Add calibrated UK enhanced CPS dataset

[MaxGhenis]

Summary

• add a public calibrated enhanced_cps_2025 dataset built from PolicyBench's public 1,000-household CPS-derived sample
• map those shared households into a UKSingleYearDataset, align additional UK-facing inputs, and recalibrate household weights against the UK target registry
• keep backward-compatible policybench_transfer aliases and add validation tests

Details

This PR introduces a public UK enhanced_cps dataset intended as the first step in a broader cross-country public-microdata strategy.

The source households come from PolicyBench's public US sample, then are mapped into UK inputs with synthetic geography plus additional UK-facing alignment for:

• council tax bands
• vehicle ownership
• pensions
• disability / PIP status
• consumption / fuel spending
• capital gains

The resulting dataset is then recalibrated to the UK national / region / country target registry used by the loss pipeline.

Files added or updated include:

• policyengine_uk_data/datasets/enhanced_cps.py
• policyengine_uk_data/datasets/policybench_transfer.py
• policyengine_uk_data/utils/reweight.py
• policyengine_uk_data/storage/enhanced_cps_source_2025.csv
• policyengine_uk_data/storage/enhanced_cps_2025.h5
• policyengine_uk_data/storage/policybench_transfer_2025.h5
• policyengine_uk_data/tests/test_policybench_transfer.py

Loss Comparison

dataset	target year	mean abs rel error	median abs rel error	share within 10%
Raw CPS transfer	2025	3.806	1.000	0.021
Enhanced CPS	2025	0.387	0.253	0.235
Enhanced FRS native-year	2023	0.458	0.120	0.476
Enhanced FRS uprated to target	2025	0.520	0.147	0.461

Interpretation:

• enhanced_cps materially improves on the raw transfer dataset.
• enhanced_frs still has tighter central fit and many more targets within 10%.
• enhanced_cps currently has fewer catastrophic outliers on the 2025 matrix, which is why its mean error is lower than the quick uprated-FRS comparison.

Validation

• python3 -m pytest policyengine_uk_data/tests/test_policybench_transfer.py
• generated the calibrated .h5 artifacts and verified dataset.validate()
• ran UK microsimulation and compared target loss before vs after calibration

Caveat

This is still not a replacement for the FRS or enhanced FRS. It is a public calibrated dataset built from open US microdata, so some UK-only supports are still missing relative to the full FRS-based pipeline.

[baogorek]

Just a few comments. I'm still not 100% sure what's going on here, regarding using CPS households from the US for UK estimation, but it looks interesting.

Also, do you really want to commit policyengine_uk_data/storage/enhanced_cps_2025.h5?

[baogorek]

I'll just be a nag here and comment that it might be easy to forget about these numbers in this file.