OECD Economics Department Working - Sipotra · ECO/WKP(2019)7 Unclassified English - Or. English ....
Transcript of OECD Economics Department Working - Sipotra · ECO/WKP(2019)7 Unclassified English - Or. English ....
Please cite this paper as:
Sila, U. and V. Dugain (2019), “Income, wealth and earningsinequality in Australia: Evidence from the HILDA survey”,OECD Economics Department Working Papers, No. 1538,OECD Publishing, Paris.http://dx.doi.org/10.1787/cab6789d-en
OECD Economics Department WorkingPapers No. 1538
Income, wealth and earningsinequality in Australia
EVIDENCE FROM THE HILDA SURVEY
Urban Sila, Valéry Dugain
JEL Classification: D31, E24, J2, J3
Organisation for Economic Co-operation and Development
ECO/WKP(2019)7
Unclassified English - Or. English
14 February 2019
ECONOMICS DEPARTMENT
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA:
EVIDENCE FROM THE HILDA SURVEY
ECONOMICS DEPARTMENT WORKING PAPERS No. 1538
By Urban Sila and Valéry Dugain
OECD Working Papers should not be reported as representing the official views of the OECD
or of its member countries. The opinions expressed and arguments employed are those of the
author(s).
Authorised for publication by Isabell Koske, Deputy Director, Country Studies Branch,
Economics Department.
All Economics Department Working Papers are available at www.oecd.org/eco/workingpapers.
JT03443161
This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the
delimitation of international frontiers and boundaries and to the name of any territory, city or area.
2 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
OECD Working Papers should not be reported as representing the official views of the OECD or of its member countries. The opinions expressed and arguments employed are those of the author(s). Working Papers describe preliminary results or research in progress by the author(s) and are published to stimulate discussion on a broad range of issues on which the OECD works. Comments on Working Papers are welcomed, and may be sent to OECD Economics Department, 2 rue André Pascal, 75775 Paris Cedex 16, France, or by e-mail to [email protected]. All Economics Department Working Papers are available at www.oecd.org/eco/workingpapers.
This document and any map included herein are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law. On 3 May 2018, the OECD Council invited Lithuania to become a Member. At the time of preparation the deposit of Lithuania’s instrument of accession to the OECD Convention was pending and therefore Lithuania does not appear in the list of OECD Members and is not included in the OECD zone aggregates. On 25 May 2018, the OECD Council invited Colombia to become a Member. At the time of preparation the deposit of Colombia’s instrument of accession to the OECD Convention was pending and therefore Colombia does not appear in the list of OECD Members and is not included in the OECD zone aggregates. © OECD (2019)
You can copy, download or print OECD content for your own use, and you can include excerpts from OECD publications, databases and multimedia products in your own documents, presentations, blogs, websites and teaching materials, provided that suitable acknowledgment of OECD as source and copyright owner is given. All requests for commercial use and translation rights should be submitted to [email protected]
ECO/WKP(2019)7 │ 3
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Abstract / Résumé
Income, wealth and earnings inequality in Australia: evidence from the HILDA
survey
This paper analyses income, wealth and earnings inequality in Australia, using the
Household, Income and Labour Dynamics in Australia (HILDA) Survey as the primary
source of data. Income inequality in Australia has risen in the last two decades, but most of
the rise occurred prior to the global financial crisis. HILDA data nevertheless show
evidence of slower income growth in the middle of the income distribution compared with
the top and the bottom. While Australia has experienced a rising inequality in wages –
mostly through rapid earnings increases among top earners - this has been offset by
increased participation and longer hours worked at the bottom of the distribution.
According to HILDA data, relative pay across different levels of education groups has not
recorded large shifts over the last 15 years. At the same time, we find evidence for job
polarisation; notably, the share of high skilled jobs versus middle skilled jobs has increased.
With respect to concerns about the casualisation of the labour force and less stable nature
of jobs amid technological change and globalisation, the incidence of casual employment
– where workers receive no paid sick leave or holiday leave - in Australia has been reported
to have risen since the 1980s, especially for females. According to HILDA data however,
the incidence of casual employment has fallen since early 2000s. Furthermore, we find no
evidence that contract duration has shortened over time.
JEL Codes: D31, E24, J2, J3
Keywords: Australia, HILDA, household panel, income distribution, inequality, wealth inequality, income
mobility, earnings inequality, job polarisation
This Working Paper relates to the 2018 OECD Economic Survey of Australia
http://www.oecd.org/eco/surveys/economic-survey-australia.htm.
****************************
Inégalités de revenu, de patrimoine et de rémunération en Australie : enseignements
de l'enquête HILDA
Nous analysons dans ce document les inégalités de revenu, de patrimoine et de
rémunération en Australie, en utilisant comme source primaire de données l'enquête sur les
ménages, les revenus et la dynamique du marché du travail en Australie (HILDA,
Household, Income and Labour Dynamics in Australia). Les inégalités de revenu se sont
creusées dans ce pays au cours des deux dernières décennies, mais cette évolution a eu lieu
essentiellement avant la crise financière mondiale. Les données de l'enquête HILDA
montrent cependant que la croissance des revenus s'est ralentie au milieu de la distribution
des revenus par rapport à ses parties supérieure et inférieure. L'Australie a connu une
augmentation des inégalités salariales (essentiellement due à la progression rapide des
rémunérations les plus élevées), mais celle-ci a été compensée par une hausse du taux
d'activité et un allongement du temps de travail dans la partie inférieure de la distribution.
D'après les données de l'enquête HILDA, les niveaux de rémunération relatifs des
différentes catégories de la population ventilée en fonction du niveau d'études n'ont pas
sensiblement changé au cours des 15 dernières années. Par ailleurs, certains éléments
mettent en évidence une polarisation des emplois ; la proportion d'emplois hautement
qualifiés a notamment augmenté par rapport à celle des emplois moyennement qualifiés.
S'agissant des préoccupations relatives à la précarisation de la main-d'œuvre et à la remise
4 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
en cause de la stabilité des emplois sur fond d'évolutions technologiques et de
mondialisation, l'incidence de l'emploi occasionnel (caractérisé par l'absence de congés de
maladie payés et de congés payés annuels) a apparemment augmenté en Australie depuis
les années 1980, en particulier parmi les femmes. Néanmoins, d'après les données de
l'enquête HILDA, l'incidence du travail occasionnel a diminué depuis le début des années
2000. En outre, aucun élément n'indique que la durée des contrats s'est raccourcie au fil du
temps.
Codes JEL : D31, E24, J2, J3
Mots clés : Australie ; HILDA ; Panel de ménages ; distribution de revenus ; inégalité ; inégalité de richesse ;
mobilité de revenu ; inégalité de revenu ; polarisation du travail
Ce document de travail est lié à l'Étude économique de l'OCDE de 2018 consacrée à
l'Australie http://www.oecd.org/fr/eco/etudes/etude-economique-australie.htm.
