Skip to content

Add the two-child limit child-poverty benchmarks, recomputed on the certified Microcosm pair - #225

Draft
MaxGhenis wants to merge 2 commits into
mainfrom
uk/two-child-limit-child-poverty
Draft

MaxGhenis wants to merge 2 commits into
mainfrom
uk/two-child-limit-child-poverty

Conversation

@MaxGhenis

Copy link
Copy Markdown
Contributor

WIP. Moves the UK two-child-limit child-poverty benchmark comparison out of the hard-coded policyengine-uk docs page (docs/book/validation/child-poverty-tcl.md, corrected by PolicyEngine/policyengine-uk#1724) into scorecard, recomputed on the current certified Microcosm pair and concept-annotated (Max, 2026-10-23: "should that info be in scorecard instead").

Plan

  • sources/dwp-two-child-limit-poverty/: DWP's Budget 2025 note, Tables 1-4 (relative and absolute low income, BHC and AHC, children / working-age adults / individuals, FYE 2027-2031), the FYE 2026 text figure and the 2 million gainers; QC against the Child Poverty Strategy note's Tables A1-A5 and the HBAI FYE 2024 methodology the note cites.
  • sources/rf-no-half-measures/: Resolution Foundation, No half measures (30 Oct 2025): 480,000 and £3.5bn, the partial-repeal options, and the LSO 2024 annex the report cites for its method. These numbers are already claims from the September AB2025 harvest; the new source vendors and QCs them, and this lane attaches new results to those claims rather than minting duplicates.
  • pipeline/compute_uk_two_child_limit_poverty.py: counterparts on policyengine.py 6.10.1's certified pair (policyengine-uk 2.132.6 + microcosm-uk-2024-25-national-20261019T083323Z-6cba5d71) through the managed route, for every concept the sources use (relative line on each world's own median and fixed at the with-limit world's, absolute line on FYE 2025 and on a constructed FYE 2011 reference, HBAI dependent children).
  • Measured decomposition of the gap between the September estimate (~253,000 children, Populace + 2.89.2) and the October one (~548,000, Microcosm + 2.148.3): an engine x data grid plus the child-count and median conventions.
  • scorecard_db/ingest_uk_two_child_limit_poverty.py, build step, lane, feeds, tests.
  • Independent review (Subfleet, hard tier), then merge on the gates.

Local national runs are held by a host reservation until 15:30Z today; compute runs then.

🤖 Generated with Claude Code

…ied-pair compute

The DWP Budget 2025 note's Tables 1-4 (relative and absolute low income,
BHC and AHC, FYE 2027-2031) plus its FYE 2026 and gainer figures, with QC
against the Child Poverty Strategy republication; and a compute script for
the counterparts on the policyengine.py 6.10.1 certified Microcosm pair.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@vercel

vercel Bot commented Oct 23, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
policyengine-scorecard Ready Ready Preview Oct 23, 2026 11:45am UTC

Request Review

Certified pair (pe 6.10.1, pe-uk 2.132.6, Microcosm) computed on Modal.
Not done: AB2025 stager must ignore this lane's DWP claims before restaging;
grid, decomposition, losers diagnostic, feeds, review.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

This branch was successfully deployed

1 active deployment
Preview — af3bc689 Deployed Oct 23, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant