Debt-to-GDP is the wrong number. What decides whether a nation survives a shock is the shape of its borrowing, not its size — as a war that should have broken Israel’s finances quietly proved.

There is a number with totemic power in 2026, and it is the wrong number. As Britain’s thirty-year gilt yield punched above 5.8 per cent this spring — a level last seen in 1998 — and Japan’s long bond hit an all-time high, the commentary reached reflexively for debt-to-GDP. The United States is past $39 trillion and 125 per cent of output, its net interest bill set to overtake the defence budget. The OECD reports a record $29 trillion of bond issuance this year, four-fifths of sovereign borrowing merely refinancing what exists — and treasuries are quietly shortening maturities to dodge a punishing long end, buying a cheaper coupon at the price of heavier rollover risk.

These figures are sobering, but they answer a question no creditor asks at the moment of truth. The level fixation rests on weaker ground than its confidence suggests: the famous Reinhart-Rogoff threshold, beyond which debt supposedly throttles growth, did not survive scrutiny once researchers found the spreadsheet error and the selective exclusions beneath it. Sovereign crises do not arrive because a ratio crosses a line. They arrive when a government can no longer place its paper at home — a function not of how much it owes but of to whom, in whose currency, and for how long. The level is the headline. The structure is the story.

Borrow the corporate frame. A company’s capital structure is its blend of debt and equity, and Modigliani and Miller taught that in a frictionless world the blend is irrelevant to value. The nearest sovereign analogue is Ricardian equivalence — an imperfect cousin, since Ricardo’s proposition concerns the timing of taxation rather than the mix of claims, but it carries the same moral: tax now, or borrow and tax later, and a rational public sees the same lifetime burden. In neither case does the financing mix matter — in theory. The interest lies in how violently the theory fails, and it fails far harder for a state.

Why? Because the frictions that break Modigliani-Miller are, for a nation, close to absolute. A firm that cannot pay is wound up; a liquidator arrives, assets are sold, claims are ranked, the residual belongs to shareholders. A nation has no liquidator. No court can seize a country, no estate is assembled, no residual claimant owns what is left. Willingness to pay diverges from ability, because sovereignty is the option to refuse. And there is no clean equity holder whose value we maximise: a state answers to heterogeneous citizens across overlapping generations whose interests openly conflict. We cannot even agree on the objective function, let alone optimise it.

So is there an optimal national capital structure? Not as a ratio. But there is a principle: a sovereign’s instruments lie on a spectrum from debt-like to equity-like, and what makes an instrument equity-like is that it pays less precisely when the country is doing badly. That property — state-contingency — is the hidden equity a nation forgets it has.

Walk the spectrum and the history writes itself. Original sin — Eichengreen and Hausmann’s term for the inability of emerging economies to borrow in their own money — is the purest debt-like trap, converting every depreciation into a balance-sheet detonation. Ask a shopkeeper in Istanbul who watched the lira shed roughly half its value against the dollar in a single brutal year what that feels like from below: dollar-priced debts doubled overnight, imported inflation gnawing at every wage, savings evaporating while the state burned reserves defending a line it could not hold. Argentina and Turkey have relived the sequence in cycle after cycle. Maturity compounds it: Mexico’s 1994 tesobonos, short-dated and dollar-indexed, looked cheap until the peso moved and the whole stock had to be rolled at once into a market that had vanished — the self-fulfilling crisis Cole and Kehoe later formalised. At the other end sit the equity instruments — Argentina’s 2005 GDP-linked warrants, which paid more when the economy outperformed, and foreign direct investment, the most equity-like of all, its returns falling with the host’s fortunes and no one demanding principal back at par in a panic. Ownership itself is structural: Japan carries a ratio that would put any emerging market on intensive care, yet finances it overwhelmingly at home — a captive domestic base that lets its bond market wobble rather than break even as yields reach multi-decade highs.

