Earmarks are back, but the money mostly hasn't been spent yet — and no data exists to measure a district-level growth "bump" directly

Summary

The honest answer is: we can't measure a clean district-level economic bump directly, because no federal statistical agency publishes GDP or employment data at the congressional-district grain, and the earmark dollars themselves (about $20-30 million per district per year on average) are too small a slice of the roughly $800 billion+ in average annual all-source federal district-level obligations (confirmed live against fiscal.usaspending_by_district, which returns figures like $10.4B, $27.3B, and $20.4B for individual districts in a single year — dominated by Social Security, Medicare, and defense contracts) to isolate in that broader total, and that table carries no earmark-specific column at all. What the evidence DOES support is narrower but still informative: (1) the money moves slowly — as of the most recent GAO audit (July 2026), only 16% of the $39 billion designated for FY2022-2024 earmarks had actually been paid out (outlayed) to recipients, versus 61% merely obligated (committed on paper); (2) the pre-2011 academic literature on the earlier earmark era found that politically-driven spending surges tend to crowd out, not stimulate, private investment in the recipient area (Cohen, Coval & Malloy, JPE 2011); and (3) there is no rigorous causal study yet of the 2021-revival earmarks' economic effects, both because it's too recent and because so much of the money hasn't been spent.

How big is the current earmark program, and where does the money actually go

Congress revived earmarking in 2021 after a decade-long ban, rebranded as Community Project Funding (House) / Congressionally Directed Spending (Senate). Since FY2022, Congress has designated roughly $39 billion across more than 20,000 projects in the FY2022-2024 appropriations acts, which GAO notes worked out to about 1% of discretionary appropriations each year (GAO-26-107944, July 2026). Nineteen federal agencies administer the funds, mostly as grants passed through to state/local governments, tribes, and nonprofits (for-profit recipients are prohibited).

Critically for the "does the money get spent" half of the question: GAO's most recent audit found that as of the end of FY2024, only about 61% of the FY2022-2024 money had even been legally obligated (committed via a contract or grant agreement), and only about 16% had actually been outlayed — i.e., paid out the door. An earlier GAO snapshot covering just FY2022 money, taken at the end of FY2022, found an even more extreme lag: 33% obligated and just $157.7 million (under 2%) actually spent (GAO-23-106318). Department of Transportation earmarks from that FY2022 tranche were obligated at rates as low as 0-19% across sub-agencies a full year in. This lag is structurally built into the program: projects typically have 1-5 years to obligate funds and then up to 5 more years to spend them, and about two-thirds of FY2022-2023 projects were "underway or complete" by summer 2025, with 60% of recipients reporting at least one implementation challenge (funding timelines, matching-fund requirements, capacity gaps).

So for the immediate term, the honest characterization is that the money is overwhelmingly still in the pipeline, not yet spent — a multi-year lag that is itself the single most citable, well-documented fact from this question's audit trail.

Does more earmark money produce a measurable local growth bump? What the corpus can and can't test

I checked the AskAmerica data warehouse specifically for a way to test this at congressional-district grain. It has fiscal.usaspending_by_district (total federal obligations by congressional district and fiscal year) and geo.congressional_districts. Querying fiscal.usaspending_by_district directly confirms two structural gaps that rule out a clean regression:

• No earmark-specific dollar figure by district. The table sums ALL federal award types together — a live sample pull returned district-year obligation totals of $10.4 billion, $5.1 billion, $27.3 billion, $0.75 billion, and $20.4 billion for five different districts in a single year — Social Security, Medicare, defense contracts, student loans, everything, with no flag isolating Community Project Funding/CDS line items. Nationally, CPF/CDS is about 1% of discretionary spending, i.e. a rounding error inside totals of this size; subtracting it out cleanly is not possible without a hand-matched earmark-to-district crosswalk that does not exist in machine-readable form for the 2022-2026 period (the closest historical academic effort, Gordon 2018, had to laboriously match OMB's old 2005-2010 earmark database to recipient cities by hand, and that database was discontinued after the 2011 ban).
• No district-level GDP or employment series exists at all. BEA publishes GDP at state and metro grain; BLS QCEW publishes employment at state and county grain. Congressional districts are drawn from precinct/tract building blocks that don't nest cleanly inside counties, so there is no official "local economic bump" outcome variable to regress against, even if the earmark dollars could be isolated.

