What Is the Use of Gold
Financial Core, Industrial Applications, and Strategic Reserve

Gold is golden because of relativistic effects on its inner electrons. The 6s orbital contracts under relativistic mass increase at about 58% of light speed, shifting the 5d-to-6s transition into the blue absorption range. The surface reflects yellow and red. Without the relativistic correction, gold would be silver-white. The same physics makes mercury liquid at room temperature. The color is computable from first principles, not a surface accident.

The physical properties: corrosion-resistant to the point where only aqua regia dissolves it, ductile and malleable beyond any other metal, a good electrical conductor, a near-perfect infrared reflector, and biologically inert. Silver beats it on conductivity but tarnishes. Copper is cheaper but corrodes. Platinum resists everything but was unworkable before industrial furnaces because it melts at 1,768°C. Gold melts at 1,064°C, within range of ancient charcoal forges. The combination of adequate performance across all five categories with no disqualifying flaw is what made gold useful to humans in the first place, and the one property that matters most to everything discussed below is chemical inertness. A metal that does not degrade over millennia is a metal that can store value across generations.

This is what gold is actually for, in economic terms. Everything else combined accounts for less than a tenth of global demand and has negligible effect on the price. The gold price is a function of the global credit system, not of how many nanometers of coating the next space telescope needs.

Gold satisfied the six requirements for sound money simultaneously: scarce enough (roughly 0.004 ppm in the crust), divisible, portable at high density, indestructible under normal conditions, uniform in composition when refined, and extremely difficult to counterfeit. Total quantity ever mined sits around 210,000 tonnes, which cast into a cube would measure about 22 meters on each side. That volume turned out to be in a narrow usable range for a monetary metal. Much less and it could not have circulated widely enough to serve as a trade medium across continents. Much more and nobody would have valued it.

Newton pegged the pound to gold in 1717 while running the Royal Mint. Bretton Woods in 1944 fixed the dollar at 35 per ounce. Nixon closed the gold window in 1971. The basic chronology is everywhere. What is not everywhere is what happened inside the gold reserve system after the peg broke, and that is where the interesting material sits.

Central banks hold approximately 36,000 tonnes of gold reserves globally. The number appears in IMF data and gets repeated in every gold overview article as though it were a hard fact. It is not. It is a reporting artifact, and the gap between what is reported and what is actually sitting in vaults accessible on short notice may be significant.

GATA raised this publicly starting around 2000. Peter Warburton, who had advised the Bank of England, laid out the dynamics of central bank gold market intervention via derivatives in his 2001 paper "The Debasement of World Currency." Kevin Crisp, who headed precious metals at Deutsche Bank, spoke at the 2019 LBMA conference about the scale of leasing activity. Neither the IMF nor any major central bank has responded by changing the disclosure framework. The question of how much of the reported 36,000 tonnes can actually be produced on demand remains unanswerable from public information. GATA's position is that the discrepancy is substantial. Central banks have not engaged with the claim.

The practical consequence is specific. Gold reserves exist to provide an asset that can be deployed when credit markets freeze and currency arrangements come under stress. That is their entire strategic purpose. If a central bank's reserves are partially composed of leased gold that has entered the market and is no longer in the vault, then the moment a crisis makes those reserves operationally relevant, the bank's options are reduced. Instead of drawing down physical metal, it must acquire metal from a market that is likely experiencing exactly the kind of dislocation that triggered the need in the first place. The buffer function degrades precisely when it should activate. Whether this degradation is marginal or severe depends on the leasing fraction, which nobody outside the central banks knows.

Fort Knox holds the United States' declared gold reserve of approximately 4,580 tonnes. The last full independent audit was conducted in 1953. In 1974 a group of members of Congress and journalists were permitted into a small number of vault compartments. Ron Paul spent years in the House pushing legislation for a comprehensive, independently supervised audit. The bills did not advance. Over seventy years have passed without a third-party verification of what is stored in the facility that holds the largest single sovereign gold stockpile on the planet.

