Household Income and Wealth, Australia

In 2019–20, compared to 2017–18, there was no statistically significant change in real (inflation adjusted), mean equivalised disposable household income (EDHI). In 2019–20, real EDHI was $1,124 per week, compared to $1,094 per week in 2017–18.

As shown in Graph 1, real EDHI increased from 2000–01 to 2007–08, with a small decline following the Global Financial Crisis (GFC) in 2008. Since then, real EDHI recovered but experienced slower growth from 2015–16.

In 2019–20, the main sources of household income were:

• employee income (62% of households), similar to 2017–18 (61%).
• government pensions and allowances (22% of households), the same as 2017–18.

In 2019–20, compared to 2017–18, real net wealth, the value of a household’s assets minus the value of its liabilities (debts) inflation adjusted, saw no statistically different change. In 2019–20, average household real net wealth was $1.04 million. This was $11,200 lower (- 1%) than in 2017–18. In 2019–20, the mean value of household assets was $1.25 million, while the mean level of household debt was $203,800.

In 2019–20 the largest household assets were from owner occupied dwellings, 40% of total household assets (a decrease from 42% in 2017–18). The second largest household assets were from superannuation, at 18% which was consistent with 2017–18. 

Graph 3 shows the average value of assets per household. Here you can see that the average value of owner occupied dwellings in 2019-20 was an estimated $502,500 per household, and superannuation funds was an estimated $229,900 per household.  

In 2019–20, 15% of household assets were from other residential and non-residential property e.g. for rent and holiday homes. The average value of other property was $191,100 per household. Just under one in four households (23%) owned property other than the dwelling in which they lived (compared to 22% in 2017–18).

In 2019–20, average household debt was statistically different compared to 2017–18. In value terms, it was $14,300 higher (7.5%) than in 2017–18, with an estimated average debt of $203,800 per household. The average value of household liabilities is shown in Graph 4. The average amount owing on home loans was $114,700, while the amount owing on loans outstanding for other property averaged $68,100. The average household debt for other liabilities was $20,700. 

To analyse the way that income and wealth are distributed across households in Australia, households can be ranked from lowest to highest income or wealth and then divided into five equal groups, with 20% of the population in each group (quintiles).

As shown in Graph 1, after taking account of the number and age of people in the household, households in the highest income quintile received 40% of total income in 2019–20 (as they did in 2017–18). By comparison, households in the lowest income quintile received 7.4% of total income (which was 7.5% in 2017–18, and statistically equivalent). The overall pattern remained relatively stable over the past two decades.

Analysis of the distribution of wealth (see Graph 1) shows that the highest wealth quintile owned 63% of total household wealth in 2019–20 (as they did in 2017–18). By comparison, the lowest wealth quintile owned less than 1% of all household wealth (0.7% in both 2017–18 and 2019–20). This indicates that the distribution of wealth is less equal than the distribution of income.

Mean equivalised disposable household income in Australia in 2019–20 was $1,124 per week. The median was lower, however, at $959 per week. This is due to the larger proportion of households with middle or low income and the small proportion of very high income households, as shown in Graph 2.

As shown in Graph 3, there is greater inequality in the distribution of wealth than income. The lowest 20% of households, in terms of net worth, had a mean net worth of $35,100. In comparison, the mean net worth of the wealthiest 20% of households was $3.27 million, or more than 92 times that of the lowest 20% of households. The mean net worth of all households in Australia in 2019–20 was $1.04 million while the median was $579,200. 

There are many summary indicators that can be used to help understand the distribution of income and wealth across the population. The ABS uses the Gini coefficient as an internationally comparable indicator. The Gini coefficient lies between 0 and 1. If everyone in the population had the same income or wealth, the Gini coefficient would be zero. Gini coefficient values that are closer to 1 represent greater inequality. Compared to other summary indicators, the Gini coefficient is not overly sensitive to low or negative incomes.
In 2019–20, the Gini coefficient for gross household income was 0.436. After taking into account household composition and income tax, the Gini coefficient for EDHI was 0.324.

The Gini coefficient for wealth is typically higher than for income, reflecting greater inequality in the distribution of wealth. The Gini coefficient for wealth in 2019–20 was 0.611

Graph 1 shows in 2019–20, just under a third (32%) of low income households also had low wealth (i.e. they were in the lowest net wealth quintile), while 11% of low income households had high wealth. This pattern is reversed for high income households. Where 40% of high income households also had high wealth whereas 7% of high income households had low wealth. 

