Tackling child poverty - progress report 2024-2025: annex C - decomposition analysis of the child poverty statistics

2. Results

This section includes the results of a three-year pool comparing 2015-18 to 2021-24. The compositional and incidence effects listed in these tables represent the impact on the total child poverty rate. The sum of compositional and incidence columns are equal to the total population change in child poverty columns.

2.1 Pooled years change 2015-18 to 2021-24

Compositional analysis of three year pooled changes allows us to understand what has driven changes in the relative child poverty rate in Scotland since the Child Poverty (Scotland) Act 2017 came into force. This enables the investigation of the underlying dynamics to understand the change, or lack of change, and the possible policy implications thereof.

2.1.1 Tenure type

Table 1 shows the decomposition of relative child poverty AHC between 2015-2018 and 2021-24 by tenure type. The analysis looked at households with children who owned their home outright or with a mortgage, those in the social rented sector, and those in private rented housing. The overall reduction in child poverty rates was driven by owners and private renters, with an offsetting effect from social renters. Of note, part of the reduction in relative child poverty driven by private renters was due to a reduction in private renters with children, relative to all households with children. As private renters have a higher than average poverty rate, with all else being equal, reductions in the size of this group reduce rates of child poverty. However, social renters saw both an increase in child poverty rate, as well as an increase in their relative group size. Due to the higher risk of child poverty in this group, this had the effect of offsetting much of the reduction in child poverty driven by owners and private renters.

Note: Tenure is estimated by household rather than benefit unit. Percentages may not sum due to rounding. Tables including poverty rate and the relative size of each subgroup are available in Appendix A.

Source: Author’s calculations based on the Households below average income (HBAI) data from the Family Resources Survey (FRS) in various years.

Relative child poverty rates Before Housing Costs (BHC) were also reviewed to understand the impact of housing costs on child poverty rates by tenure. Comparing the decomposition of child poverty rates before (table 2) and after (table 1) housing costs allows us to infer what is driving increases or decreases in poverty rates. When looking at social renting households and private renting households in particular, we see an increase in child poverty rates before housing costs over time. However, this increase over time becomes more muted for social renting households after housing costs were accounted for, moving from a 1.2% increase to a 0.2% increase in total child poverty caused by social renting households. Similarly, private renting households saw their relative child poverty rate drop after housing costs, with the corresponding impact changing total child poverty rates from an increase of 1% before housing costs from private renters to a decrease of 0.3% from private renters after housing costs are accounted for. This suggests that families experiencing child poverty have seen the impact of housing costs decrease relative to all other families, and that increases in relative child poverty are due to differences in relative income. This can be further seen when looking at households who own their house (either outright, or with a mortgage) where the change in child poverty rate over time is very similar when reviewed before and after housing costs.

Comparing the group poverty rates in table 2 with those in table 1 further shows that private renting households saw their child poverty rates increase the most after housing costs were taken into account. For example, in 2015-18 child poverty rates increased from 27% to 44% (17 percentage points), while in 2021-24 these increased from 34% to 42% (8 percentage points). Of note, the increase in child poverty rates seen for social renting households when accounting for housing costs is more muted than for private renting households, demonstrating the protective nature of social housing. This can be seen by the 8 percentage point increase in 2015-18 relative to the 17 percentage point increase seen for private renters in the same period, and only a 3 percentage point increase in 2021-24. Those who owned their homes saw almost no change in child poverty rates when comparing before and after housing costs.

Note: Tenure is estimated by household rather than benefit unit. Percentages may not sum due to rounding. Tables including poverty rate and the relative size of each subgroup are available in Appendix A.

Source: Author’s calculations based on the Households below average income (HBAI) data from the Family Resources Survey (FRS) in various years.

2.1.2 Family type

Table 3 shows the decomposition of relative child poverty AHC between 2015-2018 and 2021-24 by family type. We looked at two-adult households with children, and lone parent households. The decrease in child poverty amongst all households with children seen between periods was almost entirely due to a drop in child poverty rate for lone parent families. The composition of the family types was stable with only a negligible increase in child poverty rate for couples. We know from other data^[3] that lone parent households are more likely to do less paid work and have fewer children. We also saw reduction in poverty rates amongst household types who are workless (see table 4) or with fewer children (see table 5).

