Fix means-test income and student-loan funding routes

changelog.d/fix-2573-council-tax-reduction.fixed.md

@@ -0,0 +1 @@
+- Count claimant and partner spousal maintenance and Widowed Parent’s Allowance in Council Tax Reduction income across Great Britain, with the statutory £15 weekly disregards and Pension Credit routing.

changelog.d/fix-2573-income-support.fixed.md

@@ -0,0 +1 @@
+- Count the claimant's and partner's spousal maintenance in Income Support while continuing to disregard child maintenance.

changelog.d/fix-2573-non-dependants.fixed.md

@@ -0,0 +1 @@
+- Included spousal maintenance in Housing Benefit and Council Tax Reduction non-dependant income bands while preserving its income-tax exemption.

changelog.d/fix-2573-student-loans.fixed.md

@@ -0,0 +1 @@
+- Correct newly assigned projected student loan plans by funding nation, using household region as a proxy: Plan 4 for Scotland, Plan 1 for Northern Ireland, Plan 2 for Wales, and Plan 2 or 5 by entry cohort for England.

docs/book/assumptions/growthfactors.md

@@ -46,6 +46,24 @@ We apply the OBR's wage growth forecasts to employment-related variables.
 
 We use average earnings for these variables: `employee_pension_contributions`, `employer_pension_contributions`, `employment_income`, `employment_income_before_lsr`, `personal_pension_contributions`, `student_loan_repayments`
 
+### Newly assigned student loan plans
+
+Dataset projection holds ages fixed, so a person of a given age represents a
+later university intake in each projected year. When projection adds a loan
+to a tertiary-educated person without one, it uses current household region
+as a proxy for the funding nation, because the dataset does not record the
+funding agency. This proxy cannot identify people who moved between nations.
+
+The [official repayment-plan rules](https://www.gov.uk/repaying-your-student-loan/which-repayment-plan-you-are-on)
+give SAAS borrowers Plan 4 and Student Finance Northern Ireland borrowers
+Plan 1. For undergraduate loans, England uses Plan 1 before the 2012 intake,
+Plan 2 for 2012–2022, and Plan 5 from 2023; Wales uses Plan 1 before 2012 and
+Plan 2 from 2012. New assignments follow these routes; an unknown region
+leaves the person without an assigned loan. Take-up remains a fixed 40%
+probability with a draw keyed on person ID. This funding-nation routing
+applies to newly assigned loans; the existing-holder relabelling and
+write-off approximations are unchanged.
+
 ## Lagged average earnings
 
 We lag earnings growth by one year and use that to uprate housing service charges (including ground rent).

policyengine_uk/data/economic_assumptions.py

@@ -544,13 +544,20 @@ def uprate_student_loan_plans(
 
     As we project forward, we need to:
     1. Re-label existing holders to correct plan based on uni start year
-    2. Add Plan 1/2 holders in age bands beyond the base year's coverage
-    3. Add Plan 5 holders (new plan starting 2023)
+    2. Add holders in age bands beyond the base year's coverage
+    3. Add holders from the 2023-and-later intakes
 
     For (2) and (3), we use highest_education == TERTIARY as the signal
     for who is a graduate, then apply a flat take-up probability. Each
     person's take-up draw is keyed on person_id, so the same people take up
     a loan in every year and in any row order.
+
+    New holders follow the funding nation's undergraduate repayment plan.
+    Funding agency is unavailable in the dataset, so current household
+    region proxies for it: England, Wales, SAAS for Scotland, and Student
+    Finance NI for Northern Ireland. This cannot identify cross-border
+    movers. An unknown region leaves the plan unchanged. See
+    https://www.gov.uk/repaying-your-student-loan/which-repayment-plan-you-are-on.
     """
     # Skip if required columns don't exist yet (e.g., during
     # initial dataset creation before imputation runs)
@@ -579,8 +586,26 @@ def uprate_student_loan_plans(
 
     uni_start_year = year - age + 18
     is_england = np.isin(region, list(_ENGLAND_REGIONS))
+    is_wales = region == "WALES"
+    is_scotland = region == "SCOTLAND"
+    is_northern_ireland = region == "NORTHERN_IRELAND"
     is_tertiary = highest_ed == "TERTIARY"
 