Low income households are most likely to rely on government pensions and allowances as their main source of income, whereas employee income is the main income source for middle and high income households, as shown in Graph 2.

Some household types are more common in the low income group, as shown in Graph 3. Lone person households are more likely to be in the low income group, while couple only households are more likely to be in the high income group.

For the analysis below a retiree household is defined as a household where the reference person in the household was 65 years or older and not in the labour force.

As can be seen in Graph 4 below, high income retiree households were more likely (50%) to draw their income from superannuation income than any other income source. For middle income retiree households, superannuation income was also the most common source of household income (47%). Low income retiree households were more likely to draw their income from government pensions and allowances (89%).

Change in the distribution of income and wealth over time are an area of interest for social and economic policy analysts and researchers. Distribution analysis can indicate whether the material living standards of the community are improving evenly across the population.

In the 20 years between 2000–01 and 2019–20, the mean income (in 2019-20 dollars) of:

• Low income households increased by $155 per week to reach $456.
• Middle income households increased by $333 per week to reach $966. 
• High income households increased by $852 per week to reach $2,234.

All income groups have experienced a real increase in their income since the mid-1990s. Some of the growth in middle and high income groups was due to a broadening of the Survey of Income and Housing (SIH) income measure from 2003–04, with further improvements in 2007–08. However, there were also real increases in average incomes during this period.

In 2019–20, over a third (36%) of low wealth households (net worth less than $113,400) also had low household income, while 7% had high household income (see graph 1). For the high wealth households (net worth exceeding $1.45 million), 10% had low household income. However this group is unlikely to be at risk of experiencing economic hardship as they can draw on their wealth. 

The main assets for low wealth households in 2019–20 were:

• Other non-financial assets (including dwelling contents and vehicles) (30%).
• Superannuation (24%).
• Property (owner occupied dwellings and other property) (21%).

The main asset for middle and high wealth households was property, but in contrast to low wealth households, both these groups have property ownership rates over 92%. For middle and high wealth households, owner occupied dwellings contributed 55% and 34% respectively to their wealth, as can be seen in Graph 2.

Almost half (48%) of high wealth households have property loans, with loans for owner occupied dwellings and other properties accounting for 90% of the total value of liabilities for high wealth households.  

Just over half (54%) of middle wealth households had property loans, which made up 94% of the total value of liabilities for this group. Low wealth households were less likely to own property, with 9% of this group having property loans.

Additionally, 26% of low wealth households have debt outstanding on study loans which accounts for 18% of total liabilities for all low wealth households. This drops significantly to account for less than 2% of total liabilities owed by both middle and high wealth households due to the much higher value of other liabilities for these groups. Around 16% of middle wealth households and 18% of high wealth households have study loans.

In the following analysis, a retiree household is defined as a household where the reference person in the household was 65 years or older and not in the labour force.

High wealth retiree households are more likely (46%) to draw their household income from superannuation than any other income source. For low and middle wealth retiree households the main source of income was from government pensions and allowances (95% and 81% respectively). 

Key changes in wealth in the decade leading up to 2019–20 include:

• Middle wealth households had an average net worth of $588,400 in 2019–20 compared to $521,100 in 2009–10.
• High wealth households increased in real terms from an average net worth of $2.7 million in 2009–10 to $3.3 million in 2019–20.
• Low wealth households experienced a decrease in net worth over this time period with the average net worth of $35,100 in 2019–20 compared with $38,800 in 2009–10.

One factor driving the increase in net wealth of high income households is the value of property (owner occupied and other property). For high wealth households, average total property value increased by $683,000 between 2003–04 and 2019–20 from $1.2 million to $1.88 million. For middle wealth households average property values increased by $142,100 (from $373,600 to $515,700). Low wealth households that owned property had much lower growth of $24,200 growing to $32,300 over the last sixteen years (see graph 2).

Between 2009–10 and 2019–20 average liabilities have increased in real terms for different wealth groups (see graph 3):

• Low wealth households increased from $28,900 to $54,700.
• Middle wealth households increased from $160,500 to $213,100.
• High wealth households increased from $214,400 to $341,300.

What is household economic well-being?

When considering the circumstances of households, the key economic well-being factors that affect people’s material standard of living are income, consumption and wealth.

Income can be used to support consumption of goods and services, such as food, clothing, housing and leisure activities. Alternatively, it can be saved and invested to increase wealth which can be used at a later date to support consumption.

Some people with low incomes have considerable wealth allowing them to maintain consumption levels above their current income. People with low reserves of wealth may face financial difficulty in times of need, such as during any period of reduced income or substantial unexpected expenses.

