Anchored, not approved

On 1 September Norges Bank wrote to the Norwegian Ministry of Finance about concentration in the fund's equity index. Near the end, it raised something else. There may, the letter said, be reason to consider whether a larger allocation to unlisted assets would spread the fund's risk more widely. Any such move "would require thorough review of the benefits and drawbacks, including political and regulatory risk, and broad anchoring".

That is not a proposal. It is the sound of an institution putting a question back on the table after being told no twice.

The Government Pension Fund Global is this book's working example of a citizen stake: funded from an asset rather than a tax, ring-fenced, managed at arm's length. Chapter 8 credits the wall between the manager and the treasury. A rule is only a precommitment if the body that administers it cannot be leaned on by the body that would like to break it.

The wall is real. Chapter 8 also names who built it: the fund sits inside the central bank under a published mandate from the Ministry of Finance. What the chapter never draws out is that the wall runs both ways.

The Act places responsibility for the fund with the Ministry of Finance, which delegates day-to-day management to Norges Bank, and the rules that implement it are regulations rather than statute. Capital leaves the fund only by a resolution of the Storting. Parliament legislated the container and kept the tap that drains it. It delegated what goes inside.

What fills that gap is convention. By long practice the Ministry does not change the fund's strategy without first putting it to parliament in the annual white paper. Norges Bank's own word for this is anchoring. Anchored is a word about legitimacy, not authority.

The unlisted-equity question shows the arrangement in practice. In January 2018 Norges Bank asked to be allowed into unlisted equity. In April the Ministry said no, on reputational grounds: the finance minister said the fund's image mattered, particularly on transparency. In November 2023 the Bank asked again, this time for 3 to 5 per cent of the fund. In April 2024 the Ministry said no again, citing fees and the difficulty of measuring risk.

Both refusals reached parliament, and neither reached it as a government proposal. The Ministry set out its negative recommendation in the annual white paper. The finance committee considered it and the recommendation stood. Members were free to move something else and did not.

That is the mechanism worth naming. It is not concealment. A recommendation from the body responsible for the fund is a heavy default, and overturning it takes a majority that wants to. On this question there has not been one.

The same asymmetry shapes what the fund owns. The Ministry sets the benchmark index, and the fund has little scope to depart from it. When a company enters that index the fund buys it. Ownership arrives by construction rather than by choice, a strange thing to say about the most closely watched pool of citizen money in the world.

None of this makes the Ministry wrong. Norges Bank's own researchers found in 2023 that buyout funds have beaten listed equities by three to four percentage points a year while venture capital has trailed by one to two, a finding that cuts both ways. A ministry that refuses an expensive, opaque asset class on behalf of a fund built on transparency is arguably doing its job.

The point is not the answer. It is who gives it, and on what authority.

Chapter 8 says a fund that can update its own rule survives the changes the world throws at it. The possessive was doing more work than it could carry. This fund does not update its own rule. The Ministry rewrites the investment mandate by regulation, and the spending guideline moves by convention, as the chapter's own footnote on the 2017 cut records. The chapter offers no way to tell a disciplined refusal from a sclerotic one, because it treats the wall only as a shield against raiding and never as a filter on what the manager may do.

Norway never wrote its spending rule into law either, and has held it for a quarter of a century, which the book offers as proof that consensus can do a statute's work. It can. A convention strong enough to restrain a government is also strong enough to settle a question before the owners reach it.

That gap matters beyond Norway. Every argument for a citizen stake is an argument for an institution that will outlive the government that built it, insulated from the people who might raid it. The same wall also decides who may act. Norway shows both faces, twenty-two trillion kroner of the first and eight years of the second.

The fund is not less impressive for this. It is more precisely understood. Norwegians own it completely and at one remove: nothing in current practice takes the capital from them, and the decisions are not theirs to make.

Which is a narrower claim than the one this book has been making, and a better one. Ownership without direction is still ownership. It is just not control, and a design that confuses the two will build the wall and forget the door. The next question is who decides what the machine buys. Who owns the machine?