The Scottish Government

« Previous | Contents | Next »

Listen

CHILD POVERTY IN SOCIAL INCLUSION PARTNERSHIPS

CHAPTER THREE CHILD POVERTY IN SCOTLAND

INTRODUCTION

3.1 This chapter examines child income poverty in Scotland. The first part of the chapter focuses on the incidence of poverty in Scotland compared with elsewhere in Britain. The second part examines the incidence of child poverty among different types of household in Scotland. By 'incidence' is meant the extent or rate of poverty, that is, the percentage of children who fall below a certain level of income. The third part of the chapter examines the composition of child poverty in Scotland. That is to say, it looks at the characteristics of the families of children who were experiencing income poverty. It also compares them with the characteristics of the families of children who were not experiencing income poverty. In the fourth part of the chapter, geographical differences in the incidence and composition of child poverty within Scotland are explored. The fifth part of the chapter examines the odds of children in particular types of family being poor when other factors are taken into account.

3.2 The analysis in the first 2 parts of the chapter is based on the Households Below Average Income ( HBAI) dataset for the financial year 1999/00, which is produced by the Department for Work and Pensions. The analysis in the third, fourth and fifth parts of the chapter is based on the Scottish Household Survey ( SHS) for the calendar years 1999 and 2000.

3.3 With the exception of Table 3.1, which focuses on individuals, the unit of analysis in this and the following chapter is children. Tables 3.1 to 3.3 - which are based on the HBAI - exclude the self-employed, but all other tables in this and the following chapter include them. In both of these chapters, the term 'child poverty' is used to refer to poverty of income.

3.4 As noted in Chapter One, income is defined as disposable income after housing costs have been deducted, equivalised to take into account the size and composition of the household. The tables show results for all 3 of the income thresholds - 50 per cent, 60 per cent and 70 per cent of median income - but the discussion in the text focuses on the 60 per cent threshold. For ease of exposition, the text generally refers to children in households with incomes below the latter threshold as children in poverty.

SCOTLAND COMPARED

3.5 Although the focus of this report is on child poverty, it is helpful to place this within the wider context of poverty more generally. Table 3.1 shows, for each of the 3 nations that comprise Great Britain, the percentage of individuals below 3 different income thresholds.

3.6 In 1999/00, the incidence of poverty in Scotland was broadly similar to that in Britain as a whole. Thus, 23 per cent of individuals in Scotland had an income after housing costs that was below 60 per cent of median income, exactly the same percentage as for Great Britain (Table 3.1).

Table 3.1 Incidence of poverty among individuals by nation, 1999/00

The figures are arranged in rows and exclude the self-employed.

Source: DSS (2001) Households Below Average Income, 1994/5-1999/00, Table 3.2

3.7 Perhaps surprisingly, Scotland had a lower proportion of children in poverty than either England or Wales. In 1999/00, 29 per cent of children in Scotland were living in households with an income that was less than 60 per cent of the GB median. This compares with 32 per cent in England and 36 per cent in Wales (Table 3.2). It is not clear whether this difference is significant or reflects sampling error in the survey from which the figures are derived.

Table 3.2 Incidence of child poverty by nation, 1999/00

The figures are arranged in rows and exclude the self-employed.

Source: DSS (2001) Households Below Average Income, 1994/5-1999/00, Table 3.4

THE INCIDENCE OF CHILD POVERTY

3.8 One way to examine the data on child poverty is to look at the incidence of poverty among children in different types of household. Table 3.3 shows the percentage of children below the 3 income thresholds in 1999/00 according to different family or household characteristics.

3.9 It is clear from Table 3.3 that work is an important factor affecting the incidence of child poverty. In 1999/00, the percentage of poor children (i.e. an income below 60 per cent of the GB median) was lower:

• Among children in families where someone was in full-time work than among families that were self-employed or in part-time work; and
• Among children in families where someone was in paid work than among families where no one was in work.

