Stockpiling for the last crisis

Posted on: 12 November 2026

When Baroness Hallett published the first report of the UK Covid-19 Inquiry in July 2024, its most quoted finding was not about negligence but about aim. The country, she concluded, had prepared for the wrong pandemic. Years of planning and stockpiling had been built around influenza, the threat everyone remembered, while the one that actually arrived came from somewhere else. The finding is worth keeping in mind this week, because the same error has just surfaced in a corner of industry that most boardrooms have never had a reason to think about.

On Thursday Nikkei Asia reported that yttrium traded in Europe at $1,087 a kilogram in September, according to Argus, the London price reporting agency, which is nine times its level a year earlier. The annual comparison flatters the story. At the end of 2024 the same material cost less than $8, so in under two years its price has risen more than a hundredfold. Japanese customs figures show the human side of the number: in the first nine months of this year Japan imported 294 tonnes of yttrium oxide, 70% less than a year earlier, paying a unit price in September 64 times higher than the year before. Beijing placed yttrium and six other rare earths under export licensing in April 2025 in retaliation for American tariffs, and in January it tightened controls on dual use goods bound for Japan. When the Trump and Xi truce of last autumn led China to suspend the broader restrictions it had announced in October, the curbs on yttrium were left exactly where they were.

Few people outside materials science could tell you what yttrium does, which is rather the point. It works in tiny quantities inside objects far larger and more expensive than itself. A thin ceramic layer stabilised with yttrium allows the turbine blades of a jet engine, including the ones assembled in Derby, to operate in gas hotter than the metal underneath could survive on its own, while coatings of yttrium oxide line the chambers in which chipmakers etch circuits into silicon so that the walls are not eaten away by the plasma. Remove that film and the machine either wears out quickly or can no longer run at the temperature it was designed for.

The obvious explanation for the scramble is that yttrium was cheap and therefore nobody worried about it. Stated like that it does not quite add up. Cheap things are the easiest of all to hold in reserve, and for a company coating turbine blades a year's supply of yttrium kept on the shelf would have cost less than a few days of a consultant's time. If nobody bought ahead, it was not to save money. It was because nobody was looking.

The blind spot has a name, and it is more than forty years old. In 1983 Peter Kraljic, then a consultant at McKinsey, published an article in the Harvard Business Review under the title "Purchasing must become supply management" that went on to become a standard text for anyone running a procurement department. He proposed sorting every purchased item by two measures, how much it weighs on total spend and how risky it is to obtain, and he gave the corner where spend is low but risk is high a name that now reads like a forecast: bottleneck items. Kraljic was explicit that these deserve attention out of all proportion to their cost. The theory was there all along. What was missing was anyone with an incentive to apply it.

Procurement teams are judged on the savings they deliver, while finance directors treat idle stock as working capital to be squeezed out of the balance sheet. Inside that arrangement no buyer has ever been promoted for laying in three years of a material worth a few thousand pounds that then never ran short, because the value of a risk avoided appears nowhere in the accounts. I have watched this happen in more companies than I would care to list. Risk gets ranked by cost because cost is the only thing the ledger shows, so the same spreadsheet that overlooks a trivial line of spend also overlooks the danger hiding inside it.

There is a second mechanism, quieter still, which is that the low price was never a fact of nature. Yttrium is a by-product. It comes out of the same deposits and the same separation plants that process the heavy rare earths, concentrated mostly in southern China, which means its supply is set by the economics of other elements rather than by its own price. According to the US Geological Survey more than 90% of America's yttrium imports between 2020 and 2023 came from China, and at six or eight dollars a kilogram nobody elsewhere had any reason to separate it. Cheapness therefore did two kinds of damage at once. It took away the customer's motive to stockpile and it took away any competitor's motive to build an alternative.

The pattern becomes clearer from the other side, in a case where stockpiling worked. The earthquake and tsunami of 11 March 2011, which also triggered the Fukushima nuclear disaster, cut Toyota's supply lines and kept its plants below normal output for months. The company that had given the world just in time manufacturing, the discipline of holding almost nothing and having parts arrive at the moment they are needed, then made an exception to its own creed. It drew up a list of 500 priority items and required suppliers to hold between two and six months of chips in reserve. A decade later, as the global semiconductor shortage halted assembly lines across the industry, Toyota carried on building cars almost without interruption.

Japan also knew the rare earth problem at first hand, and still it has been caught. In September 2010, after a Chinese trawler collided with two Japanese coastguard vessels near the Senkaku Islands, which both countries claim, shipments of rare earths from China to Japan stopped. Tokyo responded seriously, backing the Australian miner Lynas together with the trading house Sojitz, and Lynas is today the leading producer of separated rare earths outside China. Yet the effort concentrated on the elements that were making headlines at the time, the ones that go into the magnets of electric motors, while yttrium stayed in Kraljic's corner where nobody reads.

The two stories make the same point from opposite ends, and it is Baroness Hallett's point. Reserves form around a trauma, which is a healthy reflex right up to the moment it leads an organisation to defend itself against the last crisis rather than the next one. Whoever wants leverage reasons the other way round and looks for the pressure point the other side does not remember having. Yttrium also suits Beijing for a practical reason, since giving up those exports costs it almost nothing in revenue while downstream they reach into chips and aero engines. A threat that is cheap to maintain is also the most credible one, which is why it stays in place while the others are traded away at the negotiating table.

What follows is more useful than any list of best practice, because it concerns what not to do. Measuring the risk of an input by what you spend on it is the original mistake, since the two quantities have no relationship whatsoever. Filling the warehouse with whatever ran short last time and calling it resilience is the second. The exercise that actually helps is to take the full bill of materials for a product and rank it not by cost but by the number of days the line would stop if each item disappeared. The top of that list rarely matches the top of the spend report, which is exactly why it is worth reading before somebody else does. Anyone looking to hurt you will not start with the largest line in your accounts but with the one nobody has ever read.