Gold reached $5,608 an ounce in January. On October 19 it changed hands near $4,112, a decline of about 27 percent from that high. Silver peaked at $121.62 and traded close to $56, a loss of roughly half the value in nine months.
Those are the numbers on the screen, measured from the January peak to the price quoted this week. What makes them strange is the other set, the one that describes physical metal.
Central banks bought 288.9 tonnes of gold in the second quarter, the largest purchase for any second quarter on record. Silver has run a supply deficit for six consecutive years, and the shortfall is projected to widen to 46.3 million ounces in 2026. Fabricators are paying premiums for prompt delivery. The people who own metal are not the ones selling it.
So the price of the metal fell by roughly a quarter while the buyers of the metal kept buying. Both facts can hold at once because they describe two different markets, and only one of them sets the price on any given afternoon.
The market that sets the price is the paper one: futures, options, exchange-traded funds, and the positions of funds that own gold as a macro trade rather than as an object in a vault. When that trade gets crowded and the price breaks, the exit is narrow. Selling produces more selling, and margin calls finish the job. None of it requires anyone to revise a view about mine supply.
Silver behaves worse in these episodes because it is the smaller market. It is thinner, it carries industrial demand from solar manufacturing and electronics that turns with the business cycle, and it attracts retail money that arrives after the story is on television. A metal that outruns gold on the way up also outruns it on the way down, and it does so without any new information about how much metal exists.
The case for more downside follows directly. If the decline is a positioning event, then positioning has to finish clearing, and crowded trades rarely clear in one move. A fall of 27 percent from a January peak is large enough that many holders are now waiting for confirmation rather than for a bottom. The enthusiasm that produced the January high has to be replaced by something more patient before a durable low forms, and patience is the one input that cannot be bought.
The case for the other side rests on the numbers at the top of this article. Central bank purchases are not momentum money. They reflect reserve policy in countries that want less exposure to a monetary system they do not control, and reserve managers do not stop buying because the price fell. They are more likely to accelerate, since the same budget buys more metal.
There is a complication in that argument, and it is the most useful thing to understand about this market. The deficit is not a new discovery. It has run for six years, and for most of that stretch the price behaved as if it did not exist. A shortfall that persists without producing a shortage in the market that sets the price is evidence that the two are connected more loosely than the supply-and-demand story suggests. Metal sits off market. Inventory data is partial and reported by parties with an interest in the number. The deficit is real. Its transmission into price is not mechanical.
Which is why the physical indicators worth watching are the ones that measure delay rather than demand. Lease rates that spike when metal is hard to borrow. Premiums for immediate delivery against later delivery. Queues at the London and New York settlement systems. These show stress in the plumbing, and plumbing is where a deficit becomes a price.
The macro frame matters as much. The January high was built on a set of assumptions: currency debasement, falling policy rates, fiscal deterioration, and demand for an asset that pays nothing in a world where the alternatives were crowded. If two of those assumptions weaken, gold reprices whether or not a central bank in Asia adds another twenty tonnes. A 27 percent drawdown is consistent with a change in the macro story. It does not require a change in the physical story.
The two explanations lead to different conclusions, and the market will not tell you which one is operating until afterwards. If the fall was positioning, the physical buyers are early and will be proved right. If the fall was the macro trade unwinding, the physical buyers are simply holding an asset whose price is set by someone else, and their purchases will look less like conviction than like a slow accumulation of a depreciating reserve.
What is not in dispute is the speed. Gold took nine months to fall 27 percent, and silver took the same nine months to lose half its value. Moves of that size in markets this liquid are usually about who is forced to sell, not about what the metal is worth. The forced sellers are nearly done. The question that remains is what the buyers do when there are no more of them to absorb.
