This post is the second of a four-part series on the topic.
If we turn to the policy or academic literature, we find that in 2005, the Australian Treasury’s Economic Roundup included an article on Net tax thresholds for Australian families (Bremner, 2005). The article defined net tax thresholds as the amount of private income a family (including single people and couples without children as well as those with children) could have before the taxes they paid were higher than the cash benefits they received. The concept of a “net tax threshold” is the same as that underlying the categorisation of some people as “zero net taxpayers” applying the second measure identified above. “Zero net taxpayers” are those at or below their net tax thresholds, and positive net taxpayers are those with incomes above these levels. This can be illustrated with a simple example involving just income tax and cash benefits for a single person of working age.
Example
Scott is a single 30 year old who has no private income and is actively looking for work. Scott is likely to be entitled to the unemployment cash benefit, Newstart Allowance, which is currently paid at a rate of $527.60 per fortnight or roughly $13,755 over the course of a year. Newstart Allowance has an “income free area” of $104 per fortnight (so that any income Scott earns up to that level does not reduce his payment). However, if Scott earns more than $104 per fortnight, Newstart Allowance is reduced by 50 cents for each dollar over $104, up to a fortnightly income of $254 per fortnight. Above $254 per fortnight or roughly $6,600 per year, Newstart Allowance is reduced by 60 cents in the dollar until the Newstart payment is zero, at a private income of $1,023 per fortnight or roughly $26,670 per year.
What about income tax? Newstart Allowance is taxable and so the normal individual marginal tax rates apply to the Newstart Allowance and to any other income earned by Scott. The current tax-free threshold for an individual is $18,200, well above the level of Newstart. In addition, the Low Income Tax Offset (LITO) has a maximum value of $445, which means that a low income individual effectively does not start to pay income tax until their income is over $20,540 per year.
Scott faces the first marginal tax rate of 19 cents in the dollar for income over that threshold. At this level, Scott’s Newstart Allowance is also being reduced at 60 cents in the dollar. This means that Scott pays income tax at a 19 % rate on 40 cents in each Newstart dollar remaining. (There is also a Beneficiary Tax Offset of 15% of any income from Newstart greater than $6,000 per year, but this stops being relevant at a disposable income a little above $22,000 per year.)
In conclusion, for every $100 per fortnight earned by Scott above $20,540 per year, he loses $60 in Newstart benefits and pays $7.60 in tax. The “net tax threshold” for Scott, at which his remaining Newstart Allowance is equal to his income tax is at a disposable income of just under $950 per fortnight, or roughly $25,000 per year. Above this point, Scott becomes a “net taxpayer” (disregarding supplementary cash benefits and any non-cash benefits and indirect taxes).
The example of a single person receiving Newstart is the least complicated case that can be illustrated. It shows that the Australian system of cash benefits is similar to Friedman’s concept of a negative income tax, in that it requires no contributions and simply withdraws payments as other incomes increase, albeit in this example at three differing rates (0, 50% and 60%) rather than a single withdrawal rate – and with many conditions attached, the most important of which is that people fall into specified categories (e.g. unemployed, with a disability, caring).
The overall system is more complicated than this because many cash benefits are designed with different payment levels, thresholds and taper rates, which can also differ by whether people are singles or living in couples. Net tax thresholds will be higher for those benefit recipients who are renting privately and receiving non-taxable Rent Assistance, and for those receiving higher rates of basic benefits such as Age Pension or Disability Support Pension (which is non-taxable) or Parenting Payment Single. Families with children who are receiving non-taxable Family Tax Benefits will also enjoy higher net tax thresholds.
The discussion above refers to benefits level, withdrawal rates and cut-out points. It is important to understand the simple arithmetic of any form of income-tested social security programme. A US political scientist, Theodore Marmor, pointed out in 1969 in a discussion of how to assess income maintenance alternatives that there are three parameters in all income tested social security schemes, which are (1) the guarantee (benefit) level (2) the tax (withdrawal) rate, and (3) the cut-off point. For example, a programme guaranteeing $2500 when earnings are zero with a “tax” rate of 50 percent entails a cut- off point of $5000. If one specifies any two of the three variables, the value of the third variable is determined. As the Henry Review of Australia’s Future Tax System pointed out more recently, with any income support payment there is an “iron triangle” associated with means testing. “The generosity of the payment (including the breadth of its coverage) needs to be balanced by how much it costs taxpayers, and the incentive for people to get off the payment by earning income. Improving one of these worsens one or both of the others.”