ECO/WKP(2019)7 │ 5
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Table of contents
Income, wealth and earnings inequality in Australia: evidence from the HILDA survey .............. 7
Introduction .......................................................................................................................................... 7 Income inequality based on the OECD Income Inequality database ................................................... 8 Income inequality based on the HILDA Survey .................................................................................. 8 Income inequality across regions ....................................................................................................... 16 Income mobility ................................................................................................................................. 18 Wealth inequality ............................................................................................................................... 21 Labour market income inequality ...................................................................................................... 25 Conclusion ......................................................................................................................................... 38 References .......................................................................................................................................... 38
Figures
Figure 1. Income inequality in Australia has risen and remains above the OECD average .................... 9 Figure 2. Gini coefficient of equivalised household disposable income ............................................... 10 Figure 3. Equivalised median disposable household income ................................................................ 11 Figure 4. Real disposable income inequality - decile ratios since 2001 ................................................ 12 Figure 5. Gini coefficient of personal disposable income ..................................................................... 13 Figure 6. Median disposable income across quintiles (2001, 2008, 2016) ............................................ 14 Figure 7. Real disposable income decile ratios ..................................................................................... 15 Figure 8. Composition of gross household income by deciles .............................................................. 16 Figure 9. Gross household income components across quintiles over time .......................................... 17 Figure 10. Equivalised disposable household income across states and remoteness ............................ 18 Figure 11. Income mobility - short-term ............................................................................................... 19 Figure 12. Income mobility - longer-term ............................................................................................. 20 Figure 13. Wealth shares of top percentiles of the net wealth distribution in selected OECD
countries ........................................................................................................................................ 22 Figure 14. Gini coefficient of net worth (HILDA) ................................................................................ 23 Figure 15. Real household median net worth (equivalence scale) ........................................................ 23 Figure 16. Composition of total household assets (2002 and 2014)...................................................... 24 Figure 17. Total household debt as a share of total assets ..................................................................... 24 Figure 18. Share of highly indebted households ................................................................................... 25 Figure 19. Median total annual pay across deciles of gross labour income .......................................... 26 Figure 20. Real earnings percentile ratios (age 15-64) .......................................................................... 26 Figure 21. Median annual hours of work across deciles of labour market income ............................... 27 Figure 22. Incidence of part-time work ................................................................................................. 28 Figure 23. Female participation and the incidence of part-time work have risen ................................. 28 Figure 24. Decile ratios of gross annual earnings (1995-2016)............................................................. 29 Figure 25. Median weekly pay across deciles of gross labour income ................................................. 30 Figure 26. Decile earnings ratios (full-time employed) by gender ........................................................ 31
6 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 27. Median earnings (full-time employed) by level of education .............................................. 32 Figure 28. Real wages by occupation .................................................................................................... 33 Figure 29. Real median wages and salaries for occupations requiring different skills levels (2001 to
2016) .............................................................................................................................................. 33 Figure 30. Employment shares by occupation (2001-2016) .................................................................. 34 Figure 31. Changes in employment shares by skill groups (2001-2016) .............................................. 35 Figure 32. Share of non-standard employment ..................................................................................... 36 Figure 33. Share of workers in casual jobs, by gender, 2001-2016 ....................................................... 37 Figure 34. Duration of current jobs, by gender, 2001-2016 .................................................................. 37
Boxes
Box 1. HILDA Survey ........................................................................................................................... 10
ECO/WKP(2019)7 │ 7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Income, wealth and earnings inequality in Australia: evidence from the
HILDA survey
By Urban Sila & Valéry Dugain1
Introduction
1. Inequality has been rising over the last 30 years in many OECD countries (OECD,
2015). By lowering employment opportunities, the financial crisis of 2008-9 and its
aftermath accelerated this trend in many countries, and social issues have come to the
forefront of public and political debates. A widening gap between rich and the poor can
have detrimental effects not only on social cohesion but also on economic growth (OECD,
2015).
2. Two other developments - globalisation and technological progress - have strongly
influenced the structure of OECD economies and the nature of their labour markets. First,
under globalisation, much of the world manufacturing production has shifted to emerging
markets with often detrimental impact on local labour markets in developed economies.
Second, technological progress has favoured certain jobs and skills. With the automation
of routine tasks, technological progress has reduced demand for medium-skill workers,
while increasing demand particularly for high-skilled, but also for low-skilled jobs (Autor
et al., 2006; Goos and Manning, 2007; and Goos et al., 2009; OECD, 2017). Indisputably,
both globalisation and technological progress have brought considerable prosperity, but
nevertheless certain groups have been - at least temporarily - stripped of jobs and
opportunities. This evolution in labour demand looks set to continue and in absence of
effective policy action more workers may struggle to secure quality jobs.
3. In this paper we focus on describing the trends in Australia’s inequality, using the
Household, Income and Labour Dynamics in Australia (HILDA) Survey as the main source
of data. The HILDA Survey is a household-based panel study that collects information
about economic and family life across Australia. It has been conducted annually since 2001
(see Box 1). The data can be used to compute various measures of inequality and the panel
structure of the data allows for assessment over time. The findings in this paper are largely
consistent with the work by other researchers on Australia (Wilkins, 2013, 2015 and 2017;
Fletcher and Guttman, 2013; Dollman et al. 2015, Greenville et al., 2013, Borland and
Coelli, 2016, and Productivity Commission, 2018).
1 Urban Sila is an economist in the Country Studies Branch of the OECD Economics Department. Valéry
Dugain served as a consultant in the OECD Economics Department when research for the paper was done. For
valuable comments and suggestions the authors would like to thank Philip Hemmings and Patrick Lenain (both
from OECD Economics Department), Michael Förster, Andrea Salvatori, Maxime Ladaique and Alexandre
Georgieff (all from OECD Employment, Labour and Social Affairs Directorate), Jonathan Coppel and Josh
Craig (both Productivity Commission). Editorial assistance from Stephanie Henry was also greatly appreciated.
8 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Income inequality based on the OECD Income Inequality database
4. According to the OECD Income Inequality database, income inequality (measured
with the Gini coefficient or 90/10 percentile ratio) in disposable income (after taxes and
transfers) in Australia increased from 1995 up to the global financial crisis, but appears to
have stabilised since then (Figure 1, panels A and B). Inequality in disposable income is
slightly above the OECD average (panel C). In Australia the contribution from transfers
and taxes to reducing inequality (for the working-age population) is close to the OECD
average (panel D). The impact of transfers and taxes on inequality reduction has however
gone down since 1995 (not shown), as in many OECD countries. In Australia,
unemployment benefits are means-tested and have strict eligibility criteria, and up to half
of displaced workers are in effect ineligible to receive them (OECD, 2016a). This reflects
a highly targeted welfare system; more than 40% of cash transfers go to low income
households compared to just below 25% on average in the OECD (Causa and Hermansen,
2017).
Income inequality based on the HILDA Survey
5. We now turn to income inequality data based on the HILDA Survey. Studies that
look at income inequality mostly base their findings on household income. The rationale
being that even if a person has low personal income they may still live in high-income
households, as for example young students or non-earning spouses of high-earning
partners, and they should therefore not be counted as "poor". In addition, the household
perspective is the most appropriate as this is the key economic and social unit where
resources are pooled and where economic, family and other decisions are taken. Following
the methodology in the OECD Income Distribution Database (see Box 1 in OECD, 2016b)
we keep the individual as a unit of observation, but we assign each individual an income
that is equal to the total household income divided by the square root of the number of
individuals (of all ages) in the household. In this way we obtain an equivalised household
income.
6. As this measure is based on household-level income, individuals of all ages should
be included. For the purposes of putting the income on equivalence scale (dividing by the
square root of the number of people in the household) everybody is counted. However, for
computing the distribution of equivalised income, the age limit is 15 and above, because
HILDA database does not provide full information and sample weights for individuals of
less than 15 years of age.
Inequality in disposable household income
7. Inequality of disposable income measured by HILDA-based Gini coefficients has
remained roughly constant over the past 15 years (Figure 2). Inequality of gross income is
significantly larger than inequality of disposable income, confirming the dampening impact
of taxes and transfers on income inequality.
ECO/WKP(2019)7 │ 9
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 1. Income inequality in Australia has risen and remains above the OECD average
Note: Panels A.-C. refer to whole population. Panel D. is for working age population and indicates the
difference between the Gini before and after taxes and transfers.
Source: OECD, Income Inequality database.
