Seventy-two hours, five decisions, no cuts

The calendar, and why every arrow points up or sideways.

Every central bank on that calendar is either hiking or holding with a bias to hike. Set the policy rates against inflation and the reason is clear: several are still below inflation in real terms, and markets expect them to keep climbing.

This is a synchronised tightening, and it is happening at the worst possible moment for government finances. That combination is what makes this week a test of the limits rather than a routine set of decisions. To see why, it helps to go back twelve months, to a moment when central banks looked as if they had finally pulled it off.

Late 2025: the soft landing that almost arrived

A regime getting where it wanted to be, until 28 February.

Rewind to the last weeks of 2025. Inflation had come down from its 2022 peak without a deep recession. The Federal Reserve had cut rates three times that autumn, to 3.50–3.75%. The ECB had brought its deposit rate down to 2.00% and stopped. Bank forecasts read like victory laps: MUFG titled its 2026 outlook “a post-peak USD world”, and J.P. Morgan saw EUR/USD at 1.22 by March. The euro sat just under 1.18 and sterling near 1.35; Japan’s ten-year yield was about 1.6% and Britain’s about 4.4% (sources: earlier Capital Dispatch coverage; Vantage; Trading Economics). Gold had just closed above $4,000 for the first time, which was treated as a curiosity rather than a warning. For central bankers, it was the closest thing in fifteen years to the world they had been steering towards.

Then, on 28 February 2026, the war with Iran began. A war does what no committee can: it changes the price of things people cannot do without. Fears for the Strait of Hormuz sent oil sharply higher, and it has stayed there; Brent traded above $117 last Friday and diesel is near $200 a barrel (source: Saxo). Inflation that had been fading turned back up, to 3.7% in the US, 3.8% in the euro area and 3.5% in the UK. The US ten-year yield has not closed below 4% since the war started (source: earlier Capital Dispatch coverage). Central banks that had been planning further cuts found themselves hiking instead: the RBA four times, the ECB in June and September, the Fed and the Bank of Japan in September, with Canada, New Zealand and the Bank of England next.

The soft landing had not been a mirage. It was real, but it rested on cheap energy, falling goods prices and a long run of low borrowing costs, none of which central banks control. When one of those supports was knocked away, the limits of what remained became visible. To understand how central banks came to be asked to carry so much, we need to look at how their role grew.

How central banks became the economy’s first responders

Each rescue worked. Each one also moved the risk somewhere else.

In February 2004, Ben Bernanke, then a Fed governor, gave a speech titled “The Great Moderation”, noting that growth and inflation had become far less volatile over two decades and crediting better monetary policy for part of it. The implication was that central banking had been largely solved. What followed was a period in which central banks were asked to do more, and did more, than at any time in their history.

Three side effects accumulated along the way. First, the water went somewhere. Holding the price of money below what the economy would choose pushed savings into whatever was scarce: houses, shares, private credit, government bonds. Second, the put was learned. Each rescue taught investors and governments that the next would come, so both took on more debt. Third, the bill moved to the public sector. After 2008, the weakest balance sheets were banks and households; today they belong to governments, and central banks themselves have taken losses on the bonds they bought, losses that in Britain are covered by the Treasury. Interest-rate policy that once affected private borrowers now runs straight through the public finances.

The rescues were not mistakes. Each prevented something worse. But each also left the system more dependent on the next one.

The question this raises is whether there is a ceiling to all this, a point at which raising or lowering the rate stops buying the outcome central banks want. Two centuries of history have a clear answer.

Every time a central bank pushed rates to the extreme, and how it ended

Twenty-eight episodes from 1825 to today, plus the long-yield extremes. The pattern is consistent enough to be a rule.

The chart below plots every major episode since 1825 in which a central bank raised its policy rate to an extreme level to defend a currency, stop a panic or break inflation. The vertical scale is logarithmic, because the range runs from 5% in Regency London to overnight rates reported as high as 7,500% in Istanbul in 2001. The colour shows the outcome. Tap any point.

Read the colours and a rule emerges. The rate held when the problem was liquidity. In 1847, 1857 and 1866 the Bank of England raised its rate to 8% or 10% during banking panics, lent freely against good collateral, and the panic passed. This was the doctrine Walter Bagehot set out in Lombard Street (1873): in a panic, lend freely at a high rate. It worked because the underlying borrowers were solvent and the state was not the problem. The rate failed when the problem was solvency or politics. Germany’s 90% in 1923, Britain’s 6% in 1931, Sweden’s 500% in 1992, Russia’s 150% in 1998 and Turkey’s overnight rates of several thousand percent in 2001 all failed because no interest rate can fix a government that cannot or will not pay, or a peg the country can no longer afford. And where it held against inflation, it held at great cost: Volcker’s 20% broke US inflation in 1981 but required two recessions and unemployment above 10%, and the political backing to tolerate them.

