – which might suggest merit in the traditional conservative focus of getting people into home ownership.

Put another way, an anti-poverty agenda would direct policy (and money) to supporting rental affordability, while an equality agenda might direct policy and resources to enabling more people to own their own home.

Of course this is a false dichotomy. Policies such as increasing Commonwealth Rent Assistance address both poverty and inequality issues, while some people will never be able to afford to buy a house so “the Australian dream” policy options are limited (and expensive).

However, the analysis does highlight the fact that even if we secure tenants’ rights so we become Europeanised and life-long renting is becomes a potentially desired option (rather than a forced option), renters will still be disadvantaged. Perhaps most challenging, this analysis also applies to those in public housing, who will be better off than they would be in the private rental market, but will nonetheless be falling behind homeowners on similar incomes.

Clearly, while current policy directions to increase renters’ rights and the provision of rental housing are absolutely necessary, we also need to change tax and other policies to reduce the difference in financial outcomes between renters and homeowners.

Debt, Interest Payments and Choices: A Surprise in the SA State Budget

As we make our way through the first month of a new financial year, political economic debates are dominated by inflation, interest rates and “full employment”. But the discussion of interest rates is generally focused on their impact on inflation at the macro-level, and on mortgages and cost of living at the micro-level. The impact of interest payments on government budgets is usually noted in budget night commentary on the deficit/surplus scorecard, but promptly forgotten for the rest of the year.

However, an unusual occurrence (at least in recent times) in the South Australian state budget should serve to illustrate why interest rates matter to government and to the community. In 2023-24 the amount of government expenditure going to servicing state debt ($1,254m) will eclipse the budget for the Department of Human Services (DHS) ($1,148m). This difference grows over the forward estimates with the DHS budget subject to real cuts (low indexation and older operational savings) while interest payments increase substantially. By 2026-27, interest payments are predicted to be $1,684m, by comparison with a DHS budget of just $1,233m.

Column graph comparing SA State Budget expenditure showing interest payments increasing from 2022-23 to 2026, while DHS expenditure remains stable.

In practice, this means that we are spending more on debt repayments than we are on the Department that is the primary provider of support services for the most vulnerable and disadvantaged people in our state. That feels wrong – but apart from the shock value, does it really matter, or it is just a statistical coincidence with no budget or social impact?

A Debt Problem?

As I pointed out in SACOSS’ post-budget analysis, government debt and deficits are not necessarily a problem – and may represent important economic stimulus or long-term investment. And there is no suggestion that this level of debt in unsustainable, although the budget papers show that a 1 percent point increase interest rates in 2023-24 would equate to an extra $203m in service payments. That would obviously be a significant imposition on the budget, but even with the debt-to-income ratio rising, the government can clearly still maintain payments. However, if debt is unchecked or interest rates continue to rise, at some point interest payments either become unsustainable, or more likely, a significant constraint on budget spending in other areas.

In that sense, the comparison of interest payments and the DHS budget is important because it reminds us of the potential impact and opportunity cost of state debt. I am not suggesting that if we were not paying that money in interest, we would be spending it on human services. That would be wishful thinking! However, all government spending has distributional impacts, and interest payments are no different. The comparison with DHS expenditure simply serves to focus the interest payment discussion on inequality.

Debt, Interest Payments and Inequality

In providing concessions and emergency supports to those on low incomes, and funding charities to provide a range of other supports, the Department of Human Services functions to transfer money and resources from the budget to those most in need. By contrast, interest payments are a transfer from the budget to government bond holders – who, by definition, are those with excess cash to afford to lend money to the government (by buying bonds).

As Piketty has pointed out, (and [as ever] some of the concerns here come from my reading of his work) it is far more advantageous to those with capital to have public deficits and receive interest on their money than to have that capital taxed to balance the budget (Capital in the Twenty-first Century, p.130). So, in creating government debt we have already chosen to favour those with capital by borrowing rather than taxing their money and creating an ongoing flow to rather than from that capital.

Seen in this light, the contrast between the DHS budget and the amount going to debt servicing is an indicator of choices about government priorities – the choice to provide relatively less to the poorest in society (via government expenditure) than to those who are better off (via taxes foregone and interest paid out).

Caveats

Of course, as with everything in economics, it is not as simple as that. The initial use of borrowed capital may be used for things which support those on low incomes, and inflation may quite separately have a counter-balancing impact by undermining the real value of bonds and the interest payable on them. Further, bond-holders may not be South Australian residents and may therefore be outside the tax ambit of the state government. In that sense the taxing v borrowing from capital argument above is not about individual bondholders. Rather it is illustrative of the options of government in general and operates at the level of class: that is, the state government has options to tax local capital rather than borrow from capital in a national or global market.

To the extent that we can look at individual bondholders, it is also worth noting that, as Piketty points out, they are no longer necessarily the wealthiest people in society (as the super-rich can invest more lucratively elsewhere). ABS wealth statistics do not record bonds as a separate category, so it is hard to confirm this. However, it is clear that the low-inflation era of the first years of this century made bonds a safe investment for middle class superannuation and investment funds, so when we talk about bondholders it is likely we are talking about middle and above-average income earners (often via superannuation or investment trusts). In that sense, the DHS/interest payment comparison is not so much about rich vs poor, but about a form of middle-class welfare instead of a transfer to those in most need of the supports that a DHS might offer them.

Finally, there is another (non-taxation) route to balancing the budget and avoiding interest flows to capital owners. That is by cutting expenditure. However, cuts to expenditure (and therefore to services) usually impact disproportionately on the poorest people. Those with the fewest economic resources are likely to be most reliant on support services and have limited or no alternative options, so expenditure cuts usually also impact most on the poor. Further, it is notable that in this SA budget, government debt continues to increase even with operational surpluses from 2023-24 onwards, so balancing the budget is not simply done by cutting expenditure. However, there is little doubt that when government services are cut (the “austerity approach”) this exacerbates inequality, so if fairness or equality is a consideration in balancing the budget, we must ultimately return to revenue issues – and our preference for borrowing rather than taxing capital.

Why Debt and Interest Payments Matter

Budget deficits and surpluses matter, but not for the reasons often touted in economic commentary (“responsible government”, “living within our means”). They matter because they determine the level of debt, which in turn, within any given interest rate regime and revenue base, impacts on the money available to spend on services. And as interest rates increase, so too does the cost of servicing government debt. The choices made to borrow rather than tax capital increasingly manifest in a distribution of government revenue to the middle and upper middle classes rather than to those on the lowest incomes who would benefit most from government service provision.

The fact that the SA state government interest payments now eclipse spending on the Department of Human Services should make us think about government priorities and the need for a stronger tax base.