Peter Whiteford Family Joblessness in Australia

Transcription

1 Social Inclusion Peter Whiteford Family Joblessness in Australia January 2009

2 Peter Whiteford Family Joblessness in Australia A paper commissioned by the Social Inclusion Unit of the Department of the Prime Minister and Cabinet January 2009 Commonwealth of Australia (PDF) (RTF) This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no partmay be reproduced by any process without prior written permission from the Commonwealth.Requests and inquiries concerning reproduction and rights should be addressed to the Commonwealth Copyright Administration, Attorney-General s Department, Robert Garran Offi ces, National Circuit, Barton ACT 2600 or posted at The fi ndings and conclusions in this report are those of the author and do not necessarily represent the views of the Australian Government. Contact for follow up Social Inclusion Unit, Department of the Prime Minister and Cabinet, ph , socialinclusionunit@pmc.gov.au. Peter Whiteford, Social Policy Research Centre, ph , p.whiteford@unsw.edu.au. Acknowledgements This paper draws on a range of previous work undertaken by the author, including Whiteford and Adema (2007), What Works Best in Reducing Child Poverty: A Benefi t or Work Strategy? OECD Social Employment and Migration Working Papers, Number 51, and Whiteford (2008), Child Poverty in OECD Countries: Trends, Causes and Policy Responses, OECD Social Employment and Migration Working Paper, forthcoming. I am grateful to John Landt of the Social Inclusion Unit, Department of the Prime Minister and Cabinet for analysis of the 2006 General Social Survey in Tables 13 and 14, and to Rob Bray of the Department of Family, Community Services, Housing and Indigenous Affairs for the data used in Tables 1 to 4. 2 A09-073

3 Contents Executive Summary 4 1 Introduction 6 2 Defi ning and measuring joblessness 7 3 Trends in family joblessness 9 4 Trends in lone parenthood 11 5 The job gap 13 6 Duration and dynamics of joblessness 15 7 The dynamics of family joblessness, child poverty and welfare receipt 2001 to How does Australia compare? Changing patterns of family formation Australia s employment performance in perspective International comparisons of family joblessness Why do some countries have low joblessness and why do some countries have high joblessness? 25 9 Consequences of joblessness child poverty and social exclusion Family and child poverty composition, risks and characteristics Family joblessness and disadvantage Types of disadvantage Intergenerational transmission of welfare receipt Tax and benefi t policies and their effects on poverty and employment Overview of scope and nature of tax/benefi t systems and support for families Adequacy of benefi ts and other support for families Does work pay? Effective marginal tax rates and average effective tax rates Are child care costs a barrier to work? The policy stance towards lone parents on income support How can joblessness be reduced? 60 References 64 Social Inclusion Family Joblessness in Australia 3

4 Executive Summary The Australian Social Inclusion Board has been asked by the Government to work as a fi rst priority on jobless families and children at greatest risk of disadvantage. The priority given to jobless families refl ects the fact that children are among the most vulnerable people in the community and that family joblessness raises the risk that children themselves may grow up to be jobless and reliant on welfare payments for a signifi cant proportion of their income. The intergenerational transmission of joblessness if it occurs undermines both equality of outcomes and equality of opportunity. While Australia has a high overall level of employment compared to other developed countries in the OECD, it also has one of the highest levels of joblessness among families with children of all rich countries. Identifying the factors behind this poor performance can contribute to improving policies to promote social inclusion. Joblessness can be defi ned in a number of ways and measured at the individual, family or household level. Much of the discussion of joblessness relies on snapshot data of the extent of joblessness at a point in time, but data on the extent of joblessness over a year or over multiple years leads to the conclusion that the experience of joblessness is more common than suggested by point-in-time data, but that long-term joblessness is concentrated on a smaller group of families at risk of severe disadvantage. For example, between 2001 and 2005 the proportion of children living in jobless households fell from around 18% to 14%, but nearly 26% of children experienced a spell of joblessness in at least one of those fi ve years, but only 5.5% of children were in jobless households for all fi ve years. Following nearly two decades of increasing joblessness, family joblessness has fallen since 1998 and is now nearly back to its level in This is mainly due to increasing overall employment in this period, but the rate of reduction in joblessness has increased since 2005, perhaps refl ecting policy changes since then requiring parents to actively look for work once their youngest child turns seven. What is the link between joblessness among families with children and child poverty? Across OECD countries, on average, only around 30% of poor families with children are jobless and most child poverty is found in families where at least one parent is in paid employment. In Australia, however, around 70% of poor children live in jobless families, making the joblessness of parents the main cause of relative low income in childhood. Moreover, for lone parents in paid work, Australia has the second lowest poverty rate in the OECD, and for couples with either one or both parents in paid employment, poverty rates are also among the lowest in the OECD. All countries with low child poverty rates (under 5%) combine both effective redistribution and low rates of family joblessness. Australia appears to have one of the most effective systems of redistribution across OECD countries, but the high level of family joblessness means that Australia has rates of child poverty that are about twice as high as the best performing countries (but less than half the level of the worst performing countries). Family joblessness is associated with a range of disadvantages including a higher incidence of poor health and disability among parents. However, while disability or poor health is likely to be a signifi cant additional barrier to employment for some families, this tends to affect a relatively small minority of jobless families. 4 Executive Summary

