In almost every major U.S. subsidy system
farm payments, clean-energy tax credits, and state economic-development deals a
small share of large, well-resourced, well-connected recipients receives most
of the money, and often receives it fastest. Scale, existing infrastructure,
and professional application capacity, not need, determine payment order.
Definition
and Scale: What "Subsidy Distribution" Actually Means
A
subsidy is any transfer of public money or public value a direct payment, a tax
credit, a loan guarantee, a grant, or a discounted resource given to a person,
farm, or company to encourage or offset an activity the government has decided
is worth supporting. That single definition covers four very different systems
that rarely get discussed together: farm payments, energy and climate credits,
corporate economic-development deals, and small-business or social-support
programs. They are usually analyzed one at a time, in separate silos, by separate
reporters and separate advocacy groups. That is precisely why the pattern
connecting them is so easy to miss.
Look
at any one program in isolation and you'll find a specific story: crop
insurance is complicated, EV credits favor buyers who can afford a new car, a
factory got a tax abatement because a mayor wanted the ribbon-cutting. Look
across all four systems at once, and a structural pattern appears. The programs
differ in purpose, in agency, in political sponsor but they share a common
allocation logic. The applicants with the most revenue, the most in-house legal
and accounting staff, the most experience with the paperwork, and the most
existing relationships with the administering agency consistently receive funds
first, most often, and in the largest amounts.
The
scale involved is not trivial. The Environmental Working Group's Farm Subsidy
Database built from USDA records obtained under the Freedom of Information Act
shows that just ten percent of America's largest and richest farms collect
almost three-fourths of federal farm subsidies. On the corporate side, Good
Jobs First's Subsidy Tracker, which aggregates state and local
economic-development awards, has documented more than 759,000 subsidy award
entries across 1,443 programs, and a landmark 2014 analysis of that database
found that three-quarters of all the economic development dollars awarded and
disclosed by state and local governments had gone to just 965 large
corporations. Meanwhile, a University of California, Berkeley analysis of
federal clean-energy tax credits found that 60% of $18 billion in credits
issued between 2006 and 2012 went to households earning more than $200,000 a
year, while only about 10% went to taxpayers earning less than $75,000.
These
are not three unrelated anecdotes. They are three independent measurements of
the same underlying phenomenon, produced by three different research
organizations using three different government data sources, over three
different time periods. That convergence is the core evidence behind this
article's central claim: subsidy systems are not neutral, and a predictable
sequence not chance, and not need determines who gets paid first.
How
Allocation Order Emerges: The Mechanism Behind the Money
Most
people imagine subsidy programs as a queue: everyone applies, everyone waits,
and money is handed out roughly in the order it's needed. The reality is closer
to a funnel with unequal openings. Four structural mechanisms consistently push
money toward large, established recipients ahead of smaller or newer ones.
1.
Design features that reward scale. Farm subsidies are frequently
calculated per acre, per bushel, or per unit of production payment formulas
built into 1930s-era programs and repeatedly re-authorized rather than
redesigned. A larger operation, by definition, produces more units, so it
collects a proportionally larger payment even when the per-unit rate is
identical for everyone. The 2025 farm bill reconciliation package illustrates
the point directly: it raised the per-person farm subsidy payment limit from
$125,000 to $155,000, with a farmer's spouse eligible for an additional
$155,000 a structural change that mechanically benefits the largest operations,
which are the only ones bumping against the old ceiling in the first place.
2.
First-come, first-served design under time pressure. When Congress
needed to move $669 billion out the door quickly during the pandemic, the
Paycheck Protection Program used a first-come, first-served model layered on
top of existing bank relationships. A peer-reviewed study using daily survey
data found that this design, combined with unequal awareness of the program,
skewed resources toward larger firms: the smallest businesses were less aware
of PPP, less likely to apply, and when they did apply they applied later,
waited longer, and were less likely to be approved. Speed-based allocation
rules look procedurally fair same rules for everyone but they systematically
favor whoever already has the infrastructure to move fast.
3.
