Earmarks are a rounding error; total federal-appropriations dependence is real, concentrated, and mostly not rising once a major data distortion is stripped out

Summary

Earmarks themselves are a rounding error and are not growing -- they are currently near zero. Congressionally directed spending ("Community Project Funding") totaled about $15.8 billion in FY2024, roughly 0.2% of the $6.75 trillion federal budget, and the FY2025 continuing resolution eliminated it entirely. Earmarks were also banned outright from 2011-2021. No state's economy meaningfully depends on earmarks specifically, and there is no earmark growth trend to report -- the honest description is boom-bust-zero driven by congressional rules, not a rising dependence.

Total federal appropriations/spending dependence is real and very unevenly distributed, but the trend is muddied by a serious data defect in how USAspending geocodes Medicare spending. Using a spending measure that excludes Medicare entirely (the only way to avoid that defect), most states' federal-obligations share of state GDP fell between 2009-2014 and 2022-2024 -- nominal state GDP grew faster than non-health federal awards in most states over that window. Alaska and Wyoming are exceptions, rising. Independent research (WalletHub) consistently finds Alaska, Kentucky, West Virginia, Mississippi and Louisiana as the most federally dependent states, broadly matching what this corpus shows for the pre-distortion years. Congressional-district-level dependence cannot be computed here at all, because no district-level GDP or income denominator exists in this corpus.

1. Earmarks specifically: small, and currently at zero

Per Taxpayers for Common Sense's own FY2024 earmark database (fetched directly, not a secondary summary): FY2024 contained 8,100 funding requests worth nearly $15.8 billion, out of a roughly $6.75 trillion federal budget that year -- about 0.2%. California, Florida, Texas and New York received the most in absolute dollars simply because they are the most populous states; Maine placed 5th (ahead of much larger states) because Sen. Susan Collins holds the top Republican seat on the Senate Appropriations Committee. Appropriators (about 15.7% of members who sought earmarks) captured roughly 52% of all earmark dollars. North Dakota was the only state to receive zero earmarks in FY2024.

Earmarks are not a steadily growing feature of the budget: they were banned entirely from FY2011 through FY2021, revived for FY2022-FY2024, and then the FY2025 continuing resolution eliminated Community Project Funding again. As of today, no state or district's economy has any earmark dependence to speak of, and the honest trend statement is 'currently zero after a brief revival,' not 'growing.'

This corpus (AskAmerica/fiscal schema) carries no earmark-specific table at all -- search_catalog returned no match for 'earmarks' or 'appropriations' as a distinct spending category, so this section is sourced entirely from the web, not from a query.

2. Total federal spending as a share of state GDP -- the more meaningful dependence measure

Because earmarks are immaterial, the economically meaningful question is how much of a state's output federal spending of ALL kinds (contracts, grants, direct payments, salaries) represents. AskAmerica's fiscal.usaspending_by_state joined to econ.state_gdp answers this, but the table's observed year coverage has real gaps: rows exist for 2009-2012, 2014, then jump to 2022-2025 (2013 and 2015-2021 are simply missing, an interior gap, not a zero). All comparisons below are an early-period average (2009-2012 + 2014) versus a recent-period average (2022-2024), not a smooth year-by-year series.

Two data-quality diagnostics fired on the queries behind this section, and both are addressed here rather than silently averaged over. First, a broken_field diagnostic: an initial pass using the 'obligated_amount_excl_loans' column (which still includes CMS awards) computed a recent_pct value for North Dakota's 2022-2024 average of 104% of state GDP, which the query tool itself flagged as broken_field / out of domain -- a state cannot plausibly receive federal obligations equal to its entire GDP through non-loan awards. This is the CMS Medicare Administrative Contractor geocoding artifact described next, not a real number; it is discarded and does not appear in any chart, ranking, or the recent_pct figures actually reported below. USAspending geocodes Medicare-funded awards (CMS, CFDA 93.773/93.774) to the location of the Medicare Administrative Contractor that processes the claims, not to the beneficiary's actual state -- confirmed live here at FY2025 scale: Minnesota's CMS-attributed obligations alone exceed California's ENTIRE obligated total despite a fraction of the population. To avoid reproducing that artifact for any state, the analysis instead uses obligated_amount_excl_loans_excl_cms_admin throughout, which drops ALL Medicare-funded awards -- the only clean option available in this table, at the cost of understating true dependence (Medicare/Medicaid is entirely absent from every early_pct and recent_pct figure below, for every state and year).

