Must States Go Into Debt?

The debate about public debt is a war of faith. Some see every loan as a sin, others regard money as unlimited. Both are wrong. We examine what holds, and in doing so we take on the theory that promises to cut the knot.

I. The Japanese puzzle

Japan has public debt of more than 230 per cent of its economic output, more than three times what the Maastricht rules allow in the euro area. Around half of its government bonds are held by its own central bank. Yet for decades there was no state bankruptcy, no hyperinflation and no debt crisis.

The explanation is simple: a state that borrows in its own currency cannot become insolvent. It creates the money in which it has to pay. That is not a theory but bookkeeping. Whoever compares the state to a thrifty housewife confuses the waterworks with the customer who has to buy the water.

But Japan, too, now shows the limit. Since 2022 inflation there has been above target, the central bank has raised interest rates, and for the 2026 fiscal year the government assumes an interest rate of 3 per cent on new bonds, the highest since 1997. Debt service costs are reaching record levels. Japan cannot go bankrupt. But every yen spent on interest is missing for something else.

II. The real limit

That a state cannot go bankrupt does not mean debt has no consequences. The real limit is not solvency but the capacity of the economy. If the state spends more money than people, machines and materials can provide, prices rise. That is experience, from the Weimar Republic to Venezuela.

What also matters is what the money is spent on. A loan for a bridge that lasts fifty years is something different from a loan for current spending used up in the same year.

III. The German obsession

In Germany the word “Schulden”, debt, carries a moral burden; it sounds like “Schuld”, guilt. The debt brake written into the Basic Law in 2009 limited the federal government’s structural new borrowing to 0.35 per cent of economic output, one of the strictest rules in the world. The result was a low debt ratio and an investment backlog: crumbling bridges, slow internet, outdated schools.

In March 2025 the CDU/CSU, SPD and Greens relaxed the rule: a special fund of 500 billion euros for infrastructure, and defence spending above one per cent of economic output outside the debt brake. The consequences are in the federal financial plan: in 2027 around 119 billion euros of new borrowing in the core budget, plus 55 billion for the special fund; in 2030 around 167 billion. Interest spending rises from 41.9 billion euros in 2027 to 80.7 billion in 2030. The Federal Court of Audit warns of 800 billion euros of new borrowing by 2029.

IV. The three camps

The advocates of thrift, from the Federal Court of Audit to parts of the CDU/CSU, consider public debt fundamentally a burden on future generations. The Keynesians, in the SPD, among the Greens and many economists, want the state to borrow in crises and save in good times, and criticise the debt brake for not distinguishing between consumption and investment. And an academic minority says: the question of financing is wrongly posed. That is Modern Monetary Theory.

V. Modern Monetary Theory put to the test

Modern Monetary Theory, MMT for short, goes back to the economist Abba Lerner, who in 1943 called for “functional finance”, and was developed by Warren Mosler, Randall Wray and Stephanie Kelton. Kelton’s book “The Deficit Myth” of 2020 made it popular. Its core: the state creates money by spending it and destroys it by levying taxes. The only relevant limit is inflation. If it rises, the state raises taxes or cuts spending.

What it gets right. It is right that a state with its own currency is not a household. And it is right that the debt ratio is an arbitrary yardstick. The 60 per cent of Maastricht was not derived but decreed. A control engineer would say: MMT replaces a wrong sensor, the debt ratio, with a better one, inflation. That is its real contribution.

Where it fails. In five places.

First, the actuator sticks. To curb inflation, the state is supposed to raise taxes or cut spending. A parliament has to decide that, in a boom, against its own voters. Months or years pass before it does. A controller that intervenes so late makes the system oscillate, like the farmer who breeds pigs when prices are high and sells them when everyone is selling. MMT presupposes a discipline that democracies demonstrably lack.

Second, the sensor reports too late. Inflation only becomes visible when overheating is already under way. The years 2021 to 2023 were the field test. The United States ran huge deficits during the pandemic, and in June 2022 inflation stood at 9.1 per cent. MMT advocates point to supply chains and energy prices, partly with justification. But this very dispute shows that nobody knew in time where the limit of capacity lay. A sensor whose reading is argued about only afterwards is no use for control.

