Add CHIP enrollment calibration targets

changelog.d/1162.added.md

@@ -0,0 +1 @@
+Add CHIP enrollment calibration targets and CPS CHIP take-up anchoring.

policyengine_us_data/build_outputs/us_augmentations.py

@@ -633,6 +633,13 @@ def _build_reported_takeup_anchors(
         reported_anchors["takes_up_medicaid_if_eligible"] = data[
             "has_medicaid_health_coverage_at_interview"
         ][time_period].astype(bool)
+    if (
+        "reported_has_chip_health_coverage_at_interview" in data
+        and time_period in data["reported_has_chip_health_coverage_at_interview"]
+    ):
+        reported_anchors["takes_up_chip_if_eligible"] = data[
+            "reported_has_chip_health_coverage_at_interview"
+        ][time_period].astype(bool)
     if (
         "receives_housing_assistance" in data
         and time_period in data["receives_housing_assistance"]

policyengine_us_data/calibration/entity_clone.py

@@ -142,6 +142,13 @@ def _build_reported_takeup_anchors(data: dict, time_period: int) -> dict:
         reported_anchors["takes_up_medicaid_if_eligible"] = data[
             "has_medicaid_health_coverage_at_interview"
         ][time_period].astype(bool)
+    if (
+        "reported_has_chip_health_coverage_at_interview" in data
+        and time_period in data["reported_has_chip_health_coverage_at_interview"]
+    ):
+        reported_anchors["takes_up_chip_if_eligible"] = data[
+            "reported_has_chip_health_coverage_at_interview"
+        ][time_period].astype(bool)
     if (
         "receives_housing_assistance" in data
         and time_period in data["receives_housing_assistance"]

policyengine_us_data/calibration/target_config.yaml

@@ -45,6 +45,12 @@ include:
   - variable: person_count
     geo_level: national
     domain_variable: medicaid_enrolled
+  - variable: person_count
+    geo_level: state
+    domain_variable: chip_enrolled
+  - variable: person_count
+    geo_level: national
+    domain_variable: chip_enrolled
   # REMOVED: is_pregnant — 100% unachievable across all 51 state geos
   - variable: snap
     geo_level: state

policyengine_us_data/calibration/unified_matrix_builder.py

@@ -2800,6 +2800,14 @@ def build_matrix(
                     reported_takeup_anchors["takes_up_medicaid_if_eligible"] = f[
                         "has_medicaid_health_coverage_at_interview"
                     ][period_key][...].astype(bool)
+                if (
+                    "reported_has_chip_health_coverage_at_interview" in f
+                    and period_key
+                    in f["reported_has_chip_health_coverage_at_interview"]
+                ):
+                    reported_takeup_anchors["takes_up_chip_if_eligible"] = f[
+                        "reported_has_chip_health_coverage_at_interview"
+                    ][period_key][...].astype(bool)
                 if (
                     "receives_housing_assistance" in f
                     and period_key in f["receives_housing_assistance"]

policyengine_us_data/datasets/cps/cps.py

@@ -579,6 +579,7 @@ def add_takeup(self):
     snap_rate = load_take_up_rate("snap", self.time_period)
     aca_rate = load_take_up_rate("aca", self.time_period)
     medicaid_rates_by_state = load_take_up_rate("medicaid", self.time_period)
+    chip_rate = load_take_up_rate("chip", self.time_period)
     head_start_rate = load_take_up_rate("head_start", self.time_period)
     early_head_start_rate = load_take_up_rate("early_head_start", self.time_period)
     ssi_rate = load_take_up_rate("ssi", self.time_period)
@@ -637,6 +638,14 @@ def add_takeup(self):
         group_keys=person_states,
     )
 
+    # CHIP: preserve current full-takeup default while anchoring CPS reporters.
+    rng = seeded_rng("takes_up_chip_if_eligible")
+    data["takes_up_chip_if_eligible"] = assign_takeup_with_reported_anchors(
+        rng.random(n_persons),
+        chip_rate,
+        reported_mask=data["reported_has_chip_health_coverage_at_interview"],
+    )
+
     # Head Start
     rng = seeded_rng("takes_up_head_start_if_eligible")
     data["takes_up_head_start_if_eligible"] = rng.random(n_persons) < head_start_rate

policyengine_us_data/db/etl_medicaid.py

@@ -129,6 +129,61 @@ def transform_administrative_medicaid_data(state_admin_df, year):
     return state_df[["ucgid_str", "medicaid_enrollment"]]
 