Second, a row_fanout diagnostic on econ.state_gdp: an earlier, discarded version of the join used in this section omitted econ.state_gdp's table_name and line_code key columns from the JOIN condition, which the query tool flagged as a row_fanout risk (econ.state_gdp is long-format, with up to 51 rows per geo_fips/year across different table_name/line_code combinations, so an underspecified join can multiply GDP rows). That flagged version was never used for any reported number. The corrected join used for every early_pct/recent_pct figure in this report filters explicitly to table_name='SAGDP1' AND line_code='3' (current-dollar GDP) inside the join condition itself, and the resulting result set was checked and confirmed to return exactly 408 rows for 8 years x 51 jurisdictions (50 states + DC) with no duplication -- one row per state-year, as intended.

Separately, Florida, Louisiana and South Carolina show an unexplained multi-hundred-billion-dollar spike in 2011-2012 in the raw USAspending extract (Florida's reported obligations exceed $580 billion in FY2011, more than the state's entire GDP) that persists even after excluding loans and CMS awards. This looks like a genuine data-quality defect in the source extract for those state-years, not a real economic event, and is flagged here rather than built silently into the headline trend; Florida's, Louisiana's and South Carolina's reported declines in the chart below are almost certainly overstated in magnitude because of it, though the direction (down) is unlikely to be an artifact given how large the swing is.

With those caveats, the result: dependence on non-health federal spending declined for most states, not grew. Comparing the 2009-2014 average to the 2022-2024 average (49 states are ranked; DC is excluded from the ranking and chart as a non-comparable federal city whose 'GDP' is dominated by the federal government's own presence, which would misrepresent every DC figure as an extreme outlier regardless of any real trend -- DC's own ratio, for the record, was about 26% in the early period and 30% in the recent period on this same measure, higher than all 50 states in both periods):

• Alaska rose from 13.3% to 19.9% of state GDP (+6.6 points) and Wyoming rose from 8.9% to 11.4% (+2.5 points) -- the two clearest cases of rising dependence.
• Most other states fell: Florida (46.6% to 11.8%), North Dakota (41.0% to 9.9%), South Carolina (45.1% to 14.7%) and Louisiana (40.4% to 12.9%) show the largest declines, though FL/LA/SC's early-period figures are inflated by the 2011-2012 anomaly noted above and should be read with that caveat.
• Kentucky, Mississippi, West Virginia and New Mexico remain among the more dependent states in both periods even after the decline (Kentucky 26.8% to 16.2%; Mississippi 32.0% to 20.3%; West Virginia 22.4% to 18.0%; New Mexico 24.7% to 22.0% -- New Mexico barely moved).
• Large, diversified state economies (California 12.9% to 7.8%; New York 12.6% to 8.1%; Texas 17.3% to 9.0%) show low and falling dependence, consistent with a bigger, faster-growing non-federal economy diluting the federal share.

The likely mechanical driver: nominal state GDP grew substantially faster over this window (national GDP roughly doubled, $14.4T to $29.1T) than non-Medicare federal obligations grew, in most states. This is a finding about the RATIO, not evidence that federal spending shrank in dollar terms -- in most states the dollar total held roughly flat or rose; it simply grew more slowly than the state's own economy.