Diagram 1 illustrates this relationship, although people’s actual well-being is affected by individual circumstances and lifestyle choices.
 

Income levels and wealth vary over a person’s life and are affected by two main factors, age and labour force participation. Incomes tend to grow until middle age. Wealth tends to be gradually accumulated during the working lives of household members and used during retirement (Graph 1).

The definitions used to measure the economic well-being of people can have a significant impact on the results. The Australian Bureau of Statistics (ABS) follows international best practice in producing micro statistics relating to household economic resources.

Income
The most comprehensive measure of income is compiled from the ABS Survey of Income and Housing (SIH) and the ABS Household Expenditure Survey (HES). This definition aligns with international standards released in 2004 and fully adopted from SIH 2007–08 and HES 2009–10:
Income consists of all current receipts, whether monetary or in kind, that are received by the household or by individual members of the household, and which are available for, or intended to support, current consumption.

Wealth
The first international wealth standards were published by the OECD in 2013.
Wealth refers to economic resources in the form of assets and liabilities. Wealth, or net worth, is the value of all the assets (e.g. property, bank accounts and shares) owned by a household less the value of all its liabilities (mortgages and other loans) at a particular point in time. Net worth may be negative when a household’s liabilities exceed its assets.

Consumption expenditure
The international definition of consumption expenditure is summarised as:
Household consumption expenditure is the value of consumer goods and services acquired, used or paid for by a household through direct monetary purchases, own account production, barter or as income in kind.

In the HES, expenditure is valued as the cost of goods and services acquired during the reference period for private use, whether or not the goods were paid for or consumed in that period. Expenditure is net of refunds and trade-ins. Consumption expenditure includes in kind income from employers, such as
subsidised housing or the use of a car for private purposes.

In recent years the ABS has made significant progress in extending its measurement of household income to reflect real world changes and enhance analytical opportunities. This includes developing new measures to allow the full economic circumstances of different types of households to be compared. In particular, the ABS has produced:
a) imputed rent (IR) estimates since 2003–04
b) social transfers in kind (STIK) allocations from SIH 2011–12 (previously only based on HES data)
c) final income estimates since 1984.

a) Imputed rent
What is it?
Income from imputed rent is allocated to owner occupiers and households living in subsidised private rentals e.g. renting from a family member. For owner occupiers, income from imputed rent is the estimated market rent of a dwelling less housing costs normally paid by a landlord such as mortgage interest, rates, insurance and repairs. For renters, it is the difference between market rent and actual rent paid.

Why include imputed rent in income?
Housing is one of the most significant living costs borne by many households. The inclusion of imputed rent in income provides a broader picture of the economic well-being of owner occupied and rent-subsidised households relative to other households, allowing more meaningful comparisons of the well-being of people living in different tenure types.

b) Social transfers in kind
What are they?
Social transfers in kind are goods and services provided by governments that benefit individuals but are provided free or at subsidised prices. Examples include free or subsidised education, health and child care.

Why include STIK in income?
STIK have a significant impact on the well-being of people and on the measurement of the distribution of income. This is important for comparisons within and across countries. In Australia, many government services have been designed to assist those most in need of financial support. The allocation of benefits differs between households, reflecting characteristics such as household composition, life cycle stages, household size and income.
The inclusion of IR and STIK increased the mean equivalised disposable household income (EDHI) from $1,062 to $1,450 per week in 2017–18 and reduced the inequality of income distribution across households (Graph 2).
 

c) Final income
What is it?
Final income is equal to household private income plus social assistance benefits in cash (e.g. age and disability support pensions, Family Tax Benefit) and STIK less income taxes and taxes on production (e.g. GST and taxes on alcohol and cigarettes). Both household income and expenditure are required to estimate final income. This data is available whenever the HES is conducted, most recently in 2015–16 (Diagram 2).

Diagram 2 illustrates the relationship between the different income concepts presented in this Fact Sheet Series.
 

Why is it important?
Final income shows the full effect of government expenditure and taxes on the distribution of income among private households in Australia. This allows policy makers to understand the effects of changes in either government revenues or spending that directly impact on the economic well-being of households.

The net effect of government benefits and taxes in 2015–16 was to increase average incomes of households in the three lowest quintiles and decrease those of the two highest quintiles (Graph 3).
 

