I delivered a short presentation recently at an event organised by Victorian Progress on the topic ‘Middle class welfare: What’s all the fuss about’. I spoke alongside David Hetherington of Per Capita and John Roskam of the IPA, and we were introduced by John Brumby. I don’t think it was recorded, and I didn’t read my remarks, but I thought I’d try and summarise my key points here, along with the slides I used on the night.


My aim in giving this presentation was to challenge what I see as the two elements of the conventional wisdom around the Australian welfare system. The first is that our welfare system is too large and is growing rapidly as a share of the economy. The second is that too much of the money we do spend goes to middle- and upper-income households. I show that neither of these is true, if you take ‘welfare’ to mean ‘cash benefits’, ie. direct transfers from government to households.


I think most people have cash transfers in mind when they refer to ‘welfare’, so that’s where I devote most of my attention in addressing the question of middle class welfare. But government also provides benefits to households through ‘in kind’ assistance (like public schools and hospitals), as well as through tax concessions or exemptions that favour particular types of activity, so I briefly address those as well.


Cash benefits in Australia represented 8.1% of our GDP in 2012, compared to an average across the OECD economies of 12.6%. Our cash benefits system is smaller, as a share of the economy, than in any other English-speaking country, including the US (9.7%). The high unemployment rates in the US and much of Europe does mean that their spending on benefits was a bit higher than usual in 2012, but our spending was still lower before the financial crisis. In 2005, we spent 7.8% of GDP on cash benefits; the OECD average was 11.3% and the US spent 7.9%.

This, to me, is a striking rebuttal to the conventional view that we have a bloated benefits system that takes an unsustainably large chunk of our national income.


The second part of the conventional view is that too much of our benefits spending goes to households in middle- and upper-income households. In fact, we ‘target’ our spending much more tightly than any other advanced economy. This chart shows the ratio of cash transfers paid to the poorest 20% of households relative to those paid to the richest 20%. In Australia, the poorest get 12.4 times as much as the richest; the OECD average is 2.1. These figures are from the mid-2000s; the latest ABS data suggests that the ratio has risen to 14.7, so our benefits system became more tightly targeted in the late 2000s.

You can see that nearly half of the advanced economies have a ratio of less than 1. This means that the richest households received more in cash benefits than the poorest households. This reflects a very different approach to welfare design – one that emphasises the ‘piggy bank’ objective of redistributing over the lifecycle rather than redistributing between the rich and poor.

We have less middle class or upper class welfare than any other advanced economy.


Here’s what the distribution of benefits across the income levels looks like. People in the second quintile receive nearly as much as the poorest 20%, on average, but a lot of this is the age pension.


(Note: unlike the previous slides, we’re now looking at benefits for working-age households)

For the past three decades, somewhere between 70 and 80% of benefits have gone to people in the poorest 30% of households. This figure was pretty stable at around 75% (give or take a few percentage points) from the mid-90s to 2003-04, with a dip in the couple of years after that. That was the period in the wake of the 2004 election, when the Howard Government created a range of new payments. A lot of these have since been abolished or means tested.

So if 70% to 75% of benefit spending goes to the poorest 30% of households, who gets the rest?


‘Upper-class welfare’, meaning transfers to the top 20% of households, has been low and fairly stable for the past few decades. Transfers to people above the median, but below the 80th percentile, was fairly stable at between 6% and 8% during the 1980s, 90s, and early 2000s, before this group benefited from the cash splash in the wake of the 2004 election. You can see that their share of spending had already fallen back a little by 2007-08, and I suspect it has fallen further since then.

Over the past few decades there has been a fairly steady rise in the share of spending going to people in households below the median, but above the 30th percentile. This has happened for a number of reasons – as Whiteford, Redmond and Adamson put it in the excellent paper from which these figures are derived,

In the Hawke-Keating period, it was the third and fourth deciles that saw
the largest increases in their share of transfers. This could be expected as the Prices
and Incomes Accord of the period was designed to use improvements in the ‘social
wage’, including higher income support for families, as a trade-off for wage restraint,
with the main beneficiaries being lower-paid working families.

The past few decades have seen a transition away from the “wage earners’ welfare state” model in which centralised wages fixation was used as a primary tool of social protection, and full-time wages were set on the basis of a family’s needs, to one in which the minimum wage is lower relative to the median and support for lower income families is delivered through government income support payments. This is part of the reason for the rise in the share of transfers to people between the 30th percentile and the median. Although their share has risen, remember that over 70% of transfers still go to the bottom 30%.

At this point, some people might say “ok, we might have less midde-class welfare than other countries, but we still have some. Why not get rid of it entirely?” To explain why I think that would be a bad idea. I’d like to make use of a version of the stylised example that I set out in a previous post.