It should also be emphasised that in the example of Scott, a single person on Newstart, the net tax threshold falls at an income between the actual income tax threshold (including the effects of the Low Income Tax Offset) and the cut-out point for benefits. By definition, the net tax threshold can’t be higher than the benefit cut-out point since at this point a person is no longer receiving payments to offset their tax liability. The same logic applies to any other tax and transfer system, a point to which the discussion will return in looking at net tax thresholds across OECD countries.
Following the 2005 Treasury article, net tax thresholds were also calculated by the Treasury and presented in subsequent Budget Papers together with information on trends in the real disposable incomes of different types of “cameo” families. This information was published in the Budget Overview as Appendix C ‑ Helping households with the cost of living, up until the 2013-14 Budget.
However, net tax thresholds are not included in Budget Papers of 2014-15 or 2015-16 under the Coalition Government. One may speculate that this is because these recent Budgets would have delivered reductions in real disposable incomes for many household types, as a result of proposed changes in social security measures.
Table 2 shows trends in net tax thresholds for different types of households between 1996-97 and 2007-08 (the period of the Howard Liberal/National Coalition Government) and between 2007-08 and 2013-14 (the period of the Rudd/Gillard/Rudd Labor Government). The end point is a projection made in the final Labor Budget.
Table 2: Real net tax thresholds, 1996-97 to 2007-08 and 2007-08 to 2013-14
| Household type |
Real net tax threshold ($2007‑08) |
Real net tax threshold ($2012‑13) |
||||
| (wage as a percentage of the average wage) | ||||||
| 1996-97 | 2007-08 | % change | 2007‑08 | 2013‑14 | % change | |
| Single person (67%) | $15,801 | $18,300 | 15.8 | $20,940 | $23,359 | 11.6 |
| Single person (100%) | $15,801 | $18,300 | 15.8 | $20,940 | $23,359 | 11.6 |
| Single person (167%) | $15,801 | $18,300 | 15.8 | $20,940 | $23,359 | 11.6 |
| Sole parent (0%) | $36,625 | $50,813 | 39.1 | $57,870 | $60,816 | 5.1 |
| Sole parent (67%) | $36,625 | $50,813 | 39.1 | $57,870 | $60,816 | 5.1 |
| One income couple (133%) | $17,026 | $31,951 | 87.7 | $37,598 | $38,960 | 3.6 |
| One income couple (167%) | $17,026 | $31,951 | 87.7 | $37,598 | $38,960 | 3.6 |
| Dual income couple (100 & 33%) | $20,607 | $34,113 | 65.5 | $38,939 | $41,281 | 6.0 |
| Dual income couple (100 & 67%) | $23,973 | $34,227 | 42.8 | $39,070 | $43,036 | 10.2 |
| One income couple with children (100%) | $35,920 | $50,813 | 41.5 | $57,870 | $60,816 | 5.1 |
| One income couple with children (133%) | $35,920 | $50,813 | 41.5 | $57,870 | $60,816 | 5.1 |
| One income couple with children (167%) | $35,920 | $50,813 | 41.5 | $57,870 | $60,816 | 5.1 |
| Dual income couple with children (100 & 33%) | $36,584 | $55,340 | 51.3 | $68,749 | $75,338 | 9.6 |
| Dual income couple with children (100 & 67%) | $36,688 | $54,820 | 49.4 | $72,947 | $82,894 | 13.6 |
| Dual income couple with children (167 & 100%) | $36,671 | $54,850 | 49.6 | $72,152 | $81,753 | 13.3 |
| Single pensioner (0%) | $21,389 | $30,444 | 42.3 | $31,450 | $34,465 | 9.6 |
| Pensioner couple (0 & 0%) | $35,879 | $53,031 | 47.8 | $51,861 | $61,049 | 17.7 |
Source: Budget Overview, Appendix C ‑ Helping households with the cost of living, 2007-08 and 2013-14.
