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

http://dx.doi.org/10.1787/cab6789d-en

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]

mailto:[email protected]

http://www.oecd.org/eco/workingpapers

ECO/WKP(2019)7 │ 3


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


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


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


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

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 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


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


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


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


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


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


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


Figure 5. Gini coefficient of personal disposable income

Persons aged 20 and above.


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


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


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


Figure 7. Real disposable income decile ratios

Individual level, respondents aged 20 and over


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


Figure 8. Composition of gross household income by deciles

A. 20 to 64 years old B. 65 years old and over


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



ECO/WKP(2019)7 │ 17


Figure 9. Gross household income components across quintiles over time

A. Persons aged 20 to 64

B. Persons aged 65 and over


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



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



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


Figure 10. Equivalised disposable household income across states and remoteness

A. States and Territories

Average Median

B. Remoteness

Average Median


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


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)


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


Quintiles in 2001






53

199 5 4

25

49

16

64

1220

45

19

6

7 7

23

48

17

3 3 6

22

69


Quintiles in 2013






67

24

10 7 5

19

43

21

116

7

22

34

21

10

48

24

38

20

3 411

24

59


Quintiles in 2013






20 │ ECO/WKP(2019)7


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


45

1812 8 6

28

45

18

85

14

19

38

19

7

912

24

41

17

5 6 8

25

65


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


Quintiles in 2007

Quintiles in 2016


3423

1811 9

2940

22

1610

16 18

26

20

13

12 11

20

28

18

8 815

26

51


Quintiles in 2001

Quintiles in 2016


62

27

13 9 8

21

35

19

11 9

9

22

27

22

12

510

27

29

21

4 614

29

50


Quintiles in 2010

Quintiles in 2016


60

2616

10 9

20

30

1716

10

10

21

26

21

13

5

15

24

25

20

5 717

27

47


Quintiles in 2007

Quintiles in 2016


59

2919 17 13

18

24

1817

16

11

20

2119

17

7

16

2121

18

511

21 2636


Quintiles in 2001

Quintiles in 2016


ECO/WKP(2019)7 │ 21


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


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


Figure 14. Gini coefficient of net worth (HILDA)


Figure 15. Real household median net worth (equivalence scale)

A. Real household median net worth B. Index 2002=100


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


2014 2010

24 │ ECO/WKP(2019)7


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.


Figure 17. Total household debt as a share of total assets


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%


2002

Superannuation Other financial assets

Property assets Other non-financial assets

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%


2014

Superannuation Other financial assets

Property assets Other non-financial assets

0%

5%

10%

15%

20%

25%

30%

35%

40%


2014 2002

ECO/WKP(2019)7 │ 25


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.


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


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)


Figure 20. Real earnings percentile ratios (age 15-64)

A. Employed, working-age population B. Index 2001=100


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


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


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.


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


2016 2001

28 │ ECO/WKP(2019)7


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).


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


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


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


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


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)


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


ECO/WKP(2019)7 │ 31


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


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


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


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


$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


90

100

110

120

130

140

150

160

170

180

2001 2004 2007 2010 2013 2016


ECO/WKP(2019)7 │ 33


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.


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).


$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


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


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


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).


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


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


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.


Figure 34. Duration of current jobs, by gender, 2001-2016

A. Females B. Males


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


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


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.

Date post:	24-Jul-2020
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

OECD Economics Department Working - Sipotra · ECO/WKP(2019)7 Unclassified English - Or. English ....

Documents