Determine whether pre-1 detailed_budget maps to ProgramStatistics

[anth-volk]

Context

policyengine-api-v2 needs to preserve the legacy single-year macro output shape when it eventually moves back to policyengine 4.x. The pre-1 Simulation.calculate_economy_comparison() result included detailed_budget, while the 4.x economic_impact_analysis() path exposes program_statistics.

Before the simulation API can safely adapt 4.x outputs back into the legacy response contract, we need to determine whether detailed_budget can be reconstructed from ProgramStatistics without semantic drift.

Questions to answer

• Does every pre-1 detailed_budget entry have a corresponding ProgramStatistics row for US and UK?
• Are signs, baseline/reform totals, changes, and recipient/count fields equivalent?
• Are labels/groupings in the old detailed_budget public contract recoverable from 4.x metadata?
• Are any old detailed_budget fields intentionally unsupported in 4.x?

Acceptance criteria

• Document the mapping from pre-1 detailed_budget to 4.x ProgramStatistics for US and UK.
• Identify any unmappable fields or sign convention differences.
• Add tests or fixtures that compare representative pre-1 and 4.x outputs if feasible.
• Recommend whether the sim API should reconstruct detailed_budget from ProgramStatistics or expose a new dedicated .py output helper.