For many types of households, net tax thresholds are the same irrespective of earnings level, although for couples they can vary depending on the split of income earned by each member of the couple and net tax thresholds are higher for pensioners than for people at younger ages.
Real net tax thresholds grew quite significantly in both periods. For single working-age people, the annual rate of change was somewhat greater under the Labor Government. However, for all other types of household, the increase was much greater under the Coalition Government. For example, for single income couples without children real net tax thresholds grew by nearly 90% – or 8% per year – in the first period, compared to 0.6% per year in the second period.
The Treasury article pointed out that changes in net tax thresholds reflected changes in the level of family payments, reductions in marginal tax rates, increases in the basic and higher tax thresholds, the liberalisation of payment income tests and changes in tax offsets for lower income taxpayers. The extent and precise nature of these changes differs between the two periods, and subsequently.
How many households receive cash benefits?
Before turning to estimates of trends in the number of “zero net taxpayers”, it is useful to put them into context by considering trends in the share of income received from cash benefits or social security payments. This is because reliance on cash benefits is likely to be one of the main reasons for being a “zero net taxpayer”.
Table 3 shows official ABS estimates from their series on Household Income and Wealth of the share of income from government pensions and allowances in household income from 1994-95 to 2013-14. The share of all households for whom social security was their main source of income fell from around 29 % in 1994-95 to 23 % in 2007-08, and rose slightly afterwards. For working age households, the decline was much stronger, from around 20 % in the mid-1990s to 12 % in 2007-08, with a slight increase afterwards.
Overall, there has been a long-term decline in the share of households with a head over 65 years of age relying on social security payments for 50 % or more of their income, from 77 % in the mid-1990s to around 64 % in the most recent year. This is likely to be due to the growth of superannuation savings in this period. The fact that the share of the total population reliant on social security has not fallen as significantly as the share of different age groups reflects the ageing of the population over this period. The share of the population over 65 (and more reliant on social security) grew from 16.8 % of income units in 1995-96 to 22.8 % in 2013-14.[1]
Table 3 also shows trends for the total population in the percentage contribution of benefits to incomes by various ranges, ranging from negligible levels of benefits received (nil or less than 1 %) to almost complete reliance (90 % or more). The share of households who get no income or practically none (less than 1%) from cash benefits has risen strongly from 41% to 49% over this period. There was a slight decline in this following the Global Financial Crisis, but the upward trend in households being virtually completely outside the social security system has resumed. The share of people who get a relatively minor part of their total income from benefits (less than 20%) has also fallen from around 21 to 16%.
At the other end of the spectrum, the share of people who get more than 90% of their income from benefits has fallen from around 22% in 1994-95 to 15% in 2013-14. In between the share of people who get between 20% and half their income from benefits has risen modestly from around 9 to 10 %, while the share who are more reliant, but not completely reliant (between 50 and 90% of their income from benefits) has risen from 6.5 to 9.3%
Table 3: Cash benefits and household income, Australia 1994-95 to 2013-14
| Main source of income: Government pensions and allowances | % contribution of government pensions and allowances to gross household income, all households | |||||||
| Working age | 65 years and over | All households | Nil or less than 1% | 1% to less than 20% | 20% to less than 50% | 50% to less than 90% | 90% and over | |
| 1994–95 | 19.3 | 77.0 | 28.5 | 41.0 | 20.7 | 9.0 | 6.5 | 21.8 |
| 1995–96 | 19.6 | 75.5 | 28.0 | 41.4 | 20.6 | 9.2 | 7.4 | 20.4 |
| 1996–97 | 20.9 | 74.1 | 28.6 | 41.7 | 19.9 | 9.0 | 8.2 | 20.3 |
| 1997–98 | 19.8 | 73.4 | 28.5 | 43.3 | 18.5 | 8.9 | 7.7 | 20.7 |
| 1999–2000 | 18.3 | 74.7 | 28.7 | 44.7 | 17.7 | 8.2 | 7.8 | 20.7 |
| 2000–01 | 17.6 | 70.6 | 28.3 | 44.4 | 16.9 | 9.5 | 7.4 | 20.7 |
| 2002–03 | 16.1 | 68.4 | 26.6 | 46.1 | 17.2 | 9.0 | 7.5 | 19.0 |
| 2003–04 | 17.4 | 69.4 | 27.7 | 44.2 | 18.6 | 9.1 | 8.8 | 18.7 |
| 2005–06 | 15.0 | 70.6 | 26.1 | 43.9 | 20.2 | 9.5 | 8.6 | 17.3 |
| 2007–08 | 11.8 | 65.5 | 23.2 | 46.4 | 20.4 | 10.1 | 8.4 | 14.3 |
| 2009–10 | 14.4 | 65.7 | 25.2 | 45.5 | 19.4 | 9.7 | 9.1 | 15.9 |
| 2011–12 | 13.8 | 64.3 | 24.8 | 47.1 | 18.0 | 9.8 | 8.9 | 15.6 |
| 2013–14 | 13.1 | 63.7 | 24.7 | 49.3 | 15.9 | 9.9 | 9.3 | 15.1 |
Source: ABS Household Income and Wealth, 2013-14.
