Test the UK labour supply response to late elasticities once the pin includes policyengine-uk#2673 #719

Follow-up to #718 (finding 2 of the review there).

After #718, elasticities that a dynamic writes after a policy modifier reach the values the modifier calculated. For the labour supply response this also needs PolicyEngine/policyengine-uk#2673. On the pinned policyengine-uk 2.132.6 the response is £0 even in a run without a modifier, a policyengine-uk limitation, so none of #718's tests covers the labour supply part. #718 closes #717, where this was recorded.

When the UK bundle pins a policyengine-uk release with #2673, add a test to tests/test_uk_modifier_cache.py:

• Use the synthetic parent and child there (the family fixture, 2027).
• Warmed run: a policy modifier stores Child Benefit £1,000 higher with set_input and then reads household net income. The dynamic sets the labour supply elasticities (income −0.05, substitution 0.15), so they are written after the modifier.
• Cold run: the same Child Benefit rise as a parameter change, with the elasticities in the construction reform.
• Expect the same employment_income_behavioral_response in both: about −£28.23 a year for the parent on the #2673 head, the income effect only (relative_wage_change is 0).
• Check that the test fails with uk/model.py from before Recalculate after each modifier step on the built UK simulation #718, where the warmed response stays £0.

🤖 Generated with Claude Code