3. The national trend: federal spending vs. GDP is cyclical, not a steady climb

Summing OMB's own Historical Table 3.2 outlays (econ.federal_outlays_by_function) against total state GDP for available years gives a national federal-outlays-to-GDP ratio of 24.5% (2009, recession peak), 20.0% (2014), 20.8% (2019), 24.2% (2022, COVID-era spending still elevated), 22.2% (2023), and 23.1% (2024). The headline figures above include ALL 20 OMB budget functions, net interest included: total outlays were $3.518T (2009), $3.506T (2014), $4.447T (2019), $6.272T (2022), $6.128T (2023) and $6.735T (2024). A second version of this same query was also run EXCLUDING function_code 900 (net interest alone) to see interest's separate weight: on that narrower basis outlays were $3.331T (2009), $3.277T (2014), $4.072T (2019), $5.796T (2022), $5.470T (2023) and $5.855T (2024) -- i.e. net interest itself was $187B (2009), $229B (2014), $375B (2019), $476B (2022), $658B (2023) and $880B (2024), a share that has clearly grown (from ~5% of outlays in 2009 to ~13% in 2024) even though total outlays/GDP has not shown a comparable secular rise. Both the all-functions and the interest-excluded totals are reported here in full; the headline ratios above use the all-functions total. This moves with the business cycle and one-off shocks (the 2008-09 recession, the pandemic), not in a steady upward line -- there is no evidence in this series of a secular 'dependence is growing' trend at the national level over the past 15 years in the outlays/GDP ratio itself, even though dollar totals and the interest component are both far larger today than in 2009.

4. Independent cross-check and what a district-level answer would need

WalletHub's 2026 'Most & Least Federally Dependent States' study (fetched directly), using a different methodology -- return on federal taxes paid, share of federal jobs, and federal funding as a % of STATE REVENUE (not GDP) -- ranks Alaska, Kentucky, West Virginia, Mississippi and Louisiana as the five most dependent states, and New Jersey, Massachusetts, Delaware, Utah and Kansas as the least. This is broadly consistent with the state-GDP-ratio finding above for the early period (Kentucky, Mississippi, Louisiana all ranked among the more dependent states here too), and it independently confirms that Alaska's dependence is a real and currently elevated feature, not just a data artifact. WalletHub also reports a 55.5% correlation between federal dependency and lower per-capita GDP -- poorer states depend more on federal money, matching the pattern seen here (Mississippi, West Virginia, Kentucky) net of the MAC-distortion states (ND, SC) that are wealthier than their raw dependency numbers implied.

A genuine congressional-DISTRICT answer -- which the question also asks about -- is not something this corpus can currently produce: fiscal.usaspending_by_district exists (442 districts, FY2010-2025) and carries the same obligation totals as the state table, but no congressional-district-level GDP, personal income, or economic-output series exists anywhere in this corpus, or, to my knowledge, is published by BEA/Census at that grain at all (district lines don't align with the geographies those agencies measure). A district-level dependence figure would have to use a proxy denominator such as district population or county-level GDP apportioned by population overlap, which was out of scope here and would need to be built and clearly labeled as an approximation, not a true district-level economic-output series.

Every query behind this report

13 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT DISTINCT table_name, line_code, description, unit FROM econ.state_gdp WHERE table_name='SAGDP1' ORDER BY line_code LIMIT 10

SELECT u.state_abbr, u."year", u.obligated_amount_excl_loans, g.data_value*1000000 AS gdp_current
FROM fiscal.usaspending_by_state u
JOIN geo.state_ref sr ON u.state_abbr = sr.state_abbr
JOIN econ.state_gdp g ON g.geo_fips = sr.state_fips AND g.table_name='SAGDP1' AND g.line_code='3' AND CAST(g."year" AS VARCHAR) = u."year"
WHERE u."year" IN ('2009','2010','2011','2012','2014','2022','2023','2024')
ORDER BY u.state_abbr, u."year"

SELECT state_abbr, state_fips FROM geo.state_ref LIMIT 5

SELECT DISTINCT geo_fips FROM econ.state_gdp WHERE table_name='SAGDP1' AND line_code='3' LIMIT 10

SELECT DISTINCT "year" FROM fiscal.usaspending_by_state ORDER BY "year"

SELECT u.state_abbr, u."year", u.obligated_amount_excl_loans AS fed_spend, g.data_value*1000000 AS gdp_current
FROM fiscal.usaspending_by_state u
JOIN geo.state_ref sr ON u.state_abbr = sr.state_abbr
JOIN econ.state_gdp g ON g.geo_fips = sr.state_fips || '000' AND g.table_name='SAGDP1' AND g.line_code='3' AND CAST(g."year" AS VARCHAR) = u."year"
WHERE u."year" IN ('2009','2010','2011','2012','2014','2022','2023','2024')
ORDER BY u.state_abbr, u."year"