The yield rows in the ledger tell the same story from the other side. The US ten-year reached about 15.8% in 1981 and Britain’s long-dated consols about 17% in 1974, both times because markets doubted that inflation would be controlled. Italy’s ten-year hit about 7.5% in November 2011 and Greece’s rose above 30% in 2012, when markets doubted those governments could pay. In every case the long end of the market moved first, and the central bank had to respond to it, not the other way round. When the Fed capped long yields at about 2.5% from 1942, it eventually had to give up the cap in the 1951 Treasury–Fed Accord. When the Bank of Japan capped its ten-year yield from 2016, it gave up in 2024. The long end, over time, belongs to the market.

Why a single lever cannot deliver everything

Five constraints that no amount of credibility can remove.

Limit one: one tool, too many targets. In 1952 the Dutch economist Jan Tinbergen, later joint winner of the first Nobel prize in economics, set out a rule now taught to every policy student: you need at least as many independent instruments as targets. Central banks have one main instrument, the short-term interest rate, and in 2026 they are being asked to hit at least five targets: inflation, employment, financial stability, the exchange rate and, implicitly, the government’s borrowing costs. Something has to give. Try it.

No setting clearly helps more than three of the seven groups, and the losers change with every choice. That is not a failure of judgement. It is arithmetic.

Limit two: rates cannot make oil. An energy shock caused by a war in the Gulf is a supply problem. Higher rates can stop it spreading into wages and other prices, but they cannot bring down the price of the oil itself. Every central bank on this week’s calendar is fighting inflation it did not create with a tool that works on demand.

Limit three: the debt arithmetic. In 1981 the economists Thomas Sargent and Neil Wallace published a paper with a memorable title, “Some Unpleasant Monetarist Arithmetic”. Its point was that when government debt is high enough, a central bank fighting inflation with higher rates can make the fiscal position worse, until the government’s need for cheap money overrides the central bank’s independence. Economists call that fiscal dominance. Today’s numbers make the paper uncomfortably relevant.

Advanced economies spent more than $3.3 trillion servicing their bonds last year, more than the world spent on defence in a war year (source: CNBC via StockAnalysis). Global debt passed $365 trillion in the first half of 2026 (source: IIF), and the IMF expects global public debt to approach 99% of world GDP by 2028 (source: IMF Fiscal Monitor, April 2026). In the US, the fiscal year to September produced a $1.99 trillion deficit and net interest above $1 trillion, roughly 19 cents of every dollar of federal revenue (derived from US Treasury data via Ground News), up from 18.5% a year earlier and projected to reach 25.8% by 2036 (source: Briefs Finance, citing CBO). In Britain, debt interest is close to 4% of national income, about twice its pre-pandemic level (source: Barclays via Vantage).

The key comparison is between the interest rate a government pays (r) and the growth of its economy in cash terms (g). When g exceeds r, debt shrinks as a share of the economy almost by itself. When r rises above g, debt grows on its own unless the government runs a surplus before interest. For most of 2010–2021, r sat far below g. Multi-decade-high yields have closed that gap. Each hike this week lifts r for every government with short-dated debt.

Limit four: the long end belongs to someone else. A central bank sets the price of money for one night. The price of money for ten or thirty years is set by investors deciding what a government’s promise is worth over that horizon. Central banks can influence it by buying bonds, but buying bonds while trying to fight inflation contradicts their own policy, as the Bank of England discovered in 2022.

Limit five: structure. Interest rates cannot fix an ageing workforce, a decade of weak productivity growth, too few homes, an energy system exposed to a single strait, or supply chains being redrawn by politics. These raise the level of interest rates an economy needs, and they make the trade-offs sharper, but they lie entirely outside the central bank’s reach. When those weaknesses dominate, the central bank can choose which group bears the pain. It cannot make the pain disappear.

Every tool in the box, and why yields kept rising

Rate moves, bond-buying changes, issuance cuts, a refused backstop and a currency intervention. Each bought days or weeks. None changed the trend.

It would be wrong to say central banks and finance ministries stood still while long-term yields climbed. Over the past ten months they have tried almost everything short of capping yields outright. The record is worth setting out in full, because it is the clearest evidence for the argument of this article: when the problem is debt, energy and credibility, the tools that work on liquidity only buy time.

