
Main Street’s trading floor
Small investors now account for a third of share trading on busy days. They have done better than the professionals expected
Bullion has fallen by more than a quarter from its January peak, even as central banks buy at the fastest pace in two years

If you had been told in January that 2026 would bring a war in the Persian Gulf, an oil shock, inflation above 3% and a downgrade-scarred American government borrowing at its highest rates in a generation, you might have bought gold. You would have lost money. Bullion peaked at about $5,608 an ounce in January. On October 14th it settled near $4,112, roughly 27% lower, and was trading at about $4,180 two days later.
That is a strange result for an asset whose whole appeal is protection against exactly this kind of year. The explanation lies partly in who has been selling and who has been buying.
Central banks have kept buying. Preliminary figures from the World Gold Council (WGC), an industry body, suggest they bought about 63 tonnes in September, the most in a month since November 2024, after 39 tonnes in August. In a WGC survey, 89% of central banks said they expected global gold reserves to rise. Their motives are well understood: since Western governments froze Russia’s reserves in 2022, many central banks, especially in emerging markets, have wanted assets that sit outside the reach of foreign courts and payment systems.
The data should be treated with care. The WGC sharply revised down its estimate of central-bank purchases in the first quarter, from 244 tonnes to about 57, and different reports give quite different totals for the second quarter. Much official buying is never publicly reported.
Investors in exchange-traded funds have been selling. Such funds shed about 45 tonnes in the second quarter, according to one industry tally, even as central banks bought several times as much. The likeliest reason is the rising cost of holding an asset that pays no interest. With the Federal Reserve raising rates and ten-year Treasuries yielding more than 5%, the income an investor gives up by holding gold instead of bonds is the highest in more than two decades.
Gold’s reputation as an inflation hedge has always been stronger than the evidence. Elroy Dimson, Paul Marsh and Mike Staunton of the London Business School, who have studied returns since 1900, find that in the 28 years in which American inflation exceeded 3%, gold produced negative returns in 13. Over the very long run it has beaten inflation, but not by much: the real price of gold rose 5.2-fold between 1900 and 2025, an annual real return of about 1.3%. American shares returned 6.6% a year after inflation over the same period, and even government bonds managed 1.6%.
Source: Dimson, Marsh and Staunton, UBS Global Investment Returns Yearbook 2026
The picture looks better since 1971, when America ended the convertibility of dollars into gold. In the 54 years since, gold has returned about 4.7% a year after inflation for American investors. But almost all of that came in a few dramatic bursts, such as the late 1970s, the decade after 2001 and the surge to January’s peak, separated by long periods of decline.
Some forecasters still expect new records within two years, citing steady central-bank demand. That may happen. But the lesson of 2026 is that gold responds to real interest rates at least as much as to fear. In a year of war, gold fell, because the Fed raised rates and bond yields climbed. For investors looking for insurance against a bad year, gold offers some protection, but not the reliable kind its reputation suggests.