Mario Draghi delivered at the Swiss Federal Institute of Technology in Zurich the tenth lecture in the Swiss National Bank’s series dedicated to the memory of economist Karl Brunner.
The former president of the EBC offers a particularly precise and hard-hitting assessment. The European monetary framework withstood every crisis, yet rested on an assumption no longer valid: growth would come from elsewhere.
Draghi starts from what he calls “the Brunner condition.” “A central bank can preserve price stability only if governments control their debt,” which implies that the gap between the rate paid on debt and growth must be positive (“r minus g”). When the first exceeds the second, the debt grows on its own. The temptation is then to have it financed by the central bank. In the 2010s, the ECB alone reversed this gap: “It is independence, properly exercised, that protected the construction.” But that era is over: “a central bank that bears growth is not a sustainable arrangement, and this arrangement does not hold today.”
According to Draghi, we have entered a new phase, where even fiscal discipline is no longer enough: “even a government with healthy finances cannot fully control the cost of its borrowings.” The cost of money is now decided in Washington and in the AI data centers, imposing a brutal imbalance for Europe which “imports all of the rate increases, but only a fraction of the growth.” One lever remains in its hands: “Of the two forces that determine debt, the interest rate is increasingly set outside Europe. Growth is the one Europe can still influence.”
It is from this reflection in monetary economics that the former ECB president delivers one of his most growth-centered speeches, in which the word or the verb “growth” recurs more than sixty times.
Draghi recalls that European firms “still grow at the pace of their home economy.” Their domestic market confines them, and national reforms will not change that, because they “cannot provide the scale on which technology now depends… Only integration can.” The completion of the single market must therefore be achieved, a goal pursued “since 1985,” along with the creation of the capital markets union and the pooling of computing power for AI, amounting to roughly 100 billion euros: “This sum is well within our reach.”
The conclusion sounds like a warning to capitals. The Maastricht framework will hold only if lawmakers “fully assume the growth objective of the Union.” On this condition alone, “growth and independence reinforce each other.”
The first decade: construction holds
Karl Brunner was one of the economists whose ideas helped shape what would become Europe’s monetary constitution.
In 1968, he introduced the term “monetarism” into the common vocabulary of economics. Alongside Milton Friedman and others, he helped restore the idea that inflation is a monetary phenomenon and that the responsibility for price levels lies with the central bank.
But Brunner was equally clear about the limits of this responsibility. The central bank could not ensure price stability alone. It required a budgetary authority whose incentives were compatible with its own.
If a government ran permanent deficits large enough for its debt to keep rising relative to output, the pressure to monetize that debt would intensify. Ultimately the two regimes could not coexist, and it would be the monetary regime most likely to yield.
These ideas appear in the Maastricht Treaty. It conferred independence on the ECB and an explicit primary objective of price stability. It also established the necessary separation between monetary policy and fiscal policy to preserve that independence: the prohibition of monetary financing, the no-bailout clause, and the excess-deficit procedure.
Whether this separation holds in practice depends, however, on more than rules. It ultimately rests on the relationship between the cost of borrowing and growth, or, in shorthand, r minus g.
If the interest rate a government pays on its debt exceeds the growth rate of the economy, the debt ratio rises on its own, unless the government runs a primary surplus. The larger the gap, the larger the required surplus, and the stronger the temptation to pressure the central bank to monetize. This same relationship was at the heart of Brunner’s own analysis.
The European framework rested on two implicit assumptions about how r and g would stay close.
The first was that growth would be secured by other policies.
Monetary policy was viewed as neutral in the long run, and fiscal policy could at best smooth the cycle. Growth was expected to come from the supply side, especially the completion of the single market, and the monetary union was supposed to amplify these gains.
The reference values for the Union’s budget rules assumed that this growth would be fairly strong. A deficit of 3% of GDP would stabilize debt at 60% of GDP only if nominal growth reached 5% per year. At the time, that corresponded to roughly 3% real growth and about 2% inflation.
But as fiscal policy was constrained by explicit rules and enforcement procedures, growth was largely assumed to be a given. After 1992, in the absence of a common timetable, progress on the single market slowed and it was never completed.
National structural policies were coordinated only through non-binding recommendations.