3.10 For example, only 8 per cent of children in families where an adult was in full-time work were poor. By contrast, 76 per cent of children living in workless households were poor. The stark contrast between these figures underlines the importance of paid work as a critical factor affecting the incidence of child poverty in Scotland.

3.11 Family type is also significantly related to the incidence of child poverty. The rate of child poverty was very much higher among lone parents than among couples. Thus, 58 per cent of children in lone parent households were poor, compared with only 18 per cent among children living with couples (Table 3.3).

3.12 Table 3.3 also suggests that there is a parental age effect in relation to the incidence of child poverty. Thus, 60 per cent of children in households headed by a young adult aged under 25 were poor. This compares with 31 per cent where the head of household was aged from 25 to 44, and 12 per cent where they were aged 45 or more. In other words, the majority of children living with young parents (that is, under 25 years of age) were experiencing poverty in 1999/00.

Table 3.3 Incidence of child poverty by family or household characteristics, Scotland 1999/00

The figures are arranged in rows and exclude the self-employed.

Source: HBAI 1999/00 (own analysis)

3.13 The child poverty rate was substantially higher among children in large families than among small families. In 1999/00, two-fifths of children in families with 3 or more children were poor in Scotland. This compares with a child poverty rate of about a quarter among children in families with only one or 2 children (Table 3.3).

3.14 Finally, there was a strong association between housing tenure and child poverty. The rate of child poverty was very high among children living in social rented housing and, to a lesser extent, among children living in private rental housing. Thus, 6 out of 10 children in social housing and 4 out of 10 in private rental housing were poor in 1999/00. The rate among children in owner-occupation was only one in 10 (Table 3.3).

THE COMPOSITION OF CHILD POVERTY

3.15 Another way to examine child poverty is to focus on the composition of the population of poor households containing children. Table 3.4 shows the socio-economic characteristics of households with children having an income below the 3 poverty thresholds, those at or above the 70 per cent threshold, and all children in Scotland.

3.16 The figures in Table 3.4 are arranged in columns and refer to the percentage of children. For example, 13 per cent of children below 60 per cent of median income were living in households headed by someone aged 45 or more. This compares with 16 per cent of all children in Scotland.

3.17 Children in poverty were disproportionately likely to be living with lone parents. Thus, while 20 per cent of all children were living in lone parent households, 37 per cent of children in poverty were living in this type of household (Table 3.4). This strong association between lone parenthood and child poverty has also been found in studies of child poverty in Britain as a whole (e.g. Adelman and Bradshaw, 1998; DSS, 2001).

3.18 Families with 3 or more children accounted for a disproportionately large share of children experiencing poverty. Families of this size made up 3 out of 10 children in Scotland but 4 out of 10 poor children (Table 3.4). Again, this association between families with 3 or more children and child poverty echoes the findings for Britain as a whole (Adelman and Bradshaw, 1998).

3.19 There was also a clear association between housing tenure and child poverty. Thus, while 34 per cent of all children were living in social rented housing, as many as 64 per cent of poor children were living in this tenure. And although 62 per cent of all children were living in owner-occupied housing, only 29 per cent of poor children were doing so (Table 3.4).

3.20 Poor children were disproportionately likely to be living in a household where someone had a limiting illness, a health problem or was disabled. In 1999/2000, almost a quarter of children experiencing poverty were living in such households, compared with only one in 6 of all children in Scotland (Table 3.4).

Table 3.4 Composition of child poverty by family or household characteristics, Scotland 1999/2000 (% of children)

HIH = highest income householder

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

3.21 Finally, children in income poverty were disproportionately likely to be living in households where the adults had no educational qualifications. Thus, one in 3 poor children were living in such households compared with one in 5 of all children (Table 3.4).

3.22 Table 3.5 confirms the importance of work as a factor helping to protect children against income poverty. In Scotland as a whole, 8 out of 10 children were living in families where the highest income householder was in employment. However, only 4 out of 10 poor children were living with highest income householders who were in paid work. Most of the remainder were either jobseekers or looking after the home. Only a small minority of children was living with highest income householders that were permanently retired, long-term sick, or disabled (Table 3.5).

