The Debt Covenant

The problem

Twenty-two proposals into this catalog, the obvious objection has been waiting, and it deserves its own entry rather than a paragraph at the bottom of someone else's.

How can a country borrowing nearly $2 trillion a year possibly justify building an endowment?

The scale is not in dispute. In fiscal year 2025 the federal deficit came to about $1.8 trillion. Debt held by the public rose by $2.0 trillion to $30.3 trillion, roughly 98 percent of GDP. Net interest reached $970 billion — an increase of $89 billion, or 10 percent, in a single year — which made interest the third-largest item in the federal budget, behind only Social Security and Medicare, and larger than national defense. The Congressional Budget Office projects interest costs will exceed defense outlays every year from 2025 through 2035, rising from about $1.0 trillion in FY2026 to $1.8 trillion by FY2035.

Against that, a permanent fund of even $1 trillion, paying out 5 percent real, produces $50 billion a year. It is a rounding error on the interest line. Anyone who tells you a sovereign wealth fund will fix the American fiscal position is not doing arithmetic.

So let us say the uncomfortable thing plainly: the American Permanent Fund is not a debt-reduction strategy, and this organization has never claimed it was. It answers a different failure. The United States government publishes, in exhaustive detail, one side of a balance sheet. It has liabilities measured to the dollar and updated daily. It has no corresponding institution on the asset side — no mechanism that accumulates, no rule that says a public asset sold is a public asset replaced. Spectrum was sold and spent. Federal research produced trillions in value and no position. Every windfall of the last fifty years arrived as revenue and departed as outlay.

A country can run a deficit and still be building. A country can run a surplus and still be liquidating. What determines which is not the deficit number. It is whether anything is being accumulated at all.

The proposal

Bind the Fund with two locks, written into its charter, so that it can never become a reason to borrow and never become an excuse to stop repaying.

Lock One — No borrowed principal. Not one dollar may enter the Fund from borrowed money. Contributions come only from asset conversions and dedicated receipts: spectrum proceeds (№ 015), the data royalty (№ 016), sovereign equity and procurement warrants (№ 020), resource royalties, and lapsed war outlays (№ 019). Congress may not appropriate deficit-financed money into the corpus, in any year, for any reason. Borrowing at 4 percent to invest at a hoped-for 6 is leveraged speculation with the public's credit, and no honest institution should promise it.

Lock Two — The dividend rises only as the debt falls. While debt held by the public exceeds 90 percent of GDP, the entire real return of the Fund is applied to debt reduction and the dividend is zero. Between 90 and 60 percent, the return is split on a published sliding schedule — more to citizens each year the ratio improves. Below 60 percent, the full real return is paid out as the dividend on Dividend Day (№ 017).

How it would work

Lock Two is the mechanism that makes the rest of this catalog fiscally coherent, and it is worth being explicit about what it does.

Every fiscal rule the developed world has tried has been suspended — Gramm-Rudman-Hollings, the European Stability and Growth Pact, statutory debt limits, sequestration. They fail for the same reason: the pain of the rule is concentrated and immediate, the benefit is diffuse and distant, and no constituency defends an abstraction. Debt reduction has never had a lobby.

Lock Two gives it one. It converts "reduce the debt-to-GDP ratio" from a technocratic preference into a line on every household's October payment. When the ratio falls, the check arrives and grows. When Congress suspends the rule, every citizen watches a payment they were promised get smaller, on a known date, with a published formula explaining exactly why. It puts three hundred million people on the side of fiscal discipline by paying them for it.

This is the same insight as Dividend Day (№ 017), applied in the opposite direction: visible, universal, calendared money is the most durable political force in a democracy. We propose to point it at the debt.

Alongside it: an audited national balance sheet, assets and liabilities together, published annually in plain language on Dividend Day, next to the Peace Ledger (№ 010). A country that cannot see its own assets will keep selling them.

The numbers

At today's ratio of roughly 98 percent, Lock Two means the American dividend is zero for a considerable time. We want that stated on the record, in the catalog, in advance — because the alternative is to promise a check in year one and quietly discover the constraint later.

