Inequality in Australia – are the rich getting richer and the poor poorer?

Cross posted from Inside Story

Australians like to think of themselves as egalitarian, and in the past Australians also liked to believe that we had a relatively equal distribution of income and wealth.  As early as the 1880s, visitors to Australia apparently remarked on the greater equality of the distribution of wealth, with the lack of a poor class, comfortable incomes being in the majority, and millionaires few and far between.

As late as 1967, Harold Holt said he did not know of any free country in the world where “what is produced by the community is more fairly and evenly distributed among the community than Australia”.  From the 1980s onwards, however, this view of Australia became subject to reappraisal.  As noted by John Hirst  “‘Egalitarianism – see under myths’: so runs the index entry in a standard sociological text on Australian society…”

Working out what has happened to inequality in Australia over the long-term is complex, although there are a number of studies by Peter Saunders, David Ingles, Ian McLean and Sue Richardson and Andrew Leigh that provide invaluable evidence on trends in income inequality since the early part of the 20^th century. While there is some disagreement about overall trends, what seems to have happened is that there was a substantial long-term decline in inequality between the 1940s and the 1970s, but the data available are limited – either based on wage trends or income tax data, so that the findings apply to the incomes of individuals rather than households.

Income inequality trends since the 1980s

The availability – and quality – of household income statistics has improved dramatically in recent decades, with the Australian Bureau of Statistics (ABS) starting to undertake income surveys from the late 1960s onwards, although it was not until 1981-82 that the surveys became regular – first every five years or so, and more recently every second year. The most recent survey which covers the 2009-10 year was issued by the ABS at the end of August. The release of high quality data has dramatically increased the number of studies of income distribution trends, also helped by the establishment of research centres such as NATSEM at the University of Canberra and the work of researchers like Ann Harding.

The ABS has gone to a lot of trouble over the years to improve the data collected in income surveys, but an unavoidable by-product of better data is that it is not always comparable over time. The figure below provides a summary picture of overall trends, but it puts together a range of estimates based on different measures of income. Between the 1980s and the middle of the 1990s the estimates of income inequality are based on annual income, but since the middle of the 1990s the ABS has published figures based on current weekly income.  The “revised” figures from 2003-04 on show a higher level of inequality than earlier estimates, as they include a broader range of income sources than earlier measures (for example, including non-cash benefits provided to employees, termination payments, irregular overtime and all bonuses, among other items).

Chart 1 shows trends in the Gini coefficient, one of the most commonly used measures of income inequality.  The Gini coefficient varies between zero and one – if everyone had exactly the same income then the Gini coefficient would be zero (perfect equality), and if one household had all the income, then the Gini coefficient would be one (complete inequality). Since the figure shows that the Gini coefficient was around 0.27 in 1981-82 and 0.328 in 2009-10, this means that income inequality has risen over the period shown.

Chart 1: Trends in income inequality (Gini coefficient) in Australia, 1981-82 to 2009-10

Source: Estimated from ABS income surveys, various years.

Some of the short term changes between surveys are not large and past ABS publications have pointed out that changes between particular years were not statistically significant.  Yet the cumulative picture is of an upward trend.  Nevertheless, it is also clear that there are periods when inequality has fallen – in the second half of the 1980s, again in the second half of the 1990s, in the early 2000s, and finally in the most recent two years (although the ABS cautions that this should also not be considered as a significant change).

It is now a common view that inequality has increased in Australia over the last 30 years – sometimes expressed as “the rich have been getting richer and the poor poorer”. The ABS surveys show, however, that this is not a fully accurate description of what has happened in Australia.

While the overall trend has been towards increasing inequality, the fact that inequality fell during some periods is a reminder that the factors producing higher or lower inequality can change. It is also worth noting that the more the overall trend is disaggregated, the more complicated the picture becomes. For example, households can have multiple sources of income, including earnings (the most important item overall), self-employment, unearned income from capital, and government benefits, and households also pay direct taxes which are subtracted from income.