5 Information on the intergenerational transmission of joblessness and welfare receipt is limited, but evidence suggest that different forms of welfare systems can lead to differing patterns of intergenerational transmission of disadvantage. One of the main explanations advanced for family joblessness is that work doesn t pay for parents that is, the combination of the withdrawal of income-tested benefi ts and the effects of the tax system produce high effective marginal tax rates (EMTRs) that punish low income families who attempt to improve their fi nancial circumstances through paid work. It is also argued that high child care costs can further reduce the rewards to paid work. While Australia operates what overall is the most targeted benefi t system in the OECD, incentives for benefi ciaries to take up part-time work are among the best in the OECD, and the only countries with better work incentives are those that provide much less generous benefi ts for families with children. While high EMTRS can be a barrier to work, it appears unlikely that they can be an explanation for Australia s high level of family joblessness compared to other countries. Child care costs also do not appear to offer an explanation for Australia s high level of family joblessness. Instead, the main explanation appears to be the entitlement conditions for payments (before 2006), with parents not required to actively look for work until their youngest child turned 16. What policies are likely to be most effective in reducing Australia s high level of family joblessness? Bearing in mind the factors discussed above, a combination of policy approaches should be considered. Maintaining strong employment growth is of key importance, but should economic conditions become more adverse it is important not to promote policies that encourage people to leave the labour force. Requiring parents to actively look for work is strongly supported, but incentives to work and adequacy of incomes - would be strengthened if assistance was provided on the basis of the Parenting Payment Single conditions rather than on the basis of Newstart Allowance conditions. Income adequacy would also be improved. Consideration should be given to promoting employment among parents with pre-school age children, perhaps on a voluntary basis, but it is important to change parental expectations to normalise paid employment. Consideration could be given to setting a target of achieving an 80% employment rate for lone parents (the level in Denmark). If the same rate of employment growth that has occurred over the past 10 years was maintained, this could be achieved by 2020, or if the rate of employment growth of the past three years was maintained, the target would be achieved by Even if such a target is achieved, this would reduce family joblessness from its current level of around 12% to around 5%. It is likely that a signifi cant proportion of those remaining jobless would have complex combinations of disadvantages. The challenge is to develop the combination of policies both cash transfers and services that promote and support employment for as many parents as possible, provide adequate income support for those with greater diffi culties in fi nding paid employment, and help to alleviate the problems of those with the most severe disadvantages. Social Inclusion Family Joblessness in Australia 5

6 1Introduction The Australian Social Inclusion Board has been asked by the Government to work as a fi rst priority on jobless families and children at greatest risk of disadvantage. The priority given to jobless families refl ects the fact that children are among the most vulnerable people in the community and that family joblessness is associated with relative low income and raises the risk that children themselves may grow up to be jobless and reliant on welfare payments for a signifi cant proportion of their income. The intergenerational transmission of joblessness if it occurs undermines both equality of outcomes and equality of opportunity. Joblessness appears to be associated with a wide range of negative outcomes for parents and children, including higher rates of income poverty, poorer health status and lower educational attainment. The association is complex, however, and causation can run both ways that is, joblessness causes disadvantage and disadvantages can also cause joblessness. This paper provides an analysis of trends in family joblessness in Australia over the last 30 years, and identifi es the main factors that have driven trends in joblessness. The paper compares the circumstances and characteristics of jobless families in Australia relative to those in other OECD countries. The paper also discusses the relationship between family joblessness and income poverty and other forms of disadvantage. On the basis of this international experience, the paper discusses broad policy directions that could be considered to reduce the problem of joblessness while also improving support for those with multiple barriers to employment. 6 1 Introduction