Germany's Bundesbank kept a substantial share of its gold reserves in vaults belonging to the New York Fed, the Bank of England, and the Banque de France, a legacy of Cold War arrangements. In January 2013 it announced a repatriation plan: bring 674 tonnes back to Frankfurt by 2020. The transfer finished three years early, in 2017. The Bundesbank described the move in administrative language. The gold market read it differently. Moving hundreds of tonnes of gold across an ocean costs money, creates logistical risk, and requires years of planning. Nobody does that as a routine housekeeping exercise. The decision to repatriate physical metal from allied vaults signals a judgment about the reliability of custodial arrangements, and the fact that Germany accelerated the timeline reinforces that reading.

After 2022, annual gold purchases by central banks broke sharply above the trend of the prior decade. The buyers were concentrated among non-Western institutions. The People's Bank of China added to reserves over eighteen consecutive months. Poland's central bank became one of the largest European buyers. Turkey and India bought steadily. In the same period, Russia, China, Brazil, and several Gulf states expanded bilateral trade settlement in currencies other than the dollar, and BRICS discussions explicitly addressed reducing dollar dependence in reserve composition.

Gold pays no yield. Holding it means forgoing the coupon on a US Treasury bond. When the dollar-based credit system is perceived as stable and US sovereign debt is treated as the global risk-free asset, there is no portfolio optimization rationale for a central bank to hold large and growing quantities of gold. The opportunity cost exceeds the benefit. The only conditions under which aggressive gold accumulation makes sense for a central bank are conditions in which the manager believes the dollar-denominated system carries increasing long-term risk that is not being priced into Treasury yields. Buying gold is the revealed preference. It says more than any communiqué.

For twenty years the standard model of gold pricing in institutional finance was built on real interest rates. Nominal rate minus inflation gives the real rate. When the real rate is high, holding gold is expensive in opportunity-cost terms, and the price tends to fall. When the real rate is zero or negative, the opportunity cost vanishes and the gold price tends to rise. From 2000 through 2020 this framework tracked the gold price reasonably well and formed the basis of most macro hedge fund gold positioning and most sell-side gold forecasting models.

The variable that changed is the composition of marginal demand. When the marginal buyer of gold is a macro fund in Connecticut or a CTA in London, that buyer responds to rate differentials, momentum signals, positioning data, and Fed forward guidance. Gold behaves like a financial asset driven by opportunity cost. When the marginal buyer shifts to the People's Bank of China or the National Bank of Poland, the decision process is entirely different. These institutions are not optimizing quarterly returns. They are making a multi-year allocation based on a geopolitical and monetary assessment that has nothing to do with the next FOMC meeting. They do not trade around data releases. They do not hedge with options. They do not liquidate on a 2% drawdown. The supply of gold is relatively fixed in the short run, so when this type of persistent, price-insensitive, large-scale buying enters the market, it overwhelms the rate signal. The old model does not break because the math changed. It breaks because the buyer changed.

For anyone running a gold book based on the real-rate model, the post-2022 period has been confusing at best and costly at worst. The structural implication is that the locus of gold price formation has migrated away from Western financial institutions and toward sovereign reserve managers in countries that are collectively running the largest current-account surpluses in the world. These buyers are not going to explain their strategy on an earnings call. Their purchase data shows up in customs statistics and IMF reserve reports with a lag, often revised. The market has to infer rather than observe. This opacity itself is a feature of the new regime and a source of model risk for anyone still pricing gold the old way.

LBMA sets the global benchmark through an electronic auction run by ICE, twice a day. Before the 2015 overhaul, five banks negotiated the price by phone conference. In 2014 it emerged that at least some participants had been exploiting advance knowledge of the fix direction to position in the derivatives market. The switch to electronic auction increased transparency, though the OTC structure of the London market remains opaque compared with exchange-traded markets.

The Shanghai Gold Exchange often trades at a premium to London. When the spread exceeds 15 dollars per ounce and persists, it indicates that Chinese physical demand is running ahead of available supply flowing into the domestic market. Since 2023 the Shanghai premium has hit unusually elevated levels repeatedly, and each spike has roughly coincided with a leg higher in the international gold price. Whether Shanghai demand is a leading indicator or simply a coincident one is debated. The correlation in the recent period is tight enough to be worth watching.