Note: Family type is estimated by benefit unit rather than household. Percentages may not sum due to rounding. Tables including poverty rate and the relative size of each subgroup are available in Appendix A.

Source: Author’s calculations based on the Households below average income (HBAI) data from the Family Resources Survey (FRS) in various years.

2.1.3 Working status

Table 4 shows the decomposition of relative child poverty AHC between 2015-2018 and 2021-24 by working status. ‘Full-time work’ refers to families where all adults are in full-time work; ‘mixed work’ refers to families where not all adults are in full-time work, but at least one adult is working; ‘none working’ refers to families where no adult is working. There was a notable divergence in child poverty rates for families in work between the two periods. Whereas the child poverty rate for families in the ‘none working’ group saw a reduction, the rate for families in the ‘mixed work’ group saw an increase. The increase in child poverty rates for families in ‘mixed work’ could be related to the reduction in number of hours worked for low income families noted in Annex A: Measurement framework. Meanwhile, there was little change in the child poverty rate among families in full-time work. The net effect of changes in the child poverty rate (incidence effect) for families in each work status group would have increased the overall child poverty rate, all else being equal.

This was, however, more than offset by the compositional effect, which was driven by an increase in families in ‘full-time work’ (who have a lower-than-average child poverty rate) and a corresponding fall in families in the ‘none working’ group (who have a higher-than-average child poverty rate). While there was also a decrease in families who had mixed work, this contributed little to the overall composition effect as the child poverty rate among this group was around average.

Although our method distinguishes between composition and incidence effects, the dynamics underlying these two effects can be interdependent. This appears to be the case for work status, since the incidence and composition effects see their impacts working in the same direction for each group. For example, the ‘none working’ group, which has the highest child poverty risk, saw both a reduction in its relative size and its child poverty rate. These two changes could be caused by distinct factors, but they could also be reflecting a shift of families experiencing child poverty moving from the ‘none working’ group to ‘mixed work’ or full-time work. This would both decrease the relative size of the ‘none working’ group but also the child poverty rate, if those moving were predominantly experiencing poverty.

Note: Working status is estimated by benefit unit rather than household. Percentages may not sum due to rounding. Full-time work refers to families where all adults are in full-time work; mixed work refers to families where not all adults are in full-time work, but at least one adult is working; none working refers to families where no adult is working. Tables including poverty rate and the relative size of each subgroup are available in Appendix A.

Source: Author’s calculations based on the Households below average income (HBAI) data from the Family Resources Survey (FRS) in various years.

2.1.4 Number of children

Table 5 shows the decomposition of relative child poverty AHC between 2015-2018 and 2021-24 by number of children in the family, grouped for families with three or more children. There was a notable divergence in poverty rates for families with one or two children and families with three or more children between the two periods. This was caused mainly by changes in child poverty rates, captured in the incidence effect, but was magnified by compositional changes.

The poverty rates for families with one child saw a decrease. This had a small impact on overall child poverty rates due to the relatively small number of families with one child. However, the largest impact came from a decrease in poverty rates for families with two children. The drop in child poverty rates for families with two children brings this group close to the 2030 overall child poverty targets. By contrast, child poverty rates for families with three or more children saw an increase, now being double that of families with one child.

Compositional changes amplified the impact on child poverty rates for these groups. The relative share of families with two children grew further, pushing down overall child poverty rates due to the below average child poverty rate for this group. Meanwhile increases in the relative share of families with three or more children pushed up poverty rates due to the higher than average child poverty rate in this group.

The change in child poverty rates and composition of the two-children families and three-or-more-children families almost entirely cancel out in their overall impact on child poverty. This divergence between the groups could be due to changes in policy such as the two-child benefit limit applied to families with children born after April 2017.

Note: Number of children is estimated by benefit unit rather than household. Percentages may not sum due to rounding. Tables including poverty rate and the relative size of each subgroup are available in Appendix A.

Source: Author’s calculations based on the Households below average income (HBAI) data from the Family Resources Survey (FRS) in various years.