7 2 Defining and measuring joblessness The jobless encompasses the unemployed and looking for work and those not in the labour force (and not actively looking for work). A signifi cant factor in family joblessness is that many mothers both lone and partnered mothers have responsibilities for caring for children which means that they may not be actively looking for work, although many may accept work if it was available. In other cases, parents are actively looking for work but are unsuccessful, or they may have short periods of employment between spells of joblessness. Joblessness exists at the individual, family and household level (and also among communities). 1 Generally speaking, the share of individuals of working age who are jobless can be expected to be higher than the share of families or households who are completely jobless, since some jobless individuals share accommodation with other family or household members who are not jobless. Family joblessness can also be looked at from the parental perspective or from the child s perspective. In Australia, a higher share of children experience joblessness than the share of families, primarily because larger families are more likely to be jobless than smaller families. One of the main points made in this paper is that there are signifi cant differences across countries in the levels of individual and family joblessness; in brief, Australia appears to have relatively low levels of overall and individual joblessness, but relatively high levels of household and family joblessness that is in Australia, joblessness is more concentrated among specifi c households where no one works than in nearly all other OECD countries. In part, this appears to be related to different patterns of household formation and composition in different countries in particular some of the countries with high levels of individual joblessness have much higher levels of multi-generation households, with jobless people being more likely to share accommodation with their parents, siblings or with adult children. In addition, these patterns are affected by the nature of the social welfare systems in each country, not only the level of welfare benefi ts, but also the conditions under which welfare payments are available and the level of support for families in employment, both in terms of cash benefi ts and support services. Joblessness has a very important time dimension for example, it is common in discussions of joblessness for statements to be made along the lines that x percent of Australian children are being raised in jobless families but these statements are potentially misleading since families can move between employment and joblessness during a year or over a number of years. The main source of data on trends in joblessness in Australia is the Labour Force Survey, and in particular the Labour Force Status and Other Characteristics of Families publication (ABS Cat. No ), which is published annually, with the most recent publication being in June The information on labour force status usually relates to the position during the week preceding household interviews that is, the estimates of joblessness refer to the circumstances of families during a one week period. In contrast, the Labour Force Experience (ABS Cat. No ) publication which is currently issued every second year presents information about the labour force experience of persons over the entire 12 month before the survey date (most recently in February 2007). 1 Household is used here to refer to households that include people both with and without dependent children, while families refer to people with dependent children. Social Inclusion Family Joblessness in Australia 7

8 A further source of data on joblessness is the Household, Income and Labour Dynamics in Australia (HILDA) Survey, a household based panel study which began in 2001, and in which panel members are traced over time, with the most recent Statistical Report presenting information on joblessness trends between 2001 and The HILDA Statistical Report defi nes joblessness as not being in employment for at least 26 weeks in the survey year, thus giving a broader measure of joblessness than the Labour Force Experience survey. As will be shown in Section 3 below each of these data sources are consistent in the picture they show of recent trends in joblessness. However, the estimated level of joblessness varies between data sources, and long duration joblessness either for the full survey year or for at least six months over the course of two, three, four or fi ve years is signifi cantly lower than is measured with point in time data from the Labour Force Surveys. The extent of joblessness also depends on the defi nition of a dependent child and the defi nition of a family or household. The discussion below focuses on trends in family joblessness where dependent children are defi ned as those less than 15 years of age; if those caring for an older dependent student were included then the absolute number of jobless families would increase, but their share of all families with dependents would decrease, as employment rates are higher for these groups of parents than for those with younger children. The international comparisons with OECD countries are based on households with at least one child less than 17 years of age. The paper also looks in some detail at the links between family joblessness, child poverty and receipt of welfare payments. The association between these factors is strong, because in Australia family joblessness is the most important single cause of child poverty, and because when families are jobless they are most commonly in receipt of social welfare payments (since they do not have access to earnings). 2 As a result it is also possible to use studies of the dynamics of child poverty and receipt of welfare payments to supplement information on these dimensions of joblessness. 2 It could be expected that nearly all jobless families receive welfare payments, but not all welfare recipients are jobless because families can combine benefi ts and earnings up to moderately high income levels. According to the 2002 General Social Survey (ABS, Cat. No ), 89% of jobless households with children under 15 were dependent on government cash pensions and allowances for their principal source of income, with three quarters of the selected persons in these households having been dependent on this source of income for more than half of the last two years. 8 2 Defining and measuring joblessness

