The Spending Rule

The problem

Every endowment dies the same way. Not by theft, usually, and not by a market crash, which it is built to survive. It dies by a series of reasonable decisions to pay out a little more than it earned, each one justified by a good cause and a good year, until the principal is quietly consumed. The people who make the decisions are rarely dishonest. They are simply closer to the good cause than to the grandchildren the fund was for.

The defence is a rule, set in advance, that says how much may leave the fund in a year, and it has to be arithmetic rather than judgment, because judgment is what the good cause will appeal to. Norway's rule, adopted in 2001, permits the government to spend the fund's expected real return, which it set at 4 percent of the fund's value. In 2017, when expected returns fell, Norway cut the rule to 3 percent, and in most years it has spent less. The rule has absorbed a financial crisis, a pandemic and an oil collapse, and the principal has never been touched.

Alaska's rule is looser and its politics show it. Since 2018 the state draws 5 percent of the fund's average value over five years, and then fights, every session, over how much of that draw goes to the dividend and how much to the budget. The 2025 dividend was $1,000, the smallest in real terms in the fund's history, not because the fund did badly but because the legislature needed the rest. The rule was written to end the argument. It only moved it.

The lesson from both is the same: the rule must set the total, it must be lower than the fund's average return so that principal grows in real terms, and it must not leave a residual for anyone to argue over. № 017 fixed the date of the dividend. This proposal fixes the arithmetic.

The proposal

The Spending Rule, in statute: each year the American Permanent Fund pays out 3 percent of its average market value over the previous five years, and nothing else. The payout goes where № 001 and № 023 direct it: to the debt while debt exceeds 90 percent of GDP, and to the citizens' dividend after. The amount may fall. The date, Dividend Day, may not move. Changing the percentage requires a law that says, in its title, that it is changing the Spending Rule.

How it would work

1. Three percent, of a five-year average. The rate is Norway's current rate, set below the long-run real return of a global index portfolio, which has been about 4 to 5 percent a year over the last century, so that the fund grows faster than inflation on average and the dividend grows with it. The five-year average is Alaska's device, and it is a good one: a crash reduces the payout gradually rather than at once, and a boom does not produce a windfall that the next year cannot match.

1. Nothing else. No supplemental draw, no special appropriation, no "one-time" transfer. The fund's only outflow in a year is the 3 percent, and it is paid on the date № 017 sets. A Congress that wants more must change the rule by name, in a law whose title says so, which is the most a statute can do to make an act visible.

1. The amount falls, the date holds. In a year after a crash, 3 percent of a lower average is a smaller cheque. It is paid anyway, on the day, and every citizen learns that the fund had a bad year in the plainest way there is. Alaska's forty years show that a dividend that shrinks in a bad year and arrives on the day is more durable than a larger one whose date is negotiable.

1. The lock comes first. While the national debt exceeds 90 percent of GDP, the Debt Covenant of № 023 directs the payout to the debt. The Spending Rule sets the size of the payout; № 023 sets where it goes. A citizen's dividend of zero under the covenant is still a payout of 3 percent, and the rule keeps the fund from being drawn faster in the years when the money is going to creditors and nobody is watching.

1. The split, downstream. The rule sets the total. What share of the total goes to the equal dividend, to Baby Shares under № 008, or to the Service Dividend of № 005 is set by those proposals and may be argued about, because it is a question of who among the citizens receives it, not of how much leaves the fund.

1. Published in advance. The trustees of № 066 publish, each January, the five-year average, the 3 percent, and the resulting per-citizen amount, so that the number on Dividend Day is known nine months before it arrives and no one can claim to have been surprised.

The numbers

At a trillion dollars. A fund of $1 trillion pays out $30 billion a year, about $90 per citizen, or, while the Debt Covenant binds, $30 billion off the debt. It is a small number and this proposal says so; the fund has to be much larger than a trillion before the dividend is a sum a household plans around, and № 001's path to that size is measured in decades.

At Norway's scale. Norway's fund is worth about $360,000 per Norwegian. The equivalent for the United States would be about $120 trillion, which no one proposes. At a tenth of that, $12 trillion, the 3 percent rule pays about $1,000 per citizen a year, in perpetuity, growing with the fund. That is the destination, and the rule is what makes it reachable, because a fund drawn at 5 percent in good years never gets there.

What the difference between 3 and 5 percent is worth. A fund earning 5 percent real and paying 3 grows about 2 percent a year in real terms; paying 5, it grows not at all. Over fifty years, the first fund is about 2.7 times the size of the second. The extra 2 percent paid out each year for fifty years buys a dividend that is, at the end, less than half of what the patient rule pays.

The honest objections

"Three percent is arbitrary, and the right number depends on returns nobody can forecast." It is a judgment, and it is Norway's after twenty years of running the experiment, set below the long-run return on purpose so that the error, if there is one, is on the side of the grandchildren. The rule can be changed, by a law that says so in its title; what it cannot be is changed quietly. If returns settle permanently lower, the change will be argued in public with the fund's own numbers on the table.

"A five-year average means the dividend keeps falling for years after a crash, when people most need it." True, and the honest alternative is worse: a payout tied to the current value swings violently, and a payout that rises in a crash to meet need is a payout drawn from principal, which is how every fund in history has been emptied. The Fund is not the country's safety net. № 004 and № 026 are. The Fund is the thing that is still there after the safety net has been used.

"Congress will change the rule in the first recession." It may, and the rule cannot stop it. What the rule does is make the change a named act, against the trustees' published objection, on the public record, with the arithmetic of what it costs the fund's future shown on the Shareholder's Report. Norway's rule has survived because every Norwegian knows what it is. That is the model.

"While the Debt Covenant binds, the dividend is zero, so why does a spending rule matter?" Because the years when the payout goes to creditors rather than citizens are exactly the years when nobody is watching the draw, and a fund can be quietly overdrawn to pay debt as easily as to pay anything else. The rule sets the ceiling whoever is receiving the money.

"Why pay anything out at all until the fund is enormous?" Because a fund that pays nothing for forty years is a fund nobody defends, and № 017 exists on the theory that the ritual of the payment is what makes the fund impossible to dismantle. The rule keeps the ritual small in the early decades and honest forever.

Sources

• Norwegian Ministry of Finance, The fiscal rule (handlingsregelen), adoption in 2001 and the 2017 revision to 3 percent, regjeringen.no; White Paper on the Government Pension Fund 2025
• Alaska Statutes 37.13.140 and 37.13.145, percent-of-market-value draw enacted by Senate Bill 26 (2018); Alaska Permanent Fund Corporation, dividend history and 2025 amount, apfc.org
• Dimson, Marsh and Staunton, Credit Suisse Global Investment Returns Yearbook, long-run real returns on global equities and bonds
• Proposals № 001 (The American Dividend); № 005 (The Service Dividend); № 008 (Baby Shares); № 017 (Dividend Day); № 023 (The Debt Covenant); № 065 (The Index Rule); № 066 (The Trustees)