0.00
0.05
0.10
0.15
0.20
0.25
0.00
0.05
0.10
0.15
0.20
0.25
ME
X
TUR
CH
L
KO
R
CH
E
JPN
NZ
L
LVA
ES
T
US
A
ISR
CA
N
ISL
SW
E
SV
K
OE
CD
GB
R
AU
S
PO
L
NO
R
ITA
HU
N
DE
U
NLD
LUX
CZ
E
ES
P
PR
T
DN
K
AU
T
SV
N
FRA
GR
C
BE
L
FIN
IRL
PointsPoints
D. Impact of taxes and transfers on inequality reduction
Reduction in market Gini due to taxation (2016 or latest)
Reduction in market Gini due to transfers (2016 or latest)
Split n/a (2016 or latest)
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
1995 2000 2005 2010 2015
A. Gini (disposable income, post taxes and transfers)
AUS
3.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
4.6
4.8
5.0
1995 2000 2005 2010 2015
B. P90/P10 disposable income decile ratio
0
0.1
0.2
0.3
0.4
0.5
0
0.1
0.2
0.3
0.4
0.5
ISL
SV
N
SV
K
CZ
E
FIN
DN
K
BE
L
NO
R
AU
T
SW
E
NLD
HU
N
PO
L
DE
U
FRA
KO
R
CH
E
IRL
LUX
OE
CD
CA
N
ES
T
JPN
ITA
PR
T
AU
S
GR
C
ES
P
ISR
LVA
NZ
L
GB
R
US
A
TUR
CH
L
ME
X
C. Gini (disposable income, post taxes and transfers)
2016 or latest
10 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Box 1. HILDA Survey
The Household, Income and Labour Dynamics in Australia (HILDA) Survey is a
household-based panel study, that started in 2001 and collects data on about 17,000
Australians each year. The data cover many aspects of life, including household and
family relationships, child care, income and employment, education, expenditure,
health and wellbeing, and other life events. At less frequent intervals the survey collects
additional information on various topics, as for example on household wealth, which
has been conducted every four years since the second wave in 2002.
As this is a panel data set, participants are surveyed every year and population weights
are provided so that statistics computed from the data can represent estimates for the
Australian population. For wave 1 of the survey, households were selected such that
representativeness of the reference population was ensured. Children born or adopted
in these households also become members of the sample. All members of the selected
households count as members of the sample, although individual interviews are only
conducted with those aged 15 years and over.
Shifts in population composition (for instance due to immigration) and sample attrition
(e.g. participants dropping out due to refusal to participate or problems in locating them)
make a sample less representative of the whole population over time. To correct for
immigration, in wave 11, a general sample top-up was conducted which allowed
immigrants who had arrived between 2001 and 2011 to enter the HILDA Survey
sample. To correct for attrition, sample weights are changed each year to adjust for
differences between the characteristics of the panel sample and the characteristics of the
Australian population.
The HILDA Survey is funded by the Australian Government through the Department
of Social Services. The Melbourne Institute is responsible for the design and
management of the Survey. For more information visit
http://melbourneinstitute.unimelb.edu.au/hilda.
Figure 2. Gini coefficient of equivalised household disposable income
Source: OECD calculations based on HILDA database.
0.30
0.32
0.34
0.36
0.38
0.40
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Gini equivalised gross household income Gini equivalised disposable household income
ECO/WKP(2019)7 │ 11
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
8. In Figure 3 we look at changes in incomes across the income distribution; it depicts
median real disposable household income across quintiles (panel A) and the index of
income across quintiles with the base year 2001 (panels B and C). Over the last 15 years,
income has grown significantly in real terms across the whole of the income distribution,
but it has grown the most - in percentage terms - for individuals in the bottom two quintiles.
The third highest increase was in the top quintile. Incomes of the rich and the poor have
therefore grown faster than that of the middle class. Interestingly, except for the bottom
quintile, the growth of incomes for all groups slowed significantly following the global
financial crisis.
Figure 3. Equivalised median disposable household income
A. Respondents aged 15 years old and over
B. Index 2001=100 C. Index 2001=100
Source: OECD calculations based on HILDA database.
9. Next we consider various income decile ratios based on household income (90/10,
90/50 and 50/10 decile ratio), shown in Figure 4. Again, no big shifts in income inequality
0
20,000
40,000
60,000
80,000
100,000
120,000
1st Quintile 2nd Quintile 3rd Quintile Sample mean 4th Quintile 5th Quintile
2016 2008 2001
100
105
110
115
120
125
130
135
140
145
1stQuintile
2ndQuintile
3rdQuintile
Samplemean
4thQuintile
5thQuintile
2016 2008
100
105
110
115
120
125
130
135
140
145
2001 04 07 10 13 16
1st Quintile
2nd Quintile
3rd Quintile
4th Quintile
5th Quintile
12 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
are apparent over the 15 year period, with a slight decline, especially at the bottom, after
the global financial crisis. We can observe a clear fall in 90/10 inequality post 2009,
primarily driven by a recent drop in inequality at the bottom (50/10 ratio), while at the top
(90/50 ratio) inequality has moved around but reached in 2016 a similar level it had in
2001. Such developments lend further evidence that the bottom and the top of the
distribution have gained against the middle.
Figure 4. Real disposable income inequality - decile ratios since 2001
Equivalised household income, respondents aged 15 and over
Source: OECD calculations based on HILDA database.
Inequality in personal income
10. In this section we briefly look at inequality based on personal income rather than
household income. This is done for comparison and also because later we move onto
earnings inequality, which more closely corresponds to personal income inequality because
earnings are based on individual level data. We analyse personal level inequality for
individuals aged 20 years or more.
11. Gini coefficients over time show roughly constant inequality over the past 15 years
(Figure 5). As expected, the level of the Gini coefficient at the personal level is higher than
at the household level.
0.0
1.0
2.0
3.0
4.0
5.0
2001 04 07 10 13 16
Ratio 90/10
Ratio 50/10
Ratio 90/50
70
80
90
100
110
2001 04 07 10 13 16
Ratio 90/10
Ratio 50/10
Ratio 90/50
ECO/WKP(2019)7 │ 13
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 5. Gini coefficient of personal disposable income
Persons aged 20 and above.
Source: OECD calculations based on HILDA database.
12. Inequality across personal income quintiles, as for household incomes, shows
fastest growth for individuals in the bottom two quintiles (Figure 6). The third highest
increase was in the top quintile. However, we can observe a peculiar evolution of personal
income of the bottom quintile over time that is not apparent in household income.
0.36
0.38
0.40
0.42
0.44
0.46
0.48
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Gini gross income Gini disposable income
14 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 6. Median disposable income across quintiles (2001, 2008, 2016)
A. Respondents aged 20 years old and over
B. Index 2001=100 C. Index 2001=100
Source: OECD calculations based on HILDA database.
13. Finally, examining decile ratios (Figure 7) reveals a similar picture of roughly
unchanged overall income inequality (90/10 ratio) from the starting point to the ending
point of the 15-year period, but quite some movement in between. The major driver of the
fall in inequality since 2011 has been the bottom 50/10 decile ratio. On the other hand,
based on personal income, top inequality (90/50 ratio) has risen since 2010.
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
1st Quintile 2nd Quintile 3rd Quintile Sample mean 4th Quintile 5th Quintile
2016 2008 2001
100
105
110
115
120
125
130
135
140
145
1stQuintile
2ndQuintile
3rdQuintile
Samplemean
4thQuintile
5thQuintile
2016 2008
100
105
110
115
120
125
130
135
140
145
2001 04 07 10 13 16
1st Quintile
2nd Quintile
3rd Quintile
4th Quintile
5th Quintile
ECO/WKP(2019)7 │ 15
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 7. Real disposable income decile ratios
Individual level, respondents aged 20 and over
Source: OECD calculations based on HILDA database.
Composition of income across the income distribution
14. Individuals at various points of the income distribution differ in their composition
of gross income. Figure 8, panel A, depicts the composition of total gross income divided
into labour income, public transfers, investment income and other income for the fiscal
year 2015-16.
15. Households at the bottom receive a large share of their income from public transfers
(mostly pensions and various allowances). Unsurprisingly, this is especially so for those
aged 65 years and above. For working-age cohorts in particular, the higher we move up the
income distribution, the higher is the share of labour market income; labour market income
is generally strongly correlated with total gross income. For the top 10th decile, however,
a large share of income comes from investment income and other income such as business
and irregular income.
0.0
2.5
5.0
7.5
10.0
2001 04 07 10 13 16
Ratio 90/10 Ratio 50/10 Ratio 90/50
70
80
90
100
110
2001 04 07 10 13 16
Ratio 90/10 Ratio 50/10 Ratio 90/50
16 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 8. Composition of gross household income by deciles
A. 20 to 64 years old B. 65 years old and over
Source: OECD calculations based on HILDA database.
