examples/us_budgetary_impact.py consumes excessive memory (peak ~7.5 GB RSS / 82 GB VMS)

Reported in the discussion on #264 (JOSS paper PR).

When run on the joss-paper branch (4.2.1), examples/us_budgetary_impact.py exhibits very high memory pressure during the baseline + reform simulations on the enhanced CPS 2024 dataset (2026, ~101k people / 41k households):

• Vahid: OOM-killed at peak ~30 GB RSS on a Mac (rc=137).
• Anthony (instrumented run): peak 7.48 GiB RSS, 82.35 GiB VMS; baseline_sim.run() 546.6s wall, reform_sim.run() 596.5s wall; VSCode crashed mid-run from memory pressure.

This makes the canonical US population example impractical for normal local development and for JOSS reviewers attempting to reproduce results.

Suggested next steps:

• Profile baseline_sim.run() and reform_sim.run() to find the dominant allocations (likely in policyengine-us core, but worth confirming).
• Determine whether intermediate arrays can be released between runs or whether the two simulations can share underlying data.
• Document peak memory requirement in the example header until the footprint is reduced.

Related: #264

Start with examples/us_budgetary_impact.py and reproduce the baseline_sim.run() and reform_sim.run() calls on the enhanced CPS 2024 dataset. Profile both runs to identify dominant allocations and determine whether intermediate data can be released or shared. Done means reducing the memory footprint or documenting the peak requirement in the example header.