+    # England introduced Plan 5 in 2023; Wales kept Plan 2. SAAS and
+    # Student Finance NI use Plans 4 and 1 respectively for all intakes.
+    # Current residence is the funding-nation proxy described above.
+    new_borrower_plan = np.select(
+        [
+            is_scotland,
+            is_northern_ireland,
+            is_england & (uni_start_year >= _PLAN5_FIRST_UNI_START),
+            (is_england | is_wales) & (uni_start_year >= _PLAN2_FIRST_UNI_START),
+            is_england | is_wales,
+        ],
+        ["PLAN_4", "PLAN_1", "PLAN_5", "PLAN_2", "PLAN_1"],
+        default="NONE",
+    )
+
     # Initialize output arrays
     new_plan = base_plan.copy()
     repayments = person["student_loan_repayments"].values.copy()
@@ -589,10 +614,10 @@ def uprate_student_loan_plans(
     draws = splitmix64_uniform(person["person_id"].values, salt=_STUDENT_LOAN_DRAW_SALT)
 
     # Helper to assign plans to eligible people
-    def assign_with_probability(mask, plan_value, prob=_GRADUATE_LOAN_TAKE_UP):
-        """Assign plan_value to the masked people whose draw is below prob."""
-        sampled = mask & (draws < prob)
-        new_plan[sampled] = plan_value
+    def assign_with_probability(mask, prob=_GRADUATE_LOAN_TAKE_UP):
+        """Assign the funding-nation plan to sampled new borrowers."""
+        sampled = mask & (draws < prob) & (new_borrower_plan != "NONE")
+        new_plan[sampled] = new_borrower_plan[sampled]
         repayments[sampled] = 0.0
 
     # === Step 1: Re-label existing loan holders ===
@@ -611,7 +636,7 @@ def assign_with_probability(mask, plan_value, prob=_GRADUATE_LOAN_TAKE_UP):
     new_plan[is_plan5_cohort & is_england] = "PLAN_5"
     new_plan[is_plan5_cohort & ~is_england] = "PLAN_2"
 
-    # === Step 2: Add Plan 1 holders in extended age bands ===
+    # === Step 2: Add holders in the extended pre-2012 age bands ===
     # Income-contingent (Plan 1) loans began with the 1998 intake, aged 44 in
     # 2024. With ages fixed, later years' Plan 1 intakes reach older ages, up
     # to the write-off.
@@ -626,9 +651,9 @@ def assign_with_probability(mask, plan_value, prob=_GRADUATE_LOAN_TAKE_UP):
         & (uni_start_year < _PLAN2_FIRST_UNI_START)
         & (uni_start_year + _PLAN1_WRITEOFF_YEARS > year)
     )
-    assign_with_probability(plan1_new_cohort, "PLAN_1")
+    assign_with_probability(plan1_new_cohort)
 
-    # === Step 3: Add Plan 2 holders in extended age bands ===
+    # === Step 3: Add holders in the extended 2012-2022 age bands ===
     # In the base year Plan 2 holders are at most the age of the 2012 intake
     # (30 in 2024). With ages fixed, by 2029 Plan 2 reaches age 35.
     # Target: NONE people who are tertiary-educated, older than the base
@@ -641,20 +666,19 @@ def assign_with_probability(mask, plan_value, prob=_GRADUATE_LOAN_TAKE_UP):
         & (uni_start_year >= _PLAN2_FIRST_UNI_START)
         & (uni_start_year < _PLAN5_FIRST_UNI_START)
     )
-    assign_with_probability(plan2_new_cohort, "PLAN_2")
+    assign_with_probability(plan2_new_cohort)
 