While specific households can move between any of these different groups, these patterns of change suggest that “independence” from social security has increased quite significantly, and complete reliance has fallen, with many of those completely reliant “moving” to partial reliance.
Further analysis by characteristics such as the age of the household head and the type of family (lone parents, couples with children) also show that the greatest reductions were in the proportion of households deeply reliant on welfare payments (i.e. receiving 90% or more of their income from this source), with the greatest reduction in deep reliance being for lone parent households (down from 40 to 22%), for households with a head aged 65 years and over (from 62 to 39%) and for households with a head aged 55 to 64 years (from 32 to 13%).
In summary, receipt of welfare has been falling for most of the past two decades; while the share of people receiving between 20 and 90% of their income from social security has risen from around 15 to 20%, complete “independence” has increased even more.
How many “zero net taxpayers” are there?
The 2005 Australian Treasury article estimated that 38 % of Australian families (singles and couples, with and without children) would receive more money in cash benefits from the Australian Government than they paid in income taxes in 2005-06. A further 5 % received cash benefits equal to their income taxes in that year. Based on the second definition, as explained above, the total of 43 % would be “zero net taxpayers”.
The article found that there were large increases in real net tax thresholds for particular types of individuals and households over this period – varying between 11.5% for single people of working age, 30% for lone parents and 71% for single income couples without children. Yet, the proportion of the total Australian population estimated to be zero net taxpayers was the same in 2005-06 as it was in 1996-97: 38 %. The proportion that paid the same amount in income tax as they received in benefits was also stable at 5 %, so overall 43% of households were zero or negative taxpayers in each year.
The 2005 Treasury estimate included households of all ages and it did not provide a breakdown by age of household head. [2] The Treasury article did estimate the proportion of different types of family households below these net tax thresholds in 1996-97 and 2005-06. Lone parents were most likely to be zero net taxpayers, but the proportion in this situation fell from 91% to 82% over this nine year period, while the share of couples with children who were zero net taxpayers fell from 35% to 33%.
Apart from Bremner (2005) and the NATSEM and ANU Centre for Social Research and Methods estimates quoted in The Australian, one other Australian article looks at this concept. In Middle class welfare in Australia: How has the distribution of cash benefits changed since the 1980s? (Whiteford, Redmond and Adamson, 2011), trends in the number of “zero net taxpayers” in Australia are examined together with a wide range of measures of the extent to which middle income groups benefit from social spending. This study provides estimates from 1981-82 up to 2007-8, but restricts its analysis to households of working age rather than the total population.
Table 4 shows estimated trends in the share of the population who are “zero net taxpayers” from these three different sources. The figures from Whiteford, Redmond and Adamson (2011) start from 1982 and show that the share of “zero net taxpayers” among people of working age jumped between the early 1980s and the mid-1990s. This change reflects the business cycle. The ABS Household Income survey for 1982 was taken at the peak of an economic boom and just before the recession of the early 1980s. Following the recession of the early 1990s, recovery started to accelerate from the mid-1990s onwards.
