diff --git a/changelog.d/cbo-medicare-net-of-premiums.fixed.md b/changelog.d/cbo-medicare-net-of-premiums.fixed.md
new file mode 100644
index 00000000000..31b17d9bb0c
--- /dev/null
+++ b/changelog.d/cbo-medicare-net-of-premiums.fixed.md
@@ -0,0 +1 @@
+The default CBO household income now counts Medicare as CBO does, at the average cost per enrollee net of premiums and excluding the Part D low-income subsidy, through a new `cbo_medicare_benefit` variable instead of gross Medicare spending.
diff --git a/policyengine_us/parameters/calibration/gov/hhs/medicare/part_d_low_income_subsidy_per_enrollee.yaml b/policyengine_us/parameters/calibration/gov/hhs/medicare/part_d_low_income_subsidy_per_enrollee.yaml
new file mode 100644
index 00000000000..6e84c8d8cac
--- /dev/null
+++ b/policyengine_us/parameters/calibration/gov/hhs/medicare/part_d_low_income_subsidy_per_enrollee.yaml
@@ -0,0 +1,54 @@
+description: The Medicare Trustees report and project these Part D low-income subsidy payments per Medicare enrollee.
+metadata:
+  unit: currency-USD
+  period: year
+  label: Medicare Part D low-income subsidy per enrollee
+  # From 2036, each value is the 2035 value times the growth in Medicare
+  # spending per enrollee, so the subsidy keeps its 2035 share of spending.
+  uprating: calibration.gov.hhs.medicare.per_capita_cost
+  reference:
+  # Each value is Table IV.B10's "Low-income subsidy" payments ($ billions,
+  # incurred basis) over Table V.B3's total enrollment (file page 196,
+  # thousands), people with Part A and/or Part B coverage, so it is per
+  # Medicare enrollee, not per subsidy recipient. 2016 to 2035 come from the
+  # 2026 report's Table IV.B10 (file page 159); 2015, which that table no
+  # longer shows, from the 2025 report's (file page 159). The two reports
+  # agree on 2016 to 2023. 2026 Medicare Trustees Report, transmitted
+  # 2026-06-09, retrieved 2026-10-11, sha256
+  # ffa56b9137006872300b0346149eae1613d09a172b6ba118aad48e66dfc48fa8; 2025
+  # report sha256
+  # 1b3de7a4faeee5ad42059e10740603182b716a5fa7b53b30d360b45b9e474d95.
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table IV.B10
+    href: https://www.cms.gov/oact/tr/2026#page=159
+  - title: 2025 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table IV.B10
+    href: https://www.cms.gov/oact/tr/2025#page=159
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table V.B3
+    href: https://www.cms.gov/oact/tr/2026#page=196
+values:
+  # Historical.
+  2015-01-01: 460.52
+  2016-01-01: 462.57
+  2017-01-01: 465.21
+  2018-01-01: 474.84
+  2019-01-01: 482.65
+  2020-01-01: 524.75
+  2021-01-01: 547.05
+  2022-01-01: 604.61
+  2023-01-01: 663.6
+  2024-01-01: 605.29
+  # 2025 is the first year of the full Inflation Reduction Act benefit
+  # redesign (file page 119). The subsidy per low-income enrollee fell from
+  # $2,833 to $1,226 while the direct subsidy per enrollee rose from $383 to
+  # $1,869 (Table IV.B9, file page 158).
+  2025-01-01: 251.12
+  # Trustees' intermediate estimates.
+  2026-01-01: 255.86
+  2027-01-01: 283.45
+  2028-01-01: 298.54
+  2029-01-01: 324.78
+  2030-01-01: 377.29
+  2031-01-01: 396.02
+  2032-01-01: 404.71
+  2033-01-01: 409.83
+  2034-01-01: 417.15
+  2035-01-01: 431.59
diff --git a/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_a.yaml b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_a.yaml
new file mode 100644
index 00000000000..5049d213595
--- /dev/null
+++ b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_a.yaml
@@ -0,0 +1,47 @@
+description: The Medicare Trustees report and project this Part A premium income per Medicare enrollee, which counts premiums paid by or on behalf of people without premium-free Part A who enroll voluntarily.
+metadata:
+  unit: currency-USD
+  period: year
+  label: Medicare Part A premium income per enrollee
+  # From 2036, each value is the 2035 value times the growth in Medicare
+  # spending per enrollee, so premiums keep their 2035 share of spending.