-    # === Step 4: Add Plan 5 holders (new plan from 2023) ===
-    # Plan 5 didn't exist in base year. Eligible: tertiary-educated NONE
-    # people in England who would have started uni 2023+.
+    # === Step 4: Add holders from the 2023-and-later intakes ===
+    # Plan 5 is for England only. Other funding nations retain their plans.
+    # Eligible: tertiary-educated NONE people who would have started uni 2023+.
     # Age constraint: must be 21+ (finished 3-year degree) to be repaying.
-    plan5_eligible = (
+    new_intake_eligible = (
         (new_plan == "NONE")
         & is_tertiary
-        & is_england
-        & (uni_start_year >= 2023)
+        & (uni_start_year >= _PLAN5_FIRST_UNI_START)
         & (age >= 21)
     )
-    assign_with_probability(plan5_eligible, "PLAN_5")
+    assign_with_probability(new_intake_eligible)
 
     # Write back to the person table (without the merged region column).
     person_out = current_year.person.copy()

policyengine_uk/parameters/gov/local_authorities/council_tax_reduction/income_disregards/maintenance.yaml

@@ -0,0 +1,24 @@
+description: Council Tax Reduction disregards this weekly amount of the applicant's and partner's combined spousal maintenance where the family includes a child or young person (the family-premium condition in the English default and Welsh schemes).
+values:
+  2013-04-01: 15
+metadata:
+  period: week
+  unit: currency-GBP
+  label: council tax reduction maintenance disregard
+  reference:
+    - title: Council Tax Reduction England pension-age Schedule 5 paragraph 20
+      href: https://www.legislation.gov.uk/uksi/2012/2885/schedule/5/paragraph/20
+    - title: Council Tax Reduction England default Schedule 8 paragraph 49
+      href: https://www.legislation.gov.uk/uksi/2012/2886/appendix/8/paragraph/49
+    - title: Council Tax Reduction Wales prescribed pension-age Schedule 4 paragraph 20
+      href: https://www.legislation.gov.uk/wsi/2013/3029/schedule/4/paragraph/20
+    - title: Council Tax Reduction Wales prescribed working-age Schedule 9 paragraph 49
+      href: https://www.legislation.gov.uk/wsi/2013/3029/schedule/9/paragraph/49
+    - title: Council Tax Reduction Wales default pension-age Schedule 5 paragraph 20
+      href: https://www.legislation.gov.uk/wsi/2013/3035/schedule/5/paragraph/20
+    - title: Council Tax Reduction Wales default working-age Schedule 7 paragraph 49
+      href: https://www.legislation.gov.uk/wsi/2013/3035/schedule/7/paragraph/49
+    - title: Council Tax Reduction Scotland pension-age Schedule 3 paragraph 19
+      href: https://www.legislation.gov.uk/ssi/2012/319/schedule/3/paragraph/19
+    - title: Council Tax Reduction Scotland working-age regulation 57(6)
+      href: https://www.legislation.gov.uk/ssi/2021/249/regulation/57

policyengine_uk/parameters/gov/local_authorities/council_tax_reduction/income_disregards/widowed_parents_allowance.yaml

@@ -0,0 +1,24 @@
+description: Council Tax Reduction disregards this weekly amount of each applicant's or partner's Widowed Parent's Allowance.
+values:
+  2013-04-01: 15
+metadata:
+  period: week
+  unit: currency-GBP
+  label: council tax reduction widowed parents allowance disregard
+  reference:
+    - title: Council Tax Reduction England pension-age Schedule 5 paragraph 7
+      href: https://www.legislation.gov.uk/uksi/2012/2885/schedule/5/paragraph/7
+    - title: Council Tax Reduction England default Schedule 8 paragraph 21
+      href: https://www.legislation.gov.uk/uksi/2012/2886/appendix/8/paragraph/21
+    - title: Council Tax Reduction Wales prescribed pension-age Schedule 4 paragraph 7
+      href: https://www.legislation.gov.uk/wsi/2013/3029/schedule/4/paragraph/7
+    - title: Council Tax Reduction Wales prescribed working-age Schedule 9 paragraph 21
+      href: https://www.legislation.gov.uk/wsi/2013/3029/schedule/9/paragraph/21
+    - title: Council Tax Reduction Wales default pension-age Schedule 5 paragraph 7
+      href: https://www.legislation.gov.uk/wsi/2013/3035/schedule/5/paragraph/7
+    - title: Council Tax Reduction Wales default working-age Schedule 7 paragraph 21
+      href: https://www.legislation.gov.uk/wsi/2013/3035/schedule/7/paragraph/21
+    - title: Council Tax Reduction Scotland pension-age Schedule 3 paragraph 6
+      href: https://www.legislation.gov.uk/ssi/2012/319/schedule/3/paragraph/6
+    - title: Council Tax Reduction Scotland working-age regulation 57(4)
+      href: https://www.legislation.gov.uk/ssi/2021/249/regulation/57