Relationship and information advantages. A related PPP study found that a
large share of borrowers received credit through existing bank relationships,
and larger banks tended to prioritize their existing bigger business customers
while smaller banks treated their smaller clients more evenly. This is not
evidence of intentional favoritism so much as evidence that subsidy delivery
usually rides on top of pre-existing financial and administrative
infrastructure and that infrastructure is not evenly distributed to begin with.
4.
Application complexity as a hidden filter. Corporate
economic-development deals, unlike direct payments, are usually negotiated
rather than formula-based. Good Jobs First's research found that big recipients
often extract subsidies through subsidiaries with names bearing no resemblance
to the parent company, and that the largest players Boeing at more than $13
billion, Alcoa at $5.6 billion, Intel at $3.9 billion, General Motors at $3.5
billion, and Ford at $2.5 billion reached those totals through hundreds of
individually negotiated deals over years, something only an organization with
dedicated site-selection and tax-incentive staff can realistically sustain.
Meanwhile, more recent Subsidy Tracker data shows firms like Amazon actively
building internal capacity for this work, having opened a dedicated tax break
office and extracted at least $82 million from nine different tax credits,
abatements, exemptions, and enterprise zone programs across four states.
None
of these four mechanisms requires bad faith. Each is defensible in isolation:
reward productive scale, move emergency money fast, work through trusted
banking relationships, let sophisticated negotiators build sophisticated deals.
Stacked together, however, they produce the same outcome every time capacity,
not need, sets the order of payment.
Who
Gets Paid First: The Ranked Hierarchy
Synthesizing
findings across farm, energy, corporate, and small-business programs produces a
consistent five-tier hierarchy, roughly in order of typical speed and volume of
benefit received.
Tier
1: Large, repeat, professionally staffed recipients. These are the
fastest and largest beneficiaries across every program category: the top 10% of
commodity farms, the roughly 1,000–2,600 corporate parents tracked by Good Jobs
First, and firms with dedicated incentive-acquisition teams. EWG's most recent
concentration data shows the top 10% of farm subsidy recipients collected 65%
of commodity subsidies in 2024, and separate EWG research found 9,526
recipients collected farm payments every single year for 40 consecutive years
between 1985 and 2024, averaging $28,000 annually and totaling $10.7 billion.
Tier
2: Established mid-size operations with existing bank or agency relationships. These
recipients don't dominate totals the way Tier 1 does, but they consistently
outperform truly small or new entrants because they already have a banking
relationship, a compliance history, or a prior award on file. Research on
state-level economic development deals found states more heavily served by
large banks saw more incentive dollars flow per employee a proxy for the
advantage that comes from being plugged into existing financial infrastructure
rather than needing to build a relationship from scratch.
Tier
3: High-income individual beneficiaries of tax-credit programs. Unlike direct
payments, tax credits require upfront capital (to buy an EV, install solar, or
renovate a home) that is then partially reimbursed. This structure inherently
favors people who can afford the upfront cost. The Borenstein-Davis research
found that for the electric-vehicle credit specifically, 90% of the benefit
went to the top income quintile, and a follow-up analysis found the bottom 80%
of filers received a little more than 10% of all credits.
Tier
4: Smaller businesses and individuals with basic banking access. This tier
receives real support, often substantial in aggregate, but later and in smaller
average amounts than Tiers 1–3. PPP data shows this tier was reached, but only
after program adjustments: 51% of all PPP loans went to businesses with fewer
than five employees and 75% went to businesses with fewer than ten, but this
outcome partly reflects a specific mid-program correction Bank Policy Institute
research notes the White House had to grant firms under 20 employees exclusive
access to the program for a two-week window because they were being left behind
under normal rules.
Tier
5: The smallest, newest, or least-connected applicants. These
recipients apply latest, wait longest, and are approved least often, when they
apply at all. The direct research finding here is unambiguous: the smallest
businesses were less aware of PPP and less likely to apply; when they did
apply, they applied later, faced longer processing times, and were less likely
to be approved.
This
is not a moral ranking of who deserves support. It is an empirical description
of sequence and volume, built from primary data. The hierarchy holds across
farm policy, energy policy, and business relief policy three politically
distinct domains administered by three different federal agencies, using three
different allocation mechanisms. That consistency is the strongest evidence
that something structural, not incidental, is producing the pattern.
