Update PolicyEngine UK to 2.88.6

[policyengine]

Summary

Update PolicyEngine UK from 2.78.0 to 2.88.6.

What changed (2.78.0 -> 2.88.6)

Added

• Support Python 3.9 and 3.10 (in addition to 3.11–3.14). On Python 3.9/3.10, pip resolves policyengine-core to a version that has been relaxed to support older Python (3.24.0+); on 3.11+ behavior is unchanged.
• • Add a microsimulation smoke test suite that runs against the unpinned latest enhanced FRS dataset and asserts plausibility bounds for UK population, UC aggregate, is_parent population, core benefit totals, and extended childcare eligibility. Catches silent model/data skew at the point the dataset is republished, not after a release.
• Add a first-pass Disabled Students' Allowance model for England higher-education students.
• Add a first-pass Travel Grant model.
• Add a first-pass Adult Dependants' Grant model.
• Add a first-pass Parents' Learning Allowance model.
• Add a first-pass 16 to 19 Bursary Fund model for vulnerable groups.
• • Add a first-pass Childcare Grant model for England full-time undergraduates.
• • Add an approximate England full-time maintenance loan model, with explicit override inputs for living arrangement and assessed household income.
• • Add an approximate England full-time maintenance loan model, with explicit override inputs for living arrangement and assessed household income.
• Add named economic-assumption parameters for the local-authority ONS income target uprating factors used by policyengine-uk-data.

Changed

• • Update .github/CONTRIBUTING.md to document the towncrier changelog.d/ workflow. The old changelog_entry.yaml + make changelog flow was deprecated some time ago; the CONTRIBUTING guide still instructed new contributors to use it, causing CI round-trips on PRs that created a changelog_entry.yaml no fragment step was looking for.
• Expose build metadata helpers for UK data artifacts, including a stable data-build fingerprint and build provenance metadata.

Fixed

• Replace new_state_pension's flat-max payout with a min(reported, max) / max * period_max formula mirroring basic_state_pension, and extend additional_state_pension to NEW-type retirees so any pre-2016 SERPS/S2P Protected Payment flows through as an add-on instead of being silently dropped. Partial-NI-record retirees now receive their actual pro-rated rate rather than the full flat max. Closes part of the ~£12 bn residual state-pension gap vs the OBR target tracked in Update PolicyEngine US to 1.19.0 #1632.
• Zero out Income Support and income-based Jobseeker's Allowance after DWP managed migration completed on 31 March 2026. New parameters gov.dwp.income_support.active and gov.dwp.JSA.income.active flip to false from 2026-04-01, matching the Tax Credits treatment. Contribution-based JSA remains active.
• Bump policyengine-core minimum to >=3.25.0 to pick up the cache-invalidation and set_input preservation fixes (Preserve set_input values across apply_reform policyengine-core#475). The 3.24.0–3.24.3 cascade left UK model tests returning zero for income_tax, UC, and other formula-driven variables when a reform is applied during simulation construction; 3.25.0 includes the regression fix.
• Migrate versioning workflow to GitHub App token (POLICYENGINE_GITHUB PAT expired).
• • Fix state_pension_type incorrectly classifying every pensioner as receiving the pre-2016 basic State Pension. The formula used values_list[0] to find when the New State Pension activated, but policyengine-core auto-extrapolates the parameter into the far future, so [0] was returning a 2040s entry instead of the 2016 activation date. Walks the list oldest-first to find the real activation instant, so post-2016 retirees are now correctly classified as NEW. Raises the modelled 2025 state pension aggregate from about £116bn to about £127bn.
• • Fix Working Tax Credit and Child Tax Credit continuing to pay out from the 2025-26 tax year onward. Working Tax Credit and Child Tax Credit ended on 5 April 2025 (HMRC/DWP). Adds a gov.dwp.tax_credits.active parameter that flips to false on 2025-04-06 and gates tax_credits on it. Removes about £1.9bn of phantom Tax Credit spending per year from 2025-26 onward while preserving the legitimate 2024-25 baseline.
• Fix maintenance loan proxy logic to require explicit higher-education evidence and improve sponsor-income assessment.
• Use modelled student loan repayments in aggregates and cap them by outstanding balance when available.
• Fixed corporate_land_value to allocate aggregate corporate land using the current weighted distribution of corporate_wealth, and refreshed the aggregate land parameters to the 2024 ONS land totals used by policyengine-uk-data.
• Corrected Universal Credit rebalancing so existing health-element claimants keep their combined standard allowance and health element award CPI-protected.

---

Generated automatically by GitHub Actions