Third, it applies only to a few. A country that has to buy energy, raw materials and goods in a foreign currency hits the limit through its exchange rate. Argentina and Turkey show how that ends. MMT works best for the United States, whose currency the whole world wants to hold. That is not a law of economics but a privilege, as we described in “The Exorbitant Privilege”.

Fourth, it does not apply to Germany. The euro countries have no central bank of their own, and Article 123 of the EU Treaty forbids the European Central Bank to finance states directly. In the euro area the Gordian knot is real, and MMT does not cut it.

Fifth, it does not answer who decides. If money is not scarce, the question of what it is spent on becomes purely a question of power. MMT proposes a state job guarantee; against the lobbies’ grip on fresh money it has nothing. What that grip looks like is shown by the bank rescues we added up in “The Bill Nobody Draws Up”: the profits stayed private, the losses became public. And what fresh money for arms brings in is shown by “Münchhausen’s Heirs”: at most fifty cents per euro.

The verdict. MMT is half right. Its diagnosis is better than the textbook view, its prescription weaker than its reputation. It is not wishful thinking where it describes how money is created. It becomes wishful thinking where it trusts politicians to brake at the right moment.

VI. The Bundesbank’s compromise

Remarkably pragmatic is, of all institutions, the Bundesbank, long the guardian of orthodoxy. In November 2025 it presented a staged plan for a new debt brake. Until 2029 high deficits for defence and infrastructure remain possible; from 2030 to 2035 they are to fall step by step. From 2036 the federal government would be allowed to borrow permanently 0.8 per cent of economic output, but only for additional investment, plus 0.35 per cent as long as the debt ratio does not exceed 60 per cent. By its own calculation the ratio would fall below 60 per cent only in the mid-2050s.

That is sensible because it distinguishes between investment and consumption. But it is a thirty-year plan that assumes governments will do until then what they promise today. This controller, too, depends on political discipline.

VII. The blind spot of the debate

What the German debate lacks is the question of what not investing costs. A closed bridge causes detours, lost time and higher transport costs. A school without technology sends graduates into a competition they will lose. These costs appear in no budget because nobody books them.

Equally missing is the question of to whom the debts are actually owed. At the end of 2021 the Eurosystem of central banks held 39 per cent of German public debt, foreign investors 30 per cent and domestic banks 16 per cent. Domestic insurers and pension funds accounted for 1 per cent; private households held practically nothing directly. The interest raised by taxpayers therefore hardly goes to German savers. And since the Bundesbank has been making losses, the Eurosystem’s share no longer flows back to the federal government as profit either; the last profit transfer was in 2019. Whoever considers a debt write-off must say whom it hits. Incidentally, the direct bill for the bank rescues amounts to only around two per cent of today’s public debt. The large sums have other causes.

VIII. The right questions

“Must states go into debt?” is the wrong question. The right ones are:

• For what? Investments that serve for decades justify debt more than current spending.
• In which currency? Debt in one’s own currency is different from debt in a foreign one. Germany no longer has a currency of its own.
• At what interest rate? New loans at one per cent are different from loans at three. Federal interest spending will almost double by 2030.
• In what situation? Debt in a recession stabilises; debt in a boom overheats.
• Who decides, and who controls? None of the three camps answers this question.

IX. What follows

All three camps start from the same point: they hope for political discipline. The thrifty want to enforce it with a rigid rule, the Keynesians rely on clever counter-steering, MMT on timely tax increases. Experience shows that all three hopes deceive.

A cybernetic approach would not demand discipline but build it in. MMT supplies the right sensor, the capacity of the economy. What is missing is an actuator that works without a parliamentary decision in each individual case. In our draft “The Cybernetic Economic Model” these are the types of money: new money arises only where something new is built, as earmarked investment money, and its quantity is coupled to a sensor, not to a vote. That would be MMT with built-in negative feedback instead of hoped-for discipline.

Public debt is neither the sin it is taken for in Germany nor the consequence-free tool some MMT supporters sell it as. It is a tool that creates prosperity when used wisely and destroys it when used foolishly. Whoever wants to use it wisely must not rely on the wisdom of those who happen to hold it.

Further reading: The Bill Nobody Draws Up · Münchhausen’s Heirs · Interest Is Not the Culprit · What Happens When the Next Crisis (Pandemic) Comes · The Primacy of Politics · The Exorbitant Privilege · Money — Tool or God? · The Cybernetic Economic Model