THE COST OF THE NEXT SHOCK

THE COST OF THE NEXT SHOCK

Four Clocks

One crisis is paid for in hours, in months, in years and in decades. The four bills arrive on different schedules, from different directions, and only the first one makes the news. The last one is already being signed.

Article 2 of 12  ·  Series I  ·  Published 20 October 2026  ·  Analysis → Forecast → Recommendations

THRIVE IN CHAOS is an AI intelligence system with human editorial oversight.

1 · The Signal

By the second week of October, a tanker captain willing to take a ship through the Strait of Hormuz could earn about $100,000 a month, with a bonus of up to $50,000 for each crossing. The insurance on the ship he was steering could run to as much as $20 million for a single voyage into the Gulf. Before this war began, the same cover cost about a quarter of a million dollars.

That is the fastest clock in the world economy, and you can read it to the day. On 11 October, two commercial ships crossed the strait. Before the closure the daily figure was about 85. In the first two weeks of October there were 23 recorded attacks on shipping in and around the strait, close to double the monthly average since the fighting started, and ten vessels were struck in four days. Two ships sank. The war-risk premium on a hull moved from a quarter of one percent of the ship's value to somewhere between seven and ten percent, and underwriters now review their quotes every twenty-four hours.

Now look at three other things that happened in the same stretch of weeks, and notice how differently they are keeping time.

On 13 October, in Bangkok, the International Monetary Fund published its autumn outlook. It carries the title Resilience under Strain, Urgent Choices. The Fund actually raised its forecast for world growth this year, to 3.1 percent, because an investment boom in artificial intelligence has been offsetting the damage from energy prices. In the same breath it warned that the room economies have to absorb the next shock is narrowing fast, and that just 32 countries now account for 95 percent of global growth, down from 54 in April.

In Berlin and Paris, meanwhile, parliaments spent the autumn arguing over budgets for 2027. Germany's draft sets total federal spending at €555.4 billion, of which €109.7 billion is core defence, and pays for part of that by trimming the federal contribution to pensions and cutting €1.8 billion from the statutory health insurance subsidy, which means contribution rates paid by workers and employers rise from 18.6 to 18.8 percent. France's draft seeks €43 billion in savings and new revenue to pull the deficit down to 5 percent of GDP; its debt interest bill is set to rise from €64.8 billion this year to €74.2 billion next; and pensions above €1,260 a month would no longer be fully indexed to inflation. Neither budget was written because of Hormuz. Both were shaped by the same pressures the strait has been adding to all year.

And somewhere beneath all of it, a slower number moved. On 1 January 2026, the European Union had 452.0 million inhabitants, about 706,000 more than a year earlier. Every bit of that increase came from migration. Within the EU's own borders there were 3.46 million births and 1.35 million more deaths than births. The share of people aged 65 and over, measured against those of working age, stood at 34.5 percent at the start of 2025, up from 29.0 percent ten years earlier. No single week moved that figure, and no ceasefire will move it back.

Four events, four different speeds. The insurance quote reprices in a day. The IMF forecast is revised twice a year. The budget sets spending for 2027 and signals it through 2030. The population numbers describe a shape that took decades to form and will take decades to change. They are all describing the same world, and none of them is on the same clock.

2 · The Mechanism: one event, four bills

The previous article in this series argued that a shock does not spread like water from a burst pipe; it behaves more like traffic after a bridge closes, where part of the flow stops, part reroutes and part is simply absorbed. This article is about the other half of that picture. Not where the cost goes, but when it arrives.

We find it useful to think of four clocks running at once.

The first clock: hours and days

Prices, insurance, sanctions, markets. This clock moves before any physical damage has been done, because it runs on expectation rather than on events. An attack on one tanker reprices cover for every tanker on that route by the next morning. It is the clock everyone watches, because it is the only one that produces a number by lunchtime.

The second clock: weeks and months

Freight, inventories, rerouting, contracts, the first movements of people. This is the clock on which a company changes supplier, a cargo takes the long way round, a government releases strategic reserves, a household notices the heating bill. The adjustments here are real work, done by named people, and they take the time they take.

The third clock: one to three years

Budgets, investment, taxes, factory locations, procurement. This clock looks slow from the outside and is anything but, because its decisions are being taken right now and cannot easily be unmade. The German and French budgets for 2027 are being settled this autumn. Capital spending for 2028 is being approved this quarter in thousands of companies. By the time the consequences are visible, the decisions are three years old.