For more information:

• OECD, 2013, OECD Framework for Statistics on the Distribution of Household Income, Consumption and Wealth 
• OECD, 2013, OECD Guidelines for Micro Statistics on Household Wealth 
• United Nations, 2011, Canberra Group Handbook on Household Income Statistics, Second Edition

Economic and social analysts and policy makers are interested in the distribution of resources and how this affects the well-being of society and individuals, particularly people’s ability to acquire the goods and services required to satisfy their needs. 

Questions that researchers ask include:

• How unequal is the distribution of income and wealth? How does this compare with earlier years, or with other countries? 
• What are the characteristics of households considered most at risk of economic hardship? Which are in greatest need of financial support?
• Do people have sufficient incomes and wealth accumulation in their working lives and to maintain an adequate standard of living in retirement?

Why is an equivalence scale used?

As household size increases, consumption needs also increase but there are economies of scale. An equivalence scale is used to adjust household incomes to take account of the economies that flow from sharing resources and enable more meaningful comparisons across different types of households.

For a lone person household equivalised income is equal to actual income. For households comprising more than one person, it is the estimated income that a lone person household would need to enjoy the same standard of living as the household in question.

How are equivalising factors calculated?
Equivalising factors are calculated based on the size and composition of the household, recognising that children typically have fewer needs than adults. The ABS uses the OECD-modified equivalence scale which assigns a value of 1 to the household head, 0.5 to each additional person 15 years or older and 0.3 to each child under 15 years.

Table 1 shows that a couple household with one child would need $1,800 weekly disposable income to have the same equivalised disposable household income (EDHI) as a lone person household with a disposable income of $1,000.
 

Relationship between equivalisation of income, consumption and wealth
Equivalence scales used for household income are equally applicable for consumption measures. There is less agreement about how to equivalise household wealth as wealth is often built up during a person’s working life and then used during retirement when the composition of the household might be quite different. However, when wealth is being used to support current consumption, particularly for households at risk of economic hardship, household wealth should be equivalised with the same scale used to equivalise household income and consumption. 

There are two common ways of presenting analysis of households:

• number of households, or
• number of people in households.

In the former, each household contributes the same regardless of its size e.g. a four person household would have the same representation as a person living alone.

To provide a better understanding of the circumstances of people it is often preferable to study people in households e.g. the number of people in Australian households experiencing economic hardship. In this analysis, each person is attributed with the characteristics of the household to which they belong e.g. household income is used to determine whether it is a low or high income household but analysis is about numbers of people experiencing hardship. This approach keeps the focus on individual circumstances while recognising that people share household resources.
 

There are several summary measures commonly used for analysing household economic well-being.

Counts
Counts provide an estimate of the total number of people or households with a particular characteristic and are derived by summing the survey weights of each observation of interest. In sample surveys the weights enable extrapolation of the survey responses to official population estimates.

Means
The arithmetic mean, or average, is the sum of all income divided by the number of observations. Advantages of the mean are that it is easy to calculate and the means of all sub components sum to the mean of all observations. Its drawbacks are the effect of extreme values and asymmetry of the distribution, both of which are relevant for income and wealth data. For example, a small number of very wealthy and a large number of relatively poor households may have the same average income or wealth as a population where there is equal distribution of resources.

Medians
Medians are calculated by ranking all observations from the lowest to the highest. The middle observation of the distribution is the median. Compared to the mean, the median is a more stable measure and is less affected by extreme values and sample fluctuations. However, median values of sub components do not sum to the median of all observations.

Measures of the distribution of income and wealth help to describe and understand how economic resources are shared across the population and households.

Frequency distribution
Frequency distributions show the proportion of people or households with a particular level of income or wealth. To produce the distribution, the item of interest is ranked by value and the population grouped into classes. The ABS currently uses $50 ranges for weekly income and $100,000 ranges for wealth.

It is useful to include the summary statistics such as the mean and median in the frequency distributions. Income and wealth distributions tend to be asymmetrical, with a small number of people having relatively high income or wealth and a much larger number having relatively low income or wealth (Graph 1).

Quantiles
Quantile is a term for groups formed by ranking the units of analysis (e.g. household or persons) in ascending order and calculating the shares of the total accruing to a given proportion of the units:

• quintiles are formed when the population is divided into five equally sized groups
• deciles into ten groups
• percentiles into 100 groups

Therefore the first quintile will comprise the first two deciles and the first 20 percentiles. The mean or the median may be used to summarise the circumstances within a quantile.

Percentile ratios
The boundary between quantiles is usually expressed as the upper value of a particular percentile. The ABS publishes the upper value of each decile (P10 to P90). This provides the range of values in each quintile e.g. the middle (3rd) quintile is formed by households with income/wealth between P40 and P60. The median of each quintile can also be determined e.g. the median of the first quintile is P10, second quintile, P30, etc. The median of the whole population is P50. 

