Where the Profit Goes

A multinational's profit has no natural location. This is not an evasion of the question; it is the question. Profit arises from a global operation (research in one country, manufacturing in a second, a brand recognised everywhere, customers in a hundred places) and assigning a share of it to each jurisdiction requires a rule, not a measurement.

The rule in use for nearly a century is the arm's length principle. Transactions between subsidiaries of the same group should be priced as though the parties were strangers. For a shipment of steel this is checkable: there is a market price, and a subsidiary charging ten times it is visibly doing something else. For the right to use a brand owned by a group company in Bermuda, there is no comparable transaction anywhere in the world, because nobody else owns that brand. The price is whatever a valuation report can be made to support.

Modern profit is concentrated precisely in those assets. Patents, software, brands, processes: the things that generate returns in a rich economy are the things with no market price to check against. The principle fails exactly where the money is.

The mechanism is simple once the licence is priced. An operating subsidiary in a high-tax country pays a large royalty to a holding company in a low-tax one. Group profit is unchanged; the country that gets to tax it has changed. Nothing moved. No factory relocated, no employee emigrated. A payment was documented.

Gabriel Zucman and colleagues estimate that something approaching forty per cent of multinational profit is booked in jurisdictions where almost none of the underlying activity occurs. Several small countries now host more reported profit than the size of their own economies, to the point where their statistical offices publish alternative measures of national income because the official ones have become uninformative.

The essential point, and the one most often lost in coverage, is that nearly all of this is legal. It is avoidance, not evasion: it uses the rules as written rather than breaking them. This explains why two decades of public outrage, boycotts and select committee hearings achieved so much less than one technical change did. Moral pressure on a company operating lawfully has limited purchase, and the responsibility for the rules sits with the governments that wrote and agreed them, which is a less satisfying target than a named firm.

The technical change came in 2021. More than 130 countries agreed a fifteen per cent minimum effective rate, enforced through an elegant device: if a subsidiary is taxed below the floor, the parent company's home country collects the difference. Booking profit in a zero-tax jurisdiction then buys nothing, because the shortfall is collected anyway. Crucially, the low-tax country does not need to cooperate or consent. The incentive is removed from the other end.

Implementation has been uneven and the rate is low enough that some structures remain worthwhile. But the logic is right, and it is the first change in a century that addresses the problem where it actually lives: in the rules for assigning profit, rather than in the character of the people using them.