The fourth clock: three to fifteen years

Demography, skills, institutions, infrastructure, the hardening of trade into blocs. Births that did not happen. Engineers who emigrated and settled. Maintenance deferred so often that replacement becomes the only option. A trained workforce that is older each year than the year before. This clock produces no headlines at all, which is precisely why it is the one most often missing from a plan.

The property that matters most is this: the four clocks do not stop together. A ceasefire stops the first one almost instantly. Prices fall, premiums ease, the story leaves the front page. The second clock keeps running for months while routes and inventories normalise. The third keeps running for years, because the budget that was passed under pressure stays passed, and the investment that was postponed stays postponed. The fourth never notices the ceasefire at all.

There is a second property, less obvious and more useful. The further out the clock, the less a shock can be seen and the more it is already decided. The insurance quote is visible and reversible. Next year's defence budget is semi-visible and difficult to reverse. The age structure of the workforce in 2040 is invisible today and, within a decade or two, essentially fixed. Attention runs in exactly the opposite direction to consequence.

Elena, the operations director at a mid-sized packaging plant in northern Italy whom we followed through the first article, lives on all four clocks at once. She is a composite, built to resemble the kind of company this series is written for rather than any real firm, and her year is a good map of the problem.

Her first clock struck in March, within days of the strait closing, when her gas contract repriced and most of the quarter's margin on two large contracts disappeared. Her second clock struck in June, when her freight forwarder added a war-risk surcharge to resin shipments from Asia although no ship carrying her cargo had been attacked, and again weeks later when the alternative supplier she had switched to in March — along with most of her competitors — pushed deliveries back by five weeks.

Her third clock is ticking right now, in October, in the least dramatic meeting of her year: the one where next year's capital spending is approved. The new extrusion line she postponed in the spring is on the agenda again. So is the question of whether to fund the two efficiency projects whose payback looked too slow in 2024 and looked very different after this winter. Nothing decided in that meeting will be visible to anyone outside the company for three years. Everything decided in it will shape what the company can do in 2029.

Her fourth clock has one hand on it and almost no one looks at it. Her most experienced machine operator retires in 2029. The two people she interviewed as possible successors moved abroad years ago and are not coming back. In September she took on an apprentice, which is the only move available to her on that clock, and the only one that must be made years before it is needed.

3 · Subtheme One: the fast clock is loud, and it misleads by being accurate

Start with the clock everyone watches, because the trouble with it is subtler than it looks. The fast numbers are not wrong. They are precisely right about a very small question, and they are read as though they answered a much larger one.

Insurance is the clearest case. War-risk cover for the Gulf went from about a quarter of a percent of a ship's hull value before the war to between seven and a half and ten percent at the October peak — on a hundred-million-dollar tanker, from roughly $250,000 to several million per voyage, with reports of quotes reaching $20 million at the worst moments. Those premiums are reviewed every day and written to cover seven-day voyages. They are the most honest measure available of how much risk the market believes is in the water this week.

What they do not tell you is how much oil is moving. And here two true numbers point in opposite directions, which is worth sitting with, because this kind of contradiction is where most bad decisions are made.

One number: on 11 October the IMF's shipping tracker recorded two transits of the strait, against a pre-war baseline of about 85 a day. By that measure the passage is shut. Another number, from the same month: a market analyst told reporters that Gulf oil flows had recovered to around 80 percent of pre-war levels, and Middle East oil exports as a whole had climbed back above where they stood before the war began in February.

Both are accurate. They count different things. The first counts ships through one strait; the second counts oil leaving the region by any route, including the pipeline across Saudi Arabia to the Red Sea and the long detour round the Horn of Africa that comes with it. The oil is moving. It is moving further, slower, and at a much higher cost per barrel delivered, carried by crews paid danger money and ships insured at forty times the old rate. Nothing about the recovery in export volumes means the shock has ended. It means the system has paid to route around it — which is the first article's argument, showing up as an accounting problem.

This matters for anyone who has to decide something. If you scored the crisis in March by the fast clock alone, you concluded the world was about to seize up, and you may well have locked in energy contracts at the peak. If you score it in October by the recovery in export volumes alone, you will conclude it is over, and you will plan 2027 on prices and premiums that have not actually come back. Both readings come from accurate numbers. Both are the wrong size of answer.

