The Spectrum Dividend

The problem

The electromagnetic spectrum belongs to the public. This is not a slogan; it is the operating premise of American law. The Communications Act of 1934 built the whole regulatory edifice on the principle that no one owns the airwaves — licensees are granted temporary use of a public resource in exchange for serving the public interest. Every carrier that sells you a phone plan is reselling something the nation owns.

Since 1994 the Federal Communications Commission has run about a hundred spectrum auctions and raised more than $233 billion for the Treasury, at an administrative cost of under one percent of receipts. It is one of the most efficient revenue mechanisms the federal government has ever built.

And essentially all of it is gone.

The proceeds flowed into the general fund, where they offset a year's ordinary spending and disappeared. The largest single auction in history — the C-band sale that cleared roughly $81 billion in 2021 — funded about six days of federal outlays. We sold a permanent asset and ate the money. Then, on March 9, 2023, the FCC's general auction authority lapsed for the first time in thirty years and stayed lapsed for more than two years, until a July 2025 statute restored it through September 30, 2034 and ordered a new pipeline of spectrum to market. So the sales are about to resume — the same way, into the same hole.

The proposal

Every dollar of future spectrum proceeds becomes principal in the American Permanent Fund (№ 001), and the licensing model shifts from one-time sales to perpetual rent.

The airwaves are the cleanest case in the entire catalog. There is no ambiguity about ownership, no need to nationalize anything, no new tax, and no new bureaucracy — the auction machinery already exists and already works. The only thing that changes is the destination of the check.

How it would work

1. The Spectrum Dividend Account. All FCC auction receipts are deposited directly into the Fund as principal rather than scored as general revenue. The principal is untouchable; only the real return is ever spent. This requires one change to the budget treatment of auction receipts and a statutory ring-fence against raiding — the provision that matters most, and the one that will be attacked first.

1. Rent, not just sales. Today a carrier pays once and holds a license that renews on strong expectation for decades. Going forward, new licenses carry a recurring annual spectrum fee, calibrated to the value the auction itself revealed, with the upfront price reduced correspondingly so the present-value burden on carriers is unchanged at issuance. The United Kingdom and Australia already levy annual licence fees on spectrum; this is imported practice, not invention. A one-time windfall funds a year. Rent funds a republic.

1. Charge the government too. Federal agencies hold enormous spectrum allocations and pay nothing for them, which means no agency has ever had to weigh what its holdings are worth to anyone else. Assess a shadow fee against federal users — paid out of their own appropriations — so that hoarding shows up on a balance sheet. The Government Accountability Office has pressed for exactly this kind of opportunity-cost accounting for years.

1. Buildout credits, not buildout mandates. Rent is reduced for licensees who meet coverage milestones in rural and tribal areas ahead of schedule. The incentive replaces a rule.

The numbers

The $233 billion already collected is spent and unrecoverable. But it makes the case vividly. Had that same $233 billion been deposited as principal and left alone, at a 5 percent real return it would today pay out roughly $11.6 billion every year, forever — about $35 per citizen per year, from the airwaves alone, without ever touching the corpus.

Thirty-five dollars is not a living. We are not going to pretend it is. That is precisely the point of a catalog of 250 proposals: the American Dividend is not built by one heroic asset, it is built by stacking dozens of them — spectrum here, public data (№ 016) next, sovereign equity and mineral royalties after. A permanent fund is an accretion, not an act. The airwaves are simply the easiest brick to lay first, and the one where public ownership is least arguable.

The forward-looking number is larger. The 2025 statute directs at least 800 megahertz of new spectrum to auction. If the coming decade's sales resemble the last, the Fund would receive well over $100 billion in principal by the mid-2030s — and unlike every previous dollar, it would still be there in 2276.

The honest objections

"That money wasn't wasted — it reduced the deficit." Partly true, and we should say so. But $233 billion spread across thirty years is roughly $8 billion a year against deficits running in the trillions; it did not change the fiscal trajectory in any measurable way. It bought a rounding error once. As principal it would buy a smaller number every year, forever, which is a better trade — and honesty requires adding that making this switch means finding roughly $8 billion a year elsewhere. That is a real cost. We are proposing to give up spendable revenue today to build an asset for people not yet born. That is what the Long Game costs.

"Auction revenue is lumpy, unpredictable, and eventually finite." Correct on all three counts. Spectrum is not infinite, and receipts arrive in irregular multi-billion-dollar lurches. But finite windfalls converted into permanent income is the exact thing sovereign funds were invented to do — Alaska did it with oil that is now declining, and the dividend outlived the boom. Lumpy deposits are fine; it is lumpy spending that wrecks budgets, and a fund that spends only its real return smooths the lumps by construction.

"Recurring fees will raise consumer prices or slow 5G buildout." The risk is real and the design has to answer it, which is why the upfront price falls as the rent rises — the intent is present-value neutrality at issuance, not a new levy. But we will concede the sharper version of this objection: a future Congress facing a deficit will be tempted to ratchet the rent, and carriers will price that risk in. The mitigation is to bind the fee to auction-revealed value by formula rather than leaving it to annual discretion. The mitigation is imperfect.

"Charging federal agencies is a shell game — one government pocket paying another." In cash terms, yes. The value is informational. Nobody can currently say what the Department of Defense's spectrum holdings are worth in alternative use, which makes every reallocation fight a contest of assertions. A price, even an internal one, produces an answer. Pair it with actual reallocation authority or it is theater.

"Why should the Fund get this rather than, say, rural broadband?" Because rural broadband is a program and the Fund is an institution, and programs are how public assets get quietly converted into current consumption with a sympathetic name attached. Buildout belongs in the licence conditions, where it is already, and where it costs the corpus nothing.

Sources

• Federal Communications Commission, Spectrum Auctions Program — 100+ auctions, $233B+ raised since 1994 (fcc.gov)
• Communications Act of 1934, public-interest licensing standard; FCC Auction 107 (C-band) results
• Congressional Research Service, History of the FCC's Spectrum Auction Authority: 1993–2025, R48861 (congress.gov)
• One Big Beautiful Bill Act (2025), restoring general auction authority through September 30, 2034 and establishing a spectrum pipeline
• Government Accountability Office, federal spectrum management and opportunity-cost reporting (gao.gov)
• Ofcom (UK) annual licence fees for mobile spectrum (ofcom.org.uk)