+  uprating: calibration.gov.hhs.medicare.per_capita_cost
+  reference:
+  # Each value is Table III.B4's "Premiums from voluntary enrollees" (file
+  # page 64, $ billions, cash basis) over Table V.B3's total enrollment
+  # (file page 196, thousands), which footnote 2 defines as the "Number of
+  # beneficiaries with HI and/or SMI coverage". The Part A premium is paid
+  # "by or on behalf of" these enrollees (glossary, file page 257), so state
+  # Medicaid buy-in payments are included. 2026 Medicare Trustees Report,
+  # transmitted 2026-06-09, retrieved 2026-10-11, sha256
+  # ffa56b9137006872300b0346149eae1613d09a172b6ba118aad48e66dfc48fa8.
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table III.B4
+    href: https://www.cms.gov/oact/tr/2026#page=64
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table V.B3
+    href: https://www.cms.gov/oact/tr/2026#page=196
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, glossary, Part A premium
+    href: https://www.cms.gov/oact/tr/2026#page=257
+values:
+  # Historical.
+  2015-01-01: 57.57
+  2016-01-01: 57.82
+  2017-01-01: 59.64
+  2018-01-01: 59.98
+  2019-01-01: 63.38
+  2020-01-01: 63.61
+  2021-01-01: 65.65
+  2022-01-01: 69.05
+  2023-01-01: 70.56
+  2024-01-01: 70.52
+  2025-01-01: 82.26
+  # Trustees' intermediate estimates.
+  2026-01-01: 97.54
+  2027-01-01: 102.32
+  2028-01-01: 106.72
+  2029-01-01: 112.68
+  2030-01-01: 117.5
+  2031-01-01: 125.13
+  2032-01-01: 131.5
+  2033-01-01: 140.4
+  2034-01-01: 149.07
+  2035-01-01: 156.27
diff --git a/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_b.yaml b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_b.yaml
new file mode 100644
index 00000000000..53b67c103aa
--- /dev/null
+++ b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_b.yaml
@@ -0,0 +1,61 @@
+description: The Medicare Trustees report and project this Part B premium income per Medicare enrollee, which counts premiums paid by or on behalf of enrollees, including income-related premiums and premiums that state Medicaid programs pay.
+metadata:
+  unit: currency-USD
+  period: year
+  label: Medicare Part B premium income per enrollee
+  # From 2036, each value is the 2035 value times the growth in Medicare
+  # spending per enrollee, so premiums keep their 2035 share of spending.
+  uprating: calibration.gov.hhs.medicare.per_capita_cost
+  reference:
+  # Each value is premium income over Table V.B3's total enrollment (file
+  # page 196, thousands), people with Part A and/or Part B coverage.
+  # - 2015 to 2026: Table III.C7 (file page 102), "Premium income", $
+  #   millions, incurred basis.
+  # - 2027 to 2035: Table III.C4 (file page 96), "Premium income", $
+  #   billions, cash basis. Footnote 7 credits premiums withheld from the
+  #   benefits due January 3, 2027 on December 31, 2026; their amount is
+  #   III.C4's 2026 cash value less III.C7's 2026 incurred value (174.8 -
+  #   173.228 = 1.572), which is added back to 2027. Cash and incurred
+  #   premium income agree in every year from 2017 to 2025 except 2020 and
+  #   2021, the other footnote 7 years.
+  # Premium income is "premium payments by (or on behalf of) enrollees"
+  # (file page 88). It includes income-related premiums: in 2019 the
+  # standard premium, $135.50 a month (Table III.C2, file page 89), times
+  # 56.019 million Part B enrollees plus Table V.E3's $8.4 billion of
+  # income-related premiums (file page 215) is $99.5 billion, against $99.4
+  # billion of premium income.
+  # 2026 Medicare Trustees Report, transmitted 2026-06-09, retrieved
+  # 2026-10-11, sha256
+  # ffa56b9137006872300b0346149eae1613d09a172b6ba118aad48e66dfc48fa8.
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table III.C7
+    href: https://www.cms.gov/oact/tr/2026#page=102
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table III.C4
+    href: https://www.cms.gov/oact/tr/2026#page=96
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table V.B3
+    href: https://www.cms.gov/oact/tr/2026#page=196
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Part B premium income
+    href: https://www.cms.gov/oact/tr/2026#page=88
+values:
+  # Historical.
+  2015-01-01: 1_214.54
+  2016-01-01: 1_296.34
+  2017-01-01: 1_389.19
+  2018-01-01: 1_554.68
+  2019-01-01: 1_615.55
+  2020-01-01: 1_729.23
+  2021-01-01: 1_772.6
+  2022-01-01: 2_009.35
+  2023-01-01: 1_969.28
+  2024-01-01: 2_054.43
+  2025-01-01: 2_169.44
+  # Trustees' intermediate estimates.