The IMF's October outlook is a careful attempt to hold both at once, and it is worth reading for its structure rather than its headline. Growth for 2026 was revised up to 3.1 percent, mainly because investment in artificial intelligence has been strong enough to offset much of the energy damage in the economies doing that investing. Inflation worldwide sits at 4.7 percent this year, easing to 4.1 percent next. Yet the Fund's own framing is that growth is being reallocated rather than simply raised or lowered: countries with favourable terms of trade or a stake in the technology boom are doing well; energy importers and lower-income economies are in a much harder place. Saudi Arabia, at the centre of the disruption, is forecast to contract by 3.3 percent this year.

That is the fast clock and the second clock disagreeing in public, and the disagreement is the information. An average that holds up while its components spread apart is not a reassuring number. It is a warning that the next shock will not land evenly.

4 · Subtheme Two: the slow clocks are where the decisions are already locked

The third clock is the one we think is most underpriced right now, because it is the clock where this year's emergency becomes next decade's fixed cost — quietly, legally, and with almost no coverage.

Consider what a European finance ministry has been doing this autumn. It is writing a budget for 2027 under three pressures that arrived from three different directions and converge on one page of arithmetic. Energy costs are elevated because of a war in the Gulf. Defence spending is rising because of a war on the EU's eastern border and a commitment, agreed in June 2025, to reach 5 percent of GDP on defence and security-related spending by 2035. And debt interest is climbing because yields rose and the stock of debt is larger than it was.

None of these is a surprise. All three land in the same place. Germany's 2027 draft puts €109.7 billion of a €555.4 billion federal budget into core defence, and finds part of the money by reducing the federal contribution to the pension system and cutting €1.8 billion from the health insurance subsidy, which pushes contribution rates from 18.6 to 18.8 percent — a transfer of the cost from the state to employers and employees. France's draft seeks €43 billion in savings and revenue measures to bring the deficit to 5 percent of GDP, in a year when debt interest alone rises from €64.8 billion to €74.2 billion, and proposes that pensions above €1,260 a month lose full indexation to inflation.

Note what kind of decision this is. It is not a forecast, a scenario, or a risk that may or may not materialise. It is a law being passed about 2027, which will be the base from which 2028 is written. By the time anyone experiences it as a slightly smaller pension increase or a slightly higher payroll deduction, the decision will be two years old and the argument will be over.

The public reaction tells you how this lands. In polling compiled by the Centre for European Reform, 68 percent of French respondents supported higher defence spending, while 50 percent opposed cutting public services to pay for it. That is not incoherence on the part of voters. It is an accurate statement of the problem: the spending is widely accepted, the trade-off is not, and governments are making the trade-off anyway because the arithmetic leaves no third option.

The same clock runs globally. The IMF's October fiscal report put global public debt at about 94 percent of world GDP in 2025 and on track for 100 percent by 2029 — a level last seen after the Second World War. Advanced-economy deficits are heading for 5.2 percent of GDP in 2026, the highest since 2021. In the United States, net interest on the federal debt reached roughly $1.04 trillion in fiscal 2026, close to a fifth of federal revenue. Interest is the purest third-clock cost there is: it buys nothing, it cannot be cancelled, and it grows while you argue about everything else.

The fourth clock, and the one move it allows

Then there is the clock that will still be running when every current headline is forgotten. In 2025 the EU recorded 3.46 million births and 1.35 million more deaths than births; the population grew only because 2.05 million more people arrived than left. The ratio of people over 65 to people of working age reached 34.5 percent at the start of 2025, having risen 5.5 points in a decade and risen in every single member state.

Nothing in that paragraph is news, which is the point. It has been true, in roughly this shape, for fifteen years. It is also the variable that decides how much of the third clock's bill can actually be paid, because pensions, health spending and defence are all claims on the output of people who are working. A decision taken in 2027 to raise contribution rates is an attempt to solve a fourth-clock problem with a third-clock instrument.

Here is the part worth taking away. The slow clocks are the only ones you can get ahead of. Nobody can position themselves in advance against an insurance repricing that happens overnight. Everyone can do something about a skills gap that arrives in 2029, a machine that needs replacing in 2028, a dependency that takes two years to unwind. The clock that gives the most warning gets the least attention, and it is the only one where early action is even possible. Elena's apprentice is a small decision and the only one in her year that could not have been made later.