Three patterns stand out. First, supply tweaks treated the symptom. Japan, Britain and the US all shifted borrowing away from long maturities, and the Bank of England stopped selling long gilts. None of that reduced how much governments need to borrow; it moved the borrowing to shorter maturities, where it will be refinanced sooner at whatever rate prevails then. The market saw through it. Second, waiting was punished and acting was only partly rewarded. The Fed’s July pause sent long yields to their highest since 2007; its September hike strengthened the dollar but did not bring long yields down. Third, the backstops came with conditions. When France’s spread reached euro-crisis levels, the ECB and the Bank of France said, in effect, that the fix lay with the French budget, not with Frankfurt. That is a central bank stating its own limit in public.

Every tool bought days or weeks. None changed the trend, because none could change the debt, the war or the budgets behind it.

Even the most direct weapon, currency intervention, worked only briefly: about $59 billion of Japanese reserves and the first joint action with Washington in fifteen years pushed USD/JPY down by nearly nine yen, and roughly a third of that has already been given back. The market is not defying central banks out of spite. It is pricing the parts of the problem they cannot reach.

Spiking yields are the market pointing at the limit

The gap between what central banks charge and what governments pay.

If central banks were fully in control, long-term yields would track expected policy rates plus a modest premium. Instead, the gap between the overnight rate central banks set and the long-term rate governments pay has widened sharply. That gap is the part of the price of money that central banks do not control, and investors are charging more for it.

Each of these gaps carries a message. In Britain, a 30-year gilt yielding more than 6% while Bank Rate is 3.75% is the market asking whether the Budget will be credible; the OBR estimates that a one-point rise in gilt yields and Bank Rate wipes about £16 billion off the government’s fiscal headroom (source: OBR via Vantage). In Japan, a ten-year yield at a 30-year high while the policy rate is 1.25% reflects worries about Prime Minister Takaichi’s spending plans (source: Trading Economics). In France, a spread over Germany at its widest since 2011 is a judgement on politics, not on the ECB. And in the US, a thirty-year yield at its highest since 2007 reflects a $1.99 trillion deficit as much as the Fed. The ECB’s Christine Lagarde has noted that rising yields are already tightening conditions, and Morgan Stanley says they provide “some degree of additional tightening” (source: Dow Jones, 25 Oct). In other words, the market is doing part of the central banks’ job for them, on terms the central banks did not choose.

This is the core of the argument. Short-term rates are still set by central banks. Long-term rates are being set by investors who have looked at the debt, the war, the structure of these economies and the history in section 04, and concluded that central banks alone cannot guarantee the outcome. When that judgement reaches currencies, it changes how they trade.

Where the limits show up first

Rate differentials still drive most days. When the limit is in sight, a higher yield can sink a currency instead of lifting it.

The textbook rule says money flows to the currency that pays more. That has largely held this year: the dollar posted its highest monthly close since March 2025 in September after the Fed hiked (source: earlier Capital Dispatch coverage), and the yen, which pays far less, needed joint Japan–US intervention in August to pull USD/JPY back from near 164 (source: Admirals).

But when investors start to doubt that a central bank can deliver, the rule can flip. They demand a higher yield to hold the country’s bonds and sell its currency at the same time. The rate rises, the currency falls. In section 04 that pattern ran through every failure from 1931 to 2018. It is the clearest sign that the market thinks a central bank has reached its limit.

That gives a practical test for this week. If a currency rises with its yields, the market still trusts the central bank. If it falls as yields rise, the market is pricing the limit. Here is where each central bank on this week’s calendar stands.

The currencies most exposed are those where the binding constraint is fiscal or political rather than monetary: the euro through France, the pound through the Budget and the yen through Japan’s debt. The dollar is strongest on rates and most exposed on debt in the longer run.

Three ways the limits could play out by year-end

Editorial scenarios with assumed weights. Not forecasts.

Five numbers will show whether the market is pushing central banks toward their limit. Each is set against the level that would signal stress. Three crossing their lines together would point to the bad-yield scenario.

A balanced trade plan: one long and one short for every market

Event-triggered setups. Only one side of each pair can activate; the trigger decides which.

When the outcome depends on whether markets believe central banks this week, a one-directional view is a guess. So each market below has two setups, built the same way: a trigger tied to this week’s events, an entry on confirmation rather than anticipation, a stop beyond the level that would prove the idea wrong, and targets at the next published reference levels. Risk and reward are of similar size on both sides.

Levels: EUR/USD, USD/JPY, USD/CAD and gold references from Admirals (21 Oct); GBP/USD references from RoboForex weekly notes (19–23 Oct); EUR/GBP and recent lows from Saxo (26 Oct); round-number levels marked as such. Reward-to-risk is to the first target. Illustrative research, not a recommendation.