16. Over time, for all groups of working age, the growth in labour income was a major
source in the rise of the dollar value of total income (Figure 9). The two bottom quintiles
also experienced an important increase in public transfers, while the two top quintiles
experienced important rises in other gross income (business income, private pensions and
regular private transfers).
Income inequality across regions
17. In this section we look at average household incomes across states and across
different types of location, from major cities to remote areas. Figure 10 shows that
differences across states can be quite significant. The median income in Australian Capital
Territory is roughly twice the income in Tasmania, the poorest state. Generally speaking,
differences across states in relative terms have not grown, but some states have recorded
faster growth in income over the last 15 years, namely Western Australia, most likely in
relation to the mining boom. The mining boom is also reflected in the fast rise in incomes
in remote areas (Panel B). Nevertheless, one should keep in mind that remote areas have
small populations, the majority of people live in major cities, and more than half live in the
two most populous states New South Wales and Victoria.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 2 3 4 5 6 7 8 9 10
Labour income Public transfers
Investment income Other gross income
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 2 3 4 5 6 7 8 9 10
Labour income Public transfers
Investment income Other gross income
ECO/WKP(2019)7 │ 17
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 9. Gross household income components across quintiles over time
A. Persons aged 20 to 64
B. Persons aged 65 and over
Source: OECD calculations based on HILDA database.
0
20
40
60
80
100
120
140
160
180
2001 2008 2016 2001 2008 2016 2001 2008 2016
3rd Quintile 4th Quintile 5th Quintile
Thousand dollars
Labour income Public transfers
Investment income Other gross income
0
5
10
15
20
25
2001 2008 2016
1st Quintile
0
10
20
30
40
50
2001 2008 2016
2nd Quintile
0
50
100
150
200
250
2001 2008 2016 2001 2008 2016 2001 2008 2016
3rd Quintile 4th Quintile 5th Quintile
Thousand dollars
Labour income Public transfers
Investment income Other gross income
0
5
10
15
20
25
2001 2008 2016
1st Quintile
0
10
20
30
40
50
2001 2008 2016
2nd Quintile
18 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 10. Equivalised disposable household income across states and remoteness
A. States and Territories
Average Median
B. Remoteness
Average Median
Source: OECD calculations based on HILDA database.
Income mobility
18. Individuals or households can move from one quintile to another over time. For
example, a new job or profitable investment can move a household from a lower quintile
to a higher one; likewise, households experiencing income loss due to unemployment spell
or retirement can move down the distributional ranks. Income mobility measures
movement of individuals or households up or down the income distribution over time. As
the HILDA data follows the same individuals and households over their life time it is
particularly suitable for analysing this.
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
2001 04 07 10 13 16
NSW VIC QLD
SA WA TAS
NT ACT
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
2001 04 07 10 13 16
NSW VIC QLD
SA WA TAS
NT ACT
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
2001 04 07 10 13 16
Major city
Inner Regional Australia
Outer Regional Australia
Remote Australia
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
2001 04 07 10 13 16
Major city
Inner Regional Australia
Outer Regional Australia
Remote Australia
ECO/WKP(2019)7 │ 19
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
19. We report in Figure 11 income mobility over a three-year time span. As can be
seen, income mobility has not changed much over the past 15 years. We report income
mobility for the spells 2001-2004 and 2013-2016. As we can see, individuals and
households primarily remain in the same quintile over a three-year period. For example, as
measured by personal income for the 2013-2016 period, 53% of individuals remained in
the bottom quintile. Interestingly, at the household level, inertia in the bottom quintile
seems even more strongly cemented: measured by the equivalent household income, 67%
of individuals remained in the bottom quintile in the period 2013-16. In contrast, for the
rich, inertia at the household level is lower than at the individual level. 69% of those in the
top quintile remained in the top quintile at the individual level, and only 59% at the
household level.
Figure 11. Income mobility - short-term
A. Personal disposable income (20 and above)
B. Household Income (equivalence scale)
Source: OECD calculations based on HILDA database.
51
2111
6 4
26
50
15
73
1218
43
17
8
7 8
23
46
16
4 38
24
69
Bottom 2nd 3rd 4th Top
Quintiles in 2001
Top Quintile in 2004
4th Quintile in 2004
3rd Quintile in 2004
2nd Quintile in 2004
Bottom Quintile in 2004
64
2312 8 3
22
41
20
12
6
8
23
37
21
13
38
20
35
22
3 410
24
56
Bottom 2nd 3rd 4th Top
Quintiles in 2001
Top Quintile in 2004
4th Quintile in 2004
3rd Quintile in 2004
2nd Quintile in 2004
Bottom Quintile in 2004
53
199 5 4
25
49
16
64
1220
45
19
6
7 7
23
48
17
3 3 6
22
69
Bottom 2nd 3rd 4th Top
Quintiles in 2013
Top Quintile in 2016
4th Quintile in 2016
3rd Quintile in 2016
2nd Quintile in 2016
Bottom Quintile in 2016
67
24
10 7 5
19
43
21
116
7
22
34
21
10
48
24
38
20
3 411
24
59
Bottom 2nd 3rd 4th Top
Quintiles in 2013
Top Quintile in 2016
4th Quintile in 2016
3rd Quintile in 2016
2nd Quintile in 2016
Bottom Quintile in 2016
20 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
20. Households in the middle of the distribution tend to be more mobile than at the
bottom or at the top. The same result is reported also by Donovan et al. (2016) for US
households. This result partly arises because households in the middle of the distribution
can move up or down quintiles, while those in the end quintiles can only move in one
direction. Nevertheless, from a policy perspective this finding is relevant as it suggests
many poor households are stuck in their low income positions. Compared to the U.S.
income mobility data reported in Donovan et al. (2016), Australia shows similar (low)
degree of mobility for the bottom quintile, however, mobility of other quintiles seems a bit
higher in Australia than in the United States.
Figure 12. Income mobility - longer-term
A. Personal income
6-year span 9-year span 15-year span
B. Household income (equivalence scale)
6-year span 9-year span 15-year span
Source: OECD calculations based on HILDA database.
45
1812 8 6
28
45
18
85
14
19
38
19
7
912
24
41
17
5 6 8
25
65
Bottom 2nd 3rd 4th Top
Quintiles in 2010
Quintiles in 2016
Top 4th 3rd 2nd Bottom
42
19 159 7
31
43
18
106
15
20
30
20
9
712
25
34
18
5 611
27
60
Bottom 2nd 3rd 4th Top
Quintiles in 2007
Quintiles in 2016
Top 4th 3rd 2nd Bottom
3423
1811 9
2940
22
1610
16 18
26
20
13
12 11
20
28
18
8 815
26
51
Bottom 2nd 3rd 4th Top
Quintiles in 2001
Quintiles in 2016
Top 4th 3rd 2nd Bottom
62
27
13 9 8
21
35
19
11 9
9
22
27
22
12
510
27
29
21
4 614
29
50
Bottom 2nd 3rd 4th Top
Quintiles in 2010
Quintiles in 2016
Top 4th 3rd 2nd Bottom
60
2616
10 9
20
30
1716
10
10
21
26
21
13
5
15
24
25
20
5 717
27
47
Bottom 2nd 3rd 4th Top
Quintiles in 2007
Quintiles in 2016
Top 4th 3rd 2nd Bottom
59
2919 17 13
18
24
1817
16
11
20
2119
17
7
16
2121
18
511
21 2636
Bottom 2nd 3rd 4th Top
Quintiles in 2001
Quintiles in 2016
Top 4th 3rd 2nd Bottom
ECO/WKP(2019)7 │ 21
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
21. Mobility is likely to be higher over longer periods of time or across generations.
Some of this mobility is to be expected, reflecting events in work and family life, such as
promotions, having children, and retirement. Figure 12 depicts income mobility for a 6-
year, 9-year and 15-year (the whole sample) time spans. As expected, the longer the time-
span the higher the mobility is, and this holds for mobility upwards as well as downwards.
Notably, with longer time-spans mobility at a personal level goes down much more
dramatically than mobility at a household level. This reinforces the finding above that
persistence of poor households in remaining poor is much higher at the household level.