3.23 Whatever the reason for the lack of employment, it is clear that there is a strong association between worklessness and child poverty in Scotland, as there is in Britain as a whole (HM Treasury, 1999). Thus, in Scotland in 1999/2000, over half of poor children were living in households where no one was in paid work, compared with one in 5 of all children in Scotland. In contrast, 9 out of 10 non-poor children were living in households where someone was in work (Table 3.5). This finding provides support for the Government's emphasis on paid employment as a route out of poverty.

3.24 Interestingly, children living with lone parents in paid work did not account for a disproportionate share of children in poverty. Only 9 per cent of poor children were living in households headed by a lone parent in work. This is exactly the same percentage as among children who were not poor (Table 3.5). This indicates that employment status is a more important child poverty risk factor than family type (Adelman and Bradshaw, 1998). This is an important finding because it means that it is not lone parenthood per se that is the critical factor underlying child poverty among this demographic group, but lack of employment.

3.25 While child poverty was clearly associated with worklessness, employment was not necessarily sufficient to prevent children from experiencing poverty. Indeed, 4 out of 10 of poor children were living in households where someone was in employment (Table 3.5). Presumably, the pay and any in-work benefit income they were receiving were insufficient to raise them above 60 per cent of median income.

3.26 In other words, work may be the surest and best way out of poverty, but it is by no means a guaranteed escape route. Low pay can be a cause of child poverty. However, the recent increases in the national minimum wage, Working Families Tax Credit, and Child Benefit may well have reduced the number of working families experiencing child poverty since this survey data was collected.

3.27 Children in poverty were much more likely than other children to live in households that had no savings or investments. Thus, in 1999/2000, 8 out of 10 poor children were living in households with no savings. This compares with only 4 out of 10 non-poor children (Table 3.6).

Table 3.5 Composition of child poverty by economic status of household or highest income householder, Scotland 1999/2000 (% of children)

The symbol -- indicates figures that are less than 0.5%

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

3.28 Where households did have savings, the amounts involved tended to be smaller among children in poverty than among other children. About 2 out of 5 poor children were living in households whose savings were less than 1,000. This compares with only one in 5 non-poor children in households with savings (Table 3.6).

3.29 Children experiencing income poverty were much less likely than children who were not in poverty to be living in households that had a bank or building society account. A third of poor children were living in households without a bank account. This compares with only one in 20 non-poor children (Table 3.7).

3.30 Finally, as Table 3.7 indicates, poor children were much less likely than other children to be living in households that had:

• access to a car,
• access to the internet at home, or
• a home contents insurance policy.

Table 3.6 Composition of child poverty by household savings, Scotland 1999/2000 (% of children)

* These percentages include only children living in households that did have savings.

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

Table 3.7 Composition of child poverty by selected household assets, Scotland 1999/2000 (% of children)

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

THE GEOGRAPHY OF CHILD POVERTY

3.31 Child poverty was not distributed in line with the distribution of the general population across Scotland. Table 3.8 shows the incidence of child poverty in urban compared with rural Scotland in 1999/2000. The percentages have been adjusted to bring them into line with the Households Below Average Income figures produced by the Department for Work and Pensions. This adjustment was necessary because of differences in the survey coverage and definition of income between HBAI and SHS, as noted in Chapter One and elaborated upon in Appendix 1.

3.32 There is a higher incidence of child poverty in urban than in rural Scotland, taking these categories as a whole. For instance, 31 per cent of children in urban areas, compared with 25 per cent in rural areas, were below the poverty threshold of 60 per cent of median GB income.

Table 3.8 Incidence of child poverty by urban or rural locality, Scotland 1999/2000

The figures are arranged in rows. They have been adjusted to bring them into line with HBAI.