As well as disaggregating the sources of income, it is possible to look at different types of households – age pensioners and the retired, couples with children, lone parent families, and single people.  Income inequality amongst lone parent families (Gini coefficient of 0.262) is lower than among couples with children (0.312), or couples without children (0.357), and inequality is highest among single person households (Gini coefficient of 0.387).

Or we can focus on geography. Over the past half-decade, income inequality has increased most significantly in Western Australia. Indeed – as Chart 2 shows – if Western Australia were a country it would be about as unequal as Portugal (although much richer), and only a little less unequal than the United States. If the Australian Capital Territory or Tasmania were countries, on the other hand, they would be about as unequal as Switzerland or Norway (although not as well-off). Inequality has risen fastest in Western Australia but so have average incomes – 76 per cent in real terms since the mid 1990s compared to the Australian average of 57 per cent. But the average income of the richest 20 per cent of West Australians has more than doubled, and this group received more than half of all the income growth that the state’s households enjoyed over the fifteen-year period.

Chart 2: Level of income inequality (Gini coefficient) among OECD countries (2005–08) and Australian states (2009–10)

Source:  OECD and ABS.

Turning back to the picture for Australia as a whole, it is certainly true that the rich have become richer – the best-off 20% of households now enjoy real incomes that are 67% higher than comparable groups in the mid-1990s, and they hold just over 40% of total household income compared to 38% fifteen years earlier. But while the income shares of the rest of the population have fallen, their real disposable incomes have also risen – by between 46% for the poorest 20% of households and nearly 50% for the median household. So the rich have become richer – but so has everyone else.

It is worth remembering, however, that this is a picture of what has happened to categories of people, not individuals – there are plenty of individuals and families who are worse-off now than they were in the past, either because they have retired, become unemployed, divorced or separated, become sick or have become disabled.  Other individuals have become better-off as they have moved into work, been promoted in their jobs or have developed businesses.

According to the latest HILDA report over the seven year period to 2008 about 25% of individuals moved up the income distribution by two or more deciles, while 19% moved down to the same degree.  Thus, all that the static figures tell us is that even if individuals have experienced these risks or these benefits, groups of individuals are financially better-off than comparable groups in the past.

 Rising inequality appears to have had different causes in different periods. Wages are the most important single source of income for households overall, and inequality in wage rates for both men and women has increased, although wage inequality between men and women fell in some periods.  Research by David Johnson and Roger Wilkins shows that the recessions of the early 1980s and early 1990s appear to have resulted in higher inequality of earnings, mainly because the share of households who had no earnings and were receiving government benefits increased significantly, while there was also an increase in the number of two-earner families at the richer end of the income distribution.

The recovery that started in the mid-1990s and continued until the Global Financial Crisis in 2008 also saw rising inequality, with the increase being most rapid after 2003. Inequality in wage rates continued to increase, but this was offset by increasing employment, so for people of working age, earnings inequality fell mainly because lower income groups benefited from higher employment rates. In particular, research by Siobhan Austen and Gerry Redmond finds that changes in women’s earnings, and partnered women’s earnings, were associated with downwards pressure on family income inequality between 1995-96 and 2007-08. In the 1980s and early 1990s, growth in women’s employment and earnings was greatest in households with relatively high male earnings. From the mid-1990s, the growth in women’s employment was concentrated in households with the lowest male earnings and this helped to stabilize family income inequality.

The impact of the GFC

What has happened since the GFC? The figure above suggests that there was a slight decline in inequality (although as the ABS points out this is most likely to represent no real change). This contrasts with earlier recessions in the 1980s and 1990s when inequality rose. Of course, Australia technically escaped recession – although unemployment did increase from around 4% to 6%, before coming down to around 5% now. Australia’s recent experience on the surface does seem to be similar to that in countries which suffered much more from the GFC – or the Great Recession as it is more commonly known in Europe and the United States. A new study, entitled The Great Recession and the Distribution of Household Income, and produced by an international research team led by Professor Stephen Jenkins of the London School of Economics looked at what has happened between 2007 and 2009 in 21 countries including Australia – although not in great detail in Australia’s case. The study found that in the two years following the financial crash of 2007, household income distributions hardly changed, even where national GDP fell sharply. Over this period, and for most countries studied, average disposable household income levels generally fell slightly and poverty rates and income inequality changed little.

