Is Gold an Inflation Hedge? What It Actually Tracks

Everyone is telling you to buy gold. Meanwhile, some of the largest pools of institutional capital on the planet hold none of it.

Norway’s sovereign wealth fund, one of the biggest single investors in the world, holds no gold at all. And not because the manager decided against it. Its mandate defines the investment universe as listed equities, fixed income, unlisted real estate and renewable energy infrastructure. Gold is simply not in it.

And yet gold has just been through one of the strongest rallies in its modern history. Somebody here is wrong.

The answer is more interesting than either camp will tell you, because almost everything retail investors believe about gold is false, including the main thing.

The main thing: gold is not an inflation hedge

This is the single most repeated claim about gold, and it does not survive contact with the data.

2022 was the cleanest experiment we are ever likely to get. US consumer price inflation peaked at 9.1%, the highest in four decades. Gold’s performance for that year was flat. Not up 9%. Not up at all.

2022	Return
US CPI (peak)	+9.1%
Gold	≈ 0%
S&P 500	−18%
US bonds	−13%

This is not a one-off. Erb and Harvey’s paper “The Golden Dilemma” examined the relationship formally and found that, over the horizons that actually matter to an investor, the link between gold and consumer prices is close to nothing.

Gold may preserve purchasing power over centuries. That is a statement about metallurgy and human psychology, not about your portfolio between now and retirement. If you want to see what inflation actually does to a sum of money over a working life, our inflation calculator will show you, and it does not need gold to make the point.

What gold actually tracks: real interest rates

Here is the mechanism, and once you see it you cannot unsee it.

Gold pays you nothing. No coupon. No dividend. No earnings. It sits in a vault and costs money to guard. So the relevant question is never “what is gold worth?” It is: what am I giving up by holding it?

That is the real interest rate, the yield on a safe government bond minus inflation.

• Real rates high and positive → a safe bond pays you a genuine return → holding a metal that yields nothing is expensive → gold struggles
• Real rates low or negative → the safe alternative pays you nothing either → gold costs you nothing to hold → gold does well

Historically the correlation has been strongly negative, and this one relationship explains most of gold’s behaviour across most of its history. It is also the reason the last two years have been so strange. Real yields have been positive. On the old model, gold should have been suffering. Instead it exploded higher.

Because something new is in the market: central banks

From 2022 to 2024, central banks bought gold at a pace far above anything in the previous decade, each year exceeding 1,000 tonnes.

Central bank gold purchases

Tonnes bought, and what "a slowdown" means

Average year, 2010–2021 473 t

The world before the shift.

2024 1,092 t

The peak of the buying.

2025 863 t

21% below 2024 — and 82% above the old average.

Both headlines are true. "Central bank buying fell 21%" is true. "Central banks are buying at nearly twice the old rate" is also true. Which one you were shown depends on who was showing you.

What matters for the price is neither headline. It is that this buyer does not care what gold does next quarter. India, Poland, Turkey, China and others are moving reserves out of dollars for reasons that have nothing to do with valuation, and a buyer who is indifferent to price behaves like a floor rather than like a market participant.

In 2025 they bought 863 tonnes, and here is where you should be careful, because this is the number gold commentary quietly rounds up. That figure is 21% lower than 2024, which came in at 1,092 tonnes. The buying has slowed.

But look at what it slowed to. The average annual central bank purchase between 2010 and 2021 was 473 tonnes. They are still buying at nearly twice the pace of the entire previous decade.

And they are not trading. India, Poland, Turkey, China and others are diversifying away from dollar reserves, a strategic decision made by institutions that do not care what the price does next quarter. That is persistent, price-insensitive demand, and it has put a structural floor under gold that the old real-rate model was never built to capture. The old model is not wrong. It is now incomplete.

The warning nobody who is selling you gold will mention

January 1980. Gold peaks at around $850 an ounce. It took roughly 28 years to get back to that price in nominal terms.

And adjusted for inflation? The 1980 peak was not reclaimed until September 2025. Forty-five years in real terms. Buy an asset at thirty, and spend your entire working life waiting to break even in purchasing power.

Because a claim like that lives or dies on how it is measured, here is the basis:

Measurement	What it uses	Time to break even
Nominal	January 1980 daily peak (~$850) against the later nominal price	~28 years
Real	Same peak, deflated by US CPI-U	45 years, cleared September 2025
Real, annual averages	1980 annual average rather than the daily peak	Shorter — the 1980 average sits well below the January spike

The choice moves the number. It does not move the conclusion: on any of these bases the wait is measured in decades.

Why? Paul Volcker took interest rates to 20%. Real yields went through the roof, and for two decades the opportunity cost of holding a metal that pays nothing was brutal. Gold can be dead money for longer than you will be investing, and anyone who does not tell you this is not describing the asset. They are selling it.

So why hold any at all?

Because nobody complains that their fire insurance failed to generate a return, as long as the house did not burn down.

Look again at 2022, from a different angle. That was the year the 60/40 portfolio broke. Stocks fell 18%. Bonds, the thing that is supposed to go up when stocks go down, fell 13%. Both, together, for the first time in a generation. That was not bad luck; it is what duration does when rates rise fast, and we took it apart in why bonds are not the safe asset you think.

Gold: flat.

Flat is not exciting. Flat, on the one day when your diversification stopped working, is exactly what insurance is supposed to do. That is gold’s job. Not to make you rich. To still be there when correlations go to one.

And the cost of that insurance, stated honestly

The cost is whatever a safe bond pays you at the time. At a 4% yield, holding gold costs you roughly 4% a year in foregone interest, before storage or fund fees. When safe yields fall toward zero, the premium falls with them, which is precisely why gold does well in that world.