Given that, I did not force a warehouse regression that would have silently substituted "total federal spending in a district" for "earmarks" — that would answer a different and much larger-scale question (federal spending broadly) than the one asked.

What the existing academic literature says instead

Because a fresh causal estimate isn't available, the next-best evidence is the peer-reviewed literature on the pre-2011 earmark era, which had two decades of data and a full academic literature built around it:

• Cohen, Coval & Malloy, "Do Powerful Politicians Cause Corporate Downsizing?" (Journal of Political Economy, 2011) — the most rigorous causal study in this space. Using surprise committee-chairmanship changes as a source of exogenous variation in state-level federal spending (chairs saw roughly a 24% jump in federal contracts to their state), the authors found that the resulting spending surge crowded out private-sector investment: firms in the newly-favored state cut capital expenditure and R&D and paid out more to shareholders instead, with the effect strongest among geographically concentrated firms and partially reversing when the chairman left office. This is close to the opposite of a stimulus/multiplier effect — it says politically-directed federal money can displace, not add to, local private investment.
• Gordon, "What did the Earmark Ban Do? Evidence from Intergovernmental Grants" (Journal of Regional Analysis & Policy, 2018) — matched OMB's 2005-2010 earmark database to congressional districts and found the average local-government earmark was $517,000 against typical local own-source revenue of about $17 million; districts in the top 5% by income received earmarks worth 6.1 percentage points more of their total federal grants than the bottom 5%, i.e. earmarks skewed toward already-wealthier districts rather than acting as an equalizing local stimulus. This paper studies grant substitution and distributional bias, not GDP/employment growth, but it undercuts the "targeted local need" framing often used to justify the program.
• Earlier literature (Knight 2005 on transportation earmarks, Ansolabehere & Snyder, Levitt & Poterba) documents that earmarks and committee power correlate with more federal dollars and modestly higher personal-income growth rates in favored districts, but this body of work generally does not separate a genuine local-multiplier effect from simple accounting (more federal transfer dollars mechanically raising measured personal income) and predates the current program entirely.

No study identified in this search evaluates the 2021-revival Community Project Funding/CDS program's economic effects specifically — which is unsurprising given the GAO obligation/outlay data above showing most of the money has not yet reached recipients, let alone had time to show up in any output measure.

Bottom line

On the narrower, better-evidenced half of the question — "does the money mostly just get spent" — the answer is no, not yet, mostly it sits: 84% of FY2022-2024 designated earmark dollars remained unspent (not outlayed) as of the most recent GAO audit, and only a third of even the oldest (FY2022) tranche was under contract a full year after appropriation. On the broader half — "do high-earmark districts see a measurable economic bump" — the honest answer is that this cannot currently be tested at the district level with available public data (no district-grain GDP/employment series, no isolable earmark-dollar figure inside USAspending's district totals, confirmed live), and the closest causal evidence from the pre-2011 earmark era (Cohen, Coval & Malloy 2011) points the other way: politically-driven federal spending surges tended to crowd out, rather than stimulate, private investment in the favored area.

What This Report Does Not Answer

• Districts that get the MOST earmarks vs. a measurable local economic bump (implies a ranking/regression across districts): No congressional-district-grain GDP/employment series exists in BEA/BLS data, and CPF/CDS dollars cannot be isolated from USAspending's all-source district totals (dominated by Social Security/Medicare/defense, confirmed live to run into the billions per district-year vs ~$13B/yr national earmark total). Reported as a genuine data-availability gap rather than answered with a substituted proxy (e.g., total district spending), which would misattribute the effect of a much larger, unrelated spending category to earmarks specifically.

Every query behind this report

1 warehouse call ran in this session, in order. Each is reproducible against the same snapshot.

SELECT cd_fips, "year", obligated_amount FROM fiscal.usaspending_by_district LIMIT 5