For other groups, on the other hand, mobility is significantly increased when increasing the
time-span. OECD work on income mobility (OECD, 2018) finds that among eight OECD
countries (including the US, Germany and France) where longer run data are available,
Australia shows high income mobility. Furthermore, Australia has among the highest
intergenerational earnings mobility across the sample of 24 OECD countries.
22. It is important to remember that (by definition) these results are based only on the
households who are in the sample in both the first and last waves of the HILDA survey for
the time-span covered. With respect to this, we can observe that the share of bottom quintile
households that drop out of the sample over time is significantly higher than for other
quintiles. This may be partly related to the higher share of older individuals in the bottom
quintile who drop out of the sample due to death. Another possibility is that poor
individuals, irrespective of age, have higher mortality due to worse health. If the way
households drop out of the sample is endogenous to the question we are asking, than the
results may be biased. Imagine an extreme case, when over a 15 year period all poor elderly
households that remained in poverty drop out of the sample due to ill-health and death,
while many of the poor young ones manage to climb the social ladder by getting jobs etc.
The data would then point to higher social mobility than actually prevails.
Wealth inequality
23. Among OECD countries, Australia appears to have relatively low wealth
inequality, despite having above average income inequality (Figure 13), according to the
OECD Wealth Distribution database. Measured by the share of wealth of the top 10%
wealthiest households, Australia is in the bottom third of OECD countries. By the same
token, Australia has a higher share of wealth that is held by the bottom three quintiles.
These data includes real wealth (housing) and private financial assets, but it does not
include any assets accrued through employer related pension schemes, or pension promises
through public pensions schemes. The apparent disconnect between income inequality and
wealth inequality is surprising. Generally, Murtin and d’Ercole (2015) report that across
OECD countries low-wealth households are typically low-income households while high-
wealth households are typically high-income households. Nevertheless, they find that for
Australia (and some other countries) this correlation is weaker, perhaps stemming from
data difficulties and comparability issues when compiling wealth data across countries.
Another possible explanatory factor for the lower correlation of wealth and income in
Australia could be high incidence of home ownership.
24. In the HILDA Survey, data on wealth are collected every four years (starting in
2002, with the latest available year 2014) and are reported at the household level2 only. We
2 Questions to individuals about wealth cover housing, incorporated and unincorporated businesses,
equity-type investments, cash-type investments, vehicles and collectibles. Respondents were also
22 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
compute wealth inequality in a similar way as income inequality, i.e. based on an
equivalence scale: each individual is assigned a value of wealth equal to the household
wealth divided by the square root of the number of household members (of all age).
Figure 13. Wealth shares of top percentiles of the net wealth distribution in selected OECD
countries
2014 or as indicated1
1. Data refers to 2012 for Canada and Spain; 2013 for Australia, Estonia, Ireland, Spain and Portugal; 2015 for
Denmark, the Netherlands and the United Kingdom. 2016 for United States.
Source: OECD, Wealth Distribution database.
25. The Gini coefficient of net worth computed from HILDA does not indicate a
sustained trend over the four available data points (Figure 14). In contrast, Dollman et al.
(2015) report that wealth inequality in Australia has been on the rise over the last 15 years
as measured by the Gini coefficient in net wealth or by the share of total wealth held by the
richest households.
26. Figure 15, panel A, depicts real household net worth across quintiles for the years
2002, 2010 and 2014. Clearly, the data for the first quintile suggest that a substantial
proportion of households have very small holdings of assets; the median net worth of the
top quintile is almost 100 times higher than the median net worth of the bottom quintile.
Net worth of the bottom quintile grew the fastest between 2002 and 2014, while from the
second quintile upwards the growth was more similar across quintiles. The rapid rise of net
worth for the bottom quintile is mostly due to a significant increase in superannuation over
the last decade. Thanks to positive valuation effects and an increase in inflows,
superannuation grew fast across the whole wealth distribution (Ryan and Stone, 2016), but
because the share of superannuation in otherwise largely asset-free bottom quintile is much
higher (see below), it contributed disproportionally to the growth of their wealth.
asked about superannuation, bank accounts, credit cards, student debt (Higher Education
Contribution Scheme (HECS)) and other personal debt.
0
10
20
30
40
50
60
70
80
90
US
A
NLD
DN
K
LVA
DE
U
CH
L
ES
T
AU
T
IRL
NZ
L
GB
R
PR
T
NO
R
FR
A
CA
N
LUX
SV
N
HU
N
AU
S
ES
P
FIN
ITA
BE
L
GR
C
PO
L
JPN
SV
K
Top 10%
Top 5%
Top 1%
Bottom 60%
OECD average (Top 10%)
ECO/WKP(2019)7 │ 23
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 14. Gini coefficient of net worth (HILDA)
Source: OECD calculations based on HILDA database.
Figure 15. Real household median net worth (equivalence scale)
A. Real household median net worth B. Index 2002=100
Source: OECD calculations based on HILDA database.
27. Figure 16 shows the composition of total assets across quintiles, where we can
observe that households in the bottom quintile own no property assets. For households in
other quintiles this is generally the largest component. For the top four quintiles
superannuation and other financial assets gain in importance the higher we go on the wealth
distribution. Furthermore, Figure 17 shows total household debt as a share of total assets.
The bottom two quintiles have the highest share of debt in total assets. Overall, the share
of debt in total assets increased significantly in the last 15 years (except for the bottom
quintile), which probably reflects increasing mortgage debt.
0.58
0.59
0.6
0.61
0.62
2002 2006 2010 2014
0
200
400
600
800
1000
1200
1400
1600
1800
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
Thousand dollars
2014 2010 2002
0
20
40
60
80
100
120
140
160
180
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
2014 2010
24 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 16. Composition of total household assets (2002 and 2014)
A. 2002 B. 2014
Note: Financial assets include bank accounts, superannuation, cash investments, equity, trust funds and life
insurance. Non-financial assets comprise property assets, business assets, collectibles and vehicles.
Source: OECD calculations based on HILDA database.
Figure 17. Total household debt as a share of total assets
Source: OECD calculations based on HILDA database.
28. The incidence of high indebtedness – as measured by the incidence of total gross
debt in excess of three times the annual household income (Figure 18) - has also been on
the rise in Australia. It should nevertheless be stressed that increasing levels of high
indebtedness do not necessarily indicate worrying developments. Increases in indebtedness
without large shifts in net debt (as for example taking out a mortgage to buy a house) or
when interest rates are low (and consequently debt servicing burden does not increase
correspondingly with higher gross debt) need not create financial problems for households.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
2002
Superannuation Other financial assets
Property assets Other non-financial assets
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
2014
Superannuation Other financial assets
Property assets Other non-financial assets
0%
5%
10%
15%
20%
25%
30%
35%
40%
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
2014 2002
ECO/WKP(2019)7 │ 25
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
High indebtedness nevertheless shows certain level of exposure, in particular when there is
a risk of large shifts in interest rates in the future. In Figure 18 we show high indebtedness
across quintiles of both income and net wealth distribution. With respect to income
distribution, households with higher incomes tend to be more highly indebted. Across net
wealth quintiles, it is the second and third quintiles that have the highest incidence, the two
quintiles that have also recorded the biggest increases over time. On both accounts, the
bottom quintiles record the lowest share of households that are highly indebted.
Figure 18. Share of highly indebted households
A. By (equivalised) household income quintiles B. By household net wealth deciles
Note: The “highly indebted” threshold has been set as household debt being three times the household
disposable income.
Source: OECD calculations based on HILDA database.
Labour market income inequality
29. Now we turn to labour market income inequality. Labour earnings are the largest
component of income for most Australians and thus the most important driver of income
inequality. In general, labour market income is positively correlated with total gross
income, as higher income individuals tend to be employed and work more hours on
average. Greenville et al. (2013) report for Australia that at the household level, the impact
of growing dispersion in hourly wages has been offset by increased employment and a
decline in the share of jobless households. This greater participation in the workforce and
longer hours worked had an especially strong impact at the low end of the labour income
distribution (part-time workers). We analyse labour income inequality at a personal level,
using HILDA data.
