The Sovereign Equity Act

The problem

Between 2010 and 2019 the Food and Drug Administration approved 356 new drugs. Research funded by the National Institutes of Health contributed to 354 of them — to the drug itself or to the biological target it acts on. The NIH investment behind those approvals came to more than $187 billion, across roughly 360,000 project-years of grant funding, producing over 22,000 patents. Industry then spent on the order of $1 billion to $1.4 billion per drug taking them through development and trials.

Both numbers are real. Both parties did indispensable work. Only one of them holds an asset at the end.

The American taxpayer is, in practice, the world's largest and most patient early-stage investor. Public money funds the decade of basic research before anyone knows whether a target exists, absorbs the failures nobody writes press releases about, and hands the surviving science to private developers who — correctly, and by design — capture the return. Under the Bayh-Dole Act of 1980, contractors and universities keep title to inventions made with federal funds. That statute worked. Before it, federally funded inventions sat unlicensed in a warehouse of dead patents, and commercialization rates were dismal. We are not proposing to repeal it and we think proposals to do so are mistaken.

But Bayh-Dole answered the question of who owns the invention and never asked the question of what the public gets for its capital. The public gets the drug, at whatever price the market sets, plus the tax revenue any profitable company generates. It gets no position. It carried the earliest and most uncertain risk in the entire chain and took the one form of compensation an investor never accepts: none.

The same pattern runs through the Defense Advanced Research Projects Agency and the internet, GPS, and touchscreens; through Operation Warp Speed's advance purchases; through the Department of Energy's loan office, which lent $34.2 billion, was repaid by Tesla nine years early, and is remembered entirely for Solyndra.

The proposal

When the federal government funds an invention or creates a market for it, it takes a small, non-voting, permanent financial position — and that position belongs to the American Permanent Fund (№ 001), not to the agency and not to the general fund.

Not price controls. Not march-in rights. Not government ownership of industry. A term sheet.

How it would work

1. A sovereign equity clause. Federal research awards above a threshold carry, as a standard condition, either a small non-voting equity stake in the commercializing entity or a low royalty on net sales, at the recipient's election. The terms are published in advance and identical for everyone — no negotiation, no discretion, no agency picking who pays what.

1. The royalty falls on the licensee, not the laboratory. Universities and nonprofit institutes elect the royalty and pass it through to the commercial licensee that actually books the sales. A university's balance sheet is never touched. This is the design detail that determines whether the proposal is workable or destructive, and it is not negotiable.

1. Procurement warrants. When the government is the first buyer at scale and its purchase is what creates the market — advance purchase commitments, first-of-a-kind energy plants, defense platforms with commercial spinouts — the contract carries warrants, priced at the time of award.

1. Silent capital, by statute. The Fund's holdings are non-voting, capped at a low single-digit percentage of any firm, and managed at arm's length by independent managers barred from taking board seats or directing operations. Norway's fund holds roughly one and a half percent of all listed equities on earth without running a single company; that is the model, and the prohibition on control should be written where a future administration cannot reach it.

1. Bayh-Dole stands. Title stays with the inventor. Nothing here slows a licence, delays an approval, or gives a bureaucrat a veto over a product.

The numbers

Be honest about the scale: this is a slow instrument, not a windfall.

A 0.5 percent royalty on the U.S. net sales of drugs whose core patents trace to NIH-funded research would, on a conservative base of $150 billion in relevant annual sales, yield about $750 million a year. Extended across federal research funding generally, a mature sovereign equity portfolio plausibly contributes low single-digit billions annually — meaningful as one of the fifty streams that build a corpus, trivial as a solution on its own.

The stronger number is the one already spent. $187 billion, from NIH alone, in one decade, in one sector, for which the public holds no position whatsoever. The Sovereign Equity Act does not recover a dollar of it. It stops the next $187 billion from going the same way.

And the honest counter-number, which we will publish because omitting it would be the exact dishonesty this catalog exists to avoid: the Department of Energy's loan portfolio, the closest thing to a live experiment in federal risk capital, has collected roughly $810 million in interest against about $780 million in defaults — a 2.28 percent loss rate that compares well to private venture lending. But once federal credit-subsidy accounting is applied, the same portfolio has been projected to lose over $2 billion. Government investing is not a free lunch, the accounting is contested, and anyone who tells you the taxpayer would have gotten rich is selling the same optimism we are criticizing.

The honest objections

"This is a tax on innovation, and it will reduce research investment at the margin." Partly true and worth conceding precisely. Any condition attached to a grant lowers its value to the recipient; some projects will decline federal funding and some firms will locate research elsewhere. The mitigations are that the terms are small, uniform, published in advance, and voluntary in the only sense that matters — nobody must take the money. An investor's term sheet is not a tax, and treating any public claim on public capital as confiscation is how the current asymmetry got normalized.

"Government picking winners is how you get Solyndra." Which is why the Fund's position is passive, diversified, and non-directive — it does not pick anything. The stake attaches automatically to whatever the research agencies were already funding on scientific merit. If those funding decisions are bad, the problem is the funding decisions, and this proposal neither improves nor worsens them. We would also note, with the DOE figures above, that the Solyndra story is told with one number and refuted with two.

"The state should not own shares in private companies. This is creeping nationalization." The strongest version of this objection is not about ownership but about temptation: a government holding equity acquires an interest in the profitability of firms it also regulates. That conflict is real and we do not think it disappears with a non-voting clause. The answer has to be structural — the Fund's managers independent and statutorily walled from regulatory agencies, holdings disclosed in full, and a hard cap per firm. Sovereign wealth funds have run this way for decades in countries with functioning rule of law. It is a risk to be managed, not a fantasy to be waved away.

"March-in rights already give the government leverage. Use those." They exist and they have never once been exercised in forty-five years, through every drug-pricing controversy of that period. A right never used is not a right; it is a rhetorical position. It is also the wrong tool — marching in attacks the price of a specific product after the fact, which is disruptive, litigable, and unpredictable. Equity is quiet, automatic, and priced up front. We would rather own a share than threaten a seizure.

"Firms will simply route around it — take private funding for the last mile and claim the federal contribution was upstream." They will try, and attribution at the boundary between basic and applied research is genuinely hard. The threshold and the published formula limit the gaming, and the patent record makes the tracing possible, as the Cleary study demonstrated at scale. But we expect leakage and we would rather collect a partial royalty honestly than design a perfect one that never passes.

Sources

• Cleary, Jackson & Ledley, "NIH funding for patents that contribute to market exclusivity of drugs approved 2010–2019," PLOS One (2023) — 354 of 356 drugs, $187B, 22,360 patents (journals.plos.org)
• Cleary et al., "Contribution of NIH funding to new drug approvals 2010–2016," PNAS (2018) (pnas.org)
• Bayh-Dole Act, 35 U.S.C. §§ 200–212, including march-in rights at § 203
• U.S. Department of Energy, Loan Programs Office portfolio data; Government Accountability Office and press reporting on interest collected, default rates, and credit-subsidy accounting (energy.gov)
• Norges Bank Investment Management, ownership share of global listed equities (nbim.no)
• Proposal № 001 (The American Dividend); № 002 (The Founders' Pledge Act)