Add Scottish Child Payment baby bonus reform

[PavelMakarchuk]

Summary

• Adds parametric reform providing £12.85/week extra for children under 1 year old on top of the standard £27.15/week Scottish Child Payment
• Reform controlled by gov.contrib.scotland.scottish_child_payment.in_effect parameter (default: false)
• Includes structural reform module for programmatic use

Changes

• New parameters at gov/contrib/scotland/scottish_child_payment/:
  • baby_bonus.yaml - £12.85/week for children under age 1
  • in_effect.yaml - boolean toggle (default false)
• Updated scottish_child_payment variable to conditionally apply baby bonus when reform is active
• Added structural reform in reforms/scottland/scottish_child_payment_reform.py
• Added 9 YAML tests covering:
  • Baseline (no reform)
  • Reform with various family compositions (1 baby, 2 babies, mixed ages)
  • Age boundary testing (1 year old gets no bonus)
  • Regional eligibility (England gets no SCP)
  • Income eligibility thresholds (£45k, £50k)

Test plan

• All 9 new YAML tests pass
• Existing 18 SCP tests still pass
• CI passes

🤖 Generated with Claude Code

[vahid-ahmadi]

@PolicyEngine review this

[policyengine]

Code Review - Scottish Child Payment Baby Bonus Reform

I've reviewed the PR implementation and found it to be well-structured with comprehensive test coverage. Here are my findings:

✅ Strengths

1. Clean parametric approach: The reform correctly uses the bracket-based parameter system (baby_bonus.yaml) with age thresholds (£12.85/week for age < 1, £0 for age ≥ 1).

2. Comprehensive tests: 9 YAML tests cover all key scenarios:

   • Baseline vs reform comparison
   • Multiple babies, mixed ages
   • Age boundary testing (1 year old = no bonus)
   • Regional eligibility
   • Income eligibility thresholds

3. Proper structural reform integration:

   • Reform is registered in reforms.py
   • Applied via simulation.py using create_structural_reforms_from_parameters
   • Toggle parameter (in_effect) defaults to false for backward compatibility

4. Logic preserved: The reform duplicates the entire scottish_child_payment variable formula, correctly preserving all eligibility logic (Scotland region check, qualifying benefits, takeup rates).

⚠️ Issues Found

1. Critical: Directory typo - policyengine_uk/reforms/scottland/ should be scotland (missing 'n')

   • File: policyengine_uk/reforms/scottland/__init__.py
   • File: policyengine_uk/reforms/scottland/scottish_child_payment_reform.py
   • Import: Line 7 of reforms/reforms.py also uses from .scottland

2. Minor: Unrelated formatting changes - docs/book/validation/student-loan-repayments.ipynb has Black formatting changes unrelated to this PR. Consider reverting or committing separately.

3. Code duplication concern: The reform duplicates the entire 167-line variable formula. If the baseline scottish_child_payment variable changes, this reform won't inherit those changes. Consider:

   • Extracting shared logic to a helper function, OR
   • Accepting this as expected behavior for parametric reforms (reform captures policy as-of implementation date)

📝 Test Validation

The test calculations appear correct:

• Base SCP: £27.15/week × 52 = £1,411.80/child/year
• Baby bonus: £12.85/week × 52 = £668.20/baby/year
• Test 2: 1 baby = £1,411.80 + £668.20 = £2,080 ✅
• Test 3: 2 babies = £2,823.60 + £1,336.40 = £4,160 ✅

🔧 Recommended Actions

1. Must fix: Rename scottland → scotland throughout
2. Should fix: Remove or separate the notebook formatting changes
3. Optional: Document the code duplication pattern in comments (if intentional)

✅ Approval Status

Once the directory typo is fixed, this PR will be ready to merge. The implementation is sound and the test coverage is excellent.

Note: I couldn't run the tests locally due to Python version mismatch (system has 3.12, package requires 3.13), but the test structure and expected values look correct based on manual calculation.

[vahid-ahmadi]

After researching the official policy sources, I found that the original implementation did not match the actual Scottish Government policy. I've now pushed a fix.

The Actual Policy

The Scottish Government policy is: "the Scottish Child Payment will rise to £40 for families with children under the age of one" (https://www.gov.scot/publications/scottish-budget-2026-2027-finance-secretarys-statement-13-january-2026-2/)

Key word: "to £40" (not "by £12.85")

This means:

• £40 total weekly payment for children under 1
• Standard rate (inflating with inflation) for children 1+
• When child turns 1, payment reverts to standard rate

The Fix (Now Applied)

I've updated the implementation in commit 845e2f7 to use:

Calculate per-child weekly amount (age-dependent)

PREMIUM_RATE_UNDER_ONE = 40.0 # £40/week fixed total
per_child_weekly = where(
age < 1,
PREMIUM_RATE_UNDER_ONE, # £40/week for under-1s (TOTAL, not bonus)
weekly_amount # Standard rate for 1+ (inflates with inflation)
)

Calculate total for all eligible children

total_weekly = benunit.sum(per_child_weekly * is_eligible_child)
annual_amount = total_weekly * WEEKS_IN_YEAR

This ensures under-1s always receive exactly £40/week regardless of base rate inflation.

Test Updates

I've also updated all test comments in scottish_child_payment_baby_bonus.yaml to explicitly document that £40 is a TOTAL payment, not a base + bonus.

Sources Confirming Fixed £40 Total

1. https://www.gov.scot/publications/scottish-budget-2026-2027-finance-secretarys-statement-13-january-2026-2/: "boosting the payment... to £40 a week"
2. https://www.which.co.uk/news/article/scottish-budget-2026-what-does-it-mean-for-your-money-atK4N0K3j4W9: "From 2027, the rate will rise to £40"
3. https://fraserofallander.org/budget-2026/a-budget-where-the-silences-were-loudest/: Base rate £28.85 in 2027-28, under-1s get £40 total (implied bonus of £11.15, not £12.85)