Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

UK: children in families with three or more children, and their low-income share, are short of HBAI

Open
  • Is something wrong?
#1,371 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
45/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
python
Domain
data, machine-learning

Research direction

Start by measuring how much the calibration work in #1334 changes the number of HBAI dependent children in families with three or more children and their low-income rates. Compare the model aggregates for period 2024 with HBAI FYE 2025 table 4.5db, then check whether adding those figures as calibration targets closes the gap. Done when the validation shows the remaining large-family and poverty shortfalls and supports a clear choice of target or release gate.

Written by the indexing model from the issue text.

Description

Summary

On the current UK national release (microcosm-uk-2024-25-national-20261023T075039Z-a133f37a, SHA-256 64f83900…, run on its pinned policyengine-uk 2.147.2), children in families with three or more children are short of HBAI FYE 2025, and children in those families with low income are shorter still. Two-child limit poverty estimates depend on exactly this group, so they come out low against published estimates even when the cost and the number of affected families match.

Children by family size, FYE 2025 (model period 2024)

HBAI figures are from the FYE 2025 data tables, table 4.5db (children, by number of children in the family, United Kingdom). The model counts HBAI dependent children (is_hbai_dependent_child) and the number of them in each benefit unit.

HBAI FYE 2025 Microcosm a133f37a Difference
All children 14.69m 14.35m -2%
Children in families with 3+ children 4.64m 4.20m -10%
All children in relative low income, AHC 4.03m (27.4%) 3.52m (24.5%) -13%
All children in relative low income, BHC 3.01m (20.5%) 2.66m (18.5%) -12%
Children in 3+ child families in relative low income, AHC 2.07m (44.7%) 1.64m (39.0%) -21%
Children in 3+ child families in relative low income, BHC 1.64m (35.4%) 1.35m (32.1%) -18%

The Universal Credit side is not short. In 2025-26 under current law the model has 467k benefit units that gain from removing the limit, with 1.64m children, against DWP's 454k households not receiving the child element for at least one child and 1.61m children in them (UC two-child limit statistics, April 2025). So the shortfall is in how many large families there are and where their incomes sit, not in UC entitlement or claiming.

Why it matters

The IFS Green Budget 2025 comment Options for reforming the two-child limit (TAXBEN, FRS 2022-23 and 2023-24, full roll-out, absolute low income after housing costs) estimates that removing a fully rolled-out limit lifts 630,000 children out of absolute poverty at about £3.6bn, and a three-child limit 420,000 at about £2.6bn. Its endnote 4 says the IFS additionally reweights the 2022-23 data by number of children.

Running the same settings on this release in 2025-26 (full roll-out by setting the UC and CTC born_before dates so that no child is protected, full Universal Credit take-up):

IFS Microcosm a133f37a
Abolition: cost £3.6bn £3.58bn
Abolition: families gaining over 800,000 households 778k benefit units
Abolition: children lifted out of absolute poverty, AHC, FYE 2011-based line 630,000 455k (-28%)
Abolition: the same on the model's FYE 2025-based line 475k (-25%)
Three-child limit: cost £2.6bn £2.51bn
Three-child limit: children lifted, AHC, FYE 2011-based line 420,000 293k (-30%)
Three-child limit: the same on the FYE 2025-based line 312k (-26%)

The FYE 2011-based line is approximated as its FYE 2021 value in the parameter file, uprated by the model's CPI index.

Cost and families line up; children lifted are a quarter to 30% short, about the size of the shortfall in low-income children in 3+ child families above. Which absolute line is used moves the figures by under 20,000. An external replication of the IFS figures using raw FRS 2024-25 also reported fewer children lifted than the IFS.

Suggested next steps

  1. Add HBAI children by number of children in the family (table 4.5db, all children column: 3.23m, 6.82m, 4.64m) as a calibration target or at least a release gate, so the large-family share cannot drift.
  2. Check the relative low-income rate for children in 3+ child families as a validation metric alongside the overall poverty rates.
  3. Measure how much of the gap the work in #1334 closes (the calibration moving weight away from low-income households, and the lone-parent UC caseload at 0.67 of DWP's at design weights) before adding a target, since the two may be the same problem.

Related: #1334, #731.

Reproduction

Aggregates only. Scripts run policyengine_uk.Microsimulation(dataset=<release h5>) for period 2024 (FYE 2025 comparison) and 2025 (two-child limit options), with is_hbai_dependent_child, in_relative_poverty_ahc/_bhc, in_poverty_ahc mapped to person, and the benefit-unit count of HBAI dependent children. Reform keys use the YYYY-MM-DD.YYYY-MM-DD form on gov.dwp.universal_credit.elements.child.limit.{born_before,child_count} and the Child Tax Credit equivalents.

🤖 Generated with Claude Code

Dominant language
Python
Stars
0
Forks
6
Avg merge
1d 14h
Merged PRs (30d)
135

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from PolicyEngine/microcosm

All issues in PolicyEngine/microcosm

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.