Can central banks still engineer an outcome everyone can accept?

Not alone. History says they never could; the last twenty years made it look as if they might.

Central banks are not powerless. They still set the price of overnight money, they can still stop a liquidity panic in an afternoon, and Volcker in 1979 and Draghi in 2012 showed that a credible central bank can change the course of an economy. But the 200-year record is unambiguous about the conditions. Those successes came when the problem was temporary or when governments moved in the same direction. When the problem was a government that could not pay, a peg that could not be afforded or a political system unwilling to accept the cost, no rate was high enough.

The situation in 2026 contains elements of both. The Iran war is a supply shock rates cannot reverse. Debt is at record levels, so every hike now raises government interest bills. The structural weaknesses (ageing, weak productivity, energy dependence, housing) lie outside the central bank’s reach. And twenty years of rescues have taught markets and governments to expect the next one. Under those conditions central banks can choose who bears the pain, but they cannot make an outcome acceptable to everyone. The bond market has started to say so, by setting long-term borrowing costs at multi-decade highs while central banks hold short rates lower. What would restore the lever’s power is not a bigger lever but help from outside it: credible budgets, an easing of the energy shock, and reforms that raise growth. Until then, the market will keep testing where the limits lie, and currencies will keep the score.

The overnight rate belongs to the central bank. The long end belongs to whoever is willing to lend for ten years, and they will do it only for a government and a central bank they believe.

Sources and notes

For informational and educational purposes only. Not financial advice. CFDs and FX are leveraged products and can result in losses that exceed expectations; trade only with money you can afford to lose. Market levels and pricing change quickly; confirm before acting. Scenario weights, stress lines, the decision-desk scoring and the outcome classification of historical episodes are editorial judgements.

This week and markets: Saxo, “Market Quick Take” (26 Oct 2026); Dow Jones Newswires via MarketScreener, “Week Ahead for FX, Bonds” (25–26 Oct 2026); Admirals, “Central Bank Super Week” (21 Oct 2026); ForexTradingCommunity week-ahead calendar; Vantage Markets, “UK Autumn Budget 2026”; Reuters via Investing.com and IndexBox (UK 30-year gilt); Trading Economics (Japan 10-year); BabyPips (UK CPI, 21 Oct 2026); AAP via Port News (RBA, 29 Sep 2026); RoboForex weekly notes (19–23 Oct 2026); earlier Capital Dispatch coverage (2025 outlooks, US 30-year, US 10-year since 28 Feb).

2026 policy actions: Advisor Perspectives/Bloomberg, US refunding (4 Feb 2026); Trading Economics, Treasury buybacks and yields (Aug–Oct 2026); ROIC/Reuters, Fed July dissents (30 Jul 2026) and ECB on France (8 Oct 2026); Northeast Times/Reuters, joint yen intervention (3 Aug 2026); CryptoBriefing, BoE QT decision (17 Sep 2026); Vantage, DMO remit (23 Apr 2026); MarketScreener/Reuters, Japan super-long issuance plan (24 Dec 2025); NBC/CNBC, BoJ taper plan (June 2025).

Debt: Institute of International Finance Global Debt Monitor via StockAnalysis (Sep 2026); IMF Fiscal Monitor (April 2026); OECD Global Debt Report 2026; US Treasury via Ground News (16 Oct 2026); Briefs Finance citing Treasury and CBO (Sep 2026); Barclays via Vantage.

History and ideas: Walter Bagehot, Lombard Street (1873); Jan Tinbergen, On the Theory of Economic Policy (1952); Thomas Sargent and Neil Wallace, “Some Unpleasant Monetarist Arithmetic”, Federal Reserve Bank of Minneapolis Quarterly Review (1981); Ben Bernanke, “The Great Moderation” (speech, 20 Feb 2004); Carmen Reinhart and Kenneth Rogoff, This Time Is Different (2009); Carmen Reinhart and M. Belen Sbrancia, “The Liquidation of Government Debt” (NBER, 2011); Bank of England Bank Rate history; Federal Reserve historical data (discount rate, federal funds, balance sheet); standard accounts of the Reichsbank (1923), sterling and gold (1931), the Treasury–Fed Accord (1951), the ERM crisis (1992), the Asian and Russian crises (1997–98), Turkey (2001), Iceland (2008), the euro crisis (2011–12), Argentina (2018), Russia (2014, 2022) and the 2022 gilt crisis. Historical figures are rounded and episode peaks are approximate.