The second assumption was that a credible central bank, fulfilling its mandate, and governments adhering to the rules, would together lower interest rates.
A credible central bank reduces the premium investors demand to compensate for inflation risk. Sound public finances reduce the premium demanded for holding government debt. With these two premia falling, r would approach g. And lower interest rates would in turn support investment and growth, so the process would sustain itself.
That is, broadly, what happened during the Great Moderation, in two stages.
In the 1990s, governments of advanced economies reduced their deficits, while central banks anchored inflation expectations. Long-term interest rates fell steadily. The term premia declined with them, and the drop was greatest where central banks strengthened their monetary policy framework.
Countries preparing for the euro benefited doubly: from this process and from converging toward a single currency. Governments determined to join ran primary surpluses averaging 2–3% of GDP by the late 1990s.
The prospect of the euro removed the risk of debt devaluation. In 1999, long-term yields in Italy and Spain had fallen by more than 8 percentage points since the start of the decade, to around 4%, while the United States saw a fall of about 3 percentage points.
Then, during the euro’s first decade, growth followed. The decline in rates stimulated activity especially on the periphery, as capital flowed from the core of the area and credit and investment expanded. From 1999 to 2007, the euro area grew by just over 2% in real terms, somewhat short of the 3% that the budget rules were built on.
This growth kept r close to g: during these years the gap between the euro-area public debt’s average interest rate and nominal growth stood at an average of 0.6 percentage points. The central bank could fulfill its mandate using only the policy rate, and its independence was not seriously tested.
But the construction showed early warning signs.
Growth came more from credit and convergence than from the single-market reforms on which the framework had bet. The late-1990s primary surpluses reflected an exceptional effort to qualify for the euro: they were well above the average seen since 1960, which had been near zero. Once these surpluses dissipated, debt stability would depend on growth’s ability to keep pace with interest rates.
The 2010s: the central bank acts to safeguard the construction
The financial crisis, followed by the sovereign debt crisis that ensued, tested this framework.
Growth based on credit vanished. And interest rates again diverged across countries, as investors doubted the public finances of several states. The r minus g gap on euro-area debt widened to 2.8 percentage points in 2008–2011.
With growth collapsing and r diverging from g, macroeconomic authorities had to step in and take charge of growth. In the first phase of the financial crisis, governments did so. In 2009 and 2010, both fiscal and monetary policies were highly expansionary.
But from the outset of the sovereign-debt crisis, governments turned to the 1990s lesson that fiscal consolidation would bring down interest rates. Some went further, convinced it would also spur growth by restoring confidence in public finances.
The fiscal stance tightened sharply. Between 2010 and 2013, the euro area’s cyclically adjusted primary balance tightened by 3.7% of GDP. Worse, this affected countries not under market pressure as well.
Yet conditions were now different. In the 1990s, the private sector expanded as interest rates fell. By 2010, banks and firms deleveraged and could not fill the gap left by the state, so fiscal multipliers were exceptionally high.
It followed that consolidation weighed on growth more than on interest rates. From the second quarter of 2011 to the first quarter of 2013, growth was zero or negative for eight consecutive quarters. Euro-area output stood about 3.5% below potential. ECB simulations indicated that if fiscal policy had offset the slowdown, the output gap would have been about half as large.
This same weakness in demand fed into monetary policy, because its effect was disinflationary for the entire euro area. Core inflation fell to 0.6% in January 2015, and longer-term inflation expectations fell to around 1.5%. Restoring price stability required monetary policy to become highly expansionary.
The weakness of growth thus became a concern for the central bank. With no other demand-supporting policy, rebooting the economy was the channel through which monetary policy could bring inflation back to target. This forced the central bank to act on a large scale, and with tools it had never used before.
Rates were cut to zero, then below. If markets believe zero is a floor, then once rates reach zero they can only rise, and this expectation is reflected across the entire yield curve. The removal of this floor thus had a greater effect than the magnitude of the cuts would suggest. At its peak, negative-rate policy is estimated to have added about 0.3 percentage points to annual GDP growth and 0.2 percentage points to inflation.
To influence longer-term rates, the central bank purchased assets, especially government bonds. This used a channel that Brunner, in his long collaboration with Allan Meltzer, helped to develop: monetary policy acts through the relative prices of all assets, not through a single interest rate.
