How Subsidies Distort Markets in Ways That Help Incumbents
A subsidy is a rule with a price tag. When Congress writes the rule, it also writes who can collect. That sentence is the whole argument of this article. The rest is documents.
The current U.S. statute book pays firms through tax credits, grants, loan guarantees, and price supports. Each program carries three parts that matter: an eligibility clause, a rate schedule, and a claims process. Read those three together and a pattern appears across programs that share no drafting history. The money tends to flow to firms that were already large. Not because anyone drafted it that way on purpose, in most cases. Because the arithmetic does it on its own.
What a Subsidy Does to a Market
Start with a definition. A subsidy distorts a market whenever it changes what firms must pay to compete. Direct grants lower one firm’s costs relative to another’s. Tax credits do the same, one return at a time. Loan guarantees cut the cost of capital for the borrower the government selects. Price supports hold a floor under revenue for the sellers the program covers. Tariffs raise the price of imports that would otherwise discipline the incumbents’ pricing. We traced who pays that particular bill in our piece on the Section 232 steel tariffs.
Every one of those instruments changes relative prices. This site asks three questions of each: who pays, who chooses, and who writes the rules. For subsidies, the short answers run: taxpayers and customers pay; program administrators choose; and the drafting record shows who shaped the eligibility text.
“Incumbent” here means something narrow. It means a firm already operating at scale in the subsidized activity: the fab operator before the fab subsidies, the grower with base acres before the crop insurance claim, the refiner with pipelines and permits before the carbon-capture credit. A subsidy does not have to name any firm to find them. The eligibility clause does the matching.
Design Choice One: Credits That Scale With Size
Section 45X of the Internal Revenue Code pays manufacturers per unit of output. The statute sets the rates: $35 per kilowatt-hour of battery cell capacity, $10 per kilowatt-hour for battery modules, seven cents per watt for certain solar modules. A pilot line and a gigafactory earn the same rate. They do not earn the same check. The credit scales with volume, and volume scales with balance sheet.
Section 45Q pays $85 per metric ton of carbon dioxide placed in secure geological storage, and $60 per ton when the ton is used for enhanced oil recovery. The rate is flat. The capture, compression, pipeline, and injection infrastructure needed to earn it is not. The firms positioned to collect are the ones that already own the reservoirs, the pipelines, and the permits. Note the second number. The code pays $60 per ton for pushing more oil out of the ground. The incumbent producer becomes the incumbent carbon handler, and the credit helps fund the change of title.
Congress knows how to cap a benefit when it wants to. The section 30D electric vehicle credit carries a price cap on the vehicle and an income cap on the buyer. Those two clauses exclude the top of the market. Section 45X carries no per-claimant cap at all. Neither does 45Q. Compare the drafting and the intent is legible: one program was written to exclude, the others to pay at scale.
Crop insurance works the same way at a different scale. USDA’s Risk Management Agency subsidizes well over half of the premium on average, and the premium scales with the insured value of the crop, which scales with acreage. The commodity programs cap payments at $125,000 per person and cut off operators above $900,000 in income. The crop insurance premium subsidy carries no payment limit at all. GAO has traced where the money lands: the largest operations collect the largest subsidies, because the subsidy is a percentage of a premium that grows with the farm.
Design Choice Two: The Application as Filter
The CHIPS incentive program did not hand out checks. It published a Notice of Funding Opportunity. Applicants filed financial statements, workforce plans, projected cash flows, the incentives offered by other governments, and, for projects seeking more than $150 million, a plan for childcare for facility workers. Each requirement is defensible on its own. Each one costs staff hours. The firms that could staff the process were the firms that already had the departments.
The results are on the record at the CHIPS Program Office. Commerce finalized awards in late 2024 to a handful of names: Intel at roughly $7.9 billion, TSMC at $6.6 billion, and awards to Micron, Samsung, and GlobalFoundries. Five recipients. All incumbents. Two of them foreign-headquartered. The stated purpose was national security. The distribution went to scale. The guardrails rule published in September 2023 (15 C.F.R. part 231) restricts expansion in countries of concern for ten years, adds clawback rights, and requires upside sharing. Those clauses bind the recipients. Nothing in the notice binds the government to a distribution that includes anyone small.
Section 48C, the advanced energy project credit, runs on an application too: a concept paper, then a full application, then an allocation. The Department of Energy screens for technical merit; the IRS allocates the credit. Every stage favors the applicant with engineers, counsel, and a finance team that can model the credit against a corporate tax return. The program does reserve part of its allocation for energy communities. That is a carve-out, and carve-outs are the exception that proves the drafting rule.
Loan guarantees filter hardest of all. Title XVII of the Energy Policy Act of 2005 lets the Department of Energy back loans for innovative energy projects, provided the borrower shows a reasonable prospect of repayment. “Reasonable prospect” translates in practice to investment-grade credit. The Loan Programs Office closed a loan of up to $9.2 billion for Ford’s battery joint venture in 2024. Ford did not need the government to teach it how to borrow. The guarantee cut its cost of capital below what a startup battery maker could ever quote.
Design Choice Three: Monetization Runs Through Someone Else’s Balance Sheet
A tax credit is worth face value only to a firm with a tax bill to offset. For two decades, renewable credits were monetized through tax equity: a small set of large banks invested in projects to absorb the credits, took a spread, and set minimum deal sizes that shut out small developers. The Inflation Reduction Act (Public Law 117-169) attacked this in two sections. Section 6418 made most credits transferable for cash. Section 6417 gave tax-exempt entities an elective payment option. Treasury finalized the regulations in April 2024.