Total daily turnover in the London OTC gold market, the core of global gold trading, expressed as physical-equivalent volume, routinely runs in the thousands of tonnes. Global mine production for an entire year is about 3,500 tonnes. The ratio speaks for itself. The overwhelming majority of daily gold "trading" is the exchange of claims on gold between financial counterparties, not the movement of metal. The system is a fractional-reserve structure applied to a commodity. It functions smoothly as long as the fraction of participants demanding physical delivery at any given moment remains small relative to the total paper outstanding. If that fraction spikes, the system cannot perform. March 2020 gave a partial demonstration of what the early stage of such a failure looks like.

ETFs hold gold on behalf of investors. The aggregate claimed gold held by all gold-backed ETFs globally substantially exceeds the quantity of eligible metal that could be delivered against those claims on short notice. For an investor using an ETF to trade the gold price, this is largely irrelevant because the intent is never to take delivery. For someone who believes the purpose of gold is to hold an asset that does not depend on the solvency of financial intermediaries, the ETF achieves nothing. The bar in a vault you control and the ETF share in your brokerage account are completely different instruments solving completely different problems, and conflating them is one of the more common errors in retail gold commentary.

The electronics industry consumes several hundred tonnes of gold a year, mostly as connector plating and bonding wire. Gold connectors do not oxidize, which keeps contact resistance stable over the life of a device. Aerospace and military specifications require them. Consumer electronics have shifted toward copper wire for cost reasons. E-waste contains gold at concentrations vastly higher than ore, 150 to 400 grams per tonne of circuit boards versus 1 to 5 grams per tonne of mined rock, but most of it is landfilled.

The Webb telescope carries 48.25 grams of gold coating across its mirror segments for infrared reflectivity. Satellites use gold-coated polyimide thermal blankets. Nanoscale gold particles discovered in 1987 by Masatake Haruta to be catalytically active have applications in exhaust treatment and CO removal in enclosed environments. Colloidal gold is the reason the line turns red on a pregnancy test or a COVID rapid test. Gold compounds were once mainstream in rheumatoid arthritis treatment and are now being reexamined for oncology applications at an early stage. Photothermal therapy using gold nanorods is in clinical trials, aiming to ablate tumors via near-infrared heating. Dental gold alloys are durable but losing share to ceramics.

Quantum computing devices use gold in wiring and packaging. The NIF fusion experiment uses a gold-bearing hohlraum. The oceans hold an estimated 20 million tonnes of dissolved gold at about 13 parts per trillion, economically inaccessible by any known method; Fritz Haber tried after World War I and failed. NASA's Psyche mission launched in 2023 toward a metallic asteroid.

All of this is interesting and none of it matters to the gold price. If every industrial and medical consumer of gold disappeared tomorrow, the price per ounce would not change by a meaningful percentage. The demand that determines the price comes from central banks, institutional and retail investors, and the savings component embedded in jewelry purchases, particularly in India and China.

The jewelry category absorbs about half of annual gold supply. The label is misleading. In India, gold jewelry functions as household savings, as a portable and concealable store of wealth, and as the primary financial asset available to rural women. Indian households hold an estimated 25,000-plus tonnes. When the Modi government's 2016 demonetization order voided 86% of circulating banknote denominations overnight, anyone whose savings were in cash lost access to most of their money. Anyone whose savings were in gold did not. Chinese consumers buying small gold beans in recent years are responding to declining deposit rates and limited attractive alternatives for household savings. In Western markets, gold jewelry is closer to a consumption good and further from a financial instrument.

Roughly 15 to 20% of gold mining still uses mercury amalgamation, releasing thousands of tonnes of mercury per year, concentrated in artisanal operations in the Amazon and sub-Saharan Africa. The Minamata Convention addresses this. About 20 tonnes of waste rock are generated per 6 grams of gold produced.

Searching "What is the use of gold" returns a list that treats jewelry, investment, electronics, aerospace, and medicine as coordinate categories. They are not. Gold is a financial asset. It has some minor industrial applications. The industrial tail does not wag the financial dog.

Gold's function for the past five thousand years and at present is to serve as a store of value and potential medium of exchange that requires no counterparty to honor a promise. In a stable credit environment, this function looks redundant. In an unstable one, it is the only function that survives when others fail. Holding gold is a position on probability: the probability that the monetary and credit system develops a serious problem during the holding period. The five-thousand-year record suggests this probability, over a sufficiently long horizon, approaches certainty. The timing is unknowable. The direction is not.