Assets are imperfect substitutes for one another. When the central bank buys long-term bonds, investors who sell them shift to other assets: bonds of other member states, bank bonds, corporate debt. The term premium falls across these assets, and funding conditions loosen throughout the economy. In the euro area, where the government-bond market is fragmented along national lines, this effect can be even stronger.
These measures achieved their aims. At their peak, they lowered long-term interest rates by about 140–150 basis points. ECB estimates attribute more than a quarter of euro-area growth between 2015 and 2019 to them. Without them, about two and a half million fewer people would have had jobs, and inflation would have become negative in 2016.
The fear at the time was budgetary dominance. It was unfounded, in principle as well as in practice.
In principle, because the treaty’s founders clearly drew the line: the central bank cannot lend to governments, but it can buy and sell their bonds on the market to conduct monetary policy. What distinguishes the two is the purpose.
With inflation well below target and falling, bond purchases were essential to fulfill the mandate. It was an exercise in monetary dominance. If inflation is a monetary phenomenon, responsibility lies with the central bank, whether inflation is too high or too low. The ECB fully assumed this responsibility, as the founders had intended.
The European Court of Justice took the same view. It ruled that these purchases fell within monetary policy and, given the safeguards attached to them, did not violate the prohibition on monetary financing.
The fear was also unfounded in practice. Purchases did not drag the central bank into the orbit of fiscal policy. They removed the condition in which budgetary dominance takes root: a debt ratio that rises on its own because r exceeds g, and with it the pressure to monetize.
By lowering borrowing costs and supporting growth, monetary policy helped reverse the r–g gap, which averaged less than 1 percentage point between 2015 and 2019. As deficits fell with the economy’s recovery, this was enough to bring euro-area debt to GDP down from 93% in 2014 to 84% in 2019.
These purchases did not bind the hands of the central bank when inflation returned, as some feared. In 2022, it raised rates at the fastest pace in its history, even as it held a substantial portfolio of bonds and incurred losses on it.
It was independence, properly exercised, that protected the construction and maintained the separation between monetary and fiscal policy.
But a central bank that bears growth is not a durable arrangement, and this arrangement no longer holds today.
Today: Global rates, local growth
Two forces are at work today that the central bank cannot, or should not, compensate for.
First, euro-area long-term rates are rising, and increasingly for reasons outside Europe’s control.
In recent months, they have reached their highest level in about fifteen years, moving in step with U.S. Treasury yields. This fits a longer-term trend: long-term yields in advanced economies are now moving in much closer tandem than when the euro was designed. And last year global factors shaped the far end of the euro-area yield curve more than in any other year of the past decade.
The construction did not anticipate this. It assumed that governments with sound finances would be rewarded with cheaper borrowing costs.
Budgetary discipline remains important, but it cannot shield a government from fluctuations in global interest rates. Going forward, even a government with healthy finances cannot fully control its borrowing costs.
Part of the rise in yields reflects inflation. Global supply shocks have again pushed up inflation, most recently through energy price increases caused by the conflict with Iran and by reduced refining capacity in Russia and the Middle East.
As long-term yields reflect expected inflation over the life of a bond, higher inflation directly pushes them higher.
The larger share reflects the supply of debt, primarily driven by procyclical fiscal policy in almost all major economies, and even more so in the United States. This year, the United States will run a deficit close to 6% of GDP and, with policy unchanged, its debt will exceed its postwar peak by 2030.
Private borrowing taps the same pool of capital. The largest U.S. tech companies alone are set to invest nearly $800 billion this year, mainly in AI data centers, about one-sixth of total annual corporate investment in the United States. Global investment in data centers could exceed $3 trillion by 2030. Much of this is financed by debt: AI-related bond issuance could reach about $300 billion this year, roughly one-third of U.S. corporate bond issuance.
Europe also contributes to the debt supply, with German fiscal expansion and a French deficit above 5% of GDP. It is unlikely this trend will reverse. If policies remain unchanged, the IMF estimates that by 2040 Europe’s advanced economies will face additional expenditures equivalent to about 4.5% of GDP for aging, defense, and climate transition. That is roughly three times faster growth than in the two decades preceding the pandemic.
With more long-term debt to hold, investors demand a higher premium for owning it.