Source: SHS 1999/2000 (own analysis)

3.33 Given the higher child poverty rate in urban compared with rural areas, it is not surprising to find that child poverty was disproportionately concentrated in urban Scotland. Thus, in 1999/2000 urban areas accounted for 68 per cent of all children in Scotland, but for 73 per cent of children in income poverty. By contrast, rural areas accounted for 32 per cent of all children but for only 27 per cent of children experiencing poverty.

Table 3.9 Composition of child poverty by urban or rural locality, Scotland 1999/2000 (% of children)

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

3.34 Even within urban and rural areas, child poverty was not uniformly distributed. For example, Glasgow had a disproportionately large share of child poverty in Scotland. It accounted for 12 per cent of all children in Scotland, but as many as 20 per cent of children in poverty. Edinburgh's share of child poverty was broadly in line with its share of all children: it accounted for 8 per cent of children and for 7 per cent of children experiencing poverty.

3.35 Table 3.10 uses the Scottish Mosaic classification of neighbourhoods, which was developed by the company Experian, to examine the distribution of child poverty across different types of residential area. It shows that children in poverty were much less likely to live in high income neighbourhoods or areas dominated by middle income owners than children who were not experiencing poverty. By contrast, children in poverty were disproportionately likely to live on disadvantaged council estates or in neighbourhoods dominated by families in council flats.

Table 3.10 Composition of child poverty by type of residential area (mosaic groups), Scotland 1999/2000 (% of children)

The figures are arranged in columns.

Source: SHS 1999/2000 (own analysis)

THE ODDS OF CHILDREN BEING IN POVERTY

3.36 So far the analysis has focused on what are called bivariate relationships, that is, relationships between child poverty and particular family or household characteristics, considered one at a time. For example, it was shown that there is a statistically significant relationship between child poverty and lone parenthood, and between child poverty and paid employment. However, paid employment and lone parenthood may themselves be related to each other, in which case it is important to disentangle the separate effect of each of these factors on child poverty.

3.37 It is possible to isolate the independent effect of each factor on child poverty using multivariate statistical techniques. One such technique, known as logistic regression, provides estimates of the odds of being in poverty for children in households having each characteristic of interest, while holding all other characteristics constant, compared to some reference category. For example, it can show the odds of a child in social housing being in poverty compared to a child in owner-occupation, while holding constant all other characteristics of interest (private renting, lone parenthood, employment, number of children in the family, etc).

3.38 Table 3.11 shows the odds of children in families or households of particular types being in poverty (i.e. below the poverty threshold of 60 per cent of GB median income). As noted above, the odds of children being in poverty for each type of household or family are calculated in comparison with a reference category, while holding all other characteristics constant. For example, holding other factors constant, in 1999/2000 the odds of children in Scotland being poor were 1.79 times higher where the ethnic background of highest income householders was not white than where it was white.

3.39 The odds for the reference category are by definition always 1.0. The asterisks indicate the level of statistical significance, if any, that can be attached to the result. No asterisk means that the odds are not significantly different from the odds for the reference category.

3.40 The odds ratios are estimates only, the value of which tends to vary to some degree according to the particular mix of variables included in the logistic regression analysis. They should therefore be seen as approximate orders of magnitude rather than precise values; and this is especially true where the odds are high.

3.41 The age of the highest income householder was divided into 4 categories and the odds of children being poor estimated for each category relative to the 16 to 24 category. Surprisingly, the odds of children being poor were greater where the highest income householder was aged from 25 to 34 than where they were aged from 16 to 24. ¹ The odds of children being in poverty were not significantly different in households where the highest income householder was aged either 35 to 44 or 45 and above from those where they were aged 16 to 24 (Table 3.11).

3.42 Table 3.11 shows that the odds of children being poor were significantly higher in families comprising 3 or more children than in families that were smaller than that. Thus, the odds were 1.9 times higher for children living in families of 3 or more children than for families comprising only one child. This confirms the importance of family size as a risk factor for child poverty.