Evidence
by Major Program Category
Farm
and Agricultural Subsidies
Federal
farm programs are the oldest and most heavily studied subsidy system in the
country, and the data trail is unusually long. EWG's concentration analysis for
2024 shows the top 10% of recipients collecting 65% of commodity subsidies, and
a separate 2025 analysis of payments tied to crop reference prices found only
40% of farms grow crops eligible for those payments at all, and the top 10% of
those eligible farmers collected nearly three-quarters of the total. The system
also has notable leakage: EWG's investigative reporting found almost 80,000
people living in some of the country's biggest metro areas nowhere near
farmland collected more than $2 billion in farm subsidies between 2020 and
2024, and separately identified 44 recipients, out of a group that received
payments for 40 consecutive years, who neither work nor live on a farm despite
an "actively engaged in farming" requirement.
Energy
and Climate Subsidies
Energy
subsidies split cleanly along the same lines as farm payments, but the
mechanism is different: instead of production-based formulas, the barrier is
upfront affordability. Beyond the EV credit findings above, reporting on
Inflation Reduction Act home-energy tax credits found that wealthier homeowners
have claimed a disproportionate share of the credits, while a parallel $8.5
billion rebate program aimed specifically at low- and moderate-income
households has been slow to launch because only a small number of states have
applied to the Department of Energy for the money. Congress has since moved to
correct part of this: reporting confirms income limits on the EV credit
households up to $300,000 and individuals up to $150,000 were imposed starting in
2023 specifically because of the earlier research on who was benefiting. That
is a genuine, documented example of transparency research changing policy a
point advocacy groups can cite as evidence that publishing distributional data
works.
Corporate
Economic-Development Subsidies
This
is the category with the highest dollar concentration of any measured in this
article. Good Jobs First's foundational analysis found at least 75 percent of
cumulative disclosed state and local subsidy dollars went to just 965 large
corporations, even though those companies accounted for only about 10 percent
of the number of announced awards meaning big companies aren't just getting
bigger checks, they're getting a wildly disproportionate share of dollars
relative to how many deals they even sign. A later database expansion found
that group's cumulative total had grown to $190 billion, still three-quarters
of the entire Subsidy Tracker database. Because state and local disclosure
remains inconsistent, academic estimates of the true annual cost of these
programs range from $45 billion to $70 billion a year a wide enough band to
show how much of this spending still isn't fully visible even to researchers
with direct database access.
Small
Business and Emergency Relief Programs
The
PPP experience is the best-documented modern case study because loan-level data
was eventually released for public research. The clearest single finding is
procedural: the program's first-come, first-served design and unequal
information about the program disadvantaged the smallest businesses
specifically, even though the same research found firms that did receive aid
reported fewer layoffs, higher employment, and improved expectations about the
future meaning the program worked well for those it reached, but reach itself
was unequal.
Geographic
and Demographic Patterns
Subsidy
concentration is not evenly spread across the map, and the unevenness tends to
track existing financial and industrial infrastructure rather than need.
Corporate subsidy dollars cluster heavily in a handful of states: Good Jobs
First's tracker shows New York, Washington, and Michigan as the top three
states by cumulative disclosed corporate subsidy dollars, at $21 billion, $13
billion, and $10 billion respectively. More recent quarterly data shows this
concentration can shift quickly around single "megadeals" one 2025–26
update found Indiana alone added $3.3 billion in new subsidies in a single
quarter, driven in large part by a single Amazon data-center complex that
secured 50 years of state sales-tax exemptions worth an estimated $4 billion
and 35 years of property-tax abatements worth another $4 billion.
On
the small-business side, geography interacts with banking infrastructure. Bank
Policy Institute research on PPP found a positive correlation between how
heavily a state was served by large banks and how many PPP dollars per
small-business employee that state received a pattern that means the strength
of a region's existing banking sector, not the severity of its pandemic
exposure, helped shape how much relief flowed there. Good Jobs First's
disclosure-quality caveat is worth repeating here directly: the organization
explicitly warns that due to uneven disclosure, it is not appropriate to make
state-by-state or jurisdiction-by-jurisdiction comparisons from its raw totals,
since some states simply report more completely than others. That caveat
matters for readers building their own analyses a bigger disclosed total
sometimes means better transparency, not more generous subsidies.