30. Figure 19 shows total annual labour income and average growth over time for all
employed individuals (employees and self-employed) of working age (15-65 years) across
deciles of labour income distribution. The chart in panel B clearly shows a U-shaped curve,
where labour incomes at the bottom and at the top grew faster than in the middle. Figure
20 depicts decile ratios, and no strong trend in inequality over the observed period is
apparent from the 90/10 earnings gap. Similar to overall income inequality, there seems to
0
5
10
15
20
25
30
35
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
2014 2002
0
5
10
15
20
25
30
35
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
2014 2002
26 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
have been a general rise in top inequality (90/50 ratio) and a reduction in the bottom
inequality (50/10) over the 15 year period.
Figure 19. Median total annual pay across deciles of gross labour income
Employed respondents, working-age population (age 15-64)
A. All employed B. Average growth rates (2001-2016)
Source: OECD calculations based on HILDA database.
Figure 20. Real earnings percentile ratios (age 15-64)
A. Employed, working-age population B. Index 2001=100
Source: OECD calculations based on HILDA database.
31. Inequality in labour market earnings stems from differences in the amount of work
that individuals do and from differences in the rate of pay (wage) they receive. To shed
$0
$20,000
$40,000
$60,000
$80,000
$100,000
$120,000
$140,000
$160,000
$180,000
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
2016 2008 2001
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
0
1
2
3
4
5
6
7
8
9
10
11
12
13
2001 04 07 10 13 16
Ratio D90/D10
Ratio D50/D10
Ratio D90/D50
80
90
100
110
120
130
140
2001 04 07 10 13 16
Ratio D90/D10
Ratio D50/D10
Ratio D90/D50
ECO/WKP(2019)7 │ 27
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
some light on the sources of labour income inequality, in the following two sections we
look at the amount of work in Australia and earnings of full-time workers across the labour
income distribution.
Earnings inequality and hours of work
32. Figure 21 depicts the distribution of total annual hours worked for all employed
individuals across the deciles of annual labour income. It is important to note that the total
annual hours variable is an indirectly obtained variable and it hence contains a lot of noise.
We derive it by first obtaining an estimate for hourly wage (by dividing HILDA weekly
earnings by weekly hours) and then dividing total annual labour income with the estimated
hourly wage. Nevertheless, the apparent evolution is consistent with the results reported in
Greenville et al. (2013). Hours worked have been growing at the bottom of the distribution,
while they have been falling or remained the same elsewhere. Greenville et al. (2013)
assign much of the increase in hours worked to part-time workers who are predominantly
concentrated at the lower end of the labour income distribution. We can observe in Figure
22 that the incidence of part-time work is indeed highest at lower deciles. Furthermore, the
incidence of part-time work has been on the rise.
Figure 21. Median annual hours of work across deciles of labour market income
Type the subtitle here. If you do not need a subtitle, please delete this line.
Source: OECD calculations based on HILDA database.
33. Longer working hours and higher participation are associated with structural
changes in the Australian labour market over recent decades, whereby female participation
and incidence of part-time work, for both sexes, have been on the rise (Borland and Coelli,
2016). Figure 23 shows these trends since 1980 based on the OECD data. Female
employment ratio has risen from below 50% in the early 1980s to close to 70%. The
employment ratio of men, on the other hand, has slightly declined. The incidence of part-
time work has risen for both groups and even more strongly for men, albeit from a much
lower base.
0
500
1000
1500
2000
2500
3000
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
2016 2001
28 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 22. Incidence of part-time work
A. 2001 B. 2016
Note: Workers are defined as employed part-time if they usually work less than 35 hours a week (ABS
definition).
Source: OECD calculations based on HILDA database.
Figure 23. Female participation and the incidence of part-time work have risen
A. Employment ratio (1980-2016), in % B. Share of part-time employment (1980-2016), in %
Note: In panel A, the employment ratio, or the employment-to-population ratio, is defined as the proportion of
the working-age population (15 to 64 years old) that is employed. Panel B shows the proportion of persons
(total and by gender) employed part-time among all employed persons.
Source: OECD, Labour Force database.
Earnings inequality and the rate of pay
34. The rate of pay of those working is another important component of earnings
inequality. Figure 24 shows OECD data on gross earnings inequality for full-time
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
Full-time Part-time
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
Full-time Part-time
0
10
20
30
40
50
60
70
80
90
1980 1984 1988 1992 1996 2000 2004 2008 2012 2016
Men Women All persons
0
5
10
15
20
25
30
35
40
45
50
1980 1984 1988 1992 1996 2000 2004 2008 2012 2016
Men Women All persons
ECO/WKP(2019)7 │ 29
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
employees across countries and over time, measured by the decile ratio (90/10). The data
suggest that Australia ranks somewhere close to the middle among OECD countries with
respect to the inequality in gross earnings of full-time employees. Over time, earnings
dispersion increased, mostly in the period before the global financial crisis.
Figure 24. Decile ratios of gross annual earnings (1995-2016)
D9/D1 Ratio, full-time dependent employees
2014: Belgium, Estonia, France, Italy, Latvia, Luxembourg, the Netherland, Slovenia, Spain, Switzerland.
2015: Chile, Denmark, Finland, Germany, Ireland, Japan, Norway and Poland.2006 for Slovenia, Spain,
Poland, Italy and Chile.
Source: OECD Earnings database.
35. According to HILDA data, weekly earnings of full-time employees grew fastest at
the top of the distribution over the last 15 years (Figure 25). Apart from the lowest two
deciles, we can observe a roughly monotonic progression in the growth of weekly earnings
across the labour income distribution. The higher is the labour income decile the higher the
growth in the rate of pay, consistent with Borland & Coelli (2016) who report that, over
the last four decades, weekly earnings for full-time employees grew significantly more for
the top earners than for bottom earners. Evolution in the rate of pay has therefore
contributed to rising income inequality in Australia.
0
1
2
3
4
5
6
ITA
BE
L
NO
R
DN
K
FIN
CH
E
NZ
L
JPN
ISL
ES
P
GR
C
AU
T
AU
S
SV
N
ME
X
OE
CD
GB
R
SV
K
DE
U
TU
R
CZ
E
CA
N
HU
N
PR
T
IRL
PO
L
CH
L
KO
R
2016 or earlier 2007 or earlier 1995
30 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 25. Median weekly pay across deciles of gross labour income
Full-time employed, working-age population (age 15-64).
A. Full-time employed B. Average growth rates (2001-2016)
Source: OECD calculations based on HILDA database.
36. The evolution of earnings decile ratios (90/10, 90/50 and 50/10) for full-time
employees by gender is depicted in Figure 26. The figures confirm the findings from above.
For both male and female workers top-end inequality (90/50 ratio) has been on the rise,
while bottom-end inequality (50/10 ratio) fell (especially for women), as bottom earners
recorded faster growth in earnings over the observed period.
37. Borland & Coelli (2016) report that the rising earnings inequality in Australia over
the last four decades is linked to changes in the occupation composition of employment. At
the same time, however, they report that the earnings differentials between workers with
different levels of education attainment have remained stable, which they interpret as
reflecting a large recent rise in the numbers of university graduates. Chatterjee et al. (2016)
report that the rising wage inequality in Australia cannot be explained by observable factors
such as education or experience but unobservable or residual factors reflecting idiosyncratic
risk and unexpected labour market outcomes.
38. We explore HILDA data to better understand the evolution of wages across
education groups and occupations over the last 15 years. Figure 27 shows median labour
market income across different education levels for full-time employed workers. The data
are rather erratic, probably due to the sample size getting small at this level of
disaggregation. For both men and women it appears that the rate of pay of the highest
education group (Masters and PhDs) has grown the least over the last 15 years, particularly
in the last several years. On the other hand, median wages of those with certificate III or
IV appear to have grown the fastest. It is not obvious from the chart what has happened
with overall inequality across education levels. However there are indications that wage
gaps have generally been slightly reduced. Further analysis shows that the percentage gap
between highest and lowest median wage has been slightly reduced, and variation across
education groups (measured by the coefficient of variation) has also declined.