3.43 The odds of children being in poverty were higher where the highest income householder had no educational qualifications than where they had some, though the difference was relatively small. This highlights the importance of parental educational attainment as an important risk factor for income poverty among children. ²

Table 3.11 The odds of children being below 60 per cent of median income, Scotland 1999/2000

Odds are based on logistic regression 'best-fitting' model estimates.

Significance levels: * <0.05; ** <0.01; *** <0.001.

Source: SHS 1999/2000 (own analysis)

3.44 The odds of children being in poverty are also significantly related to housing tenure. Children in social housing were 1.98 times more likely to be poor than were children in owner-occupation, when other factors were held constant. The odds of being in poverty were even greater (2.6 times) for children in private renting.

3.45 Child poverty was also higher where the household did not have access to a car compared with where they had access to one or more cars (Table 3.11). This result may reflect the fact that access to a car can enable people who might otherwise be unemployed to take up jobs that are located some distance from their home.

3.46 By far the most important factor affecting the odds of child poverty was the employment status of the household. Children in couples where only one partner was in paid work were nearly 6 times more likely to be poor than children in couples where both worked. The odds of children in workless couples being poor were 38 times higher (though see Paragraph 3.40). For children of lone parents who were in paid work, the odds were 3 times higher. Finally, for children of lone parents who were not in employment the odds were 21 times higher.

3.47 These results clearly indicate that paid work crucially affects the likelihood of children being in poverty, even when other factors are held constant. When it comes to reducing the odds of child poverty among couples, 2 partners in employment is better than one, and one in employment is much better than none.

3.48 The difference in odds between lone parents in work and those who were not in work was very large. However, the odds of children being poor (relative to the reference group of a couple where both partners worked) were greater for couples where one partner worked than for lone parents in work, holding other factors constant. This implies that the relationship between lone parenthood, paid employment and child poverty is more complex than is suggested by examination of the bivariate tables alone.

3.49 Finally, the odds of children being poor were lower where the highest income householder or their partner (if any) was in receipt of disability benefits than where they were not. The same was also true of children living in households where someone had a limiting illness, health problem or disability; and also where someone in the household required regular care or help (not included in the 'best-fitting' model). This at first sight surprising result may reflect the higher level of social security benefits paid to people with disabilities compared to other claimants.

3.50 However, it is important to note that, in line with the HBAI reports, the SHS income data were not equivalised to take into account the considerable extra expenses that disabled people incur compared with other people (DSS, 1998). Had these extra costs been taken into account, the true odds of children who live in households containing disabled people being poor would almost certainly have been much higher than the estimates shown in Table 3.11.

CONCLUSIONS

3.51 This chapter has examined child poverty in Scotland. It has shown that the incidence or rate of child poverty in Scotland in 1999/2000 varied according to family or household type. It was especially high among:

• adults with no educational qualifications
• workless households
• lone parents
• families with 3 or more children
• parents under 25
• tenants in social and private housing.

3.52 The rate of child poverty was particularly high among workless households - that is, households in which no adults were in paid employment or self-employment. Three quarters of all children in workless households in Scotland were living in poverty.

3.53 The rate of child poverty was found to be higher in urban Scotland than in rural Scotland. Correspondingly, urban areas accounted for a disproportionately large share of children in poverty. The City of Glasgow was also found to account for a larger proportion of children in poverty than its share of all children in Scotland. Residential areas characterised as 'disadvantaged council estates' and 'families in council flats' also accounted for a large amount of children in poverty. Thus, child poverty is not evenly spread across Scotland but is disproportionately concentrated in certain types of area.

3.54 Finally, the odds of children in Scotland being poor varied by family or household type, even when other factors were held constant. However, by far the most important determinant of child poverty was the employment status of the adults in the household. It is notable that, when other factors were controlled for, children living with lone parents in work had lower odds of being poor than either lone parents not in work or couples where only one partner was in work. Thus, employment status was more important than family type in determining whether children were living in poverty. These results confirm that work is the surest route out of poverty, even if - in 1999/2000 - it was not always a sufficient one.

« Previous | Contents | Next »