9 3 Trends in family joblessness As shown in Table 1 and Figure 1, the extent of family joblessness has fl uctuated over time, generally rising up until 1998, but falling since then. Long-term trends in family joblessness refl ect a number of underlying factors, including family change and labour force changes. Lone parents have increased as a share of families with children (Table 2), and since lone mothers and fathers have lower employment rates than partnered fathers, family joblessness has also increased. In addition, during the recessions of the early 1980s and early 1990s, rising unemployment disproportionately affected parents, and in the initial stages of economic recovery family unemployment did not fall at the same pace as overall unemployment. The last decade (since 1998), however, has seen a sustained fall in family joblessness, which is now close to its level in This is likely to refl ect the strong aggregate employment performance during this period, and possibly some of the more recent policy changes affecting Parenting Payments. In 1980 around 11% of Australian families with children less than 15 years of age were jobless, comprising close to 60% of lone parent families and just under 5% of couple families. Total joblessness among families with children reached a peak of 19% in 1998, but has since fallen to 12% in Joblessness among couples with children is strongly related to the state of the labour market and consequent trends in unemployment with joblessness increasing from 4.8% to 8.5% between 1980 and 1983, and from 6.3% to 10.8% between 1990 and 1993 (its highest level). Joblessness among couples with children in 2008 (4.3%) was at its lowest level since Joblessness among lone parents has fl uctuated over time, but has generally been falling since 1998 from 58% to 42%. But in relative terms, the decline in lone parent joblessness has not been as rapid as the decline in couple joblessness so that currently close to 72% of jobless families are lone parent families, compared to a low of 53% in Table 1: Families with children less than 15 years by family type and joblessness status, 1980 to 2008 Couple families with children less than 15 years One parent families with children less than 15 years All families with children less than 15 years Total Jobless Jobless Total Jobless Jobless Total Jobless Jobless 000s 000s % 000s 000s % 000s 000s % Social Inclusion Family Joblessness in Australia 9

10 Note: Jobless families are those where no parent is employed. Source: ABS and ; Data provided by FaHCSIA. Figure 1: Trends in Family Joblessness in Australia, 1980 to Couples Lone parents All families Source: Calculated from Table Trends in family joblessness

11 4 Trends in lone parenthood Lone parents as a percentage of all families with children increased from 12.2% in 1980 to 23.1% in 2004, but fell to 20.9% in 2008, the most sustained fall in 30 years (Table 2). The increase in the number of lone parent families as a share of all families with children is related to a number of underlying factors, including trends in widowhood, trends in divorce and separation, and trends in births outside marriage and cohabitation, as well as developments in age of mother at fi rst marriage and age at fi rst birth. Table 2: Trends in lone parenthood, 1980 to 2008 Number of lone parent families Lone parents as % of all families with children Lone parents as % of jobless families with children 000s % % Source: ABS and ; Data provided by FaHCSIA. Social Inclusion Family Joblessness in Australia 11

12 There was a sharp increase in the divorce rate immediately following the 1975 Family Law Act, which then fell but not back to the pre-1975 level. However, divorce rates are only an indirect indicator of lone parenthood currently just half of all divorces involve children under 18 years, a proportion which has fallen since the 1970s (ABS, Australian Social Trends, 2008). Moreover, an increasing share of families is cohabiting, and if these relationships end the result is separation rather than divorce. Further, the share of births outside marriage among all births has increased signifi cantly to around one-third in Births outside marriage, however, are only an indirect indicator of lone parenthood the share of births outside marriage acknowledged by the father 3 has increased over time to 90% of these births, although not all of these joint registrations are for couples who are actually cohabiting. An indicator of very long term trends in the levels and characteristics of lone parenthood is given in the Family Characteristics and Transitions Survey (ABS Cat. No ). This survey found that among people aged 18 to 34 years, around 24% had experienced the divorce or separation of their parents before the age of 18 years; this proportion was lower for older age groups, with 18% of persons 35 to 44 years having experienced divorce or separation in childhood, 12% of those aged 45 to 54 years, and 8% or less of older age groups. In contrast, however, 19% of persons aged 75 years and over and 14% of those aged 65 to 74 years had experienced the death of a parent during their childhood, compared to only 5% of persons less than 34 years of age and roughly 10% of persons aged between 35 and 64 years. Putting these two experiences together shows that around 25% of persons over 75 years had the experience of lone parenthood or being an orphan during their childhood, compared to 30% of persons less than 35 years of age. In fact, the age group with the lowest share of those experiencing lone parenthood in childhood were those aged 55 to 64 years (18%) the fi rst cohorts of the baby boom generation because of their lower experiences of bereavement in childhood than their parents generation and because of their lower experience of divorce and separation than subsequent age groups. 3 This is measured by the presence of the father s name on the birth certifi cate Trends in lone parenthood