In the 2010s, central-bank purchases compressed the term premium, keeping euro-area long-term yields well below those in the United States. Today, with inflation above the target, it can no longer do so. On the contrary, unwinding past purchases will raise the term premium, by about 20 basis points in the euro area by the end of the decade.
Second, euro-area growth now lags behind the rate of interest it faces on a global scale.
The immediate impetus from higher deficits and AI investment is strongest in the United States, which also has more to gain from faster productivity growth if AI delivers on its promises. Europe bears all of the rate increases but only a fraction of the growth.
This gap does not yet show up in the debt service burden. Debt issued during the era of low rates still carries low coupons, so the average rate paid by governments, about 2.25%, remains well below a nominal growth rate of around 3.5%.
Yet each year, roughly one-sixth of this debt is refinanced at current rates, and the average rate rises accordingly. The Commission projects that by the early 2030s the euro-area’s average interest rate on debt will surpass the growth rate.
If current trends continue, growth will not rise enough to close the gap. Year after year, the reforms intended to deliver growth have remained unfinished.
The single market in services has been a goal since 1985; yet over the past twenty years, barriers to services within Europe have not fallen faster than those faced by foreign firms.
The capital markets union has been on the agenda for more than ten years, and since 2014, frictions for a saver investing in stocks in another euro-area country have fallen only half as much as those for investing in the United States.
Finally, productivity growth in the euro area has remained well behind that of the United States. Since 1999, output per hour has risen at roughly half the pace of the United States, and since 2019 growth rates have diverged further: 0.3% per year in the euro area versus 1.8% in the United States.
The need for growth, and what monetary policy can do
The IMF has recently laid out what Europe can expect if it stays on its current trajectory.
If governments finance their increased spending needs without new consolidation or reforms, Europe-wide debt ratios would reach about 130% of GDP by 2040. Weighted by the size of economies, they would reach about 155%.
Some medium-term fiscal consolidation is inevitable. But if it must come solely through spending cuts and tax hikes, it is unlikely to yield the necessary result.
Over the past three decades, a typical consolidation episode in Europe has amounted to a total of 3–4% of GDP. Weighting the countries by the size of their economy, those that must adjust would today need to improve their primary balance by about 6–7% of GDP, roughly twice as much as before. And even the smallest effort has never been politically easy.
The ability of governments to undertake this consolidation thus depends on growth. Of the two forces that determine debt, the interest rate is increasingly set outside Europe. Growth is the one Europe can still influence.
A half percentage point more growth per year, maintained through 2040 with some of the extra revenue saved, would allow Europe to travel about one-third of the way toward a sustainable debt path. The remaining effort would then be close to what Europe has already managed to achieve.
Such growth is within reach. Rapid AI adoption could add up to 0.4 percentage points to the growth of total factor productivity over the next decade, and national reforms together with reforms of the single market could collectively add about 0.5 percentage points per year.
Without this growth, debt ratios will continue to rise on their own, and the temptation will be to finance them through the central bank. History suggests that when faced with a choice between significant fiscal tightening and monetary financing, governments have more often turned to the central bank than taken the hard policy decisions themselves.
So what can monetary policy do? The first priority is to control inflation
In a phase of deteriorating debt dynamics, markets will begin to test the central bank’s commitment. If they come to doubt that price stability takes precedence over government financing, the inflation premium will reappear, borrowing costs will rise, and budgetary positions will worsen.
This cycle can become self-reinforcing. Fragile public finances heighten the expectation of future debt monetization. This expectation lifts inflation expectations, which again pushes up borrowing costs, further weakening public finances. To the point, the central bank might not be able to raise rates without fueling monetization expectations.
The second priority is to avoid weakening growth more than necessary. In the face of inflation shocks, monetary policy cannot now drive growth, but it can ensure that the trajectory of the policy rate does not rise more than what is required to stabilize inflation over the medium term. This is compatible with the priority given to price stability.
The stakes are high, because growth is sensitive to the expected path of rates. Estimates suggest that the anticipated rise in short-term rates since the end of last year will reduce total growth by about 1.1 to 1.3 percentage points over 2026–2028. These estimates predate the more recent rate hikes since mid-August.
