Directing more federal money to a state does not produce a detectable, distinct boost to its economic growth

Summary

The best-identified evidence in the political-economy literature says no — when a member of Congress gains the power to direct more federal money to their state (e.g. by becoming a committee chair), the state's private economy does not grow faster afterward, and the highest-quality quasi-experiment on this exact question finds the opposite: firms in that state cut capital expenditure by roughly 15% in the years after their senator/representative gained a powerful chairmanship (Cohen, Coval & Malloy 2011, Journal of Political Economy). Separately, the standard fiscal-stimulus literature (Feyrer & Sacerdote 2011) finds federal money does create jobs on average — but the effect depends entirely on what the money is spent on (infrastructure and low-income transfers work; state block grants and education aid showed close to zero net job creation), which is a different question from whether political direction of funds grows a state's economy. We also built an original state-panel test on this corpus's own data (USAspending obligations by state vs. BEA state GDP growth); it finds no statistically detectable relationship in either direction, but the test has only two usable years of overlapping data and is not a strong test on its own — it is presented as a corroborating check, not the primary evidence.

What 'directing federal money' does to a state's economy — the literature

The best causal test: committee power as a natural experiment. Cohen, Coval and Malloy's NBER/JPE paper (NBER w15839, published in the Journal of Political Economy) is the paper built specifically to answer this question causally. The problem with a naive comparison is reverse causality and confounding: states that are already growing attract more federal contracts, and members from richer/growing states are more likely to become powerful. Cohen, Coval & Malloy solve this by using an event that is exogenous to a state's economic trajectory — a member of that state's congressional delegation unexpectedly ascending to the chairmanship of a powerful committee (which mechanically triggers a jump in federal spending directed to that state, largely through seniority rules unrelated to the state's economy). Over a 40-year panel, they find that in the years following such an ascension, the average firm headquartered in that state cut capital expenditures by about 15% and reduced R&D and employment growth — the opposite of what a simple 'more federal money = more growth' story would predict. The effect reverses partially when the chairman leaves office, is present in both House and Senate, in large and small states, and is strongest among firms whose operations are concentrated in that state (i.e., that can't easily substitute activity to another state) — all patterns consistent with a real crowding-out/uncertainty mechanism (increased regulatory attention, corporate taxes redirected to accommodate earmarked programs, and resource competition) rather than an artifact of one industry or firm.

Stimulus-spending multipliers say something different, and it matters which question you're asking. Feyrer and Sacerdote's ARRA study (NBER w16759) finds that stimulus dollars DID create jobs on net — roughly one job per $170,000 in a cross-state comparison, implying a fiscal multiplier of about 0.5-1.0 overall, but close to 2.0 when education grants (which showed no job effect) are excluded and only infrastructure/low-income support spending is counted. This is a genuinely different question: ARRA dollars were allocated mostly by formula, not primarily by which member sat on which committee, so this measures the average effect of federal dollars landing in a state, not the effect of political influence specifically directing them there.

Presidents target money too, and the targeting itself is not growth-driven. Kriner and Reeves' work (The Particularistic President, Cambridge 2015; working paper version) documents that federal spending is disproportionately routed to swing states and core partisan constituencies for electoral reasons — reinforcing that the allocation itself is political rather than need- or opportunity-driven, which is exactly the confound the committee-chair natural experiment above is built to strip out.

Earmarks specifically: the post-2011-ban literature. A ban on congressional earmarks (2011-2021) measurably reduced grants that Congress could direct to state governments ("What did the Earmark Ban Do? Evidence from Intergovernmental Grants," JRAP), which is consistent with earmarks being a real, meaningful channel by which individual members move money to their states — but that same literature finds the ban weakened Congress's governing capacity without producing offsetting savings, and does not find that the earmark-driven grants themselves produced measurable extra state growth. Earmarks returned in 2021 capped at roughly 1% of discretionary spending, too small a share, and too short a post-return window, for growth effects to be separately testable yet.

An original test on this corpus's data

We built a state-panel test using this connector's own tables: fiscal.usaspending_by_state (obligated federal dollars by state and fiscal year, excluding loans, which mislocate to loan-servicer states) joined to econ.state_gdp (BEA real state GDP, SAGDP1 table) and census.pep_population for per-capita normalization. This is a correlational check, not a causal test — it cannot separate 'money causes growth' from 'growing states attract more federal contracts' — but a null result here at least says the raw data doesn't show even the naive positive association.

Important data-coverage caveat, disclosed rather than hidden: fiscal.usaspending_by_state declares a 2017-2025 window, but only 2022-2025 are actually loaded when scanned (data_coverage check). Real state GDP from BEA is available only through 2024. That leaves exactly two usable spending-year -> next-year-growth pairs (FY2022 spending -> 2023 growth, FY2023 spending -> 2024 growth) across 50 states — 100 state-year observations, but only 2 independent time periods. This is a genuinely thin panel and the result below should be read as a corroborating data point, not conclusive evidence on its own.

Two-way fixed-effects panel (state + year fixed effects, clustered standard errors, log federal dollars per capita predicting next-year real GDP growth in percentage points): coefficient = +1.87, p = 0.81 — not distinguishable from zero. Pooled cross-section (no fixed effects, the more naive comparison): coefficient = +0.38 percentage points of extra growth per log-point of spending, p = 0.31, and a leave-one-out sensitivity check shows the sign of this already-insignificant estimate flips negative if North Dakota (an extreme outlier at over $90,000 per capita in federal obligations, largely energy-royalty and MAC/Medicare-administrator driven) is excluded. In short: in the only two years this corpus's USAspending table actually holds, there is no statistically detectable relationship between how much federal money lands in a state and how fast that state's economy grows the following year, in either direction.

Bottom line

Combining the literature and the data check: the honest answer to "how much does a state's economy actually grow after its members of Congress direct federal money to it" is not measurably more than it would have anyway, and the best-identified causal study finds a negative effect on private investment (roughly -15% in capital expenditure among firms in the newly-favored state, per Cohen/Coval/Malloy). The positive-sounding 'stimulus multiplier' numbers often cited (1.3x-2.0x local income multipliers) describe a different mechanism — broad, largely formula-driven federal spending during a recession — not the effect of political seniority or earmarking directing money to a specific state. There is no strong evidence in the literature, and no evidence in this corpus's own state-level GDP and spending data, that congressional direction of federal money is itself a growth engine for a state's economy.