The asset side matters too. Precautionary reserves are self-insurance — the lesson the emerging world drew, expensively, from the Asian crisis, though Dani Rodrik measured their carrying cost — and sovereign wealth funds go further, as Norway’s oil fund turns an exhaustible endowment into permanent state equity. Neither lowers the debt number; both raise the country’s distance from distress, which is the variable that matters.

This is why the “BIFs” — Britain, Italy and France, the market’s newly minted fiscal misbehavers, cast as heirs of the old PIIGS — are misdiagnosed when scolded merely for borrowing too much. Their vulnerability is rigidity, not size: hard nominal claims, a lengthening share of foreign ownership, thinning domestic demand as central banks retreat and price-sensitive funds step into the gap. They are levered without enough equity in the structure, and the spread they pay over Germany prices that brittleness, not the ratio alone. Two compound it by treaty: Italy and France borrow in a currency they cannot print, having surrendered the depreciation valve to Frankfurt. It is original sin chosen rather than inflicted — structure that flatters in calm weather and punishes in a squeeze.

For the mirror image, look to Israel. Through the multi-front war that began in October 2023, its level deteriorated on every metric the ratio-watchers prize: public debt climbed from about 61 per cent of output at the end of 2023 to roughly 69 per cent a year later, the 2024 deficit blew out to 6.8 per cent, war costs ran toward sixty-odd billion dollars, and all three major agencies downgraded the sovereign — Moody’s twice, once in a two-notch cut to Baa1. By the headline number, a country in trouble. Yet Israel never lost the ability to finance, because its structure was built to bend. It entered the shock with a low debt ratio — fiscal space is optionality banked in advance — and it borrows overwhelmingly at home and in its own currency: of the roughly 360 billion shekels raised after October 2023, about three-quarters came from the domestic market, where deep pools of pension and provident money behave like patient, captive equity. When it chose to tap foreigners, it still could, placing a record $8 billion dollar issue in March 2024. The level screamed; the structure absorbed.

None of this is to pretend the level-watchers are fools. Composition can betray you quickly: a captive base can lose its appetite, a comfortable maturity profile shortens with every auction, and the equity-like instruments remain thin, dear and easily gamed. Markets do punish ratios, sometimes before the structure has shifted at all. But that is an argument for watching structure more closely, not less. The level is a lagging summary of decisions already taken; the shape is the live variable that decides whether the next shock is absorbed or amplified. By the time the ratio screams, the composition has usually been whispering for years.

The honest reframing is one of contingent claims. Treat a sovereign’s equity as a call option on national assets and its default as the embedded put — the credit-risk logic Robert Merton built for firms, later stretched over whole states — and “optimal capital structure” stops being a target ratio and becomes a posture: maximise distance-to-distress, buy state-contingency where it is cheap, lengthen maturity and deepen the domestic base where rollover risk concentrates. A debt office that internalised this would stop reporting a single headline ratio and start reporting the convexity of its liabilities — how the burden behaves not at the mean but in the tail, where currencies gap and growth collapses together.

One caveat dominates — hierarchy. The reserve-currency issuer enjoys the “exorbitant privilege” Gourinchas and Rey dissected: it goes long the world’s risky assets while issuing the safe debt the world craves, an insurer paid to hold the float. America’s structure is sui generis precisely because demand for its liabilities rises in the very crises that would shut a peripheral borrower out. No middle power can replicate it; the optimum is conditional on where you sit in the monetary order.

There is, then, an optimum — but it is a verb, not a number, and a verb most governments decline to conjugate until the choice is no longer theirs. It is the discipline of borrowing in a shape that shares risk rather than concentrating it, that bends with the cycle rather than amplifying it. The nations now learning this the hard way are not those with the largest ratios but those who confused leverage with strength. Israel came through its own test with the bills still mounting and the ratio still elevated — no triumph, and the strain is real — but with a structure that shared the blow rather than amplifying it. When the tail arrives, that is what answers for a state, because no liquidator ever will.