Table 4 Summary of decomposition framework and indicators of interest

From: Intergenerational persistence of poverty in five high-income countries

Component

Parameter

Description

IGPov

β1

The association of childhood poverty with adult poverty, equivalent to the sum of F, M, T and R

Family background (F)

ρ1 − δ1

The influence of the following indicators on IGPov: share of childhood in home with no [woman/man] present; share of childhood in single-parent home; age of mother at birth; mean maternal employment rate during childhood; highest educational attainment of mother; and average number of children in home during childhood

Mediating benchmarks (M)

δ1 − γ1

The influence of the following indicators, observed in young adulthood, on IGPov, conditional on F: has high school degree; has more than high school degree; is employed; is employed and works more than 30 h per week; is a single parent with children present at home; is married; is married and spouse has more than a high school degree; and has another person in the home who is employed (at some point in adulthood)

Tax and transfer insurance (T)

(β1 − ρ1) − (θ1 − γ1)

The effect of taxes and transfers on insuring against observable family background characteristics and benchmark attainment

Residual (R)

θ1

The persistent association of childhood poverty with adult poverty that is not channelled through F or M and is not offset by T

  1. See equations (1)–(10) for formal descriptions of the parameters. All of the mediating benchmarks are measured as mean or maximum values over 25–35 years of age (see the Supplementary Information for precise variable definitions). In all models, we include a base set of controls for maximum and minimum age observed, sex, maximum year observed, whether one is living with one’s parents and the share of ages 0–5, 6–10 and 11–17 years, respectively, during which the individual was observed in our data. See section C of the Supplementary Information for descriptive statistics by country.