UK bus fare and subsidy consumer targets from Chronicle facts

[juaristi22]

Context

PolicyEngine/chronicle#200 audited the bus fare, bus subsidy, and Capital Gains Tax inputs inherited from policyengine-uk-data. The ownership boundary is:

• Chronicle preserves publisher artifacts and emits publisher facts.
• Microcosm owns target profiles, fact selectors, measurement bindings, period and geography transformations, activation rules, calibration, and diagnostics.

The bus imputation stage is already scoped in #685, but Microcosm has no consumer target declarations for bus_fare_spending or bus_subsidy_spending. This issue owns that missing target-surface work. It does not duplicate the CGT issues listed below.

Chronicle #202 also introduces 14,656 monthly private-rent facts from January 2023 through June 2026, including 316 England and Wales local authorities in every month, plus 841 monthly Northern Ireland Universal Credit claimant facts from January 2023 through May 2025 for 18 PCON24 constituencies and 11 local government districts. Their Microcosm-owned selectors, measurement bindings, geography-vintage resolution, and support-aware activation belong in #759 after the contract migration in #708; Chronicle emits the facts without owning those consumer profiles.

Those local facts are not part of this issue's bus implementation, but #759's compile membership must account for them explicitly as active or deferred rather than leaving the new Chronicle rows unwired.

Verified source-to-target translation

Chronicle #202 exposes these DfT facts:

Source fact	Consumer interpretation
BUS05ai passenger fare receipts, year ending March 2025: England £3,417.389m; London £1,347.435m; England outside London £2,069.954m	Direct amount targets for bus_fare_spending over the matching England, London, and England-outside-London populations
BUS05bi total estimated net support, year ending March 2025: England £3,024.904m; London £1,130.214m; England outside London £1,894.690m	Direct amount targets for bus_subsidy_spending over the matching populations
NTS0705a 2024 local-bus trips per person by real household-income quintile: 47.610, 37.451, 24.842, 18.452, 12.978	A possible proxy for the within-population fare-spending shape, not a direct fare-spending target
BUS0415 quarterly local-bus fare indices	A possible explicit period-alignment input for passenger fare receipts, not for net support

These facts replace uk-data's hardcoded England anchors and distribution constants. They do not publish an exact UK fare-receipt or subsidy total. BUS04 operating revenue is not the same measure, and DfT does not cover Northern Ireland.

Scope

1. Add Microcosm-owned target declarations and PolicyEngine measurement bindings for the direct BUS05ai and BUS05bi facts.
2. Bind each direct fact to the matching geographic support. Do not apply an England fact to all UK households.
3. Decide whether to activate the London and England-outside-London partition in addition to the England total. If all three are active, declare and test their reconciliation rather than treating overlapping constraints as independent evidence.
4. If NTS0705a is used, encode the normalization and income-quintile mapping as an explicit consumer-side proxy operation. The target metadata and diagnostics must say that trips per person are being used to shape spending; they must not call the result a publisher-stated spending share.
5. If BUS0415 is used for period alignment, encode the index operation in Microcosm with the source quarter, destination period, selected geography series, and deterministic tests. Do not apply the fare index to bus subsidy.
6. Wire the profiles into the UK national registry, parity fixture, compilation tests, and calibration diagnostics with Chronicle package and fact provenance.
7. Sequence the bus amount stage from Net-new UK stages: UC deduction attributes, bus fares, WAS debt #685 before the national calibration seam from Calibrate the UK national build from Ledger-backed targets #623, so target-fit evidence measures the exported values rather than pre-imputation placeholders.

Geography rule

The default direct targets should stop at the publisher's England support. An exact UK total remains inactive until matching measures exist for Scotland, Wales, and Northern Ireland.

If a release needs an estimated UK total before then, define it as a separate Microcosm-derived target with an explicit method and provenance. Do not overwrite or relabel the England publisher fact, and do not silently carry forward uk-data's England-to-UK population uplift.

Acceptance

• No bus target value is copied as a literal from policyengine-uk-data; direct values resolve from a hash-pinned Chronicle consumer artifact.
• bus_fare_spending and bus_subsidy_spending have separate concepts and measurement bindings.
• England, London, and England-outside-London predicates compile against the intended household support, with a test that the two subareas reconcile to England within publisher precision.
• Any NTS-based spending-shape profile is marked as proxy-derived, reports the original trips-per-person facts, and has a deterministic normalization test.
• Any BUS0415 period transform reports the original and aligned values and cannot be selected for the subsidy target family.
• The registry and fixture represent each bus target's active or inactive state explicitly; they do not substitute inline values or invented provenance when a required Chronicle fact is absent.
• Target compilation fails closed when the required Chronicle package, fact, geography, period, or measurement binding is absent or ambiguous.
• Terminal diagnostics report initial and final fit for every active bus target and preserve package ID, fact ID, profile ID, and transformation provenance.
• An exact UK total is either inactive with a machine-readable reason or explicitly classified as a Microcosm estimate; it is never presented as a DfT fact.

Dependencies

• Add 2023-2026 UK facts and restore source vocabulary chronicle#202 source packages for BUS05i, NTS0705, and BUS0415.
• Net-new UK stages: UC deduction attributes, bus fares, WAS debt #685 for the bus fare and subsidy output stages.
• Calibrate the UK national build from Ledger-backed targets #623 for the national calibration seam and target-fit evidence.

CGT disposition from the same audit

Do not duplicate the existing CGT issues here:

• UK: banded CGT gains targets (by size of gain) to constrain the distribution, not just count + total #467 owns binding the nine HMRC Table 2.1a size-of-gain bands. Chronicle already supplies the 18 count-and-gains facts; the stale "publisher table not pinned" exclusion should be removed there.
• UK CGT top-tail age and geography joints inherit from incidence and look off — benchmark against HMRC marginals #725 owns benchmarking and possible promotion of the 42 HMRC Table 5 region facts and 30 Table 6 age facts now packaged by Chronicle Use target materialization cache by default #202.
• The 142 numeric HMRC Table 3 facts now packaged by Chronicle Use target materialization cache by default #202 remain fenced under UK capital gains: imputed distribution has no top tail, missing 55% of CGT #552 unless their rounding, suppression, taxable-income proxy, and policy-endogenous taxpayer definition are separately adjudicated. Chronicle retains suppressed count cells as guards rather than inventing values. UK rebuild: sequence calibration after the amounts imputations; staging seam has no CGT target parity evidence #704 owns the requirement that CGT amounts run before calibration.

HMRC Tables 4, 7, 8, and 9 remain out of scope because the current UK model exposes aggregate capital_gains, not the relief-, asset-, residential-property-, or carried-interest-specific variables those tables measure.

Out of scope

• Authoring Chronicle source packages.
• Reimplementing the bus amount stages from Net-new UK stages: UC deduction attributes, bus fares, WAS debt #685.
• Inventing or sourcing missing Scotland, Wales, or Northern Ireland totals.
• Adding new Capital Gains Tax subtype variables.