policyengine_uk/tests/test_council_tax_reduction_maintenance_wpa.py

@@ -0,0 +1,184 @@
+"""Statutory CTR income and disregards, independently calculated at 52 weeks.
+
+England: SI 2012/2885 Sch 1 para 16, Sch 5 paras 7/20; default SI
+2012/2886 Sch para 54, Sch 8 paras 21/49. Wales: WSI 2013/3029 Sch 1
+para 10, Sch 6 para 17, Sch 4 paras 7/20, Sch 9 paras 21/49; default
+WSI 2013/3035 Sch 5 paras 7/20, Sch 7 paras 21/49. Scotland: SSI
+2021/249 reg 57(1)(b)(vi), (d), (4), (6); SSI 2012/319 reg 27(1)(j),
+(o), Sch 3 paras 6/19. Wales's family premium applies to a family with
+a child or young person (3029 Sch 1 para 1(1)(c), Sch 7 para 3).
+"""
+
+import pytest
+
+from policyengine_uk import Simulation
+
+YEAR = 2026
+AREAS = {
+    "ENGLAND": ("LONDON", "WESTMINSTER"),
+    "WALES": ("WALES", "CARDIFF"),
+    "SCOTLAND": ("SCOTLAND", "GLASGOW_CITY"),
+}
+pytestmark = pytest.mark.usefixtures("cloned_uk_tax_benefit_system")
+
+
+def family(
+    country,
+    *,
+    age=40,
+    recipient="claimant",
+    maintenance=0,
+    wpa=0,
+    child=False,
+    child_maintenance=0,
+    route="general",
+    head_applies_alone=False,
+):
+    people = {
+        "claimant": {
+            "age": age,
+            "is_claimant_or_partner": True,
+            "is_household_head": True,
+            "state_pension": 0,
+        },
+        "partner": {
+            "age": age,
+            "is_claimant_or_partner": True,
+            "is_household_head": False,
+            "state_pension": 0,
+        },
+        "other": {
+            "age": 22,
+            "is_claimant_or_partner": False,
+            "is_household_head": False,
+        },
+    }
+    if child:
+        people["child"] = {
+            "age": 5,
+            "is_claimant_or_partner": False,
+            "is_household_head": False,
+        }
+    people[recipient].update(
+        spousal_maintenance_income=maintenance,
+        widowed_parents_allowance_reported=wpa,
+        maintenance_income=child_maintenance,
+    )
+    region, authority = AREAS[country]
+    # Pin the route and other awards to isolate the statutory income list.
+    # Reported WPA is deliberate even at pension age: its statutory income
+    # treatment also applies to an under-pension-age partner's payment.
+    benunit = {
+        "members": list(people),
+        "council_tax_reduction_pensioner": age >= 67,
+        "council_tax_reduction_head_applies_alone": head_applies_alone,
+        "child_benefit": 0,
+        "tax_credits": 0,
+        "universal_credit": 0,
+        "income_support": 0,
+        "in_receipt_of_guarantee_credit": route == "guarantee",
+        "in_receipt_of_savings_credit_only": route == "savings",
+        "pension_credit_income": 13_000 if route == "savings" else 0,
+        "pension_credit": 200 if route == "savings" else 0,
+        "pension_credit_war_pension_income": 0,
+    }
+    return Simulation(
+        situation={
+            "people": {
+                name: {key: {YEAR: value} for key, value in inputs.items()}
+                for name, inputs in people.items()
+            },
+            "benunits": {"family": benunit},
+            "households": {
+                "home": {
+                    "members": list(people),
+                    "region": region,
+                    "local_authority": authority,