Under plausible assumptions, the contributing streams in this catalog might build a corpus in the low hundreds of billions across a decade and a half. Its real return would go entirely to debt service. The first actual dividend under this rule is a 2040s event at the earliest, and only if the ratio falls, which requires primary-balance improvements this proposal does not itself deliver.

What the corpus does in the meantime is not nothing. A $200 billion fund returning 5 percent real retires $10 billion of debt a year without a tax increase or a spending cut, permanently, and compounds against a liability that is otherwise compounding alone. More importantly, it establishes the accounting identity that America currently lacks: a public asset sold is a public asset replaced.

The honest summary: this proposal makes the American Dividend smaller and slower and later than the version anyone would prefer to campaign on. We think a promise that survives contact with a bond market is worth more than one that doesn't.

The honest objections

"A household with credit-card debt should not open a brokerage account. Pay the debt first." The best version of the objection, and Lock One concedes it entirely — that is precisely why borrowed money is barred from the corpus. Where we part company is on asset conversions. Selling the airwaves to fund one year of operating expenses while carrying $30 trillion in debt is not the prudent household paying down its card; it is the household selling the car to make the minimum payment and calling it discipline. The choice is not between an endowment and repayment. It is between converting assets into consumption or into assets — and Lock Two ensures the assets service the debt.

"Debt-to-GDP thresholds are arbitrary, and the 90 percent figure is discredited." Correct, and we should be the ones to say so. The Reinhart-Rogoff finding of a growth cliff above 90 percent did not survive replication — there was a spreadsheet error and coding choices that drove the result, and the profession's considered view is that no sharp threshold exists. We use 90 and 60 anyway, and we are not claiming they are empirical constants. They are commitment devices: round, memorable, defensible numbers around which a durable political rule can be built, with a sliding schedule between them precisely because there is no cliff. A rule needs a number. This one is honest about where its number comes from.

"Fiscal rules never bind. This one will be suspended in the first recession, like all the others." Probably it will be suspended at some point, and it should include an explicit supermajority escape for genuine emergencies rather than pretending otherwise — unsuspendable rules get evaded instead of amended, which is worse. The claim is narrower: this rule has an enforcement mechanism the others lacked, because suspending it cuts a visible universal payment rather than triggering an abstraction. That is a stronger defense than any previous rule enjoyed. It is not an impregnable one.

"You have just admitted the dividend pays nothing for twenty years. Why would anyone support this?" Because the alternative on offer is not a dividend in five years; it is no dividend ever, and no asset column ever, and another fifty years of selling public property to fund current spending. Alaska's fund was established in 1976 and paid its first dividend in 1982 — six years — but Alaska started with an oil boom and a small population. We are starting with a $30 trillion liability. Institutions that outlive their founders are built on this timescale. We named the pillar The Long Game on purpose.

"Interest costs are driven by interest rates, not by an absence of assets. This proposal addresses the wrong variable." Largely true in the short run — the FY2025 jump owes much to rates repricing existing debt. But rates are not a policy instrument available to Congress, and the stock of debt is. This proposal does not claim to lower rates. It claims that the asset side of the national balance sheet should exist, and that its earnings should be pointed at the liability until the liability is manageable.

Sources

• U.S. Treasury and Congressional Budget Office, FY2025 results — deficit of approximately $1.8 trillion; debt held by the public $30.3 trillion, about 98% of GDP (cbo.gov)
• Committee for a Responsible Federal Budget, "Interest Costs Just Surpassed Defense and Medicare" — net interest of $970 billion in FY2025, up from $881 billion (crfb.org)
• Congressional Budget Office, The Budget and Economic Outlook: 2025 to 2035 — interest exceeding defense outlays 2025–2035, rising to $1.8 trillion by FY2035 (cbo.gov)
• Peter G. Peterson Foundation, interest costs on the national debt (pgpf.org)
• Herndon, Ash & Pollin, "Does High Public Debt Consistently Stifle Economic Growth? A Critique of Reinhart and Rogoff," Cambridge Journal of Economics (2014)
• Proposals № 001 (The American Dividend), № 010 (The Peace Ledger), № 015, № 016, № 017, № 019, № 020