13 5 The job gap As discussed below, lone parents and lone mothers in particular experience much higher unemployment rates than partnered mothers, with the unemployment rates of lone mothers consistently being twice as high as the unemployment rate of partnered mothers over the last decade. Having said this, the main contributing factor to joblessness among lone parent families is their higher rates of not participating in the labour force either for part or whole of the year. Table 3 shows job gaps for lone mothers and partnered mothers, simply calculated as the difference between the employment rate of lone mothers and partnered mothers, according to the age of their youngest dependent child and the number of children they have. Job gaps are consistently wider for lone parents with two or more children compared to those with one child, but the gaps widened for both family groups. At the beginning of the period shown (1994), job gaps were widest for lone parents with a youngest child aged less than fi ve years, and fell for those with older children. However, over this particular period, job gaps widened for those with a youngest child less than fi ve years and also for those with a youngest child aged 10 to 14 years, and it was only for those with a youngest child aged 5 to 9 years that job gaps narrowed. Using data from the 2001 Census, it is possible to identify the differences in employment rates of lone mothers and partnered mothers in more detail. Table 4 shows that for both partnered mothers and lone mothers full-time employment rates increase signifi cantly with the age of children from less than 10% when the youngest child is under one year of age to close to 40% when the youngest child is 14 years of age. Part-time employment rates also increase with age, although for partnered mothers they are highest in the early school years (6 to 11 years), and decline thereafter, as mothers are then more likely to work full-time. 4 It is also apparent that up until the age of around 10 years that difference in part-time employment rates are much more signifi cant than differences in full-time employment. 4 Within the partnered group, there is also a job gap between married and de facto partners, with a larger gap in part-time work for all ages of children and in full-time work for those with a youngest child less than 3 years of age, but de facto wives with children over 3 years of age are more likely to be working full-time than corresponding married mothers. Social Inclusion Family Joblessness in Australia 13

14 Table 3: Job gaps between lone and partnered mothers by age of youngest child and number of children, 1994 to 2005 Percentage point difference between the employment rates of lone and partnered mothers Age of youngest child Number of children Aged up to 4 years Aged 5 to 9 years Aged 10 to 14 years One child Two or more Source: ABS and ; Data provided by FaHCSIA. Table 4: Employment rates and job gaps between lone and partnered mothers by age of youngest child and full-time and part-time employment status, 2001 Full-time and part-time employment rates and percentage point difference between the employment rates of lone and partnered mothers Age of youngest Partnered mothers Lone mothers Difference child Years Full-time Part-time Full-time Part-time Full-time Part-time Total Source: 2001 Census; Data provided by FaHCSIA The job gap

15 6 Duration and dynamics of joblessness As noted above, the extent of family joblessness depends on the time period over which it is measured. Both family composition and joblessness and other labour force characteristics need to be seen as dynamic processes and not fi xed states. The data used in Sections 3 to 5 comes from an ABS survey that looks at the labour force status of families during one week in a specifi c month in each year. If joblessness is defi ned as referring to the entire year or to longer periods then the rate of joblessness is lower. Table 5: Labour force experience by family status, 1997 to 2007 % of group In labour force part of year Worked part of year Unemployed part of the year Not in labour force part of the year Never in the labour force Never worked 2007 Female lone parents All lone parents Other mothers Female lone parents All lone parents Other mothers Female lone parents All lone parents Other mothers Female lone parents All lone parents Other mothers Female lone parents All lone parents Other mothers Female lone parents All lone parents Other mothers Source: Calculated from Labour Force Experience, Australia, ABS Cat. No , various years. Social Inclusion Family Joblessness in Australia 15

16 For example, in 2007 around 45% of lone parents were jobless at the time of the monthly survey (Table 1), but Table 5 shows that 37% were jobless for the entire year. 5 Between 1997 and 2007 the proportion of female lone parents jobless for the entire year fell from 47% to 39%. This decline of around 17%, however, was not as pronounced as the decline of around 27% in the joblessness rate for lone parents measured in the monthly survey. Table 5 also shows the unemployment experience of lone parents is signifi cantly higher than that of partnered mothers in 2007, 8.0% of partnered mothers were unemployed for part of the year compared to nearly 20% of lone mothers (and slightly higher for lone fathers). As noted earlier, in terms of the total joblessness of lone parents non-participation in the labour force is more signifi cant than unemployment, with 28% of lone mothers not having been in the labour force at all during the year, so that non-participation accounted for close to 75% of their experience of joblessness. However, in terms of the difference in joblessness between lone and partnered mothers, it is unemployment that is more signifi cant, accounting for roughly 60% of the difference in joblessness between the two groups of mothers. 6 5 It is important to note that the total number of lone parents during any year will be greater than the number at a point in time during the year since people move into and out of lone parenthood. 6 This can be calculated by comparing the differences between the share of each group who never worked and the share who were never in the labour force, since by defi nition this difference is due to people who were participating in the labour force but were unable to fi nd a job Duration and dynamics of joblessness