+  2026-01-01: 2_448.69
+  2027-01-01: 2_549.32
+  2028-01-01: 2_755.79
+  2029-01-01: 2_946.87
+  2030-01-01: 3_178.93
+  2031-01-01: 3_409.4
+  2032-01-01: 3_662.85
+  2033-01-01: 3_985.68
+  2034-01-01: 4_334.37
+  2035-01-01: 4_652
diff --git a/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_d.yaml b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_d.yaml
new file mode 100644
index 00000000000..1c8af7fff7f
--- /dev/null
+++ b/policyengine_us/parameters/calibration/gov/hhs/medicare/premium_income_per_enrollee/part_d.yaml
@@ -0,0 +1,57 @@
+description: The Medicare Trustees report and project this Part D premium income per Medicare enrollee, which counts premiums withheld from benefits, paid to the federal government or paid directly to plans, including income-related premiums.
+metadata:
+  unit: currency-USD
+  period: year
+  label: Medicare Part D premium income per enrollee
+  # From 2036, each value is the 2035 value times the growth in Medicare
+  # spending per enrollee, so premiums keep their 2035 share of spending.
+  uprating: calibration.gov.hhs.medicare.per_capita_cost
+  reference:
+  # Each value is Table III.D3's "Premium income" (file page 116, $
+  # billions, cash basis) over Table V.B3's total enrollment (file page 196,
+  # thousands), people with Part A and/or Part B coverage. Footnote 1:
+  # "Premiums include both amounts withheld from Social Security benefits or
+  # other Federal payments, direct payments to the Federal government, and
+  # those paid directly to Part D plans." Footnote 6 (file page 117) credits
+  # one January's withheld premiums on December 31 of the year before: about
+  # $0.2 billion in 2015, $0.1 billion in 2020 and an estimated $0.2 billion
+  # in 2026, which are moved to 2016, 2021 and 2027. Premium income includes
+  # income-related premiums, which are deposited into the Part D account
+  # (footnote 73, file page 151). So each value equals the Part D enrollment
+  # share times calibration.gov.hhs.medicare.part_d.premiums_per_enrollee,
+  # plus Table V.E4's aggregate income-related premiums (file page 218) per
+  # Medicare enrollee. 2026 Medicare Trustees Report, transmitted
+  # 2026-06-09, retrieved 2026-10-11, sha256
+  # ffa56b9137006872300b0346149eae1613d09a172b6ba118aad48e66dfc48fa8.
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table III.D3
+    href: https://www.cms.gov/oact/tr/2026#page=116
+  - title: 2026 Annual Report of the Boards of Trustees of the Federal Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds, Table V.B3
+    href: https://www.cms.gov/oact/tr/2026#page=196
+values:
+  # Historical.
+  2015-01-01: 224.86
+  2016-01-01: 245.3
+  2017-01-01: 264.13
+  2018-01-01: 264.91
+  2019-01-01: 255.14
+  2020-01-01: 249.65
+  2021-01-01: 267.27
+  2022-01-01: 270.08
+  2023-01-01: 274.75
+  2024-01-01: 282.08
+  # 2025 premiums fell with the voluntary Part D premium demonstration for
+  # stand-alone prescription drug plans (file page 109).
+  2025-01-01: 215.04
+  # Trustees' intermediate estimates. The 2030 jump is the Inflation
+  # Reduction Act's reset of the base beneficiary premium percentage
+  # (file page 154).
+  2026-01-01: 281.3
+  2027-01-01: 356.74
+  2028-01-01: 389.05
+  2029-01-01: 404.32
+  2030-01-01: 674.95
+  2031-01-01: 708.2
+  2032-01-01: 720.06
+  2033-01-01: 752.61
+  2034-01-01: 776.68
+  2035-01-01: 809.85
diff --git a/policyengine_us/parameters/gov/household/cbo_social_insurance_benefits.yaml b/policyengine_us/parameters/gov/household/cbo_social_insurance_benefits.yaml
index ac597fb7156..550ea17a0f1 100644
--- a/policyengine_us/parameters/gov/household/cbo_social_insurance_benefits.yaml
+++ b/policyengine_us/parameters/gov/household/cbo_social_insurance_benefits.yaml
@@ -2,7 +2,7 @@ description: Social insurance benefits included in CBO household income.
 values:
   0000-01-01:
     - social_security
-    - medicare_gross_cost
+    - cbo_medicare_benefit
     - unemployment_compensation
     - workers_compensation
 
@@ -10,3 +10,16 @@ metadata:
   unit: list
   period: year
   label: CBO social insurance benefits
+  reference:
+    # "Social insurance benefits consist of benefits provided through Social
+    # Security (Old Age, Survivors, and Disability Insurance); Medicare
+    # (measured as the average cost to the government of providing those
+    # benefits, net of offsetting receipts); regular unemployment insurance
+    # ...; and workers' compensation." (Appendix A, note 3.)