The effects of monetary tightening can last longer than the tightening itself. Productivity growth is typically seen as the remit of structural and budgetary policies, largely insulated from monetary policy. Yet in the United States and in the euro area, about half of productivity growth since 1999 has come from capital deepening, which depends on the cost of capital. By weighing on investment and innovation, a restrictive policy can depress productivity and potential output for a period. With potential growth only slightly above 1%, Europe has little room for error.
When shocks are frequent and complex, finding the right rate path is largely a matter of communication: markets must understand how the central bank will react in each scenario.
The reason is that markets do not price in a single trajectory for interest rates. They price in all trajectories they deem possible, weighted by their probability, and the rate path observed today reflects the average.
If markets know how the central bank will react, most of the weight concentrates on a single trajectory. Otherwise, they must contemplate alternative responses: for example, in the face of supply shocks, central banks may simply raise rates in line with overall inflation. This possibility lifts the average, and the expected rate path may end up higher than required for price stability.
It is important, especially as global factors play an increasing role in driving European yields, to ensure that the expected path of policy rates reflects the central bank’s intentions, not uncertainty about how it will react.
The emphasis on communication today makes clear that the most essential skill is the ability to explain.
A path toward European renewal
The principal engine of long-term growth, namely the development and diffusion of new technologies, lies beyond the influence of any single central bank. Growth should become an explicit objective shared by all governments.
The usual European answer, with each country acting on its own, will not suffice in the world it now faces. National reforms matter, but they cannot provide the scale on which technology—and thus growth—now depends. Only integration can.
Recent work shows what the cost of lacking scale looks like. Between 2008 and 2023, the gap between the value of American and European listed companies widened from about 3,000 billion to 34,000 billion dollars. This gap is not due to a handful of flagship firms or sectoral differences. It is most pronounced among young firms and in sectors with the highest returns to scale.
These studies also identify the cause. European firms still grow at the pace of their home economy: a 1% rise in a country’s GDP is associated with a 0.8% rise in its firms’ revenues, whereas the revenues of American firms bear no relation to the size of their home state.
Supra-national reforms have thus become the most important lever for growth. A detailed plan for implementing these reforms was laid out in the European Competitiveness Report, and much of it now appears in the Union’s common roadmap. The first step is to implement this roadmap.
This will also help finance the investments Europe needs in technology, clean energy, and defense, at a time when public finances are constrained.
Completing the single market and the capital markets union would raise the return on capital in Europe, which for large European firms is about one-fifth below that observed in the United States. Combined with a monetary policy that does not keep rates higher than necessary, this would enable the private sector to shoulder a larger share of investment.
A larger share of Europe’s abundant savings could then be invested domestically, and more foreign savings could be attracted, as investors seek to diversify away from American assets. Last year, euro-denominated international debt issuance reached its highest level since the introduction of the single currency, and portfolio capital inflows approached their historic highs.
A faster growth path would likewise improve the balance between growth and interest rates, freeing space for public investment without compromising debt sustainability. The fiscal room would be even greater if this investment were directed at the level, national or European, where it is most effective.
Today AI stands as one of the clearest examples. Secure access to computing capabilities benefits all Europeans, yet those capacities are insufficient when every country acts alone. The Union hosts less than 5% of global AI computing capacity. A recent study suggests that raising that share to 15% by 2030—that is roughly the Union’s share of the global economy—would require about €1.3 trillion of investments.
The public portion needed to reduce the risk of this investment, estimated at about €100 billion, is particularly well-suited to financing through the European budget. Even that amount would amount to only about 5% of the upcoming budget currently under negotiation, and would relieve national budgets during consolidation. This sum is well within our reach.
The Brunner condition remains valid: a central bank can preserve price stability only if governments control their debt. What has changed is the way this condition is met. A budgetary framework cannot achieve it alone when growth is weak.
As integration advances, growth can stay firmly above interest rates, and the central bank will be better shielded from political pressures. Growth and independence then reinforce each other.
One generation ago, Europe endowed itself with a monetary constitution that has withstood every crisis since. It rests on an independent central bank with a clear priority of price stability, and on fiscal discipline. This framework remains the right one.
But for this constitution to endure, lawmakers, at both national and European levels, must now fully assume growth as an objective of the Union. If they do, Europe will be set on the path to renewal.