UC Rebalancing Analysis dashboard

policyengine.py-based validation of the Universal Credit rebalancing package legislated by the Universal Credit Act 2025: an above-CPI uplift to the standard allowance and a fixed monthly health element of £217.26 for new claimants from April 2026, both toggled together via gov.dwp.universal_credit.rebalancing.active. Static, first-round net fiscal and distributional impact for 2026-27 → 2029-30, plus a per-claimant grid (220 SA + 440 HE archetypes) covering family type × children × age × employment income × LCWRA claim timing.

Repository layout

src/uc_rebalancing/        Python package (analysis + pipeline + CLI)
tests/                Pytest suite for the analysis functions
dashboard/            Next.js dashboard (Reform + Methodology tabs)
data/                 Pipeline output (uc_rebalancing_results.json)
docs/                 Additional documentation
uc_rebalancing_analysis.ipynb   Original notebook

Quick start

conda activate python313
pip install -e '.[simulation,dev]'
pytest
uc-rebalancing-build --sync-dashboard
cd dashboard && npm install && npm run dev

uc-rebalancing-build runs the enhanced-FRS Microsimulations (counterfactual = rebalancing OFF, reform = current law, plus single-leg sims for decomposition), evaluates each financial year 2026-27 → 2029-30, builds the per-claimant grid via batched single-Simulation packing, and writes data/uc_rebalancing_results.json. Adding --sync-dashboard also copies the JSON to dashboard/public/data/uc_rebalancing_results.json so the dashboard picks it up. Full build time is around 2–3 minutes.

Validation

Per-claimant chart

Check	Expected	policyengine.py	Verdict
Single 25+, £0 income, SA gain	≈ £247 (IFS)	£254	✅ +3% (CPI denominator difference)
Couple 25+, £0 income, SA gain	≈ 1.57 × single (SA ratio)	£398	✅ matches couple SA ratio
Single under-25, £0 income, SA gain	≈ 0.79 × single 25+	£201	✅ matches under-25 SA ratio
Single 25+, £10k, SA gain	partial (UC near zero)	£47	✅ SA £5,538 − taper £5,491 = £47
Single 25+, £20k, SA gain	£0 (UC zeroed)	£0	✅ taper exceeds SA
Single 2 kids, SA stays positive longer	yes (bigger UC envelope)	£254 to £40k	✅ child element widens taper window
Single 4 kids, £0, SA gain	£0 (benefit cap binding)	£0	✅ cap reduction +£253.64 absorbs uplift
HE new claim, single 25+, £0	−£2,983 cut + £254 SA = −£2,729	−£2,729	✅ exact
HE new claim, single under-25	−£2,983 + £201 = −£2,782	−£2,782	✅
HE new claim, couple 25+	−£2,983 + £398 = −£2,585	−£2,584	✅ (£1 rounding)
HE pre-2026 = SA-only positive	matches SA leg values	matches	✅
Sign convention (gain +, loss −)	consistent	consistent	✅
Monotonic in income (no kinks)	gradual taper-driven fade	yes	✅

Aggregate 2029-30

Metric	policyengine.py	Published	Gap	Source
SA leg cost	£1.67bn	£1.85bn	−9.6%	DWP IA Table 4
HE leg saving	−£2.33bn	−£2.10bn	−10.7%	DWP IA Table 9
Net package	−£0.66bn	−£0.21bn	legs amplify	DWP IA Tables 4+9
SA gainers	5.99m	6.69m	−10.5%	DWP IA Table 2
Single 25+ nominal increase 2025-26 → 2029-30	£736	£725	+1.5%	DWP IA evidence base

Aggregate gaps of ≈10% are expected: the pipeline is static while the DWP IA tables include behavioural responses (≈8k extra UC take-up, ≈40k fewer WCAs). The CPI denominator the DWP IA uses also differs slightly from PE-UK's parameter-driven uprating, which shows up as a 3% gap on the IFS comparison.

Edge cases verified as real policy interactions

1. 4-kids single, £0 income → £0 SA gain. Benefit cap binding. The reform increases UC by the SA uplift; the cap reduction increases by the same amount, so household net income is unchanged. Confirmed against the is_benefit_cap_binding and benefit_cap_reduction variables.
2. Single 25+, £10k → £47 (not £254). Counterfactual UC fully tapered to zero; reform retains a £47 residual. Most of the £254 uplift is absorbed by the taper at that earnings level.
3. 4-kids single transition £0→£32→£254 across £0/£10k/£20k. Cap stops binding once earnings reduce raw UC entitlement below the cap, then the full uplift comes through. Real UC × benefit-cap interaction.
4. HE leg loser count ≈ 2.06m vs DWP 750k. Different metrics. DWP's 750k is the new-claim flow cohort through 2029-30. The pipeline's n_losing counts every LCWRA benunit whose UC moves when the HE leg is toggled (≈full LCWRA stock). Both are legitimate; the DWP figure is the policy count, the pipeline figure is the simulation footprint.

External sources cross-referenced

• Universal Credit Act 2025 (legislation.gov.uk) — schedule of cumulative SA uplifts (2.3% / 3.1% / 4.0% / 4.8%) and the £217.26 new-claimant LCWRA freeze.
• DWP UC rebalancing Impact Assessment, August 2025 (withdrawn) — aggregate fiscal cost, gainers and losers counts, single 25+ £725 nominal increase.
• IFS: April's universal credit changes mean giveaways now, takeaways later — £247/year above-inflation slice for single 25+.
• UC and PIP Bill (parliament.uk) — page 16 confirms the £217.26 freeze.
• DWP UC standard allowance rates 2025-26 — match gov.dwp.universal_credit.standard_allowance.amount (Single 25+ £400.14/mo, Couple 25+ £628.10/mo, Single under-25 £316.98/mo).
• DWP UC taper rate (55%) and work allowance (single no kids no LCWRA → £0; with LCWRA → ~£404/mo) — match PE-UK behaviour.
• DWP benefit cap 2024-25 (£22,019.92/yr single outside London) — CPI-indexed to ~£25,323 in 2029-30 — matches the cap binding observed in the 4-children archetype.