+    - title: Congressional Budget Office, The Distribution of Household Income in 2021, Appendix A
+      href: https://www.cbo.gov/system/files/2024-09/60341-income.pdf#page=27
+    # "The measure of Medicare benefits in The Distribution of Household
+    # Income includes the federal government's average cost per enrollee of
+    # providing Medicare, net of premiums paid by enrollees."
+    - title: Congressional Budget Office, Sensitivity of the Distribution of Household Income to the Treatment of Health Insurance From 1979 to 2021
+      href: https://www.cbo.gov/publication/61910
diff --git a/policyengine_us/tests/core/test_cbo_medicare_benefit_series.py b/policyengine_us/tests/core/test_cbo_medicare_benefit_series.py
new file mode 100644
index 00000000000..c3da92714a1
--- /dev/null
+++ b/policyengine_us/tests/core/test_cbo_medicare_benefit_series.py
@@ -0,0 +1,220 @@
+"""The per-enrollee series that ``cbo_medicare_benefit`` subtracts.
+
+CBO counts Medicare in household income at the government's average cost
+per enrollee, net of premiums paid by enrollees, and counts the Part D
+low-income subsidy separately. ``cbo_medicare_benefit`` takes CMS Medicare
+spending per enrollee (``calibration.gov.hhs.medicare.per_capita_cost``) and
+subtracts four series under ``calibration.gov.hhs.medicare``, each an
+aggregate from the 2026 Medicare Trustees Report over Table V.B3's total
+enrollment (people with Part A and/or Part B):
+
+- ``premium_income_per_enrollee.part_a``: Table III.B4 premiums from
+  voluntary enrollees.
+- ``premium_income_per_enrollee.part_b``: Table III.C7 incurred premium
+  income through 2026, then Table III.C4 cash premium income, with the
+  December 31, 2026 credit moved back to 2027.
+- ``premium_income_per_enrollee.part_d``: Table III.D3 premium income with
+  footnote 6's December 31 credits moved back a year.
+- ``part_d_low_income_subsidy_per_enrollee``: Table IV.B10 low-income
+  subsidy (the 2025 report for 2015).
+
+These tests read parameters only; the YAML tests cover the variable.
+
+Invariants:
+
+1. Transcription: every value equals its formula applied to the Trustees'
+   published cells, rounded as stored.
+2. Conservation: each series times total enrollment returns its published
+   aggregate, within the cells' rounding.
+3. Differential: the Part D series equals the Part D enrollment share times
+   ``part_d.premiums_per_enrollee`` (which excludes income-related premiums,
+   per Part D enrollee) plus Table V.E4's income-related premiums per
+   Medicare enrollee: two derivations from different cells agree.
+4. Part B premium income includes income-related premiums: in 2019 the
+   standard premium times Part B enrollment plus Table V.E3's aggregate
+   income-related premiums reproduces it within 0.5 percent, while the
+   standard premium alone falls more than 5 percent short.
+5. Bounds: every series is positive, and net Medicare spending per enrollee
+   stays between 0 and gross spending, in every year from 2015 to 2200.
+6. Extension: from 2036 to 2100 each series grows with Medicare spending
+   per enrollee, so each keeps its 2035 share of spending; later years hold
+   the 2100 value.
+7. Calendar years: every date in a year reads the January 1 value.
+"""
+
+import pytest
+from hypothesis import given
+from hypothesis import strategies as st
+from policyengine_core.parameters.operations.get_parameter import get_parameter
+
+from policyengine_us.system import system
+
+MEDICARE = "calibration.gov.hhs.medicare"
+SERIES = {
+    "part_a": f"{MEDICARE}.premium_income_per_enrollee.part_a",
+    "part_b": f"{MEDICARE}.premium_income_per_enrollee.part_b",
+    "part_d": f"{MEDICARE}.premium_income_per_enrollee.part_d",
+    "lis": f"{MEDICARE}.part_d_low_income_subsidy_per_enrollee",
+}
+SPENDING = f"{MEDICARE}.per_capita_cost"
+FIRST_YEAR, LAST_PROJECTED_YEAR, END_YEAR = 2015, 2035, 2100
+YEARS = range(FIRST_YEAR, LAST_PROJECTED_YEAR + 1)
+
+
+def _cells(*values):
+    return dict(zip(YEARS, values))
+
+
+# 2026 Medicare Trustees Report (sha256 ffa56b91...8fa8), file pages.
