Improve rent uprating by tenure

[MaxGhenis]

Summary

• uprate private rents directly from ONS private rental price growth by region instead of backing them out from aggregate OBR rent
• update social rent growth to observed RSH stock-average growth through 2025, then the England policy cap forecast
• add a regression test covering the tenure-specific 2025 rent uprating path

Validation

• uv run pytest policyengine_uk/tests/test_rent_uprating.py -q
• uv run pytest policyengine_uk/tests/test_fiscal_year_parameters.py -q

Notes

• private rent forecasts after observed ONS data use the March 2026 OBR aggregate rent forecast as the documented fallback
• uv.lock is intentionally excluded from this PR

[MaxGhenis]

Impact check for the rent forecast fix in this review thread.

Method: I compared the pre-fix PR behavior against current head 72b88cb14, using the latest private HF dataset policyengine/policyengine-uk-data-private/enhanced_frs_2023_24.h5 at repo revision d9c736887f7bb4ed34f01c000e4924b1131bd094 (2026-04-19 22:34 UTC). I started from the same 2025 household rents and recursively applied 2026-2030 rent growth under:

• pre-fix: private rent growth = obr.rent; social rent growth = obr.social_rent
• fixed: private rent growth backed out from aggregate OBR rent growth after applying social rent growth

This isolates the review-fix delta, not the full PR-vs-main delta.

year	aggregate rent, pre-fix	aggregate rent, fixed	delta	delta %
2026	£150.382bn	£149.092bn	-£1.290bn	-0.858%
2027	£155.626bn	£154.038bn	-£1.588bn	-1.020%
2028	£160.175bn	£157.890bn	-£2.285bn	-1.427%
2029	£165.013bn	£162.072bn	-£2.941bn	-1.782%
2030	£170.246bn	£166.805bn	-£3.441bn	-2.021%

Downstream aggregate effects are modest because lower rents reduce UC/HB entitlement slightly, while AHC income rises because rent paid falls by more than benefits fall.

year	Housing Benefit delta	UC delta	household net income delta	AHC household net income delta
2026	-£0.039bn	-£0.064bn	-£0.103bn	+£1.188bn
2027	-£0.048bn	-£0.077bn	-£0.124bn	+£1.463bn
2028	-£0.044bn	-£0.108bn	-£0.152bn	+£2.133bn
2029	-£0.032bn	-£0.136bn	-£0.169bn	+£2.773bn
2030	-£0.037bn	-£0.146bn	-£0.183bn	+£3.257bn

Poverty impact is small:

• Absolute BHC poverty: no detectable change in this comparison.
• Absolute AHC poverty: all people falls by 0.04-0.15pp depending on year; child poverty falls by 0.09-0.35pp.
• Relative poverty moves by at most about 0.33pp in either direction, mostly from threshold effects.

[MaxGhenis]

Clarifying the parameter/index change from the review fix: the changed values are the 2026-2030 private-rent forecast fallback entries under ons.private_rental_prices. Pre-fix, these were equal to the OBR aggregate rent YoY forecast. Post-fix, they are backed out so the weighted private/social path reconciles to the OBR aggregate rent forecast.

Cumulative private-rent index shown with 2025 = 1:

year	pre-fix private YoY	fixed private YoY	YoY delta	pre-fix private index	fixed private index	index delta
2026	3.3400%	2.0664%	-1.2736pp	1.033400	1.020664	-0.012736
2027	3.0200%	2.7757%	-0.2443pp	1.064609	1.048995	-0.015614
2028	2.3000%	1.6894%	-0.6106pp	1.089095	1.066716	-0.022379
2029	2.3800%	1.8391%	-0.5409pp	1.115015	1.086335	-0.028680
2030	2.5800%	2.2136%	-0.3664pp	1.143783	1.110382	-0.033401

So by 2030, the fixed private-rent fallback index is about 3.34 index points lower than the pre-fix path, or roughly 2.9% lower relative to the pre-fix private-rent index.