Table 3 showed that the share of working-age households for whom social security was the principal source of income fell from 19.3 % in 1994-95 to 11.8 % in 2007-08. Yet, Table 4 shows that the share of zero net taxpayers was virtually unchanged at around 27 % over this period. This is consistent with the results of the 2005 Treasury estimates which showed no change in the share of the overall population who were “zero net taxpayers” between 1996-97 and 2005-06. The Treasury figures for 2005-06 were very similar to NATSEM’s estimates for the same year.
A more recent analysis by the ANU Centre for Social Research and Methods estimates suggest an increase from around 43 % to 48 % for 2013-14 and 2015. It is interesting to investigate why this increase might have occurred.
Table 4: Estimated numbers of zero net taxpayers,% of working age and all households, Australia, 1982 to 2015
| Whiteford et al. (2011) | Bremner (2005) | NATSEM/Centre for Social Research and Methods, various years | |
| 1982 | 21.9 | .. | .. |
| 1990 | 22.0 | .. | .. |
| 1994–95 | 26.7 | .. | .. |
| 1995–96 | 28.3 | .. | .. |
| 1996–97 | 27.7 | 43.0 | .. |
| 2000–01 | 27.4 | .. | .. |
| 2002–03 | 25.7 | .. | .. |
| 2003–04 | 26.0 | .. | .. |
| 2005–06 | 25.8 | 43.0 | 43.9 |
| 2007–08 | 27.1 | .. | |
| 2013–14 | .. | .. | 48.0 |
| 2015 | .. | .. | 47.4 |
Notes: Figures from Whiteford, Redmond and Adamson (2011) refer to households of working age only; figures for Bremner (2005) and NATSEM/ANU refer to working age plus pension age households.
The first reason concerns income tax and benefit settings for individuals or households over the age of 65. Table 3 showed that reliance on social security payments is much higher among households with a head aged 65 and over – 64 % in 2013-14 compared to 13 % of households with a working age head. The ANU Centre for Social Research and Methods is reported to have estimated that the share of “zero net taxpayers” among those aged 65 and over increased from 87 % in 2005 to 90 % in 2015.
The earlier discussion of the Grattan Institute analysis of tax data suggests that this is mainly due to the Howard Government’s decision in 2007 to exempt people over 60 from paying income tax on their superannuation withdrawals.
It is also the case that in 2009, rates of Age Pension were significantly increased, which other things being equal could be expected to reduce the share of net taxpayers. However, as part of this reform the withdrawal rate on pensions was increased from 40% back to 50% (the rate which had applied between 1969 and 2000). The effect of this was to keep the pension cut-out point at the same income level, so that there should not have been a major impact on the number of zero net taxpayers, remembering that the net tax threshold in the Australian system must be below the payment cut-out point.
For working age households, Table 3 showed that the share that get more than half of their income from social security payments fell from 19.3% in 1994-95 to 11.8% in 2007-08, but Table 4 shows that over this period, the share of zero net taxpayers was broadly stable, going from 26.7% to 27.1%.
It is likely that this is the result of a combination of factors. The withdrawal rates on unemployment payments were reduced in the mid-1990s and those on pensions including the Disability Support Pension, Parenting Payment Single and Carers Payment were also reduced in 2000 with the introduction of the Goods and Services Tax. Withdrawal rates on family payments were also reduced at this time, and again in 2003. As noted by Harding, Vu, Payne and Percival (2009) a key reason for the reduction in withdrawal rates for these benefits was concern about the implications for work incentives of the very high effective marginal tax rates (EMTRs) which result from the interaction of benefit withdrawal rates and income tax.[3]
A second reason is increases in the tax-free threshold and changes in the first marginal tax rate over this period. The basic tax threshold rose from $5400 in 1995-96 to $6,000 in 2007-08; while this was a fall of around 20 % in real terms, the Beneficiary Tax Offset ensured that those completely reliant on Newstart were not liable for income tax. The real value of the Low Income Tax Offset also increased significantly. Moreover, the first tax rate was cut from 20 to 15 cents in the dollar in this period, which could be expected to push up net tax thresholds.