+# Table V.B3 (page 196), total enrollment, thousands.
+TOTAL_ENROLLMENT = _cells(
+    *(55_589, 57_073, 58_683, 60_020, 61_535, 62_887, 63_980, 65_166),
+    *(66_606, 68_066, 69_289, 70_743, 72_322, 74_026, 75_436, 76_598),
+    *(77_521, 78_327, 79_058, 79_827, 80_632),
+)
+# Table III.B4 (page 64), premiums from voluntary enrollees, $ billions.
+PART_A_PREMIUMS = _cells(
+    *(3.2, 3.3, 3.5, 3.6, 3.9, 4.0, 4.2, 4.5, 4.7, 4.8, 5.7),
+    *(6.9, 7.4, 7.9, 8.5, 9.0, 9.7, 10.3, 11.1, 11.9, 12.6),
+)
+# Table III.C7 (page 102), premium income, incurred basis, $ millions.
+PART_B_INCURRED = dict(
+    zip(
+        range(2015, 2027),
+        (
+            *(67_515, 73_986, 81_522, 93_312, 99_413, 108_746, 113_411),
+            *(130_941, 131_166, 139_837, 150_318, 173_228),
+        ),
+    )
+)
+# Table III.C4 (page 96), premium income, cash basis, $ billions.
+PART_B_CASH = _cells(
+    *(69.4, 72.1, 81.5, 93.3, 99.4, 111.2, 111.0, 130.9, 131.2, 139.8, 150.3),
+    *(174.8, 182.8, 204.0, 222.3, 243.5, 264.3, 286.9, 315.1, 346.0, 375.1),
+)
+# Table III.D3 (page 116), premium income, cash basis, $ billions, and
+# footnote 6's December 31 credits (page 117).
+PART_D_CASH = _cells(
+    *(12.7, 13.8, 15.5, 15.9, 15.7, 15.8, 17.0, 17.6, 18.3, 19.2, 14.9),
+    *(20.1, 25.6, 28.8, 30.5, 51.7, 54.9, 56.4, 59.5, 62.0, 65.3),
+)
+PART_D_DECEMBER_31_CREDIT = {2015: 0.2, 2020: 0.1, 2026: 0.2}
+# Table IV.B10 (page 159), low-income subsidy, $ billions; 2015 from the
+# 2025 report's Table IV.B10 (page 159).
+LOW_INCOME_SUBSIDY = _cells(
+    *(25.6, 26.4, 27.3, 28.5, 29.7, 33.0, 35.0, 39.4, 44.2, 41.2, 17.4),
+    *(18.1, 20.5, 22.1, 24.5, 28.9, 30.7, 31.7, 32.4, 33.3, 34.8),
+)
+# Table V.E4 (page 218), Part D income-related premiums, $ billions.
+PART_D_IRMAA = _cells(
+    *(0.9, 1.0, 1.2, 1.4, 1.6, 1.8, 1.8, 1.7, 2.2, 2.2, 2.4, 3.1, 3.5),
+    *(4.0, 4.6, 10.8, 11.9, 12.6, 13.7, 14.7, 15.9),
+)
+# 2019: Table III.C2 (page 89) standard monthly premium, Table V.B3 Part B
+# enrollment (thousands), Table V.E3 (page 215) income-related premiums
+# ($ billions).
+STANDARD_PART_B_PREMIUM_2019 = 135.50
+PART_B_ENROLLMENT_2019 = 56_019
+PART_B_IRMAA_2019 = 8.4
+
+
+def _part_b_billions(year):
+    if year in PART_B_INCURRED:
+        return PART_B_INCURRED[year] / 1e3
+    # III.C4 footnote 7: January 2027 withholdings credited December 31, 2026.