Percentile ratios summarise the relative distance between two points on the income or wealth distribution. Percentile ratios will be less volatile than measures based on means, particularly at each end of the distribution. To illustrate the full spread of the income distribution, the percentile ratio should use points near the extremes e.g. the P90/P10 ratio. The P80/P20 ratio better illustrates the magnitude of the range  or the majority of the population. The P90/P50 and P10/P50 ratios compare the ends of the distribution with the median and these are commonly used to understand how the wealthier compare to average and the poorer to average. 

Table 2 shows that income is more equally distributed than wealth. In 2019–20, the equivalised income of households at the top of the 80th percentile (or fourth quintile) was 2.6 times higher than that of households at the top of the 20th percentile (or lowest quintile), whereas wealth was 11 times higher (P80/P20). 

Shares of income or wealth
Income or wealth shares can be calculated and compared for each quantile of a population. The aggregate income/wealth of units in each quantile is divided by the total aggregate of the entire population to derive quantile share. 

Graph 3 shows income and wealth shares by decile. Household wealth is more unequally distributed than household income. People in the three lowest equivalised income deciles received 13% of all income, whilst people in the three lowest equivalised wealth deciles held only 3% of all wealth in 2019–20.

Gini coefficient
The Gini coefficient is a single statistical indicator of the degree of inequality. It equals zero when all people have the same level of income and equals one when one person receives all the income.
In general the smaller the Gini coefficient, the more equal the distribution of income or wealth. Any increase in the income of a person with income greater than the median will always lead to an increase in the Gini coefficient, while an increase in the income of a person with income lower than the median will always lead to a decrease in the coefficient.

A time series analysis of the Gini coefficient can be found in fact sheet 4.

Sampling Error
Household survey estimates are based on a sample of possible observations and are subject to sampling variability. The sampling error is a measure of the variability that occurs by chance because a sample, rather than the entire population, is surveyed. One measure of  the likely difference is given by the standard error (SE). Another measure of the likely difference is the relative standard error (RSE), which is obtained by expressing the SE as a percentage of the estimate. The RSE is a useful measure in that it provides an immediate indication of the percentage errors likely to have occurred due to sampling, and thus avoids the need to refer also to the size of the estimate. 

The ABS annotates estimates with a RSE between 25% and less than 50% by a preceding asterisk (e.g. *3.4) to indicate they are subject to high SEs and should be used with caution. Estimates with RSEs of 50% or more are preceded with a double asterisk (e.g. **0.6),  indicating that these estimates are considered unreliable for most purposes.

Another measure of sampling error is the Margin of Error (MOE). This describes the distance from the population value that the sample estimate is likely to be within and is particularly useful to understand the accuracy of proportion estimates. It is specified at a given level of confidence. Confidence levels typically used are 90%, 95% and 99%.

The ABS calculates MOEs at the 95% confidence level. Proportion estimates with a MOE greater than 10% are annotated with a preceding hash sign (e.g. #3.4%) to indicate they have a high margin of error and should be used with caution.

Significance Testing
To compare estimates between surveys or between populations within a survey it is important to determine whether apparent differences are ‘real’ or simply the product of differences between the survey samples. A common approach is to determine whether the difference  between the estimates is statistically significant by calculating the standard error of the difference between two estimates (x and y) and using that to calculate the test statistic using the formula below:

If the value is greater than 1.96 there is good evidence of a statistically significant difference at 95% confidence levels between the two populations for the characteristic being tested. Otherwise, it cannot be stated with confidence that there is a real difference between the populations.

For more information:

• United Nations, 2011, Canberra Group Handbook on Household Income Statistics, Second Edition, United Nations Economic Commission for Europe

People living in low economic resource households are of particular policy and research interest because of their greater risk of experiencing economic hardship. This fact sheet summarises different methods available to identify these households and provides guidance on methods of analysing them.

There are many factors influencing whether people are experiencing economic hardship. The analysis of household economic well-being is enhanced significantly when the income, consumption and wealth dimensions are studied jointly, recognising they vary over the life cycle:

• income is affected by workforce participation
• wealth tends to be accumulated during people’s working life and then consumed in retirement
• younger people may have higher expenditure needs e.g. to buy a home or start a family.