It should be remembered, however, that the reductions in income tax rates and in withdrawal rates on payments as well as the introduction of the Senior Australians Tax Offset were all part of the compensation package for the introduction of the Goods and Services Tax, with the aim of broadening the tax base and reducing income taxes. Inclusion of the effects of higher indirect taxes would reduce the number of “zero net taxpayers”.
Overall, the substantial increase in real net tax thresholds over this period is the result of deliberate government policy decisions about the structure of the income tax and not the result of extensions of welfare assistance. The fact that the share of “zero net taxpayers” does not appear to have increased in line with the increase in net tax thresholds is also a consequence of growth of earnings and family incomes over the period from the 1990s up to the GFC. Real average weekly earnings grew by 15 % between 1996-97 and 2007-08, and by a further 8.5 % between 2007-08 and 2013-14. This strong growth in wages was further enhanced by increases in employment and the growth of two earner households. As real incomes grow, the operation of income tests means that the role of social security benefits shrinks – a finding confirmed by research from Whiteford (2014) and Herault and Azpitarte (2014).
For example, because Newstart benefits are indexed to prices and real wages have increased both benefit levels and cut-out points have fallen in Australia relative to average wages. In 2001, Newstart plus Rent Assistance for a single person was 27 % of the average wage (using OECD data) but by 2014 payments had fallen to 21 % of the average wage, with the cut-out point falling from 44 to 39 % over the same period. Given that the net tax threshold must be lower than the benefit cut-out point, we would expect the contribution of Newstart payments to the number of zero net taxpayers to be falling.
What about the apparent increase in the share of “zero net taxpayers” in the most recent ANU estimates? The tax-free threshold was significantly increased in 2012-13 as compensation for the introduction of carbon pricing[4], and reflecting a recommendation of the Henry Tax Review. This may also have contributed to the increase in the share of “zero net taxpayers” in the latest estimate. Correspondingly, when real earnings do not grow the size of the social security system will tend to be stable, while increases in unemployment will tend to increase the share of “zero net taxpayers”. Real average weekly earnings have risen by only 0.5 % since 2013-14.
It is also worth noting that not all “zero net taxpayers” are in the bottom half of the income distribution. In Middle class welfare in Australia: How has the distribution of cash benefits changed since the 1980s? we estimated that of the 27.7 % of working age households who were “zero net taxpayers” in 1996-97 when the Howard government was elected about 1.9 percentage points were in the top 50% of the population, but by 2007-08 when the Rudd Government was elected – and the overall figure was still 27.7 % – this had risen to 2.8 percentage points, implying a marginal shift upwards in the location of zero net taxpayers. (This can also be seen as a measure of a very small increase in “middle class welfare”.) Even within the bottom half of the income distribution there are differences in the share of “zero net taxpayers”. In 2007, 89.5% of the lowest income decile were zero net taxpayers, reducing to 78% of the second decile, 47% of the third decile, 24% of the fourth decile and 16% of the fifth decile.
Part three of this series will discuss international comparisions of net tax thresholds .
The full article on which this post is based can be found at the TTPI website.
[1] These estimates differ from estimates for the number of taxpayers due to a different starting point, and because not all persons are in scope for household surveys; for example they do not include people in hospitals or nursing homes.
[2] It is also worth noting that an aspect of their methodology might be expected to overestimate the number of zero net taxpayers. This is because they used a microsimulation model to estimate household’s entitlements to cash benefits as well their income tax obligations, instead of using data of individuals who were actually receiving cash benefits. In the case of a single person entitled to Newstart Allowance (as in the example above), this approach assume that any person who satisfies the income test for payments is actually receiving a payment. However, this may not be true in practice. For example, if Scott has been on Newstart Allowance and then moves into part-time or low-paid work, he could be expected to continue to receive part payments of Newstart Allowance. In contrast, a person who finishes study and then gets a low paid job at the same income level may or may not apply for a part-rate of Newstart (particularly given the onerous reporting requirements that such a person would incur).
[3] See Tax and Transfer Policy Institute Policy Brief 1/2016 for a discussion of EMTRs.
[4] This was partly offset by a reduction in the Low Income Tax Offset.
Recent Comments