 
+def transform_administrative_chip_data(state_admin_df, year):
+    reporting_period = year * 100 + 12
+    print(f"Reporting period is {reporting_period}")
+    state_df = state_admin_df.loc[
+        (state_admin_df["Reporting Period"] == reporting_period)
+        & (state_admin_df["Final Report"] == "Y"),
+        [
+            "State Abbreviation",
+            "Reporting Period",
+            "Total CHIP Enrollment",
+        ],
+    ].copy()
+
+    state_df["FIPS"] = state_df["State Abbreviation"].map(STATE_ABBREV_TO_FIPS)
+
+    state_df = state_df.rename(columns={"Total CHIP Enrollment": "chip_enrollment"})
+
+    problem_states = state_df[state_df["chip_enrollment"].isna()][
+        "State Abbreviation"
+    ].tolist()
+
+    if problem_states:
+        print(
+            f"Warning: States with missing CHIP enrollment in {reporting_period}: "
+            f"{problem_states}"
+        )
+        print("Attempting to use most recent non-zero values...")
+
+        for state_abbrev in problem_states:
+            state_history = state_admin_df[
+                (state_admin_df["State Abbreviation"] == state_abbrev)
+                & (state_admin_df["Final Report"] == "Y")
+                & (state_admin_df["Total CHIP Enrollment"] > 0)
+                & (state_admin_df["Reporting Period"] < reporting_period)
+            ].sort_values("Reporting Period", ascending=False)
+
+            if not state_history.empty:
+                fallback_value = state_history.iloc[0]["Total CHIP Enrollment"]
+                fallback_period = state_history.iloc[0]["Reporting Period"]
+                print(
+                    f"  {state_abbrev}: Using {fallback_value:,.0f} "
+                    f"from period {fallback_period}"
+                )
+                state_df.loc[
+                    state_df["State Abbreviation"] == state_abbrev,
+                    "chip_enrollment",
+                ] = fallback_value
+            else:
+                print(f"  {state_abbrev}: No historical data found, keeping 0")
+
+    state_df["ucgid_str"] = "0400000US" + state_df["FIPS"].astype(str)
+
+    return state_df[["ucgid_str", "chip_enrollment"]]
+
+
 def transform_survey_medicaid_data(cd_survey_df):
     cd_df = cd_survey_df[
         ["GEO_ID", "state", "congressional district", "S2704_C02_006E"]
@@ -145,7 +200,7 @@ def transform_survey_medicaid_data(cd_survey_df):
     return cd_df[["ucgid_str", "medicaid_enrollment"]]
 
 
-def load_medicaid_data(long_state, long_cd, year):
+def load_medicaid_data(long_state, long_cd, year, long_chip_state=None):
     DATABASE_URL = f"sqlite:///{STORAGE_FOLDER / 'calibration' / 'policy_data.db'}"
     engine = create_engine(DATABASE_URL)
 
@@ -175,6 +230,30 @@ def load_medicaid_data(long_state, long_cd, year):
             "state": {},
         }
 
+        if long_chip_state is not None:
+            nat_chip_stratum = Stratum(
+                parent_stratum_id=geo_strata["national"],
+                notes="National CHIP Enrolled",
+            )
+            nat_chip_stratum.constraints_rel = [
+                StratumConstraint(
+                    constraint_variable="chip_enrolled",
+                    operation="==",
+                    value="True",
+                ),
+            ]
+            nat_chip_stratum.targets_rel.append(
+                Target(
+                    variable="person_count",
+                    period=year,
+                    value=long_chip_state["chip_enrollment"].sum(),
+                    active=True,
+                    source="CMS CHIP",
+                )
+            )
+            session.add(nat_chip_stratum)
+            session.flush()
+
         # State -------------------
         for _, row in long_state.iterrows():
             # Parse the UCGID to get state_fips
@@ -215,6 +294,40 @@ def load_medicaid_data(long_state, long_cd, year):
             session.flush()
             medicaid_stratum_lookup["state"][state_fips] = new_stratum.stratum_id
 
+        if long_chip_state is not None:
+            for _, row in long_chip_state.iterrows():
+                geo_info = parse_ucgid(row["ucgid_str"])
+                state_fips = geo_info["state_fips"]
+                parent_stratum_id = geo_strata["state"][state_fips]
+
+                new_stratum = Stratum(
+                    parent_stratum_id=parent_stratum_id,
+                    notes=f"State FIPS {state_fips} CHIP Enrolled",
+                )
+                new_stratum.constraints_rel = [
+                    StratumConstraint(
+                        constraint_variable="state_fips",
+                        operation="==",
+                        value=str(state_fips),
+                    ),
+                    StratumConstraint(
+                        constraint_variable="chip_enrolled",
+                        operation="==",
+                        value="True",
+                    ),
+                ]
+                new_stratum.targets_rel.append(
+                    Target(
+                        variable="person_count",
+                        period=year,
+                        value=row["chip_enrollment"],
+                        active=True,
+                        source="CMS CHIP",
+                    )
+                )
+                session.add(new_stratum)
+                session.flush()
+
         # District -------------------
         if long_cd is None:
             session.commit()
@@ -291,9 +404,15 @@ def main():
 
     # Transform -------------------
     long_state = transform_administrative_medicaid_data(state_admin_df, year)
+    long_chip_state = transform_administrative_chip_data(state_admin_df, year)
 
