Add bus_fare_spending input variable

[vahid-ahmadi]

What

Adds a Household input variable bus_fare_spending (COICOP 7.3.2 — bus & coach fares), the companion to the LCFS bus fare imputation in PolicyEngine/policyengine-uk-data#428.

Why

The data PR writes a bus_fare_spending column, but without a registered variable the model silently skips it on load (simulation.py ignores unknown columns). This variable lets the model consume it. bus_fare_spending is the passenger fare households pay, distinct from bus_subsidy_spending (the ETB government-subsidy benefit-in-kind) — the building block for modelling bus fare reforms (£2 cap, free travel for young people), as the bus analogue of the existing rail fare modelling.

Details

• Mirrors petrol_spending / diesel_spending, which are likewise COICOP sub-components of transport_consumption.
• CPI-uprated (gov.economic_assumptions.indices.obr.consumer_price_index), matching transport_consumption and the data-side uprating.
• Not added into the consumption total — transport_consumption already includes bus fares, so adding it would double-count.

Verified: registers in the tax-benefit system and the uprating parameter resolves.

Resolves #1779. Companion to PolicyEngine/policyengine-uk-data#428 (tracked in PolicyEngine/policyengine-uk-data#427).

🤖 Generated with Claude Code

[vahid-ahmadi]

I found one integration issue in the first pass: bus_fare_spending declared CPI uprating on the Variable, but extended datasets use policyengine_uk/data/uprating_indices.yaml for growth-factor uprating, so the new column stayed flat while transport_consumption CPI-uprated.

I pushed a420b3ea to this branch to add bus_fare_spending to the CPI growth-factor map and add a focused regression test. Verified locally:

uv run pytest policyengine_uk/tests/test_road_fuel_volume_uprating.py policyengine_uk/tests/test_no_economic_assumptions.py

Result: 4 passed.