17 7 The dynamics of family joblessness, child poverty and welfare receipt 2001 to 2005 As discussed earlier, a further source of data on joblessness is the Household, Income and Labour Dynamics in Australia (HILDA) Survey, a household based panel study which began in 2001, and in which panel members are traced over time, with the most recent Statistical Report presenting information on trends between 2001 and The HILDA Statistical Report (Headey and Warren, 2008) defi nes joblessness as not being in employment for at least 26 weeks in the survey year. This is a broader defi nition than used in the Labour Force Experience Survey and is therefore likely to give higher estimates than if permanent joblessness was measured (i.e. not in employment at any time in the year). Table 6 summarises a range of relevant results from the latest Hilda Statistical Report: overall, between 2001 and 2005: Between 9 and 12% of the working age population was jobless during each year, with an apparent decline in this percentage over the fi ve-year period; but close to 19% of the population of working age individuals experienced a spell of joblessness at least once in these fi ve years; 3% were jobless for three years, 2% for four years, and a further 3% for fi ve years. The joblessness rate for lone mothers was much higher, but fell more signifi cantly from around 45% to 30% between 2001 and 2005; but 55% of lone mothers experienced a spell of joblessness at some stage in the fi ve years; around 14% of lone mothers were jobless for three years, 7% for four years and 20% for all fi ve years. Between 2001 and 2005, the share of all children who experienced joblessness fell from 18% to 14%, but more than one quarter of all children experienced a period of joblessness in the 5 year period; with 10% being in jobless families for four or fi ve years. Between 2001 and 2005, the share of children in lone mother households who experienced joblessness fell from 52% to 44%, but more than two-thirds experienced a spell of joblessness during the fi ve year period; with 8% being in jobless families for four years and a further 29% for all fi ve years. The risk of long-term joblessness is thus much higher for children in lone parent families than for children generally the ratio of long-term joblessness among children in lone parent families was more than fi ve times as high as for children generally, whereas in any one year the ratio was around three to one. Social Inclusion Family Joblessness in Australia 17

18 The extent of welfare reliance (i.e. government transfers accounting for more than 50% of gross household income) among lone mothers is higher than is joblessness, since families can still receive payments with relatively high levels of earnings, but like joblessness, welfare reliance fell signifi cantly between 2001 and 2005 from 54% to 39%; only around one-quarter of lone mother households were never welfare reliant, with 17% being reliant for four years and 37% for all fi ve years. 7 Among all children under 15 years, the prevalence of welfare reliance was broadly unchanged over these fi ve years at roughly 20% of all children; just over two-thirds of all children were never in welfare-reliant households, but close to 10% were reliant for all fi ve years. Rates of relative income poverty (at 50% of median equivalent income) for lone mothers are signifi cantly lower than either welfare reliance or joblessness among lone mothers, although there was only a slight decline in relative poverty in this group over a fi ve-year period. Around 44% of lone mothers never experienced a spell of poverty in this period and only 4% spent all fi ve years in relative poverty. For couples with children and for all children, poverty rates were lower again, but also showed no clear trend but with perhaps a slight increase for couples with children. Around 85% of couples with children and 75% of all children never experienced poverty, although this implies that the proportion of each group who ever experienced some poverty was about twice as high as suggested by annual data. 8 In summary, the last decade has seen a signifi cant improvement in family joblessness after many years of signifi cant worsening. This picture is confi rmed by monthly data on labour force status, by annual data on labour force experience, and by longitudinal data on long term joblessness. When looked at over a longer period, family joblessness is more common overall than when measured using annual data, but a signifi cant proportion of this is short term. However, there are also a number of very disadvantaged children and families more than 5% of all Australian children lived in households which were jobless for all 5 years, 20% of lone mother families were jobless for fi ve years, and close to 30% of children in lone mother households experienced family joblessness for 5 years. Thus, within a picture of strong improvements in the circumstances of the majority of the population and also in the circumstances of some of the disadvantaged, it is clear there is a smaller but still signifi cant group who appear to have remained in extreme disadvantage experiencing long-term joblessness and welfare reliance. 7 A further source of information on welfare reliance is the longitudinal survey of Centrelink customers, based on administrative data. For example, Gregory and Klug (2002, 2003) looking over the period 1995 to 2001, found that around 16% of lone parents who claimed parenting payment single (PPS) in 1995 left after one spell and around 18% were still on PPS in 2001, around 20% had multiple spells on PPS in the six years and 46% had multiple spells of PPS and other payments; of those who had two spells around 40% also received parenting payment partnered. People with single spells tend to have entered the system when they were older, were more likely to be home-owners and were more likely to have earned income; while people with multiple spells entered when young, were renters, and less likely to have earnings. Looked at in cross-section, the average length of time a lone parent spends on PPS is between 2 and 3 years, but over a long duration (the time with dependent children), the cumulative length of time could be as much as 12 years. 8 It should be noted that estimates of the extent of income poverty can be quite sensitive to where the line is drawn because Australia s targeted social security system is associated with a high degree of clustering around benefi t rates which are quite close to a 50% of median income poverty line The dynamics of family joblessness, child poverty and welfare receipt