In recognition of the importance of this, the ABS has collected both income and wealth in every Survey of Income and Housing (SIH) from 2003–04 (apart from 2007–08). The ABS Household Expenditure Survey (HES) has been conducted six yearly since 2003–04 on a sub sample of SIH households. Expenditure, financial stress, income and wealth data are available for HES households.

Socio-Economic Indexes for Areas (SEIFA) 
Census data (including education, employment, occupation, income and housing) has been used by the ABS to identify the relative socio-economic advantage and disadvantage of geographic areas in Australia compared with other areas.

The 2016 SEIFA includes an Index of Relative Socio- Economic Disadvantage (IRSD) and an Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD). 

As well as being used to analyse Census data, the IRSD and IRSAD by decile and/or quintile have also been added to survey files (including CURFs) for household surveys such as the SIH, HES and GSS from 2002 onwards.

For more information: Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia, 2016 (2033.0.55.001).

Low economic resource measure
The ABS has developed a low economic resource measure (LER) that includes people who are simultaneously in the lowest four deciles of both equivalised disposable household income (EDHI) (including imputed rent) and equivalised household net worth (LER40). This measure therefore excludes people with either relatively high incomes or relatively high wealth. As a result it is more likely to correctly classify people at risk of experiencing economic hardship compared to measures using income or wealth alone.

The LER is a relative measure that classifies around 20% of people in low income, low wealth households. It does not identify whether these people are actually experiencing economic hardship. The actual proportion will vary over time as the joint distribution of income and wealth changes. One of the strengths of this measure is its ability to contrast the characteristics of the LER population with those in the low income and low wealth quintiles. Table 1 compares selected characteristics of LER households to households with low income or low wealth only. The proportion of couple or lone person households where the reference person is 65 and over, reduces from 18% of low income households to 8% of LER households, reflecting their ability to draw on accumulated wealth.

In contrast, whilst 36% of low income households are private renters, this group accounts for 53% of LER households.
 

The LER measure can be broadened by considering experiences of ‘financial stress’ or ‘missing out’. The indicators used to define these measures are listed in Table 3 of this fact sheet. 

Graph 1 shows examples of LER measures by:

• varying the cut-off for low income and low wealth (40th percentile (LER40) or 30th percentile (LER30)), then
• adding whether the household experienced ‘financial stress’ or ‘missing out’.

In 2015–16, 24% of people lived in LER40 households and 16% in LER30 households. When experiences of ‘financial stress’ were also considered this reduced to 14% of LER40 and 10% of LER30 households.

When measuring economic well-being it is preferable to consider multiple dimensions, particularly income and wealth, however both measures are not always available. This section describes several commonly used single dimension measures of economic well-being.

Income
Income is the most frequently available measure of economic well-being. For most households, it is the main resource used to meet daily expenses. However, analysis using income alone has significant limitations. Income can be volatile for people who are making transitions between study, jobs, into retirement or changing their hours of work e.g. to care for children. At these times, households may draw on other resources, such as using savings or increasing their debt.
Being able to identify households with accumulated wealth to supplement low incomes is desirable as these people are less likely to experience economic hardship than households without alternative resources to fall back on.

a) Relative poverty measures based on income 
Many developed countries use relative poverty to measure the economic well-being of households. These measures identify the proportion of people with an income below a certain fraction of median EDHI. The OECD publishes various analyses based on poverty lines below 40%, 50% or 60% of median incomes (50% used most often), while Eurostat commonly uses 60% as the cut-off.
Limitations of relative poverty measures include:

• the number of people in poverty is determined by an arbitrary fraction of income (which may not reflect actual hardship).
• the proportion of people identified can change dramatically e.g. in Australia, real median incomes have risen significantly in recent years and the thresholds identified at 40% and 50% of the median are very sensitive to changes in single and couple pension payment points relative to the median.
• the definition and measurement of income can have a significant impact e.g. imputed rent (IR) and social transfers in kind (STIK) are often excluded from income definitions. However, the benefits received from either owning a home or receiving subsided rent (valued by imputing an equivalent rental income), or from receiving services from the government, impact significantly on the economic well-being of particular groups e.g. a person able to access free or subsidised health care can be better off than a person with similar income but not able to access these social provisions.

Table 2 shows that the proportion of the Australian population below a relative poverty line varies between 19% (using 60% of median EDHI) and 2% (using 40% of median EDHI including IR and STIK).

b) Low income households
While it is tempting to label all households in the lowest income decile as ‘low income’, ABS analysis suggests there are variable economic circumstances for households in this group. The ABS defines low income households as households in the lowest equivalised disposable household income quintile, excluding the 1st and 2nd percentiles (i.e. the 3rd to 20th percentiles inclusive). The 1st and 2nd percentiles are excluded due to the high wealth and expenditure characteristics those household exhibit, and the prevalence of income types other than employee income and government pensions and allowances.

