Data
Microdata pipeline
PolicyEngine constructs its representative household dataset through a multi-stage pipeline, starting from the Family Resources Survey and producing two weight matrices: one for 650 parliamentary constituencies and one for 360 local authorities.
Constituency (national)
Local authority
Family Resources Survey
Base microdata
Input
~20,000 households
→
Output
~20,000 households
DWP Family Resources Survey 2023-24
The Family Resources Survey (FRS) 2023-24 provides the baseline microdata, covering income, benefits, housing, and demographics for ~20,000 UK households. Variables are mapped to PolicyEngine's system, and benefit take-up rates are stochastically assigned.
- Demographics: age, gender, marital status, employment status, education, region (12 UK regions)
- 18+ reported benefit types: child benefit, universal credit, PIP, DLA, ESA, JSA, state pension, pension credit, housing benefit, carer's allowance, and more
- Income: employment, self-employment, private pension, savings interest, dividends, property, maintenance
- Housing: rent, council tax (with imputation for ~75% missing), mortgage interest/capital, tenure type, accommodation type
- Stochastic take-up: child benefit, UC, pension credit, marriage allowance, Scottish Child Payment, tax-free childcare, extended/universal/targeted childcare
- Grossing weights from FRS GROSS4 field