Scottish households' water charges are uprated by the England and Wales water bills series #2501

Problem

#1863 adds water_and_sewerage_charges to uprating_indices.yaml under gov.economic_assumptions.yoy_growth.ofwat.water_bills. That series is the average household bill in England and Wales, set by Ofwat. The uprating applies one rate to every household, so Scottish households get the England and Wales path too. Scottish Water's charges are set separately, under limits from the Water Industry Commission for Scotland (WICS), and followed a different path: far lower in 2025-26, higher in 2026-27:

Scottish Water sources:

• 2025-26: "Average monthly household bills will rise by £3.68 per month or £44 per year from 1 April 2025 ... This equates to a rise of 9.9% in charges." (Scottish Water, "Charges 2025 to 2026", 28 January 2025). Original, archived.
• 2026-27: "The increase represents an 8.67% rise, in line with limits set by the independent economic regulator, the Water Industry Commission for Scotland (WICS)." Average annual charge £532, up £42 (Scottish Water, "Average Water Bills to Rise £42 in 2026", 22 January 2026). Original, archived.

Size

In a microcosm_uk_2024_25 build, Scottish households (2.63m weighted) have a mean water_and_sewerage_charges of £396 in 2024-25, £1.04bn in total. By 2026-27 the England and Wales path (1.27 x 1.054) takes the mean to about £531. The Scottish Water path (1.099 x 1.0867) gives about £473. That overstates Scottish water charges by about £57 per household, or about £150m in total, in 2026-27. The error lowers Scottish hbai_household_net_income_ahc by the same amount, so it tends to raise measured Scottish poverty after housing costs. These are back-of-envelope figures from the dataset, not a model run.

Northern Ireland

Northern Ireland has no domestic water charges. In the same build every Northern Ireland household has water_and_sewerage_charges of zero, so any growth rate leaves it at zero. Nothing is wrong in the numbers today. It should still be explicit, so a future dataset that gives Northern Ireland households a water charge does not silently take the England and Wales rate.

Proposal

1. Add a Scottish series, for example gov.economic_assumptions.yoy_growth.scottish_water.water_bills, with 0.099 at 2025-01-01 and 0.0867 at 2026-01-01, cited to the statements above. For 2027-28 on, use the next WICS price determination, covering 2027-28 on, once it is final, and record the vintage.
2. Uprate water_and_sewerage_charges by country, the way uprate_council_tax in policyengine_uk/data/economic_assumptions.py does: England and Wales regions take ofwat.water_bills, Scotland takes the new series, and Northern Ireland takes no growth (None, as council tax does), with a comment saying why. Move water_and_sewerage_charges out of the single-index list in uprating_indices.yaml, and update the code-health test that Uprate three more columns, and guard the gap that hid them #1863 adds so it accepts this case.
3. Check policyengine_uk/utils/create_ahc_deflator.py, which reads the ofwat.water_bills index. If it is meant to be UK-wide, it should use a household-weighted UK water path rather than the England and Wales one.

Related: #1863 (adds the uprating), #2384 (England and Wales 2027-29 projections cannot be reproduced).

🤖 Generated with Claude Code