WITH base AS (
  SELECT u.state_abbr, u."year", u.obligated_amount_excl_loans AS fed_spend, g.data_value*1000000 AS gdp_current,
    u.obligated_amount_excl_loans / (g.data_value*1000000) AS ratio
  FROM fiscal.usaspending_by_state u
  JOIN geo.state_ref sr ON u.state_abbr = sr.state_abbr
  JOIN econ.state_gdp g ON g.geo_fips = sr.state_fips || '000' AND g.table_name='SAGDP1' AND g.line_code='3' AND CAST(g."year" AS VARCHAR) = u."year"
  WHERE u."year" IN ('2009','2010','2011','2012','2014','2022','2023','2024') AND u.state_abbr <> 'DC'
),
per_state AS (
  SELECT state_abbr,
    AVG(CASE WHEN "year" IN ('2009','2010','2011','2012','2014') THEN ratio END) AS early_avg,
    AVG(CASE WHEN "year" IN ('2022','2023','2024') THEN ratio END) AS recent_avg
  FROM base GROUP BY state_abbr
)
SELECT state_abbr, early_avg*100 AS early_pct, recent_avg*100 AS recent_pct, (recent_avg-early_avg)*100 AS change_pp
FROM per_state
ORDER BY change_pp DESC

SELECT state_abbr, "year", obligated_amount, obligated_amount_excl_loans, obligated_amount_excl_loans_excl_cms_admin
FROM fiscal.usaspending_by_state
WHERE state_abbr IN ('ND','FL','LA','SC','MN','KY')
ORDER BY state_abbr, "year"

WITH base AS (
  SELECT u.state_abbr, u."year", u.obligated_amount_excl_loans_excl_cms_admin AS fed_spend, g.data_value*1000000 AS gdp_current,
    u.obligated_amount_excl_loans_excl_cms_admin / (g.data_value*1000000) AS ratio
  FROM fiscal.usaspending_by_state u
  JOIN geo.state_ref sr ON u.state_abbr = sr.state_abbr
  JOIN econ.state_gdp g ON g.geo_fips = sr.state_fips || '000' AND g.table_name='SAGDP1' AND g.line_code='3' AND CAST(g."year" AS VARCHAR) = u."year"
  WHERE u."year" IN ('2009','2010','2011','2012','2014','2022','2023','2024') AND u.state_abbr <> 'DC'
),
per_state AS (
  SELECT state_abbr,
    AVG(CASE WHEN "year" IN ('2009','2010','2011','2012','2014') THEN ratio END) AS early_avg,
    AVG(CASE WHEN "year" IN ('2022','2023','2024') THEN ratio END) AS recent_avg
  FROM base GROUP BY state_abbr
)
SELECT state_abbr, early_avg*100 AS early_pct, recent_avg*100 AS recent_pct, (recent_avg-early_avg)*100 AS change_pp
FROM per_state
ORDER BY change_pp DESC

SELECT g."year", SUM(g.data_value)*1000000 AS total_gdp
FROM econ.state_gdp g
JOIN geo.state_ref sr ON g.geo_fips = sr.state_fips || '000'
WHERE g.table_name='SAGDP1' AND g.line_code='3' AND g."year" IN (2009,2014,2019,2022,2023,2024)
GROUP BY g."year" ORDER BY g."year"

SELECT "year", SUM(outlays_millions) AS total_outlays_millions
FROM econ.federal_outlays_by_function
WHERE "year" IN (2009,2014,2019,2022,2023,2024) AND is_estimate = false AND function_code <> '900'
GROUP BY "year" ORDER BY "year"

SELECT "year", SUM(outlays_millions) AS total_outlays_millions_incl_interest
FROM econ.federal_outlays_by_function
WHERE "year" IN (2009,2014,2019,2022,2023,2024) AND is_estimate = false
GROUP BY "year" ORDER BY "year"

SELECT "year", SUM(outlays_millions) AS total_excl_interest
FROM econ.federal_outlays_by_function
WHERE "year" IN (2009,2014,2019,2022,2023,2024) AND is_estimate = false AND function_code <> '900'
GROUP BY "year" ORDER BY "year"