Tackling child poverty - progress report 2023-2024: annex C - decomposition analysis of the child poverty statistics

Executive Summary

The purpose of this decomposition analysis is to better understand what has driven changes in the relative child poverty rate after housing costs in Scotland since the Child Poverty (Scotland) Act 2017 came into force. This allows for the investigation of the underlying dynamics to understand the change, or lack of change, and the possible policy implications thereof.

The analysis follows the same method as used in previous decomposition studies by the Institute for Fiscal Studies and the Joseph Rowntree Foundation. The decomposition breaks down the overall change in child poverty by changes caused by subgroups, depending on the characteristic of interest. This is presented as the i) incidence effect, which represents the change in the risk of being in poverty for different groups, and ii) the compositional effect, which represents the change in the relative size of different groups. The analysis focuses on four main characteristics and their subgroups; household tenure type, family type, working status of families, and the number of children in each family.

Two separate time comparison analysis were undertaken. Pooled 3-year average changes were analysed, reviewing the change in relative child poverty after housing costs observed between 2015-18 and 2021-24. Single year changes were also observed between 2022-23 and 2023-24 to review changes in the latest available data. However, due to the limited sample sizes, the single year data could not be sufficiently broken down by subgroups to show meaningful trends. As such, only the pooled 3-year average analysis has been presented.

The analysis found that the number of children in the family and the family work status had notable effects between 2015-18 and 2021-24. First, there was a divergence in the child poverty rates for families with one or two children, which saw a decrease, and the child poverty rates for families with three or more children, which saw an increase. The combination of increasing child poverty rates for families with three or more children, and the increase in the relative size of this subgroup, worked to dampen the reduction in overall child poverty rates caused by families with one or two children. The divergence over time could be explained by policy differences such as the two child benefit cap which was implemented in 2017, as well as the legacy impact of benefit freezes.

Second, there was a divergence in child poverty rates for families according to their work status. There was a small increase in child poverty rates for families that had at least one adult working less than full-time. This was in contrast to a decrease in the poverty rate for families where no adults were in work, while families with all adults in full-time work saw almost no change. Of particular note was the slight change in makeup of the groups, with a small increase in families working full-time, and decreases in numbers in other groups. This helped to push down child poverty rates due to a reduced risk of child poverty for families in full-time work. These changes suggest that limited work intensity, alongside barriers to full-time work, are important factors in keeping relative child poverty rates from falling.

The evidence from the analysis of groups under the tenure and family type characteristics was less clear. This could imply that these characteristics were not central to driving changes in aggregate child poverty rates over the periods considered here, but it could also reflect limitations with the method, such as the relatively high level at which the characteristics are disaggregated. It is also important to remember that the groups are not mutually exclusive, and the effect of driving forces on poverty rates are therefore complex.