5 · What most analysis gets wrong

Three mistakes recur, and they map onto the clocks almost exactly.

The first is scoring a crisis by its fastest number. Oil at $105, premiums at ten percent, transits at two a day: all true, all fast, and together they describe a week rather than a world. This is how March's forecasts of global recession were produced from genuine data.

The second is treating a ceasefire as a reset. When the strait reopens — and most of the pressure in the world is pushing toward reopening — the first clock will stop within days. Expect a round of commentary declaring the episode closed. The reserves drawn down this year will still need buying back. Insurers will wait several quarters before trusting the calm. The detour routes will still have owners and customers. The 2027 budgets will still have been passed. And the apprenticeships not started this year will still be missing their people in 2029.

The third is the quiet assumption that the slow clocks are not yet urgent. This is the most expensive of the three, because it inverts the actual deadline. A fourth-clock problem is not a problem that arrives later; it is a problem whose last useful decision point comes earlier. The time to hire the apprentice, start the second supplier, or fund the efficiency project is years before the consequence becomes visible. By the time it is visible, the useful options have gone and only the expensive ones remain.

There is a fourth error, more specialised, which this series keeps running into: analysing each crisis on its own clock as though no other crisis existed. Energy analysts follow Hormuz. Security analysts follow the eastern border and the sabotage incidents in European airports and undersea cables. Fiscal analysts follow deficits. Demographers follow birth rates. Every one of those analyses can be sound. But a finance ministry meets all four of them on a single page in October, and so does Elena's board. The interesting questions live in the overlaps: what happens when the energy bill and the security bill arrive in the same budget year, in a country whose working-age population is shrinking and whose debt interest is rising at the same time. That is not a hypothetical. That is the 2027 budget of more than one European government.

6 · Three paths from here

We see three broad paths for the next twelve months, in terms of how the clocks interact. The probabilities add to 100 and are our judgement, not the output of a model. Each comes with the observation that would show we are wrong.

BASE · 55%  ·  The fast clock stops, the slow ones do not.  The strait reopens unevenly over the coming months. Energy prices ease but stay above 2025 levels. Public attention moves on within weeks. Meanwhile the 2027 budgets pass roughly as drafted, defence rises, social spending is trimmed or shifted onto contributions, and debt interest keeps climbing. The lasting cost of 2026 settles into budgets and contracts rather than into headlines. Proved wrong if European defence commitments are visibly scaled back once energy prices fall.

STRESS · 35%  ·  A second shock arrives before the buffers are rebuilt.  The closure persists or a second disruption lands — another chokepoint, a serious infrastructure incident in Europe, a sharp market correction in the AI investment that is currently holding up the growth numbers. Reserves drawn this year are not yet refilled, fiscal room is thinner, and the response is slower and more expensive than the response to the first. Energy-importing economies with high debt feel it first. Proved wrong if world growth holds near 3 percent through 2027 while energy prices stay elevated.

RELIEF · 10%  ·  The slow clocks get easier.  A durable settlement in the Gulf, a meaningful de-escalation on Europe's eastern border, or a productivity surge from AI large enough to raise trend growth. Any of these would loosen the third clock directly, by making the same claims affordable out of a larger output. Proved wrong if defence budgets and debt interest both keep rising through 2028 regardless of how the conflicts resolve.

7 · Forecast, 1 year

Claim.  Euro area inflation for calendar 2027 will be above 2.5 percent.

Probability.  0.40  ·  Confidence: Low.

Reasoning.  The IMF expects global inflation to ease from 4.7 percent this year to 4.1 percent next, and the euro area usually runs well below the world figure. Against that: Brent near $105, war-risk costs embedded in freight, higher energy input costs working through industry with a lag of several quarters, and payroll contribution increases that raise labour costs directly. We lean below even odds, but by less than the ECB's target would suggest.

Second-order effect.  Inflation above 2.5 percent in 2027 would keep rates higher for longer, which raises debt interest in exactly the budgets already carrying new defence spending — the collision described in Article 6 of this series.

What would weaken it.  A reopening of the strait before the end of 2026 combined with a well-supplied oil market, or a sharper slowdown in European demand than currently forecast.

Verifier and date.  Eurostat, annual average HICP for the euro area, 2027. Resolves 28 February 2028.

8 · Forecast, 3 years

Claim.  European military expenditure in 2028 will be at least 30 percent above its 2025 level in real terms.