Why
the Disparity Persists
Three
forces keep reproducing this hierarchy year after year, even across changes in
political administration.
Programs
are rarely redesigned from scratch. Farm subsidy formulas trace back to
Depression-era legislation and get renewed, patched, and re-authorized rather
than rebuilt around current farm structure. A payment-per-unit formula written
when the "average farm" was a fraction of today's size will always
favor today's largest operations, regardless of who is in the White House or
which party controls Congress.
Speed
and scale requirements favor whoever is already prepared. Emergency
programs like PPP have to move enormous sums in days or weeks. There is no time
to build new infrastructure, so the money necessarily flows through whatever
infrastructure banks, accountants, state economic-development offices already
exists, and that infrastructure was not built with equal geographic or
firm-size reach.
Negotiated
subsidies reward negotiating capacity. Unlike a farm payment or a tax
credit, a corporate incentive deal is bespoke a mayor, a governor's
economic-development office, and a company's site-selection team negotiate an
individually tailored package. That process inherently favors whichever side
has more institutional capacity, and companies large enough to build dedicated
tax-incentive offices as Amazon and Samsung have both done, according to Good
Jobs First reporting will always out-negotiate a first-time applicant working
from a template.
Implications
for Equity and Growth
The
public conversation about subsidies is usually framed as a binary subsidies are
either good economic policy or wasteful giveaways. The distributional data
suggests a more precise framing: a subsidy program can be sound in its overall
economic goal (supporting farm income stability, encouraging clean-energy
adoption, cushioning small businesses through a shock) while still delivering
its actual dollars in a pattern that undercuts that stated goal. A farm program
meant to protect family farms that instead sends 65% of its money to the
largest 10% of operations is not achieving its own stated purpose as
efficiently as its design would suggest. A clean-energy incentive meant to
accelerate broad adoption that instead sends 90% of its benefit to the top
income quintile is subsidizing purchases that, in many cases, would likely have
happened anyway.
This
has a direct fiscal-growth consequence too. GAO's improper-payments research
shows the government continues to lose enormous, avoidable sums even before
concentration effects are considered: agencies reported $162 billion in
improper payments across 68 federal programs in fiscal year 2024, with 75
percent of that concentrated in just five program areas, and cumulative
improper payments since fiscal year 2003 now total roughly $2.8 trillion.
Improper payments and concentrated payments are different problems one is an
error, the other is a design outcome but both point to the same underlying
weakness: allocation systems that were never built with rigorous, real-time
monitoring of where the money actually lands.
Risks,
Limitations, and Counterarguments
A
fair accounting of this topic requires acknowledging where the concentration
story is more complicated than it first appears.
Scale-based
concentration is not automatically unfair. If farm
subsidies are explicitly designed to be proportional to production, then a
large farm receiving a large payment is the program working exactly as
designed, not evidence of capture. The more precise critique is not that large
farms get large payments, but that the formula itself, and the payment caps
meant to limit it, have been repeatedly loosened rather than tightened a policy
choice, not an inevitability.
Disclosure
gaps cut both ways. Good Jobs First's own transparency caveat that
state-by-state comparisons are not appropriate given uneven disclosure means
some of the apparent "winners" in corporate subsidy totals may simply
be states that report more completely, while genuinely large but poorly-disclosed
subsidies elsewhere go undercounted. Any ranked hierarchy built on today's data
is a hierarchy of what's visible, not necessarily a complete hierarchy of
what's real.
Not
every concentration finding replicates cleanly across time. Bank Policy
Institute research specifically pushed back on the "large banks favored
large borrowers" narrative in PPP, finding that 75% of all PPP loans went
to businesses with fewer than 10 employees and that large banks were not
disproportionately represented in the largest loan-size category. Other
researchers, using different data cuts and different time windows, found the
opposite that larger banks did prioritize their bigger existing customers. Both
conclusions come from credible researchers using real loan-level data; the
disagreement reflects genuine measurement complexity different snapshots in
time, different bank samples, and different definitions of "large"
not obvious error on either side. Readers should treat PPP bank-behavior
findings as contested rather than settled.