     # Load (state admin only, no CD survey) ---
-    load_medicaid_data(long_state, long_cd=None, year=year)
+    load_medicaid_data(
+        long_state,
+        long_cd=None,
+        year=year,
+        long_chip_state=long_chip_state,
+    )
 
 
 if __name__ == "__main__":

policyengine_us_data/db/validate_hierarchy.py

@@ -197,11 +197,13 @@ def validate_demographic_strata(session):
     # because CD-level survey data is disabled pending 119th Congress
     # district code remapping (see etl_medicaid.py TODO).
     # the national medicaid target actually uses the `medicaid` (expense) variable
+    # chip_enrolled has national and state targets but no district targets.
     expected_counts = {
         "age": 18 * 488,
         "adjusted_gross_income": 9 * 488,
         "snap": 1 * 488,
         "medicaid_enrolled": 1 * 51,
+        "chip_enrolled": 1 * 52,
         "eitc_child_count": 4 * 488,
     }
 

policyengine_us_data/parameters/take_up/chip.yaml

@@ -0,0 +1,12 @@
+description: Percentage of people who enroll in CHIP, if eligible.
+metadata:
+  label: CHIP takeup rate
+  unit: /1
+  period: year
+  reference:
+    - title: CMS Medicaid and CHIP enrollment data
+      href: https://data.medicaid.gov/dataset/6165f45b-ca93-5bb5-9d06-db29c692a360
+# Keep the current PolicyEngine-US default behavior until state-specific CHIP
+# rates are derived from the new CMS enrollment targets and modeled eligibility.
+values:
+  2018-01-01: 1.0

policyengine_us_data/storage/README.md

@@ -42,6 +42,12 @@
   • Location: https://data.medicaid.gov/dataset/6165f45b-ca93-5bb5-9d06-db29c692a360?conditions%5B0%5D%5Boperator%5D=%3D&conditions%5B0%5D%5Bproperty%5D=reporting_period&conditions%5B0%5D%5Bvalue%5D=202512&conditions%5B1%5D%5Boperator%5D=%3D&conditions%5B1%5D%5Bproperty%5D=preliminary_or_updated&conditions%5B1%5D%5Bvalue%5D=U
   • Notes: Uses `total_medicaid_enrollment`, not combined Medicaid and CHIP enrollment.
 
+- **chip_enrollment_2024.csv, chip_enrollment_2025.csv, chip_enrollment_2026.csv**
+  • Source: Medicaid.gov performance indicator dataset, Applications, Eligibility, and Enrollment Data
+  • Date: December 2024 final reports, December 2025 final reports, and January 2026 preliminary reports
+  • Location: https://data.medicaid.gov/dataset/6165f45b-ca93-5bb5-9d06-db29c692a360
+  • Notes: Uses `total_chip_enrollment`, not combined Medicaid and CHIP enrollment. The 2026 file mirrors the reporting period and preliminary status used by `medicaid_enrollment_2026.csv`.
+
 - **district_mapping.csv**
   • Source: created by the script `policyengine_us/storage/calibration_targets/make_district_mapping.py`
   • Notes: this script is not part of `make data` because of the length of time it takes to run and the

policyengine_us_data/storage/calibration_targets/chip_enrollment_2024.csv

@@ -0,0 +1,52 @@
+state,enrollment
+AK,14814
+AL,183595
+AR,82091
+AZ,120881
+CA,1232909
+CO,135402
+CT,23444
+DC,17485
+DE,12528
+FL,172021
+GA,234732
+HI,22067
+IA,85253
+ID,20533
+IL,291267
+IN,143162
+KS,77160
+KY,133062
+LA,147957
+MA,197970
+MD,197870
+ME,28061
+MI,173347
+MN,5012
+MO,131476
+MS,84484
+MT,19812
+NC,347586
+ND,5221
+NE,37691
+NH,19337
+NJ,253484
+NM,47202
+NV,52007
+NY,675709
+OH,241124
+OK,83925
+OR,188738
+PA,289619
+RI,0
+SC,104711
+SD,16387
+TN,171595
+TX,356109
+UT,36723
+VA,200323
+VT,5196
+WA,60356
+WI,80976
+WV,37499
+WY,5961

policyengine_us_data/storage/calibration_targets/chip_enrollment_2025.csv

@@ -0,0 +1,52 @@
+state,enrollment
+AK,12976
+AL,194447
+AR,83783
+AZ,112634
+CA,1233169
+CO,140354
+CT,25353
+DC,16392
+DE,13989
+FL,160955
+GA,190017
+HI,24762
+IA,85504
+ID,19848
+IL,322451
+IN,126144
+KS,72305
+KY,133266
+LA,137824
+MA,191701
+MD,198218
+ME,27038
+MI,169340
+MN,4361
+MO,133259
+MS,83921
+MT,20272
+NC,346463
+ND,5542
+NE,37597
+NH,19186
+NJ,259759
+NM,43713
+NV,47076
+NY,662825
+OH,242337
+OK,79005
+OR,184481
+PA,281068
+RI,33661
+SC,114661
+SD,14418
+TN,183297
+TX,346787
+UT,34982
+VA,195480
+VT,5432
+WA,65056
+WI,100494
+WV,39681
+WY,5945