0
500
1000
1500
2000
2500
3000
3500
1 2 3 4 5 6 7 8 9 10Gross labour income deciles
2016 2008 2001
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
1 2 3 4 5 6 7 8 9 10
Gross labour income deciles
ECO/WKP(2019)7 │ 31
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 26. Decile earnings ratios (full-time employed) by gender
A. Males
Earnings decile ratios (90/10, 50/10 and 90/50) Index 2001=100
B. Females
Earnings decile ratios (90/10, 50/10 and 90/50) Index 2001=100
Source: OECD calculations based on HILDA database.
39. Globalisation and technical change can impact the relative rate of pay and
employment opportunities among different occupations. Technological progress has
favoured certain skills compared to others, and information and communication technology
can replace certain tasks better than others. This in turn is reflected in the labour market
where we can observe job polarisation: a decline in the share of middle-skill, middle-pay
jobs relative to jobs with higher or lower skill levels (OECD, 2017). Coelli and Borland
(2016) find evidence of job polarisation in Australia, although they report that the effect
1.6
2.0
2.4
2.8
3.2
3.6
4.0
4.4
4.8
2001 2004 2007 2010 2013 2016
Male 90/10 ratio
Male 50/10 ratio
Male 90/50 ratio
80
85
90
95
100
105
110
2001 2004 2007 2010 2013 2016
Ratio 90/50
Ratio 50/10
Ratio 90/10
1.6
2.0
2.4
2.8
3.2
3.6
4.0
4.4
4.8
2001 2004 2007 2010 2013 2016
Female 90/10 ratio
Female 50/10 ratio
Female 90/50 ratio
80
85
90
95
100
105
110
115
120
125
130
2001 2004 2007 2010 2013 2016
Ratio 90/50
Ratio 50/10
Ratio 90/10
32 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
was mostly concentrated in the 1980s and 1990s. All this also has a bearing on the earnings
distribution.
Figure 27. Median earnings (full-time employed) by level of education
A. Males
Level Index 2001=100
B. Females
Level Index 2001=100
Source: OECD calculations based on HILDA database.
40. Using HILDA data, we depict differences in earnings across different skill and
occupation groups (Figure 28 and 29). While there have been no large shifts in relative pay
across occupations, a clearer picture emerges when the occupations are grouped into high-
skill, medium-skill, and low-skill occupations. This is shown in Figure 28, where it is
clearly seen that higher skills attract higher salaries. The gap in wages between high and
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
$80,000
$90,000
$100,000
$110,000
$120,000
2001 2004 2007 2010 2013 2016
Masters or doctorateGrad diploma, grad certificateBachelor or honoursAdv diploma, diplomaCert III or IVYear 12Year 11 and below
90
100
110
120
130
140
150
160
170
180
2001 2004 2007 2010 2013 2016
Masters or doctorateGrad diploma, grad certificateBachelor or honoursAdv diploma, diplomaCert III or IVYear 12Year 11 and below
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
$80,000
$90,000
$100,000
$110,000
$120,000
2001 2004 2007 2010 2013 2016
Masters or doctorateGrad diploma, grad certificateBachelor or honoursAdv diploma, diplomaCert III or IVYear 12Year 11 and below
90
100
110
120
130
140
150
160
170
180
2001 2004 2007 2010 2013 2016
Masters or doctorateGrad diploma, grad certificateBachelor or honoursAdv diploma, diplomaCert III or IVYear 12Year 11 and below
ECO/WKP(2019)7 │ 33
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
low skilled was rising between 2007 and 2014. It has started to close since then, but it still
remains larger than in the early 2000s.
Figure 28. Real wages by occupation
Full-time employed persons aged 15 and over, median wages
Note: Full-time employed respondents reporting positive values for earnings.
Source: OECD calculations based on HILDA database.
Figure 29. Real median wages and salaries for occupations requiring different skills levels
(2001 to 2016)
Full-time employed individuals aged 15 and over
Note: Occupations are ranked by wage level following Autor and Dorn (2013) and Goos et al. (2014). High-
skill occupations include jobs classified under the ISCO-88 major groups: legislators, senior officials, and
managers (group 1), professionals (gr. 2), and technicians and associate professionals (gr. 3). Middle-skill
occupations include the ISCO-88 major groups: clerks (gr. 4), craft and related trades workers (gr. 7), and plant
and machine operators and assemblers (gr. 8). Low-skill occupations include the ISCO-88 major groups:
service workers and shop and market sales workers (gr. 5), and elementary occupations (gr. 9).
Source: OECD calculations based on HILDA database.
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
$80,000
$90,000
2001 2004 2007 2010 2013 2016
Managers
Professionals
Technicians and Trades Workers
Community and Personal Service Work
Clerical and Administrative Workers
Sales Workers
Machinery Operators and Drivers
Labourers
30,000
35,000
40,000
45,000
50,000
55,000
60,000
65,000
70,000
75,000
30,000
35,000
40,000
45,000
50,000
55,000
60,000
65,000
70,000
75,000
2001 2004 2007 2010 2013 2016
Low skill Medium skill
High skill
34 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
41. Finally, we consider changes in employment over time across various occupation
and skill groups. Figures 30 and 31 show supporting evidence for job polarisation - the
gradual decline in the share of middle skilled jobs - jobs where routine tasks can
increasingly be performed by computers and robots. Figure 30 shows changes in
employment shares by occupations and Figure 31 changes in employment shares across
skill groups. Employment share of professionals (high skilled) and personal services
workers (medium skill, non-routine) has increased significantly, while that of clerical
workers (medium skill, routine), service and sales workers (low skill, non-routine) and
labourers (medium skill, routine) has decreased. The changes in these shares are however
not only related to automation and decline of the manufacturing sector, but also to other
structural changes such as population ageing and greater demand for personal services.
Figure 30. Employment shares by occupation (2001-2016)
In percentage
Source: OECD calculations based on HILDA database.
0
5
10
15
20
25
Managers Professionals Technicians andTrades
Community andPersonal Service
Clerical andAdministrative
Sales MachineryOperators/
Drivers
Labourers
2001 2008 2016
ECO/WKP(2019)7 │ 35
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 31. Changes in employment shares by skill groups (2001-2016)
Percentage change
Note: Occupations are ranked by wage level following Autor and Dorn (2013) and Goos et al. (2014). High-
skill occupations include jobs classified under the ISCO-88 major groups: legislators, senior officials, and
managers (group 1), professionals (gr. 2), and technicians and associate professionals (gr. 3). Middle-skill
occupations include the ISCO-88 major groups: clerks (gr. 4), craft and related trades workers (gr. 7), and plant
and machine operators and assemblers (gr. 8). Low-skill occupations include the ISCO-88 major groups:
service workers and shop and market sales workers (gr. 5), and elementary occupations (gr. 9).
Source: OECD calculations based on HILDA database.
Non-standard workers and duration of contracts
42. According to OECD data, Australia has quite a high share of workers in so called
non-standard employment, which OECD (2015) categorises as workers that work part-
time, on temporary contracts or as self-employed (Figure 32). In Australia, it is part-time
work that dominates, while the shares of self-employment and temporary work are
comparatively low (these groups are not mutually exclusive, workers can work part-time
and be on a temporary contract, for example). Across the OECD, non-standard employment
has been on the rise since 1990s. Non-standard employment can be associated with poorer
labour conditions (wages, working time, job security, leave entitlements, etc.), less training
and poorer job prospects. Yet, at the same time, part-time, temporary and self-employment
arrangements may be attractive to certain workers as they offer higher flexibility, and might
have been chosen voluntarily (OECD, 2015).
-6
-4
-2
0
2
4
6
8
Low skill Medium skill High skill
36 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Figure 32. Share of non-standard employment
Source: OECD, Labour Force Statistics database.
43. Borland and Coelli (2016) report that the incidence of casual employment (those
that receive no paid sick leave or holiday leave – a narrower definition to the non-standard
work above) in Australia has risen since the 1980s, especially for females, but it peaked in
the 2000s and then started falling. We find support for this in the HILDA data; there have
been no large increases in casual employment in the last 15 years. As shown in Figure 33,
the share of men in causal jobs fell until about the time of the global financial crisis, but
has risen since. For females, similarly, the falling trend stalled with the crisis, and it shows
tentative signs of reversal in the most recent periods. The incomes of casual workers are
significantly lower than of other workers (panel B), and this has to do partly with lower
hours and lower experience, but the gap seems to be rising.