EV
credit critiques are also contested on methodology. Some analysts
argue that studies focused only on direct tax-credit claimants understate how
many moderate-income buyers benefit indirectly through leases, since dealers can
apply the credit to lower monthly lease payments even for buyers who never
personally claim it on a tax return. This is a legitimate methodological
objection, and it means the "90% went to the top income quintile"
finding, drawn from IRS filing data, may somewhat overstate concentration for
the pre-2023 credit though it does not change the finding that filers who did
claim the credit directly skewed heavily toward high incomes, nor does it
change the fact that Congress itself acted on the original research by imposing
income caps.
Correlation
between government spending and improper payments is not evidence of
favoritism. The
GAO's improper-payment totals are a measure of administrative error and fraud
risk, not proof that money is being deliberately steered toward large
recipients. It's included here for fiscal-scale context, not as evidence of the
concentration pattern itself.
What
Can Be Tracked and What Could Change
Readers
who want to monitor this pattern themselves, rather than rely on periodic news
coverage, have more direct access to primary data than most people realize.
·
EWG's Farm Subsidy Database (farm.ewg.org)
publishes annual concentration tables, state and county rankings, and
individual recipient lookups built directly from USDA records obtained through
FOIA.
·
Good Jobs First's Subsidy Tracker (subsidytracker.goodjobsfirst.org)
is a free, searchable database of more than 759,000 state, local, and federal
subsidy awards, downloadable by company, state, or program.
·
GAO's PaymentAccuracy reporting tracks
improper-payment rates by federal program annually and is required reading for
anyone assessing administrative as opposed to distributional waste.
·
IRS Statistics of Income and Congressional
Research Service reports provide the underlying tax-filing data
behind income-distribution findings for credits like the EV and home-energy
programs.
On
the reform side, two mechanisms have already shown they can shift outcomes when
the underlying data becomes public. The EV credit income caps enacted starting
in 2023 are a direct, documented policy response to the Borenstein-Davis
concentration findings. Separately, the Governmental Accounting Standards
Board's Statement No. 77, which now requires local governments to disclose
tax-abatement revenue losses in their financial reports, is beginning to close
some of the corporate-subsidy disclosure gap that Good Jobs First has flagged
for over a decade Indiana's new transparency portal, cited above, is a direct
product of that rule. Neither fix eliminates concentration, but both
demonstrate that publishing distributional data, consistently and specifically,
is one of the few interventions with a proven track record of changing the
payment order itself.
Key
Takeaways
·
Across
farm, energy, and corporate subsidy systems, a small share of large,
well-resourced recipients consistently receives the largest and fastest share
of benefits — a pattern documented independently by USDA-based, IRS-based, and
state-disclosure-based research.
·
The
mechanism is structural, not conspiratorial: production-based formulas,
first-come-first-served speed requirements, existing banking relationships, and
negotiation capacity each independently favor scale.
·
The
clearest counter-example EV tax-credit income caps enacted after concentration
research was published shows that transparency can change allocation outcomes,
but only when it's specific enough to act on.
·
Disclosure
remains incomplete enough that any "who gets paid first" ranking,
including this one, describes what is visible in current data rather than the
full picture.
Frequently Asked Questions
What
does equity allocation mean?
Equity allocation refers to distributing a
resource money, ownership, or benefits in a way that accounts for fairness, not
just equal shares. In the subsidy context used throughout this article, it
means examining whether benefits are distributed according to need,
contribution, or some other justifiable principle, versus simply flowing to
whoever has the most capacity to claim them. In corporate finance, the same
phrase describes how ownership stakes (equity) in a company are divided among
founders, employees, and investors a related but distinct concept from subsidy
distribution.
What
is the formula for a subsidy?