It appears that countries have been able to cushion households from the immediate effects of the Great Recession by means of benefits and other social safety nets.  However, the study predicts that as governments cut public spending and raise taxes to confront structural deficits, household incomes will be hit, and for up to five or 10 years or even longer, depending on when economic growth returns.  Cuts to a country’s welfare system as part of fiscal consolidation are likely to lead to greater instability in households’ income levels, poverty rates, and income inequality.   The study points out that instability is less likely in countries that already have relatively health fiscal balances and are not facing large tax rises or benefit cuts (such as Sweden and Germany) compared to those, such as Ireland, saddled with both large cuts and deficits or the UK, which lies between the two extremes.

Given that the latest ABS Income survey was only published at the end of August there are as yet no studies that analyse in any detail what happened in Australia between 2007 and 2009, although Peter Martin gives a useful summary of broad trends. Some fairly broad level comparisons are also possible. For example, both average and median household disposable income fell by about $11 a week between 2007-08 and 2009-10, the first fall in 13 years.  The average number of employed persons per household also fell, although it still remains at the second highest level recorded in the income surveys. The proportion of households whose principal source of income is government pensions and benefits increased from 23.2% to 25.1% – although that is the second lowest level since the mid-1990s and remains below the level in 2005-06.

But while the general picture given by the income surveys is of gradual or incremental change, this is not the case for all groups.  As mentioned above, income inequality in Western Australia has continued to increase significantly.  For particular groups and specific income sources, trends are also disparate.  For example, Chart 3 shows trends in inequality among households with a head aged 65 years and over since 2000-2001. In contrast with the general picture of relatively small changes, for older people there have been very large shifts in inequality.  Up until 2000, income inequality among older households was much lower than among the population generally, but it then accelerated dramatically, so that by 2007-08, the Gini coefficient had risen from about two-thirds that of the overall population to equal overall income inequality, and for older couples to be higher.

Chart 3: Trends in income inequality (Gini coefficient) among households with a head aged 65 years and over, Australia, 2000-2001 to 2009-10

Source: Estimated from ABS income surveys, various years.

The rise in inequality among older households appears to have been driven mainly by rapidly increasing income from capital, and is likely to have been associated with the boom in the share market prior to the GFC – between 2000 and late 2007, the All Ordinaries Index rose by around 120%.  Preliminary analysis that Gerry Redmond and I are undertaking that looks at trends in the incomes of working-age families since the 1980s also finds that in the period of the Howard Government, the main factor driving increasing income inequality after 2003 was increasing income from increasingly unequal capital sources, which more than offset the equalising impact of employment growth.

Correspondingly, Figure 2 shows that income inequality among older households fell in the period that corresponds to the GFC, particularly for older single person households.  It is plausible that part of this decline in inequality was a consequence of the fall in the stock market– which fell by about 30% between its peak and late 2009 (and by another 20% since then).  But another factor that seems likely to have lead to this significant decline in income inequality for older single people is the very large increase in the value of the single age pension that was made in 2009-10 Budget following the recommendations of the Harmer Pension Review.

Inequality – where to from here?

There are legitimate disagreements about what constitutes acceptable and unacceptable inequality. Some might argue that any increase in inequality should be countered by higher redistribution through more progressive taxes and more generous social spending, while others would argue that increasing inequality is part of the process of generating higher incomes for all Australians.

From a narrower statistical perspective, one can be concerned about increasing inequality simply because “averages mislead” – higher average incomes do not necessarily benefit all Australians if the income gains only accrue to higher income groups.  To date, this has not been the experience in Australia – the benefits of strong average income growth over the past decade have been more widely shared than in the United States.

Nevertheless, it is important to monitor trends in inequality, since without reliable statistics we would not know what is really happening to different groups in Australian society.  In order to understand actual trends we need reliable statistics on distributions and not just on averages, which is why we should be grateful to the Australian Bureau of Statistics for continuing to collect high quality data on income and wealth.