+    december_31_credit = PART_B_CASH[2026] - PART_B_INCURRED[2026] / 1e3
+    return PART_B_CASH[year] + (december_31_credit if year == 2027 else 0)
+
+
+def _part_d_billions(year):
+    return (
+        PART_D_CASH[year]
+        - PART_D_DECEMBER_31_CREDIT.get(year, 0)
+        + PART_D_DECEMBER_31_CREDIT.get(year - 1, 0)
+    )
+
+
+AGGREGATES = {
+    "part_a": PART_A_PREMIUMS.get,
+    "part_b": _part_b_billions,
+    "part_d": _part_d_billions,
+    "lis": LOW_INCOME_SUBSIDY.get,
+}
+
+
+def _read(name, date):
+    return get_parameter(system.parameters, name)(date)
+
+
+def _value(series, year):
+    return _read(SERIES[series], f"{year}-01-01")
+
+
+def _spending(year):
+    return _read(SPENDING, f"{year}-01-01")
+
+
+def _per_enrollee(billions, year):
+    return billions * 1e9 / (TOTAL_ENROLLMENT[year] * 1e3)
+
+
+@pytest.mark.parametrize("series", SERIES)
+@pytest.mark.parametrize("year", YEARS)
+def test_series_is_the_trustees_aggregate_per_medicare_enrollee(series, year):
+    expected = _per_enrollee(AGGREGATES[series](year), year)
+    assert _value(series, year) == round(expected, 2)
+
+
+@pytest.mark.parametrize("series", SERIES)
+@pytest.mark.parametrize("year", YEARS)
+def test_series_times_enrollment_returns_the_aggregate(series, year):
+    rebuilt = _value(series, year) * TOTAL_ENROLLMENT[year] * 1e3 / 1e9
+    # Stored to the cent; enrollment of at most 81 million moves the
+    # aggregate by at most $0.0005 billion.
+    assert rebuilt == pytest.approx(AGGREGATES[series](year), abs=1e-3)
+
+
+@pytest.mark.parametrize("year", YEARS)
+def test_part_d_series_matches_share_times_premium_plus_irmaa(year):
+    share = _read(f"{MEDICARE}.part_d.enrollment_share", f"{year}-01-01")
+    premium = _read(f"{MEDICARE}.part_d.premiums_per_enrollee", f"{year}-01-01")
+    irmaa = _per_enrollee(PART_D_IRMAA[year], year)
+    # The share is stored to four decimals and the premium to the cent, so
+    # the product can be off by up to about $0.05.
+    assert share * premium + irmaa == pytest.approx(_value("part_d", year), abs=0.1)
+
+
+def test_part_b_premium_income_includes_income_related_premiums():
+    standard = STANDARD_PART_B_PREMIUM_2019 * 12 * PART_B_ENROLLMENT_2019 * 1e3 / 1e9
+    published = _part_b_billions(2019)
+    assert standard + PART_B_IRMAA_2019 == pytest.approx(published, rel=0.005)
+    assert standard < published * 0.95
+
+
+@given(st.integers(FIRST_YEAR, END_YEAR + 100))
+def test_series_are_positive_and_net_spending_is_between_zero_and_gross(year):
+    deductions = sum(_value(series, year) for series in SERIES)
+    assert all(_value(series, year) > 0 for series in SERIES)
+    assert 0 < _spending(year) - deductions < _spending(year)
+
+
+@pytest.mark.parametrize("series", SERIES)
+def test_extension_keeps_the_2035_share_of_spending(series):
+    share_2035 = _value(series, LAST_PROJECTED_YEAR) / _spending(LAST_PROJECTED_YEAR)
+    for year in range(LAST_PROJECTED_YEAR + 1, END_YEAR + 1):
+        assert _value(series, year) / _spending(year) == pytest.approx(
+            share_2035, rel=1e-9
+        )
+
+
+@given(st.integers(END_YEAR, END_YEAR + 100))
+def test_series_hold_the_2100_values_afterwards(year):
+    for series in SERIES:
+        assert _value(series, year) == _value(series, END_YEAR)
+
+
+@given(st.integers(FIRST_YEAR, END_YEAR), st.integers(1, 12), st.integers(1, 28))
+def test_every_date_in_a_year_reads_the_january_value(year, month, day):
+    date = f"{year}-{month:02d}-{day:02d}"
+    for series in SERIES:
+        assert _read(SERIES[series], date) == _value(series, year)
diff --git a/policyengine_us/tests/policy/baseline/gov/hhs/medicare/costs/cbo_medicare_benefit.yaml b/policyengine_us/tests/policy/baseline/gov/hhs/medicare/costs/cbo_medicare_benefit.yaml
new file mode 100644
index 00000000000..f3e1652a12b
--- /dev/null
+++ b/policyengine_us/tests/policy/baseline/gov/hhs/medicare/costs/cbo_medicare_benefit.yaml
@@ -0,0 +1,111 @@
+# cbo_medicare_benefit = Medicare spending per enrollee - premium income per
+# Medicare enrollee (Parts A, B and D) - Part D low-income subsidy per
+# Medicare enrollee, the same for every enrollee.
+# 2025: $17,751 - $82.26 - $2,169.44 - $215.04 - $251.12 = $15,033.14.
+
+- name: Case 1, enrolled person gets the average net of premiums and the low-income subsidy.