At its simplest, a subsidy is calculated as
the per-unit subsidy amount multiplied by the quantity subsidized: Total Subsidy = Subsidy per Unit ×
Number of Units. In tax-credit programs, it's typically a
percentage of qualifying spending up to a cap (for example, a home-energy
credit reimbursing 18% of costs up to a limit). In price-support programs, it's
often the gap between a guaranteed reference price and the market price,
multiplied by production volume. The exact formula varies by program and is set
in each program's authorizing statute or regulation.
Who
qualifies for a subsidy?
Eligibility depends entirely on the specific
program. Farm subsidies generally require the recipient to be "actively
engaged in farming" and to grow an eligible crop, though enforcement of
that requirement has been inconsistent, as this article notes. Energy tax
credits require an eligible purchase (an EV, solar panels, efficiency upgrades)
and, for some programs, fall under an income cap. Corporate
economic-development subsidies are typically negotiated case by case, based on
job-creation or investment commitments rather than a fixed eligibility test.
Because rules differ so widely, the only reliable way to confirm qualification
is to check the specific program's guidelines through the administering agency.
What
are the three most common sources of equity funding?
Outside the subsidy context, "equity
funding" usually refers to how businesses raise capital by selling
ownership stakes rather than borrowing. The three most common sources are:
(1) personal
savings and funding from friends and family, typically used at
the earliest startup stage; (2) angel
investors, individuals who invest their own capital in exchange
for equity, usually at seed stage; and (3) venture capital firms, which invest
pooled institutional funds in exchange for equity, typically at later growth
stages. Private equity and public stock offerings (IPOs) are additional sources
used once a company is more established.
How
do you calculate the level of a subsidy?
The subsidy level is generally the difference
between what a recipient would have paid or received without government support
and what they actually pay or receive with it. For a price-support program,
that's the gap between the guaranteed price and the market price, multiplied by
quantity. For a tax credit, it's the dollar value of the credit actually
claimed. Analysts also express subsidy levels as a percentage of a good's
market price, or as total dollars per recipient, per acre, or per unit
whichever metric best fits the comparison being made.
Does
a subsidy have to be paid back?
No. A subsidy is a grant, payment, discount, or tax reduction, not a loan, so it does not need to be repaid under normal program rules. That's the key distinction between a subsidy and a loan or loan guarantee, which the recipient is expected to repay (sometimes on favorable terms). There are exceptions: some programs include clawback provisions that require repayment if a recipient fails to meet conditions, such as job-creation targets tied to a corporate economic-development deal, or if a recipient is later found ineligible.
Conclusion:
A System That Rewards Readiness, Not Just Need
The
evidence assembled here from USDA payment records, IRS filing data, and state
disclosure databases points to one consistent conclusion: subsidy programs are
not handing money out randomly, and they're not handing it out purely by need
either. They're handing it out, disproportionately and repeatedly, to whoever
is most ready to receive it: the largest farm, the highest earner, the
corporation with a dedicated incentives office, the business with an existing
banker. That is not a partisan claim. It is a pattern visible in data collected
under Republican and Democratic administrations alike, across agricultural,
energy, and emergency-relief policy.
None
of this means subsidies are inherently indefensible, or that every large
recipient is gaming the system. It means the actual order of who gets paid
first deserves the same scrutiny as the total amount spent because a program's
stated purpose and its measured outcome are, in every category examined here,
two different things.
If this kind of evidence-based tracking is useful to you, subscribe to get notified when this analysis is updated with new USDA, IRS, and Good Jobs First data each year, or download the full sourced data table referenced throughout this piece to run your own comparisons. Transparency has already changed subsidy policy once the EV credit income caps prove it. The more people who can see the actual payment order, the more likely that pattern is to keep shifting.
Disclaimer: This article is for informational and educational purposes only and does not constitute financial, legal, or policy advice. Subsidy data changes as new government releases become available; figures reflect the most recent verified data at time of publication and are scheduled for review annually or upon major legislative changes. All statistics are drawn from and attributed to primary sources, including the USDA (via the Environmental Working Group's Farm Subsidy Database), the U.S. Government Accountability Office, the Congressional Research Service, Good Jobs First's Subsidy Tracker, and peer-reviewed academic research. Readers seeking guidance on a specific subsidy application should consult the administering agency directly or a qualified professional.