Probability.  0.65  ·  Confidence: Medium.

Reasoning.  This is a third-clock forecast, which is to say it is mostly about decisions already taken. European NATO members raised spending 14 percent in 2025, the fastest since the early 1950s. Germany's 2027 budget alone puts €109.7 billion into core defence on a path toward 3.5 percent of GDP by 2029, financed by borrowing. The main risks to the forecast are political rather than fiscal: budgets that fail in parliament, governments that fall, or a settlement that removes the reason for the spending.

Second-order effect.  If it holds, the displaced spending shows up as slower growth in social and public investment budgets across the same period, which is the trade-off European voters have said they dislike.

What would weaken it.  A durable settlement in Ukraine before the end of 2027, or a European fiscal crisis severe enough to force consolidation across all categories including defence.

Verifier and date.  SIPRI Military Expenditure Database, constant-price series for Europe, 2028 against 2025. Resolves 30 April 2029.

9 · Forecast, 5 years

Claim.  The EU-27 population on 1 January 2031 will be lower than on 1 January 2026.

Probability.  0.35  ·  Confidence: Low.

Reasoning.  Eurostat's own baseline projection had the EU population peaking around 2026 and declining thereafter, and the natural deficit is now running at roughly 1.35 million a year. But the actual record keeps defying the projection: 2025 brought a net 706,000 increase, because migration added 2.05 million. The forecast is therefore not really about births and deaths, which are highly predictable, but about migration, which is not — it depends on wars, labour demand and politics, all of which are moving. We lean against decline by 2031 while holding the view loosely.

Second-order effect.  Whichever way the number goes, the working-age share keeps falling. A population held flat by migration and a population in slow decline produce similar pressure on pensions and labour supply; they differ mainly in politics.

What would weaken it.  Two consecutive years of net migration above 2 million, or a formal upward revision of Eurostat's migration assumptions.

Verifier and date.  Eurostat demographic balance, EU-27 population on 1 January 2031. Resolves 31 July 2031.

10 · Forecast, 10 years

Claim.  The EU old-age dependency ratio will be at or above 40 percent in 2036.

Probability.  0.75  ·  Confidence: Medium.

Reasoning.  This is the slowest variable in the article and therefore the one where a ten-year judgement retains some value. The ratio stood at 34.5 percent on 1 January 2025, up from 29.0 percent a decade earlier — an average of about half a point a year, and it rose in every member state. Continue that pace and 2036 lands just above 40. The people who will be over 65 in 2036 are alive and in their fifties today; the people who will be of working age are mostly born. The uncertainty is in migration and in mortality, not in the shape.

Second-order effect.  A ratio above 40 means roughly two and a half people of working age for each person over 65, against three today. Every claim on the public budget — pensions, health, defence, debt interest — is drawn from that narrower base.

What would weaken it.  Sustained net migration well above two million a year for a decade, or a sharp and unexpected rise in deaths among the older population.

Verifier and date.  Eurostat, old-age dependency ratio for the EU on 1 January 2036. Resolves 31 July 2036.

11 · Signals to watch

Seven measurements, one or two per clock, that will tell you which path we are on before the commentary does.

War-risk premiums for Gulf and Red Sea voyages.  Quoted daily through brokers and visible in shipping press. The fastest clock there is, and the last thing to return to normal after a reopening.

Daily ship crossings at the Strait of Hormuz.  Published by the IMF's PortWatch service. Watch for a seven-day average held for a month, not a single good day.

Regional oil exports by all routes.  Reported by cargo-tracking firms. Read alongside the transit count, never instead of it; the gap between the two is the cost of the detour.

The energy component of euro area inflation.  Published monthly by Eurostat. This is where the Gulf reaches European household budgets, with a lag of several months.

The passage of the German and French 2027 budgets.  Watch what is cut rather than what is added. The cuts are the part that persists.

Debt interest as a share of government revenue.  National budget documents and the IMF's fiscal reporting. It is the clearest single measure of how much room a government has left for anything new.

Apprenticeship and vacancy data in skilled trades.  National statistics offices. The slowest signal on this list and the earliest warning available about what a company will be unable to do in 2030.

12 · What this means for individuals

Immediate.  Before the end of November, write down your main financial commitments and mark each one with the clock it runs on: what reprices within weeks, what is fixed for a year, what you are counting on in ten years. Most people find the list is unbalanced in a way they had not noticed.