+  period: 2025
+  absolute_error_margin: 0.01
+  input:
+    age: 70
+    is_medicare_eligible: true
+    medicare_enrolled: true
+  output:
+    medicare_gross_cost: 17_751
+    cbo_medicare_benefit: 15_033.14
+
+# CBO assigns the average to every enrollee, so the person with 20 quarters
+# of coverage (a $6,216 Part A premium in 2025) and the person paying the
+# income-related premiums get the same value as the others, though each one's
+# medicare_cost differs.
+- name: Case 2, every enrollee gets the same value whatever premiums they pay.
+  period: 2025
+  absolute_error_margin: 0.01
+  input:
+    people:
+      premium_free_part_a:
+        age: 70
+        is_medicare_eligible: true
+        medicare_enrolled: true
+        medicare_quarters_of_coverage: 40
+      pays_part_a_premium:
+        age: 70
+        is_medicare_eligible: true
+        medicare_enrolled: true
+        medicare_quarters_of_coverage: 20
+      pays_irmaa:
+        age: 70
+        is_medicare_eligible: true
+        medicare_enrolled: true
+        medicare_quarters_of_coverage: 40
+      disabled_under_65:
+        age: 50
+        is_medicare_eligible: true
+        medicare_enrolled: true
+        medicare_quarters_of_coverage: 40
+      not_enrolled:
+        age: 70
+        is_medicare_eligible: true
+        medicare_enrolled: false
+    tax_units:
+      tax_unit:
+        members: [premium_free_part_a]
+        adjusted_gross_income:
+          2023: 30_000
+      tax_unit_2:
+        members: [pays_part_a_premium]
+        adjusted_gross_income:
+          2023: 30_000
+      tax_unit_3:
+        members: [pays_irmaa]
+        # IRMAA uses income from two years before.
+        adjusted_gross_income:
+          2023: 1_000_000
+      tax_unit_4:
+        members: [disabled_under_65]
+      tax_unit_5:
+        members: [not_enrolled]
+    households:
+      household:
+        members:
+          [
+            premium_free_part_a,
+            pays_part_a_premium,
+            pays_irmaa,
+            disabled_under_65,
+            not_enrolled,
+          ]
+  output:
+    cbo_medicare_benefit: [15_033.14, 15_033.14, 15_033.14, 15_033.14, 0]
+
+- name: Case 3, first year of the series.
+  period: 2015
+  absolute_error_margin: 0.01
+  input:
+    age: 70
+    is_medicare_eligible: true
+    medicare_enrolled: true
+  output:
+    # $11,929 - $57.57 - $1,214.54 - $224.86 - $460.52.
+    cbo_medicare_benefit: 9_971.51
+
+- name: Case 4, last Trustees projection year.
+  period: 2035
+  absolute_error_margin: 0.01
+  input:
+    age: 70
+    is_medicare_eligible: true
+    medicare_enrolled: true
+  output:
+    # $31,764.35 (the 2034 CMS value grown at the 2031-2032 rate) - $156.27
+    # - $4,652 - $809.85 - $431.59.
+    cbo_medicare_benefit: 25_714.64
+
+- name: Case 5, not enrolled.
+  period: 2025
+  input:
+    age: 70
+    is_medicare_eligible: true
+    medicare_enrolled: false
+  output:
+    cbo_medicare_benefit: 0
diff --git a/policyengine_us/tests/policy/baseline/household/income/household/cbo_household_income.yaml b/policyengine_us/tests/policy/baseline/household/income/household/cbo_household_income.yaml
index 07d9f4f1d8e..d52fdf17f34 100644
--- a/policyengine_us/tests/policy/baseline/household/income/household/cbo_household_income.yaml
+++ b/policyengine_us/tests/policy/baseline/household/income/household/cbo_household_income.yaml
@@ -14,6 +14,7 @@
 
 - name: CBO before-tax income adds social insurance and alias matches
   period: 2025
+  absolute_error_margin: 0.01
   input:
     employment_income: 40_000
     employer_sponsored_insurance_premiums: 6_000
@@ -35,10 +36,14 @@
     cbo_household_market_income: 49_102
     # CMS Medicare spending per enrollee, 2025.
     medicare_gross_cost: 17_751
-    # $12,000 + $17,751 + $2,000 + $1,000.
-    cbo_household_social_insurance_benefits: 32_751
-    cbo_household_income_before_transfers_and_taxes: 81_853
-    cbo_household_income: 81_853
+    # $17,751 less 2025 premium income per Medicare enrollee ($82.26 Part A,
+    # $2,169.44 Part B, $215.04 Part D) and the $251.12 Part D low-income
+    # subsidy per Medicare enrollee.
+    cbo_medicare_benefit: 15_033.14
+    # $12,000 + $15,033.14 + $2,000 + $1,000.