While income and wealth statistics describe the economic resources available to people and expenditure statistics describe their consumption patterns, there are other issues relevant to understanding living standards e.g. a person with poor health and high health care costs may have reduced income for other purchases. In attempting to identify which households have the lowest economic well-being, other indicators of poor economic outcomes can be considered. Data relating to experiences of financial stress and missing out are collected in the HES.

Indicators of financial stress in the last 12 months

Financial stress experiences:

• Unable to raise $2000 in a week for something important
• Spend more money than received
• Could not pay gas, electricity or telephone bill on time
• Could not pay registration or insurance on time
• Pawned or sold something
• Went without meals
• Unable to heat home
• Sought assistance from welfare/community organisations
• Sought financial help from friends or family

Missing out experience:

• Could not afford holiday for at least one week a year
• Could not afford a night out once a fortnight
• Could not afford friends or family over for a meal once a month
• Could not afford special meal once a week
• Could only afford second hand clothes most of the time
• Could not afford leisure or hobby activities

Financial stress information can provide insight into people’s economic well-being although analysis needs to consider overall circumstances. Some individuals may have consumption priorities which differ from socially accepted norms of the ‘basics of life’. In 2015–16, 14% of household in the highest EDHI quintile reported at least one financial stress indicator (Graph 2). As in previous years, the most commonly reported indicators of financial stress were the inability to afford a holiday for at least one week a year (23%), followed by an inability to afford a night out once a fortnight (17%).

Another key policy interest is people experiencing longterm and persistent economic hardship as distinct from those experiencing short-term hardship.

Longitudinal datasets, such as the Household Income and Labour Dynamics Australia Survey (HILDA) and the ABS Australian Census Longitudinal Dataset (ACLD), are important sources for identifying people experiencing long-term economic hardship. The HILDA has been tracking the economic circumstances of more than 17,000 respondents since 2001. The (ACLD) will provide a five-yearly snapshot of the income and housing circumstances of  people from 2006.

The SIH measures the short-term persistence of economic hardship by comparing income from the previous financial year with current year income. The circumstances of people with low incomes in both periods can be identified. Combined with wealth data which is more stable over time, this provides a more accurate picture of whether hardship is persistent.

As well as financial stress experiences, the HES also collects data on people’s perception of their current financial circumstances compared to two years ago and their ability to save money. 

For more information:

• Australian Bureau of Statistics (ABS) 2013, Household Income and Income Distribution, Australia, 2011–12, (cat. no. 6523.0), Feature article: Low Economic Resource Households
• Australian Bureau of Statistics (ABS) 2017, Household Expenditure Survey, Australia: Summary of Results, 2015–16, Financial stress and spending
• McLachlan, R., Gilfillan, G. and Gordon, J. 2013, Deep and Persistent Disadvantage in Australia, Productivity Commission Staff Working Paper, Canberra
   

Changes in the levels and distribution of economic resources in a society over time are key concerns of social and economic analysts. 
This fact sheet presents time series analysis of the three dimensions of household economic well-being – income, consumption and wealth.

The analysis uses data from the Survey of Income and Housing (SIH) and Household Expenditure Survey (HES).

• Median equivalised disposable household income (EDHI) has increased in real terms from $585 in 1994-95 to $959 in 2019–20 (up 64%)
• Wages and salaries increased 52% in real terms between 1994–95 and 2019–20.
• Government pensions and allowances increased 24% in real terms between 1994–95 and 2019–20.
• Average weekly household expenditure between 1984 and 2015–16  increased in real terms by one third from $1,065 to $1,425. 
• Median net worth has increased in real terms from $519,300 in 2009–10 to $579,200 in 2019–20. 
   

Income data has been collected in the HES since 1984 and in the SIH since 1994–95. Since 1994–95, median equivalised disposable household income (EDHI) has increased in real terms from $585 to $959 (up 64%). Low income households have had a slightly lower real increase in their average income (59% at top of P10) than high income households (69% at top of P90) (Graph 1).

Average wages and salaries and government pensions and allowances both increased significantly in real terms between 1994–95 and 2011–12 (52% and 24%, respectively). A small part of the measured increase since 2003–04 was due to improvements in the compilation of income introduced in SIH 2007–08 and recompiled where data was available for 2003–04 and 2005–06.

