A predicted catastrophe that never arrived proves the warning was a hoax and the money spent preventing it was a scam

Why the suspicion is not stupid

The strongest thing that can be said for this reasoning is that the evidence genuinely is absent. This is not a case of people refusing to look. There is nothing to look at. You are being asked to accept that a catastrophe was real on the strength of the work that stopped it, and the work is technical, dispersed and mostly invisible, while the bill is a single large number.

Institutions also do overstate risks, and people who noticed that are not being paranoid. Anyone who has watched an official warning inflate has a reasonable prior, and this argument feels like applying it.

And there is one case in this family where the sceptics landed a real hit. Not every country spent heavily on Y2K remediation. Italy, South Korea and Russia allocated relatively modest budgets and came through with few problems too. That is genuine evidence that the scale of spending in some countries exceeded what the problem required, and no honest treatment can skip it. Note what it argues, though. It is an argument about proportionality, not about whether the bug was real, and it was reached by comparing responses rather than by pointing at the absence of a disaster.

The one case where the counterfactual can be measured

The usual objection to this whole argument is that it cannot be settled: you cannot rerun 2000 without the remediation. Ozone is the exception, and it is worth dwelling on because it is the closest thing to a controlled result this subject offers.

The Antarctic ozone hole was identified in 1985. The Montreal Protocol was signed in 1987. And then the hole kept growing, for nineteen more years: 22.4 million square kilometres in 1987, peaking at 29.6 million in 2006. For two decades, anyone wanting to argue the treaty had accomplished nothing had an excellent-looking chart to do it with.

In 2018 NASA published a study that did not rely on the trend line at all. Instead of asking whether the hole was shrinking, it measured the chemical composition inside the hole and asked why. The decline in depletion tracked the decline in chlorine from CFCs, not weather variation. That is direct attribution: the intervention was linked to the recovery by mechanism rather than by correlation.

The lesson generalises even though the measurement does not. Where the mechanism can be checked, the prevention holds up. The reason this argument keeps winning elsewhere is not that the interventions are weaker; it is that most of them cannot be instrumented the way the stratosphere can.

August 2026: the same move, eleven months after the warning

The most recent instance compresses the whole pattern into under a year. In August 2025 the IPC confirmed famine in Gaza Governorate, and four UN agencies jointly called for an immediate ceasefire and an unimpeded, large-scale humanitarian response. Nutrition coverage for children under five stood below 1 per cent.

A ceasefire came in October. Assistance scaled up. By May 2026 nutrition support reached 60 per cent of children under five, and a survey in July recorded acute malnutrition between 0.2 and 0.8 per cent by governorate. In August those figures were circulated as proof that the famine had been fabricated.

So: agencies asked for an intervention, the intervention was made, the intervention worked, and the evidence that it worked was produced eleven months later as proof that the request had been dishonest. UNICEF answered it plainly, saying both findings stand and the improvement is a product of the assistance delivered in between rather than evidence it was unnecessary. Our file on the UNICEF famine numbers takes that case on its own terms.

What is new here is the speed. In the classic examples the gap between warning and dismissal runs to decades, long enough that the people arguing have forgotten the intervention or never knew about it. This one turned around inside a year, with the paperwork still in public view.

Why this particular shortcut is so hard to shake

Part of it is simple availability. The warning was loud and public; the response was quiet and technical. What people retain is a memory of alarm followed by a memory of nothing, and the comparison between those two memories is devastating to the alarm.

Part of it is the lag. Interventions in slow systems look useless for as long as the system takes to turn, which for ozone was almost twenty years and for vaccination is a generation. During that stretch, the evidence available to an ordinary observer genuinely does point the wrong way.

Part of it is that unfamiliarity is self-reinforcing. Measles vaccination averted an estimated 59 million deaths between 2000 and 2024. The result is that most people in high-coverage countries have never seen measles, and never seeing it makes the vaccine look disproportionate. The intervention manufactures the very ignorance that argues against it. WHO still recorded around 95,000 measles deaths in 2024, and first-dose coverage in 2025 was 84 per cent, below where it stood in 2019.

And part of it is that the conclusion is flattering. Deciding the panic was manufactured puts you among the people who were not fooled. It costs nothing, it requires no technical engagement, and it is available immediately. Believing the prevention worked requires accepting that you were right to be worried, which most people find less appealing than having been right to be sceptical.