19 Table 6: Family joblessness, welfare reliance and relative income poverty: prevalence and persistence, 2001 to 2005 Working age individuals Lone mother households Joblessness Welfare reliance Relative income poverty Children in all households Children in lone mother households Lone mother households All children under 15 Couples with children Lone mothers Children under 15 years Cross-sectional results: prevalence in given year Persistence: Number of years in situation Never year years years years years Notes: Joblessness is defi ned as living in a household in which no member was in paid work for at least 26 weeks in the preceding year; welfare reliance is defi ned as living in a household where more than 50% of gross income came from government income support payments or family payments; income poverty is defi ned as living in a household with a disposable income adjusted for household size (equivalized) less than 50% of the median equivalized income. Source: Headey and Warren, Social Inclusion Family Joblessness in Australia 19

20 8How does Australia compare? Comparisons of the extent of family joblessness in different countries needs to take account of at least two main sets of factors: fi rst, how patterns of family and household formation and composition differ across countries and second how levels and patterns of employment differ. 8.1 Changing patterns of family formation As in Australia, family life and the nature of partnerships between adults have changed in most OECD countries. Fewer marriages are taking place than in the past; on average across the OECD marriage rates have fallen from 8.1 marriages per 100,000 persons in 1970 to 5.1 in There is considerably variation across countries; marriage rates remain highest in Turkey, Denmark and the US and lowest in Belgium, Greece and Italy; Australian marriage rates are above average and eighth highest in the OECD. Over the same period, most OECD countries recorded signifi cantly higher crude divorce rates, except Australia, Luxembourg and the US (Stevenson and Wolfers, 2007), in fact, since 1970, Australia has had the lowest increase in crude divorce rates of any OECD country. In 2004, on average across the OECD the crude divorce rate was 2.3 per people, twice the level recorded in 1970, and this upward trend has contributed to an increase in the number of single parent families and the number of children living in reconstituted families. Australia is just above the OECD average. In 2004, on average across the OECD, there were just over 4 divorces per 10 marriages. In Belgium there were 3 divorces for 4 new marriages; at the other extreme, in Mexico there was only 1 divorce per 10 newly formed marriages. The decline in the marriage rate has been accompanied by a tendency to defer the age at which it occurs. On average, the mean age of women at fi rst marriage has increased from 23.3 in 1980 to 27.7 years in In Denmark, Iceland and Sweden a women who gets married for the fi rst time is on average over 30 years of age (Table 7). In many Nordic and continental western European countries and in Australia and New Zealand, cohabitation is increasingly regarded as an alternative to marriage rather than merely a trial leading up to formal marriage, and in Norway and Sweden there are now more adults in the age group who are cohabiting than in a formal marriage (OECD, 2007b). In fact, across the OECD the mean age of women at fi rst marriage is now higher than the mean age of women at fi rst childbirth. Therefore, it is not surprising that the number of children being born out-of-wedlock has increased rapidly. At least 4 out of 10 children are born outside marriage in Denmark, Finland, France, New Zealand and the UK, whereas the rate is over half in Norway and Sweden and two out of three children in Iceland. At around a third of all births, the out of wedlock birth rate in Australia is only slightly above average, but the distribution is very skewed, ranging from less than 2% of all births in Korea and Japan and only 5% in Greece to more than half in some of the Nordic countries, as noted earlier. Teenage motherhood has become relatively rare in most countries, but teenage birth-rates remain relatively high in Mexico, Turkey and the United States. The teen birth rate in Australia is below the OECD average, but again the distribution is very skewed, ranging from 3.5% in Korea to 67% in Mexico. Overall, the picture of family relationships and composition in OECD countries suggests that at the two extremes there are societies with conservative family values low lone parenthood, relatively low divorce rates and low rates of births outside marriage and at the other end of the 20 8 How does Australia compare?