44. Borland and Coelli (2016) furthermore report that employment contracts have,
contrary to popular belief, on average become of longer duration. HILDA data again
support this finding; between 2001 and 2016, contracts have become of slightly longer
duration for both men and women (Figure 34). Nevertheless, Australia has recorded rising
0
10
20
30
40
50
0
5
10
15
20
25
30
35
40
45
50
HU
N
SV
K
PO
L
CZ
E
LVA
ES
T
GR
C
TUR
SV
N
PR
T
ES
P
FIN
US
A
ITA
FRA
LUX
CA
N
OE
CD
CH
L
IRL
NZ
L
ISR
SW
E
ME
X
JPN
ISL
BE
L
NO
R
DN
K
GB
R
DE
U
AU
T
AU
S
CH
E
NLD
A. Part-time employment
% of total employment, 2017
0
5
10
15
20
25
30
AU
S
GB
R
JPN
NO
R
IRL
AU
T
BE
L
OE
CD
DE
U
DN
K
CH
E
CA
N
ITA
FIN
SW
E
FRA
KO
R
NLD
ES
P
B. Temporary employment
% of dependent employment, 2017
0
5
10
15
20
25
30
NO
R
DN
K
CA
N
SW
E
AU
S
DE
U
JPN
FRA
AU
T
FIN
BE
L
CH
E
GB
R
IRL
ES
P
NLD
OE
CD
ITA
KO
R
C. Self employment
% of employment, 2017
ECO/WKP(2019)7 │ 37
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
underemployment - workers who work part-time but would prefer to work longer hours if
they had a chance. From about the mid-1990s underemployment has risen, in particular for
young workers (15-24) among whom the incidence of part-time work has risen rapidly
(Borland and Coelli, 2016).
Figure 33. Share of workers in casual jobs, by gender, 2001-2016
A. Share of all workers in casual jobs B. Median labour income
Note: ABS definition, no paid holiday leave and no paid sick leave.
Source: OECD calculations based on HILDA database.
Figure 34. Duration of current jobs, by gender, 2001-2016
A. Females B. Males
Source: OECD calculations based on HILDA database.
10
15
20
25
30
35
40
2001 04 07 10 13 16
Female Male
$0
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
2001 04 07 10 13 16
Casual Non-casual
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2001 04 07 10 13 16
Less than 1 year 1 to 5 years
5 to 10 years 10 plus years
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2001 04 07 10 13 16
Less than 1 year 1 to 5 years
5 to 10 years 10 plus years
38 │ ECO/WKP(2019)7
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Conclusion
45. This paper analyses income, wealth and earnings inequality in Australia. Evidence
from OECD data shows that income inequality in Australia has risen in the last two decades
- as in many other OECD countries - and is above the OECD average. However, most of
the rise in income inequality in Australia was concentrated before the global financial crisis
in 2008. Since then, income inequality has been roughly constant. Despite no shifts in
overall inequality, there is evidence in the HILDA data that growth of incomes in the
middle of the distribution has been slower than in the tails of the distribution.
46. Echoing income inequality, there has been no clear trend in the overall inequality
in labour-market income over the last 15 years. While Australia has experienced a rising
inequality in wages - that grew most quickly for top earners - this has been offset by
increased participation, longer hours worked and a decline in the share of jobless
households, with most of the effect at the bottom of the distribution (Greenville et al.,
2013). We observe most of these developments also in the HILDA data over the last 15
years.
47. We depict differences in earnings across education, occupation and skill groups and
we show changes in employment. According to HILDA data, relative pay across education
groups has not recorded large shifts over the last 15 years, but we find evidence for job
polarisation. Notably, the share of high skilled jobs versus middle skilled jobs has
increased. Considering changes in employment by occupation, we find that employment of
professionals and personal services work has increased significantly, while that of clerical
workers, sales workers and labourers has decreased.
48. Australia’s labour market has changed markedly over the recent decades, with a
growing participation rate, especially of women, and higher incidence of part-time work.
The incidence of casual employment - that receive no paid sick leave or holiday leave - in
Australia has been reported to have risen since the 1980s, especially for females, but
according to the HILDA data it has fallen since early 2000s. We also find no evidence for
a popular belief that contracts have become of shorter duration over time, echoing results
reported by Borland and Coelli (2016).
References
Autor, D., L. Katz and M. Kearney (2006), “The Polarization of the US Labor Market”, American
Economic Review, Vol. 96, No. 2, pp. 189-194.
Borland, J. and Coelli, M (2016). ‘Labour Market Inequality in Australia’. Economic Record, 92, pp 517-
547.
Causa, O. and Hermansen, M. (2017), “Income redistribution through taxes and transfers across OECD
countries”, OECD Economics Department Working Papers No. 1453, OECD Publishing, Paris.
Chatterjee, A., Singh, A., and Stone, T. (2016), "Understanding Wage Inequality in Australia", Economic
Record, vol. 92, no. 298, pp. 348-360. http://onlinelibrary.wiley.com/doi/10.1111/1475-
4932.12263/full
Coelli, M. and Borland, J. (2016), ‘Job Polarisation and Earnings Inequality in Australia’, Economic
Record, 92, 1–27.
ECO/WKP(2019)7 │ 39
INCOME, WEALTH AND EARNINGS INEQUALITY IN AUSTRALIA: EVIDENCE FROM THE HILDA SURVEY Unclassified
Dollman, R., Kaplan, G., La Cava, G. and Stone, T. (2015). Household Economic Inequality in Australia.
Research Discussion Paper. Reserve Bank of Australia.
Donovan, S., Labonte, M. and Dalaker, J. (2016). ‘The U.S. Income Distribution: Trends and Issues’.
Congressional Research Services. R44705.
Fletcher, M., & Guttmann, B. (2013). ‘Income inequality in Australia’, Economic Round-Up (2), pp 35-
54.
Goos, M. and A. Manning (2007), “Lousy and Lovely Jobs: The Rising Polarization of Work in Britain”,
Review of Economics and Statistics, Vol. 89, No. 1, pp. 118-133.
Goos, M., A. Manning and A. Salomons (2009), “Job Polarization in Europe”, American Economic
Review, pp. 58-63.
Greenville, J., Pobke, C., and Rogers, N. (2013). ‘Trends in the Distribution of Income in Australia’.
Productivity Commission Staff Working Paper. March 2013.
Murtin, F., and d’Ercole, M. M. (2015), "Household wealth inequality across OECD countries: new
OECD evidence", OECD Statistics Brief, June 2015, No.21.
OECD (2018), A Broken Social Elevator? How to Promote Social Mobility, OECD Publishing, Paris.
OECD (2017), "How technology and globalisation are transforming the labour market", Chapter 3,
Employment Outlook 2017.
OECD (2016a), Back to Work: Australia. Improving the Re-employment Prospects of Displaced
Workers.
OECD (2016b), Income inequality remains high in the face of weak recovery, Income Inequality Update,
Centre for Opportunity and Equality (COPE).
OECD (2015), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris.
Productivity commission (2018), Rising inequality? A stocktake of the evidence. Productivity
Commission Research Paper, August 2018.
Ryan, P. and Stone, T. (2016) ‘Household Wealth in Australia: Evidence from the 2014 HILDA Survey’.
Reserve Bank of Australia. Bulletin, June Quarter.
Wilkins, R. (2013). ‘Evaluating the Evidence on Income Inequality in Australia in the 2000s’. Melbourne
Institute Working Paper Series (No. 26/13).
Wilkins, R. (2015). ‘Measuring Income Inequality in Australia’. The Australian Economic Review (vol
48, No.1, pp 93-102.
Wilkins, R. (2017), The Household Income and Labour Dynamics in Australia Survey: Selected Findings
from Waves 1 to 15, Melbourne Institute of Applied Economic and Social Research.