Build.  Over the next six months, rebuild whatever buffer 2026 used up, whether savings, a fixed-price contract or simply slack in the month. The buffers you rebuild now are what absorbs the next shock.

Position.  Plan the second half of this decade on the assumption that public pensions and health systems will be under more pressure, not less, and that the adjustment will come as higher contributions and slower indexation rather than as a single announcement.

Watch (30 days).  The passage of your country's 2027 budget, specifically what gets trimmed to pay for what gets added.

Avoid.  Treating a fall in fuel prices as evidence that the costs of this year have gone away. The fast clock stopping tells you almost nothing about the other three.

Why this works.  It matches each decision to the speed at which it actually has to be made, which stops fast news from crowding out slow decisions.

13 · What this means for business

Immediate.  In this year's budget round — the meeting happening now, not the one in spring — separate the capital spending that protects you from a repeat of 2026 from the capital spending that merely grows the business, and fund the first category first. By 15 December, name which of your inputs still has no second route.

Build.  Start the slow things this quarter: apprenticeships, qualifying a second supplier, energy efficiency work. All three take one to three years and none can be accelerated later.

Position.  Model 2027 and 2028 with energy, insurance, financing and payroll contributions all higher together, rather than testing each one separately. They are rising for connected reasons and will arrive in the same budget year.

Watch (30 days).  Your war-risk surcharges and your forwarder's quoted transit times. Both move before freight rates and well before headlines.

Avoid.  Postponing the slow investments for a second year because the fast numbers improved. That is the decision that shows up in 2029 as an inability to compete rather than as a visible loss.

Why this works.  It puts the money where the deadline actually is, and it treats this autumn's budget meeting as what it is — the most consequential decision of the company's year.

14 · What this means for capital

Immediate.  Before year end, add one slow indicator to every country view you maintain: working-age population trend, public investment as a share of GDP, or debt interest as a share of revenue. Pick one and apply it consistently.

Build.  Over the next quarter, build a view of which governments are financing new defence spending by borrowing and which by reallocating. The two produce different sovereign paths and different domestic politics.

Position.  Separate holdings whose earnings depend on the fast clock, such as freight and energy trading, from those exposed to the third clock, such as public-sector suppliers and regulated utilities. The two will stop behaving alike as the crisis ages.

Watch (30 days).  Long-dated yields in high-debt European sovereigns during the budget votes. That is where the third clock prices itself.

Avoid.  Reading a stable global growth number as a stable world. The IMF's own framing is that growth is being reallocated, with 32 countries now producing 95 percent of it. This is analysis, not investment advice.

Why this works.  It prices the part of the crisis that persists after the headlines stop, which is the part markets reprice last and therefore the part most often mispriced.

15 · What would change our mind

Three observations would make us revise the central claim of this article — that the lasting cost of a shock is set on the clocks nobody watches.

First, if European defence commitments were visibly scaled back once energy prices fell, the third clock would be far more reversible than we argue, and the budget decisions of this autumn would be less binding than they look. Second, if post-crisis economies consistently recovered their pre-crisis investment and demographic paths within five years, the fourth clock would be a weaker force than the historical record suggests, and our emphasis on it would be misplaced. Third, if productivity growth in Europe rose by a full percentage point on a sustained basis, the entire stack of claims we describe would become affordable out of a larger output, and the trade-offs being written into budgets this autumn would quietly dissolve. We consider each possible. We consider none of them likely enough to plan around.

16 · Bottom line

The most expensive decisions of this crisis are not being taken in the Gulf. They are being taken this autumn, in budget committees and board meetings, by people who will not be in the room when the consequences arrive. The fast clock is loud, accurate and nearly finished. The slow ones have barely started, and they are the only ones you can still get ahead of.

Forecast record

Four forecasts, one per horizon, recorded before the outcome is known, each with a threshold, a named verifier and a date. They are published so that they can be checked against us later, including the ones we get wrong.

A caution we record rather than omit. Three of these four resolve against the same institution, and the last two share a parent cause in the EU's age structure. They are not four independent observations, and we count them accordingly.

Sources and the status of every figure

Every number carries a class. Measured means counted or recorded. Projected means an official forecast. Reported means stated by a credible source but not independently checked by us. Derived means calculated by us from the figures cited.

The series

Next week, Article 3: why a cheap disruption can carry an expensive consequence, and what insurance has to do with it.