Imagine we have an unemployment benefit that’s worth a maximum of $300 a week. As you start to earn income from work, you lose the unemployment benefit at the rate of 50c in the dollar. This means that by the time you are earning $600 a week, you don’t get any unemployment benefit.

Now say that someone came along and wanted to tighten this payment. It’s needless middle-class welfare, they might say, to have people on a little less than $600 a week who are still eligible to receive the payment. Let’s say that they wanted to change the payment structure so that it cut out at $500 a week rather than $600.

There are two ways you could do that. Both of them involve potentially damaging trade-offs.


The first option, which I’ve labelled here as Scenario B, would involve maintaining the $300 a week maximum payment, but reducing the benefit at a rate of 60c in the dollar as people earn an income. This would mean that the poorest recipients still get the same amount, but people on the benefit who go out and work get to keep less out of every extra dollar that they earn. If they earn an extra dollar, they’ll pay tax on it, and lose 60c of their benefit. This means that under Scenario B, people face higher ‘effective marginal tax rates’ than in the original scenario. The incentive to work will be lower, and it will be harder for people to get ahead.

The second option, Scenario C, would keep the 50c taper, so as not to increase effective marginal tax rates, but would cut the maximum payment rate to $250 a week. This option means that everyone on the payment now receives less money, including the poorest beneficiaries.


We started off with Scenario A, which sees people on up to $600 a week receiving some benefit. This is less targeted than the other two options. But to target the payment more tightly means either reducing work incentives (raising EMTRs) or reducing the adequacy of the payment for everyone. Both of these trade-offs involve real costs. Given that we already have the most tightly means tested cash payments system in the developed world, I don’t think there’s much to be gained by tightening the screws any further at the expense of adequacy or work incentives.

So far I’ve only been talking about cash benefits. But, of course, that’s only one part of the system.


Governments also provide benefits to households by directly funding services. We don’t means test access to public education, or public hospitals, so the distribution of benefits from those things is much flatter than with cash benefits. This slide shows the ABS estimate of the monetary value of some of these services that households receive.

This is a big part of the “tax-welfare churn” that the conservative Centre for Independent Studies complains about. Here, the churn involves government taxing middle-income households and then effectively giving the funds back to them in the form of heart surgery or school education for their children. I think there are pretty solid reasons to pool risks and have services from which most people derive a benefit. I don’t think public hospitals or schools are what critics generally have in mind when they’re criticising middle class welfare.

I also think that if we’re going to talk about the distribution of benefits from social services, we need to also include the distribution of benefits from other services as well. Who benefits most from public roads, or from police services or the defence force? The rule of law and the courts? These questions are no less relevant to the question of tax-spending ‘churn’ than the distribution of social services.


The third and final element of ‘welfare’ to consider is tax expenditures. These are concessions or exemptions from tax that benefit particular types of activity over others. For the beneficiaries, the net effect is not different to receiving a payment from government. They also reduce revenue, so they come at a cost to government. This is why they’re called tax expenditures.

The biggest single area of tax expenditures relates to superannuation. Treasury estimates that these concessions are worth around $32 billion this financial year, and will rise to around $45 billion by 2015-16. Concessions and exemptions associated with housing are the next biggest form of tax expenditure, followed by a grab bag of other measures, like the exemption of food from the GST.

A lot of these tax expenditures exist for good reasons, and I wouldn’t want to see them abolished. But, unlike the cash benefits system, I think there is potential to reduce a range of these expenditures in a way that would raise revenue and improve equity at the same time. You could use that revenue either to cut taxes or increase service provision, whichever your preference.


(This slide depicts the situation without the low-income contribution, which the Coalition has pledged to repeal.)

One example is superannuation contributions tax. This is a 15% flat tax. This means that a low paid worker would pay more on their super contributions than they would if they received that same money as regular income – they have a negative concession. The typical full-time worker would pay 34% on an extra dollar earned (including Medicare Levy), but 15% if that dollar went into super – that’s a concession of 19 percentage points. High income earners get a concession of over 30 percentage points.

The Government has improved one end of the scale with the low-income contribution, which effectively refunds the tax that low-income people receive on their super. It remains a flat tax for everyone else, although plans have been announced to increase the super tax rate for people on $300 000 or more. It would still be flat in between.

I’d favour a situation in which your super contributions were taxed at your marginal rate, minus a fixed percentage point concession (say 20%). They concessional treatment would still be there, but the benefits of the concession would no longer skew so heavily towards high-income earners.

When you include tax expenditures on things like super or negative gearing, I think there is room for improvement.


But as we saw a few months ago during the debate around taxes on super, any suggestion of reducing or removing tax expenditures is met with cries of ‘class war’. I think the critics of Australia’s welfare system don’t have tax expenditures in mind, but rather our payments system. Here I think they’re misguided or disingenuous.