Use shared childcare inputs and derive the child category in DC CCSP

[hua7450]

Summary

A sweep of all 48 state child care subsidy programs found DC CCSP to be the outlier on two counts. This PR fixes both.

1. DC defined its own attending-days input. dc_ccsp_attending_days_per_month duplicated the shared childcare_attending_days_per_month that 17 other state programs read (AK, AR, AZ, FL, KY, MA, MN, MO, NE, OK, PA, TX, VA, WV, WY and others). DC now reads the shared variable as period.this_year, matching the convention in those states (int ⇒ STOCK, so no monthly division).

2. DC's reimbursement age category was a pure input. dc_ccsp_child_category defaulted to PRESCHOOL for every child. Of the 48 programs, 45 derive the age band in a named variable and CA derives it inline — DC was the only pure input.

Following the established pattern

31 of the 52 state child care age variables derive the band with .calc() on a months-keyed single_amount bracket, and age, period.this_year * MONTHS_IN_YEAR is the dominant age source. Alabama is a direct template: al_ccsp_age_category is a clean three-member enum with thresholds at 0, 36 and 60 months — the same boundaries DC uses — and AL keeps provider type, region and quality tier in their own enums rather than fusing them into the age one.

So DC now has a clean age enum, and the fused rate-table row is derived from it:

dc_ccsp_age_category      INFANT_AND_TODDLER / PRESCHOOL / SCHOOL_AGE
                          p.months.calc(age * MONTHS_IN_YEAR)
age_category/months.yaml  0 -> 0, 36 -> 1, 60 -> 2
        |
        v
dc_ccsp_child_category    the seven-row rate-table key

Going through a three-member enum also avoids mapping onto non-contiguous indices of the seven-member category enum, which would break silently if anyone reordered it.

Age bands

The FY25 rate sheet names three bands but does not define them; the definitions come from the licensing regulations and the policy manual.

Band	Boundary	Source
Infant and Toddler	under 36 months	5-A DCMR § 199 defines "Infant" as under 12 months; QIN infant/toddler eligibility runs "until Sept. 30 of the year when the child turns 3 years old" (policy manual p. 34)
Preschool	36 to under 60 months	5-A DCMR § 199: "Preschool or preschooler — a child thirty-six (36) to sixty (60) months of age but younger than school age"
School-Age	60 months and over	Policy manual p. 5: "School-age children (ages 5 and older) may only participate in the Child Care Subsidy Program for OST care"

Which rate rows the formula produces

The rate sheet fuses the age band with the service authorized, so only three of the seven rows follow from age:

Row	Produced when
INFANT_AND_TODDLER	under 36 months
PRESCHOOL	36 to under 60 months
SCHOOL_AGE_BEFORE_OR_AFTER	60 months and over

The other four are reachable only by setting dc_ccsp_child_category directly, and the formula documents why:

• INFANT_AND_TODDLER_SPECIAL_NEEDS, PRESCHOOL_AND_SCHOOL_AGE_SPECIAL_NEEDS — the special needs rate is authorized at the facility, not the child. Per the FY26 Level I Subsidy Agreement §(4), "in order to receive the special needs rate, the Provider shall complete the Special Needs Rate Request Form by facility," renewed annually. No household characteristic implies it, so it is deliberately not derived from is_disabled.
• PRESCHOOL_BEFORE_AND_AFTER — wraparound care around a public pre-K day is not in the household data, and it carries the same rate as PRESCHOOL wherever both are published.
• SCHOOL_AGE_BEFORE_AND_AFTER — care on both ends of the school day. Care on a single end is the more common placement.

There is no plain school-age row on the sheet, because school-age children may participate only for out-of-school-time care — every school-age placement is before or after school care by definition.

Known limitation: SCHOOL_AGE_BEFORE_OR_AFTER is priced only in the two traditional columns, so a school-age child whose dc_ccsp_schedule_type is set to an extended day or nontraditional column reaches no published rate and needs SCHOOL_AGE_BEFORE_AND_AFTER set directly. That case is pinned in a test. dc_ccsp_schedule_type defaults to FULL_TIME_TRADITIONAL, where both school-age rows carry the same amount.

Impact

Under the old PRESCHOOL default, a child development center full-time traditional day was priced at $63.00 regardless of age — 36% below the $98.65 infant rate and 41% above the $36.74 school-age rate. Microsimulation output is unchanged, because childcare_attending_days_per_month is 0 for everyone in the Populace microdata, as with every other attending-days state.

Files

variables/gov/states/dc/dhs/ccsp/payment/dc_ccsp_age_category.py                new
variables/gov/states/dc/dhs/ccsp/payment/dc_ccsp_child_category.py             formula added
variables/gov/states/dc/dhs/ccsp/payment/dc_ccsp_maximum_subsidy_amount.py
variables/gov/states/dc/dhs/ccsp/copay/dc_ccsp_first_child_copay.py
variables/gov/states/dc/dhs/ccsp/copay/dc_ccsp_second_child_copay.py
variables/gov/states/dc/dhs/ccsp/payment/dc_ccsp_attending_days_per_month.py   deleted
parameters/gov/states/dc/dhs/ccsp/age_category/months.yaml                     new

Not addressed here

Left for separate PRs so this one stays reviewable:

• dc_ccsp_maximum_subsidy_amount is gated only on StateCode.DC, and dc_ccsp sums it over every SPM unit member — an adult or over-age child with nonzero attending days inflates the subsidy. The copay side correctly gates on the two youngest eligible children.
• reimbursement_rates.yaml encodes only the Developing/Progressing quality tier; OSSE publishes five designations (Developing, Progressing, Quality, High-Quality, Preliminary), and there is no facility-quality dimension in the breakdown.
• The second-child copay amount at 2.31 FPG is $0.60, between $1.90 (2.21) and $2.59 (2.41) — likely a transcription slip from the sliding fee scale.
• CHILD_HOME_AND_EXPANDED_HOME PRESCHOOL_BEFORE_AND_AFTER full-time traditional is encoded as $53.03; the FY25 rate sheet says $53.04.

Test plan

• New dc_ccsp_age_category cases covering a four-child family and both band boundaries at exact months (36 and 60)
• dc_ccsp_child_category cases assert the daily rate each band reaches, including the school-age part-time rate ($18.37/day) and the unpriced nontraditional cell
• policyengine-core test .../states/dc — 332 passed
• gov/hhs/ccdf, household_state_benefits.yaml and tests/policy/baseline/partners — 699 passed
• make format clean
• CI passes

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (66cefde) to head (0de5ea4).
⚠️ Report is 15 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main     #9285   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         6    +5     
  Lines           22       115   +93     
=========================================
+ Hits            22       115   +93     

Flag	Coverage Δ
unittests	100.00% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.