Improve corporate land allocation

[MaxGhenis]

Summary

• allocate corporate_land_value using the current weighted distribution of corporate_wealth instead of a stale corporate-wealth parameter total
• refresh the aggregate household and corporate land parameters to the 2024 ONS land totals used by policyengine-uk-data
• add a regression test covering weighted allocation and zero-corporate-wealth behavior

Why

household_land_value is already on the newer regional-intensity path, but corporate_land_value was still being scaled by aggregate_corporate_land_value / household.wealth.corporate_wealth.

That made the result depend on an out-of-sync parameter total rather than the actual weighted corporate-wealth distribution in the dataset. In local measurement on enhanced_frs_2023_24 for 2025, that left corporate land materially high versus the ONS target.

With this change, local weighted aggregates moved to:

• household_land_value: £4.737tn
• corporate_land_value: £2.058649tn
• land_value: £6.796tn

The key fix is the corporate component: it now matches the policyengine-uk-data ONS corporate-land target level exactly instead of drifting above it.

Testing

• uv run pytest -q policyengine_uk/tests/test_corporate_land_value.py policyengine_uk/tests/test_parameter_metadata.py
• uvx ruff check policyengine_uk/variables/household/wealth/corporate_land_value.py policyengine_uk/tests/test_corporate_land_value.py