21 spectrum, countries with liberal family values above average lone parenthood, high divorce and high rates of birth outside marriage. Australia is towards the liberal end of this spectrum, but is not among the most liberal countries. It is also worth noting that with the exception of the lowest income OECD countries (Turkey and Mexico), conservative countries also tend to have the lowest fertility rates. Table 7: Selected family statistics, OECD countries Total fertility rate Mean age of women at first birth Mean age of women at first marriage Births out-of-wedlock Teenage birth rate** * Australia Austria Belgium Canada Czech Republic Denmark Finland France Germany Greece Hungry Iceland Ireland Italy Japan Korea Luxemburg Mexico Netherlands New Zealand Norway Poland Portugal Slovak Republic Spain Sweden Switzerland Turkey United Kingdom United States OECD * Years of reference for total fertility rates Canada and the U.S. 2003; Korea, Mexico and New Zealand, 2004 ** Adolescence fertility rate; births per 1,000 women age Source: D Addio and Mira d Encole (2005), Trends and Determents of Fertilty Rates in OECD Countries: The Role of Policies; and OECD (2007), Society at a Glance. ( Social Inclusion Family Joblessness in Australia 21

22 8.2 Australia s employment performance in perspective Australia s overall employment performance compared to the other 29 OECD countries is very strong. Adjusting for statistical differences (Productivity Commission, 2007), Australia s workforce participation rate in 2005 was 65.5 per cent, with Australia s ranking in the OECD at 5th place, behind Iceland, New Zealand, Canada and Switzerland, and in 2008 Australia had the 8th lowest unemployment rate in the OECD. But relatively low participation rates are recorded for prime aged males (25 to 54 years), child-bearing aged women (25 to 44 years), particularly women with children less than six years of age, and older men and women (55 to 64 years). In contrast, for youth (15 to 24 years), Australia ranked 2nd highest among OECD countries. The Australian workforce is also characterised by a high prevalence of part-time work. In 2005, 2.8 million Australians or around 29 per cent of the workforce were employed part-time. Most part-time workers are women 72 per cent in Around 46 per cent of female employees and 15 per cent of male employees participate on a part-time basis. The level of part-time employment is the third highest in the OECD after the Netherlands and Switzerland. Thus, while Australia s overall employment performance is strong this is largely due to higher levels of youth employment and high levels of part-time employment. 8.3 International comparisons of family joblessness Table 8 shows the relationship between individual, household and family joblessness in On the basis of individual joblessness, Australia ranks eighth lowest in the OECD 9 and is 6 percentage points below the OECD average. When looking at the share of the working age population living in jobless households, in contrast, Australia is the fi fth highest in the OECD and more than 4 percentage points above the OECD average. On the measure of jobless households with children, Australia is again the fi fth highest in the OECD and 5.5 percentage points above average. The disparity between individual or overall joblessness and household or family joblessness can be assessed by looking at the ratios of these measures: on this basis, only three OECD counties have higher ratios than Australia (Germany, Norway and the United Kingdom), and when the focus is on families with children only Germany and the United Kingdom have higher ratios. The ratio of family to individual joblessness in Australia is more than twice the OECD average. In part, this refl ects the fact that some countries have very low ratios between family joblessness and overall individual joblessness notably, the Southern European countries (Greece, Italy, Portugal and Spain), the East Asian countries (Japan and Korea) and the low income OECD countries (Mexico and Turkey), but also a disparate group of high income OECD countries (the Nordic welfare states, except Norway), Luxembourg, Switzerland, Canada and the United States. Using a measure of polarisation how does the actual level of family joblessness compare with the predicted level (based on the overall employment rate in each country) 10 Figure 2 shows that Australia actually has the second highest level of polarisation for families with children in the 9 These fi gures are not adjusted for the statistical differences referred to earlier for details, see Productivity Commission (2007). 10 Polarisation is calculated as follows: for each country, the non-employment rate among the working-age population is the starting point; if this is 30%, for example, then it is assumed that if non-employment were random, then 30% of lone parent households would be predicted to be jobless, while for couples with children the predicted level of joblessness would be the joint probability of both adults being jobless, i.e. 30% multiplied by 30% or 9%. For each country, the predicted level of family joblessness is then the weighted average of these probabilities (depending on the share of lone parent and two-parent families). The predicted level of family joblessness is then compared to the actual level of family joblessness, with the percentage point difference being characterised as polarisation. See Gregg (1996), Dawkins, Gregg and Scutella (2002) and Gregg, Scutella and Wadsworth (2008) for details How does Australia compare?