That is the premium. It is real. It should be on the table before you buy, not discovered afterwards.

Institutional consensus sits around 5–10% as a strategic allocation. Above that is a conviction call, not diversification. But the percentage matters less than the discipline that almost nobody applies:

Institutions sell gold when it runs hot. If your 10% becomes 13% — you trim.

That is not market timing. It is rebalancing, and it is the mechanism that converts a non-yielding insurance policy into an actual source of return. It is also the part almost nobody actually does. If you want to see what skipping it costs over a couple of decades, our behaviour gap simulator runs the race for you.

Silver is not cheaper gold. It is a different asset.

This is where retail investors get genuinely hurt, so read this part twice.

Gold is essentially monetary, a store of value and a central bank reserve asset. Silver is roughly 58% industrial: 657.4 million ounces of a 1.13 billion ounce market in 2025. More than half of all demand comes from factories, meaning solar cells, electronics, electrical connections and increasingly AI data-centre infrastructure. Only a minority is investment or jewellery.

Silver demand, 2025

Who actually buys silver

Industry · 58%

Investment & jewellery

657.4 million ounces 472.6 million ounces

In a financial crisis, gold

Rises. People move toward it. Every buyer in the market wants it for the same reason you do.

In a financial crisis, silver

Often falls. Factories slow, solar installations get postponed, and the larger half of the market stops buying at exactly the wrong moment.

The industrial half drags the monetary half down, precisely when you needed the monetary half to save you.

Now think about what that means in a crisis. Gold goes up, because people flee to it. Silver goes down, because factories slow, solar installations get delayed, and industrial demand collapses. The industrial half drags the monetary half into the ground, precisely when you needed the monetary half to save you.

What just happened, and what it teaches

Silver more than doubled during 2025, its strongest year since 1979, on a genuine story: the Silver Institute now forecasts a sixth consecutive annual supply deficit for 2026. There is real scarcity here.

Then, in early 2026, it fell hard, as recession fears hit industrial demand expectations and the speculative money that had piled into the rally ran for the exit.

And here is the detail that ought to be printed on every silver marketing page: when the price spiked, solar manufacturers simply cut the silver content of each panel by around a fifth. That is the commodity cycle doing what the commodity cycle always does. High prices are their own cure. A structural deficit is a real thing, and it is not the same as a one-way bet.

If you buy silver, you are not buying insurance against a financial crisis. You are making a bet on industrial demand.

That may be a good bet. But know which one you are making.

And the tax asymmetry, which is enormous

Under EU law, investment gold is exempt from VAT, through a specific special scheme in the VAT Directive. Silver gets nothing. In many member states, physical silver attracts the full standard VAT rate at purchase.

That is a hole you are climbing out of before you have made a single cent, and it is one of the least discussed differences between the two metals. (Tax rules vary by country and change. Verify yours.)

A German footnote worth more than most of this article

If you are taxed in Germany, there is a second asymmetry, and it is larger than the VAT one.

Shares, funds and bonds are hit by Abgeltungsteuer on the gain whenever you sell, however long you held them. Physical gold is not, because it is not a capital investment in the eyes of the law. It is an “other asset” under §23 EStG, which taxes a private sale only if you bought and sold within one year. Hold it longer and the gain is simply not taxable income. There is also a €1,000 exemption limit per calendar year for gains inside that window.

There is a small joke buried in the statute. The same paragraph extends the period from one year to ten for any asset that produced income in at least one calendar year. Gold never does. The thing critics hold against it, that it yields nothing, is exactly what keeps it in the one-year window instead of the ten-year one.

And this is not confined to coins in a safe. In May 2015 the Bundesfinanzhof ruled on Xetra-Gold, a listed bearer note securitising a claim to physical delivery, and held that redeeming it was not a taxable capital transaction under §20 EStG. The court’s reasoning was that a claim to receive metal is not a claim to receive money, so it is treated as the gold itself and falls under the §23 regime instead. For the investor, the practical effect is that the right kind of gold product held for more than a year is treated like the bar, not like the share.

The words “the right kind” are doing real work in that sentence. This turns entirely on the specific product’s structure and terms, and there are gold ETCs on the market for which it does not apply. It is also the kind of rule that gets changed by a finance ministry that would rather it were not there. Check the specific product, and take advice on your own situation. But if you are German and you have been buying a gold product without ever having asked this question, it is the most valuable half hour available to you in this whole article.

What to actually do

Use physically-backed products, not miners. Gold mining shares are equities. They carry management risk, cost inflation, and political risk in the countries they dig in, and, decisively, they crash with the stock market in a crisis.

That is the exact opposite of the property you were buying gold for. Miners can be a legitimate leveraged bet on the gold price. They are not a substitute for the metal, and using them as your insurance is a category error.

Rebalance with discipline. Trim when it runs. Add when it lags. Process, not conviction.

And be honest about what you are buying. The same discipline applies to the asset most often sold as gold’s successor, and we looked at that claim in why bitcoin is not digital gold.

The thing worth remembering

Gold is not a get-rich scheme. It is not an inflation hedge. It is not a must-own at any price.

Gold is an insurance policy that pays out on the financial system’s worst day — and charges you a premium every other day.

In a world of record sovereign debt, geopolitical fracture, and central banks accumulating at nearly twice the pace of the previous decade, that premium may well be worth paying. But pay it knowingly, in a size you have chosen deliberately, with a rule for when you sell. Not because a stranger on the internet told you the world was ending.

Educational content only — not investment advice, and not a personal recommendation. Speak to a qualified, licensed professional before acting.