Improvements in the SIH since 2003–04

The ABS has implemented improvements to the SIH to ensure the survey accurately measures the distribution of economic resources among households in Australia, including:

2003–04

• Integration of the SIH with the HES 
• Computer assisted personal interviewing (CAPI) introduced
• Sample design improved
• Extra income questions (incl. non-cash and irregular income; salary sacrificed income specifically collected)
• New benchmarking methods
• Wealth data and imputed rent for first time

2007–08

• Further improvements to income incl. lump sum payments, financial support from family and trusts
• Implementation of new income definition incl. recompiling 2003–04 and 2005–06 where possible

2009–10

• Wealth data every SIH
• SIH income and wealth comparison with Australian System of National Accounts (ASNA) published in appendices of 6523.0 and 6554.0 

2011–12

• Social transfers in kind (STIK) allocated in every SIH

2013–14

• Previous HES only items incl. disability and health care cards in every SIH to improve STIK allocations
• More detailed superannuation information

2015–16

• Improved processing of government payments information in the SIH and HES through the introduction of an eligibility-based model. Missing or anomalous government payments values are now produced by the model
• Improved methodology for imputed rent which are outlined further in the 'Imputed rent' chapter of this publication

2017–18

• Credit card and HECS/HELP debt information is now collected at the person level and aggregated to the household level. Previously collected at the household level

2019–20

• Computer assisted web interviewing (CAWI) introduced

The Gini coefficient is a single statistic between zero and one and is a summary indicator of the degree of inequality, with values closer to 0 representing less inequality, and values closer to one representing greater inequality. Since 1994–95, the Gini coefficient for EDHI has been lowest in 1996–97 (0.292) and highest in 2007–08 (0.336). It decreased by 4% between 2007–08 and 2019–20 (Graph 3).

As incomes have risen, consumption expenditure has also risen. Between 1984 and 2015–16 average weekly expenditure of all households increased in real terms by one third from $1,065 to $1,425. The increase in expenditure was greatest for households in the fourth and fifth gross income quintiles. In these quintiles, average income exceeded average consumption expenditure by 22% and 44%, respectively in 2015–16. By comparison, households in the lowest two income quintiles had average expenditure higher than their average disposable income (Graph 4).

Consumption patterns of households have changed since 1984. Current housing costs increased from 13% of total household expenditure on goods and services in 1984 to 20% in 2015–16. The proportion of expenditure on food and non-alcoholic beverages declined gradually in the same period (from 20% to 17% of total consumption expenditure), while spending on clothing and footwear has halved (from 7% to 3% of total) (Graph 5).

Improvements in the HES since 1998–99

1998–99

• Household Expenditure Classification (HEC) replaced HES Commodity Code List for classifying expenditure
• Financial stress indicators collected for first time

2003–04

• HES and SIH integrated (HES for a subsample of SIH respondents)
• Expenditure, income, wealth and financial stress available for all HES households

2009–10

• Non-cash benefits from employers included in consumption expenditure
• Expenditure also classified by the international Classification of Individual Consumption by Purpose (COICOP)
• Extra metropolitan sample of households with main source of income government pensions and allowances added to HES for development of a Pensioner and Beneficiary Living Cost Index
• HES expenditure comparison with the ASNA published in Appendix 3 of Household Expenditure Survey (6530.0)

The distribution of wealth in Australia is less equal than income. Comprehensive information on the composition of the assets and liabilities held by households has been collected in the SIH and HES since 2003–04. Previously, the value of owner occupied dwellings and loans on those dwellings were the only wealth data collected in these surveys.

Median net worth has increased in real terms from $519,300 in 2009–10 to $579,200 in 2019–20. The average net worth of high wealth households has increased by more than the net worth of low wealth households e.g. the net worth of households at the top of the fourth quintile (P80) increased by 25% (to $1.4m) while the net worth of households at the top of the lowest quintile (P20) increased by 7% (to $113,400) in the ten year period to 2019–20 (Graph 6).

The composition of assets has remained relatively stable between 2009–10 and 2019–20. There was a slight reduction in the proportion for property assets (own dwelling and other property), decreasing from 60% in 2009–10 to 56% of total household assets in 2019–20. Superannuation rose from 14% to 19% of total household assets in the same period (Graph 7).

Property loans made up a slightly higher proportion of liabilities in 2019–20 (90%) than in 2009–10 (88%).

For more information:

• ABS, 2009, Household Income and Income Distribution, Australia 2007–08, (cat. no. 6523.0), Appendix 4: Improvements to income statistics, ABS, Canberra