+                }
+            },
+        }
+    )
+
+
+def income(sim):
+    return float(sim.calculate("council_tax_reduction_applicable_income", YEAR)[0])
+
+
+@pytest.mark.parametrize("country", AREAS)
+@pytest.mark.parametrize("age", [40, 80])
+@pytest.mark.parametrize("recipient", ["claimant", "partner"])
+def test_general_rules_count_wpa_after_fifteen_pounds_weekly(country, age, recipient):
+    # £156.65 * 52 = £8,145.80; less £15 * 52 = £780 gives £7,365.80.
+    sim = family(country, age=age, recipient=recipient, wpa=8_145.80)
+    assert income(sim) == pytest.approx(7_365.80, abs=0.01)
+
+
+@pytest.mark.parametrize("country", AREAS)
+@pytest.mark.parametrize("age", [40, 80])
+@pytest.mark.parametrize("child, expected", [(False, 5_200), (True, 4_420)])
+def test_maintenance_disregard_depends_on_the_family(country, age, child, expected):
+    # £100/week counts in full without a child; (£100 - £15)*52 otherwise.
+    sim = family(country, age=age, maintenance=5_200, child=child)
+    assert income(sim) == pytest.approx(expected, abs=0.01)
+
+
+@pytest.mark.parametrize("country", AREAS)
+def test_maintenance_payments_share_one_family_disregard(country):
+    sim = family(country, maintenance=2_600, child=True)
+    sim.set_input("spousal_maintenance_income", YEAR, [2_600, 2_600, 0, 0])
+    assert income(sim) == pytest.approx(4_420, abs=0.01)
+
+
+@pytest.mark.parametrize("country", AREAS)
+def test_small_payments_are_not_over_disregarded(country):
+    assert income(family(country, maintenance=500, wpa=500, child=True)) == 0
+
+
+@pytest.mark.parametrize("country", AREAS)
+def test_only_applicant_partner_income_counts_and_child_maintenance_is_excluded(
+    country,
+):
+    assert (
+        income(family(country, recipient="other", maintenance=5_200, wpa=8_145.80)) == 0
+    )
+    assert income(family(country, child_maintenance=5_200, child=True)) == 0
+    assert (
+        income(
+            family(
+                country, recipient="child", child=True, maintenance=5_200, wpa=8_145.80
+            )
+        )
+        == 0
+    )
+
+
+@pytest.mark.parametrize("country", AREAS)
+def test_head_applying_alone_has_no_partner_or_child_disregard(country):
+    assert (
+        income(
+            family(
+                country, recipient="partner", maintenance=5_200, head_applies_alone=True
+            )
+        )
+        == 0
+    )
+    assert income(
+        family(country, maintenance=5_200, child=True, head_applies_alone=True)
+    ) == pytest.approx(5_200)
+
+
+@pytest.mark.parametrize("country", AREAS)
+def test_guarantee_credit_still_passports_the_whole_income(country):
+    sim = family(country, age=80, maintenance=5_200, wpa=8_145.80, route="guarantee")
+    assert income(sim) == 0
+
+
+@pytest.mark.parametrize("country", AREAS)
+@pytest.mark.parametrize("child, expected", [(False, 13_200), (True, 12_420)])
+def test_savings_credit_uses_assessed_income_without_adding_maintenance_again(
+    country, child, expected
+):
+    # The Secretary of State's £13,000 already contains £5,200 maintenance.
+    # Add £200 savings credit and subtract only CTR's child-family disregard.
+    sim = family(country, age=80, maintenance=5_200, child=child, route="savings")
+    assert income(sim) == pytest.approx(expected, abs=0.01)