Transferability is a real improvement. It is also a toll road. Sellers of credits trade in the secondary market at a discount to face, and brokers take a fee in the middle. The buyer, often an insurer or a bank, keeps the spread. A firm with its own tax appetite uses the credit at 100 cents. A firm without one sells at a discount. The code calls it transferability. The market calls it a haircut. Either way, the fee falls on whoever is too small to owe enough federal income tax to matter.
Design Choice Four: Grandfather Clauses and Date Lines
The Inflation Reduction Act’s prevailing wage and apprenticeship rules multiply a credit fivefold for compliant projects. Projects that began construction within 60 days of enactment were treated as compliant without showing anything. The clause paid the shovel-ready. Shovel-ready is a synonym for incumbent.
The hydrogen credit, section 45V, drew one of the largest dockets in recent Treasury history. The final rules issued in January 2025 kept the three-pillar framework for emissions accounting and attached transition relief keyed to construction start dates. The pattern repeats across programs. The earlier your project, the softer your obligations. New entrants meet the full test. Firms already standing on the favored side of a date line meet a lighter one.
The oldest grandfather in the book is the sugar program. USDA’s Commodity Credit Corporation lends to sugar processors at a statutory rate, 18.75 cents a pound for raw cane, and takes the sugar as collateral. If prices fall, processors can forfeit the sugar and keep the loan. To protect the loan, the department also manages marketing allotments and tariff-rate quotas that hold U.S. sugar prices well above world prices. The rates sit in the Farm Service Agency’s loan schedules. Consumers pay the gap at the grocery shelf. The benefit concentrates among a small number of growers and processors. The program has run in some form since the 1930s. That is what a locked-in baseline looks like after nine decades.
Who Writes the Rules: Read the Docket
None of these design choices wrote themselves. The 45X regulations proposed in December 2023 had to decide where “production” ends and “assembly” begins, a line that determines which firm collects the per-unit credit. The large manufacturers and their trade associations filed long comments on component definitions. A small firm with three engineers can file a comment. It cannot file a hundred pages of supply-chain modeling. If you have never worked a docket, start with our guide to reading a Federal Register comment file.
The 45V docket drew tens of thousands of comments from oil majors, utilities, hydrogen startups, and environmental groups, each arguing for a different emissions test. The September 2024 proposed rules on foreign entities of concern would restrict credit eligibility for supply chains with certain foreign ties. The stated purpose is national security. The compliance effect is another fixed cost: supply-chain audits, contract restructuring, certification. The firms that can absorb the audit keep the credit. The entrant importing cathode material restructures or exits.
The courts changed this game in 2024. Loper Bright Enterprises v. Raimondo ended judicial deference to agency readings of ambiguous statutes, which means eligibility disputes over words like “production” will now be decided by judges reading the text. Corner Post, Inc. v. Board of Governors held that the six-year clock on challenging a rule runs from when a firm is first injured, not from publication. That reopened the window for late-formed entrants, the one change of the term that runs in the small firm’s favor.
Congress retains the Congressional Review Act, which lets it strike a final rule within roughly 60 session days of its arrival. The tool has been aimed at agency rules of every description in recent years. It has rarely reached the plumbing of credit eligibility. The definitions that allocate the money tend to survive.
What You Can Verify Yourself
Every claim above rests on a public document. The statutory rates sit in the code, at 26 U.S.C. § 45X, with the per-unit figures in the text. The CHIPS award amounts and the award agreements are posted by the program office; the clawback terms sit in the annexes. Read them before you read the press release. Grant awards, unlike tax credits, are published by recipient on USAspending.gov. Tax return confidentiality under section 6103 keeps the credit claims secret, which is why the grant record and the credit record are not symmetric. You can find out who received a CHIPS award. You cannot find out who claimed 45X. That asymmetry is itself a design choice, and it serves whoever benefits from the absence of a ledger.
The Short Version
Scale the benefit with volume, and the largest firms collect the most. Filter with an application, and the staffed firms apply. Monetize through a balance sheet, and the unstaffed pay a discount. Grandfather by date, and the firms already there keep their terms. Then write the definitions in a docket that only the staffed can work. Each choice is defensible in isolation. Stacked, they form a machine with one output. You do not need to prove intent. The eligibility clause does the matching on its own.
Frequently Asked Questions
How do subsidies help incumbents?
Through design. Benefits that scale with volume, applications that cost staff time, monetization that requires tax appetite or a bank, and grandfather clauses that favor firms already operating all sort by size. The eligibility text does the sorting without naming a single firm.
Which current federal subsidies most favor large firms?
Section 45X per-unit manufacturing credits, section 45Q per-ton carbon capture payments, the CHIPS incentive awards, and Title XVII loan guarantees all pay or lend at scale. Each requires either volume, staff, or a balance sheet to collect.
Are there subsidies designed to favor smaller players?
A few. The section 30D vehicle credit caps vehicle price and buyer income, which excludes the top of the market rather than the bottom. Section 48C reserves part of its allocation for energy communities. Section 6417 lets governments and nonprofits elect direct payment. Caps and carve-outs show that Congress knows the difference.
Why can’t I find out who claimed a tax credit?
Because section 6103 of the Internal Revenue Code makes tax returns and return information confidential. Grant recipients are published on USAspending.gov. Credit claimants are not. The transparency gap is structural, not accidental.
What should I read first in a subsidy program?
The eligibility clause, then the rate schedule, then the claims or application procedure. Ask who can satisfy all three at once. That intersection is where the money actually goes.