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

1. Introduction

Aggregate child poverty rates can change for many reasons. Factors such as housing tenure, family size, and employment patterns are all known to affect the risk of being in poverty, but the poverty risk associated with any given characteristic can change over time due to shifts in policy or the wider economy. For example, increased expenditure on in-work benefits might reduce the child poverty rate for families in work. Child poverty rates can also be impacted by demographic changes. For example, an increase in average family size would be expected to increase the overall child poverty rate, since families with three or more children tend to be at higher risk of experiencing poverty.

This analysis seeks to identify what has driven changes in relative child poverty rates in Scotland across a selection of household characteristics. For each characteristic, the analysis employs a method previously used by the Institute for Fiscal Studies and the Joseph Rowntree Foundation to decompose^[1] changes in aggregate child poverty rates into two effects: ^[2]

• The incidence effect represents the change in the risk of being in poverty for different groups, the impact of which will depend on the size of each group relative to the total population.
• The compositional effect represents the change in the relative size of different groups, the impact of which will depend on the initial poverty risk faced by each group.

The analysis focuses on four key characteristics: housing tenure, family type, work status, and number of children in the household. All of these characteristics are related to the drivers of poverty and are known to influence poverty risk. Future analysis could look at the characteristics of the other priority groups which are: disability, age of the mother, age of the youngest child, and ethnicity. Although, the ability to analyse these groups will depend on sample sizes.

The analysis uses the same data as of the official child poverty statistics: the Households Below Average Income (HBAI) dataset. This approach is in line with the official statistics which also use three years of data, but differs in that the analysis is based on the pooled dataset rather than on averages of the individual years.

For the pooled analysis, the financial years 2015-16 to 2017-18 were used as a baseline period before the Child Poverty (Scotland) Act 2017 came into force, to be compared against the latest three years of data, 2021-22 to 2023-24. The analysis focuses on one of the four target measures - relative child poverty after housing costs (AHC), with results also produced on a before-housing-cost (BHC) basis to help explain and contextualise the findings.

Meanwhile, the single-year analysis compared child poverty rates in the latest available years of data, 2022-23 and 2023-24. However, due to the limited sample sizes available in each individual year, there was insufficient data to break down the characteristics into meaningful subgroups and be confident that the outcomes were robust. For example, families with three or more children as a group had insufficient sample sizes to be analysed separately and have a distinct experience to families with one or two children. As such, the single-year analysis is not presented in this report.

While the decomposition analysis helps to highlight proximate causes, these causes must themselves be explained in light of the economic and policy context in order to understand what is ultimately driving changes in child poverty.