The indivisibility of deterrence

The National Review's lead proposition — that a weakened NATO reduces the disincentive for Beijing to test American resolve elsewhere — is, at its heart, an argument about credibility as a public good. Let me translate it into terms I find congenial: the security architecture of the liberal world is not a collection of bilateral bargains that can be repriced one by one without affecting the others. It is a system, and systems have the disagreeable property that when one load-bearing element is quietly removed, the structure does not politely announce its distress until it is already falling.

I spent the better part of 1944 at Bretton Woods arguing precisely this in the monetary sphere. The lesson of the interwar years — the competitive devaluations, the beggar-thy-neighbour tariffs, the collapse of the gold exchange standard — was that no nation could secure its own prosperity by defecting from the common architecture, even when defection looked locally rational in the short run. The same logic governs military alliances. A country that signals it will honour its commitments selectively invites adversaries to search, with some care, for the seam between the commitments it means and the ones it does not.

National Review's argument, as reported, is that the stronger and more united NATO remains, the less incentive Beijing has to test American resolve in the Pacific. I would sharpen that slightly: it is not merely the strength of NATO that matters but the perceived unconditional quality of the guarantee. Animal spirits — to borrow a phrase I coined for investment behaviour but which applies equally well to the calculations of revisionist powers — are governed by confidence and narrative quite as much as by sober balance-of-force accounting. If the narrative shifts to one in which Washington's commitments are contingent on domestic political weather, the deterrent loses potency regardless of the hardware behind it.

I should be honest about the counter-argument. There is a respectable case, which I do not dismiss, that resources are finite and strategic priorities must be ranked; that an overstretched guarantor is worse than a selective one; that Europe, prosperous and populous, ought to shoulder a larger share of its own defence. These are not foolish propositions. But they argue for a rebalancing of burden-sharing within the alliance, not for a signalled indifference to the alliance's survival. The distinction matters enormously. One message reads: we expect more from our partners. The other reads: the guarantee is negotiable. Only one of those messages is heard in Beijing as reassuring.

The economic consequences of political settlements — the phrase is my own, from a different context and a different war — are routinely underestimated by those who negotiate them in the heat of the moment. The architects of Versailles did not intend to fertilise the politics that followed. Those who casually discount the Atlantic alliance today may not intend the strategic vacuum they are creating either. Intentions are beside the point; consequences are not. What I would urge on the present administration — and I acknowledge I am reasoning from principles rather than from any detailed knowledge of 2026's particular diplomatic geometry — is that the cost of appearing to abandon Europe is not a cost that falls only on Europe. It falls, with compound interest, wherever the United States next needs to be believed.