+    cbo_household_social_insurance_benefits: 30_033.14
+    cbo_household_income_before_transfers_and_taxes: 79_135.14
+    cbo_household_income: 79_135.14
 
 - name: CBO means-tested transfers include only the listed programs
   period: 2026
diff --git a/policyengine_us/variables/gov/hhs/medicare/costs/cbo_medicare_benefit.py b/policyengine_us/variables/gov/hhs/medicare/costs/cbo_medicare_benefit.py
new file mode 100644
index 00000000000..1a4e1d96df5
--- /dev/null
+++ b/policyengine_us/variables/gov/hhs/medicare/costs/cbo_medicare_benefit.py
@@ -0,0 +1,36 @@
+from policyengine_us.model_api import *
+
+
+class cbo_medicare_benefit(Variable):
+    value_type = float
+    entity = Person
+    label = "CBO Medicare benefit"
+    documentation = (
+        "Medicare benefit as the Congressional Budget Office counts it in "
+        "household income: the government's average cost per enrollee of "
+        "providing Medicare Parts A, B and D, net of premiums paid by or on "
+        "behalf of enrollees. CBO assigns the same average to every "
+        "enrollee and counts the Part D low-income subsidy separately as a "
+        "means-tested transfer, so this subtracts per-enrollee averages of "
+        "premium income and of the low-income subsidy from Medicare "
+        "spending per enrollee, rather than the person's own premiums."
+    )
+    unit = USD
+    definition_period = YEAR
+    reference = (
+        "https://www.cbo.gov/publication/61910",
+        "https://www.cbo.gov/system/files/2024-09/60341-income.pdf#page=27",
+        "https://www.cbo.gov/system/files/2018-07/54234-workingpaper.pdf#page=46",
+    )
+    defined_for = "medicare_enrolled"
+
+    def formula(person, period, parameters):
+        p = parameters(period).calibration.gov.hhs.medicare
+        premiums = p.premium_income_per_enrollee
+        total_premiums = premiums.part_a + premiums.part_b + premiums.part_d
+        net = (
+            p.per_capita_cost
+            - total_premiums
+            - p.part_d_low_income_subsidy_per_enrollee
+        )
+        return max_(net, 0)
diff --git a/policyengine_us/variables/gov/hhs/medicare/costs/medicare_gross_cost.py b/policyengine_us/variables/gov/hhs/medicare/costs/medicare_gross_cost.py
index 48533e92fb4..1baeeaf188a 100644
--- a/policyengine_us/variables/gov/hhs/medicare/costs/medicare_gross_cost.py
+++ b/policyengine_us/variables/gov/hhs/medicare/costs/medicare_gross_cost.py
@@ -7,12 +7,19 @@ class medicare_gross_cost(Variable):
     label = "Gross Medicare cost"
     documentation = (
         "Annual Medicare spending on behalf of the beneficiary before "
-        "subtracting beneficiary premiums. This matches the gross Medicare "
-        "benefit concept used in the CBO household-income framework."
+        "subtracting beneficiary premiums: CMS Medicare spending per "
+        "enrollee, which covers Parts A through D, including the Part D "
+        "low-income subsidy. CBO's standard household income counts "
+        "Medicare net of premiums instead (cbo_medicare_benefit); CBO adds "
+        "the premiums back only in an expanded sensitivity measure, which "
+        "it calls gross Medicare spending."
     )
     unit = USD
     definition_period = YEAR
-    reference = "https://www.cms.gov/about-cms/what-we-do/medicare"
+    reference = (
+        "https://www.cms.gov/about-cms/what-we-do/medicare",
+        "https://www.cbo.gov/publication/61910",
+    )
     defined_for = "medicare_enrolled"
 
     def formula(person, period, parameters):
diff --git a/policyengine_us/variables/household/income/household/cbo_household_social_insurance_benefits.py b/policyengine_us/variables/household/income/household/cbo_household_social_insurance_benefits.py
index b39c09e5db5..82ecc087d38 100644
--- a/policyengine_us/variables/household/income/household/cbo_household_social_insurance_benefits.py
+++ b/policyengine_us/variables/household/income/household/cbo_household_social_insurance_benefits.py
@@ -7,8 +7,8 @@ class cbo_household_social_insurance_benefits(Variable):
     label = "CBO household social insurance benefits"
     documentation = (
         "Social insurance benefits included in CBO household income: Social "
-        "Security, gross Medicare benefits, unemployment insurance, and "
-        "workers' compensation."
+        "Security, Medicare benefits net of enrollee premiums, unemployment "
+        "insurance, and workers' compensation."
     )
     definition_period = YEAR
     unit = USD