Key takeaways
- The NYSE Arca Gold Miners Index returned 16.45% a year over the decade to 31 May 2026, against 14.13% for the London gold PM fix.
- Strip out the final twelve months and gold beat miners across the previous nine years, 11.7% a year against 10.8% on the same two series.
- Barrick's margin per ounce widened from $1,123 to $3,115 in a year, up 177%, while its realised gold price rose 66% to $4,823.
- Industry all-in sustaining costs hit a record US$1,785 an ounce in the first quarter of 2026, up 16% year on year on royalties and overheads.
- Sovereigns minted from 1837 and gold Britannias are exempt from capital gains tax as sterling currency; gains on a gold ETC are taxed at 18% or 24%.
Three people decide in the same week that they want gold. The first buys Britannias and puts them in a safe. The second buys a physically backed ETC inside her ISA. The third buys a fund full of mining shares. All three would tell you they own gold. Do they?
Not the same gold, no. Over the ten years to 31 May 2026, the NYSE Arca Gold Miners Index returned 16.45% a year. Gold itself, measured by the London PM fix in dollars, returned 14.13%. Miners won by 2.3 percentage points a year.
So the third one was right, and the other two were sentimental? Take out the last twelve months and see. Over the nine years to May 2025, gold compounded at 11.7% a year and the miners index at 10.8%. One extraordinary year didn't merely create the decade's advantage. It created all of it, and covered a nine-year deficit as well.
That's the case for treating bullion, a physically backed exchange-traded commodity and a basket of mining shares as three separate exposures. They're correlated. They aren't substitutes, and which one you hold changes what you own. Each one tracks something different, costs you something different, and is taxed differently in the UK.
Three ways to invest in gold, three sets of moving parts
Physical bullion is the simplest. You own metal. Its value is the spot price of gold times the weight you hold, and nothing else. There's no counterparty, no management company, and no income. What you pay is the dealer's spread when you buy and again when you sell, plus whatever storage and insurance cost.
An exchange-traded commodity looks like a fund and isn't one. The iShares Physical Gold ETC is, in its own Key Information Document dated 9 April 2026, "a series of secured debt securities issued by iShares Physical Metals public limited company" that "are structured as debt securities and are not units in a collective investment scheme". The metal sits in allocated form with JPMorgan Chase Bank N.A., London Branch, and its sub-custodians. The trustee is State Street Custodial Services (Ireland) Limited.
Miners are equities. Their revenue is the gold price times ounces sold, and everything below that revenue line is a business: labour, diesel, royalties, sustaining capital, tax, and a government somewhere that can change its mind. Buy the shares and what you own is the residual after all of it.
The decade across three ways to invest in gold, and the nine years inside it
If you want to check the comparison yourself, here's exactly what it's built from. Gold is the LBMA afternoon fix in US dollars. Miners are the NYSE Arca Gold Miners Index on a net total return basis, as disclosed in ASA Gold and Precious Metals' semi-annual report to the SEC. Everything is measured to one fixed date, so nothing here depends on a convenient cut-off.
| Annualised, to 31 May 2026 | Gold (LBMA PM, USD) | NYSE Arca Gold Miners (NTR) |
|---|---|---|
| 1 year | 38.7% | 81.8% |
| 3 years | 32.3% | 45.6% |
| 5 years | 19.1% | 20.2% |
| 10 years | 14.1% | 16.5% |
Two caveats belong right here, not in a footnote. The gold column is a price series, so it ignores storage, insurance and dealing costs. The miners column is a net total return index including reinvested dividends, and no investable product tracks it for free. The comparison therefore flatters miners somewhat before a single fee is charged.
Read the table from the bottom up and miners look like the better vehicle. Read it from the top and you can see why. The five-year gap is 1.2 points a year, which is noise. The one-year gap is 43 points, which is the whole story. Over the four years to May 2025, gold returned 14.6% a year and the miners index 8.4%.
Sterling investors did slightly better in gold than the dollar figures suggest. In pounds, the London fix compounded at 15.1% a year over the same decade. The allocation question that sits underneath all this — how much gold you hold, if any — is covered separately in the evidence on a 5-10% gold weight and its 45-year real drawdown.
The counter-argument: operating leverage is real
The strongest objection to all of this is straightforward. If what you want is exposure to a higher gold price, miners give you more of it per pound invested, because their costs are broadly fixed and their revenue isn't. Buy the leverage, not the metal.
The arithmetic supports it, and the numbers are large. Barrick's realised gold price in the first quarter of 2026 was $4,823 an ounce, against $2,898 a year earlier. That's a 66% rise. All-in sustaining costs over the same comparison went from $1,775 to $1,708 an ounce.
So the margin on each ounce went from $1,123 to $3,115. The gold price rose 66% and the margin rose 177%. Barrick's adjusted earnings per share rose 180%. That's what operating leverage looks like when it works for you, and anyone telling you miners are simply a worse gold is ignoring it.
It also explains the 81.8% year in the table above. When gold moves this far this fast, a business with a fixed cost base and a variable selling price is a better instrument than the metal. The evidence for the objection is the last three years.
Where the leverage leaks
The problem is that the cost base isn't fixed. Industry-wide all-in sustaining costs hit a record US$1,785 an ounce in the first quarter of 2026, up 5% on the quarter and 16% on the year, according to the World Gold Council's supply data. The council attributes the rise to royalties and corporate overheads.
Royalties are the important word. They're usually levied on revenue, so a higher gold price mechanically raises the cost line. In the fourth quarter of 2025 the same series hit what was then a record US$1,706 an ounce, up 20% year on year, and the council named royalties tied to a gold price up 55% as one of the drivers. A slice of every gold rally you were counting on is handed straight back.
Barrick's own numbers show the same thing under the headline. Cost of sales rose 18% year on year to $1,922 an ounce even as all-in sustaining costs fell. Total cash costs rose 9%. Gold production fell 5% to 719,000 ounces. The margin per ounce widened, but the ounces sold didn't cooperate, and as a shareholder you get both halves.
Then there's the country risk, which has no equivalent on the bullion side. Barrick's Loulo-Gounkoto complex in Mali was under a provisional administration for part of 2025. The dispute ended with a settlement announced on 24 November 2025 that dropped all charges, released four detained employees, terminated the administration and withdrew Barrick's arbitration claims at ICSID. Barrick regained operational control on 16 December 2025.
That's a single asset at a single company, and it resolved. But imagine the same headline landing while you hold the fund. A bar in a vault can't be placed under provisional administration, and that difference doesn't show up in any correlation coefficient.
The council also flags two things still building. Higher energy prices from the conflict in the Middle East appeared at some operations during the first quarter, with a larger effect expected afterwards. And legacy hedges struck far below spot are still hurting margins at a number of operations.
What a gold ETC charges, and what it doesn't promise
The iShares Physical Gold ETC carries ongoing costs of 0.12% a year and estimated transaction costs of 0.00%. On a $10,000 holding that's $12 in year one and $93 over five years, on the issuer's own moderate-scenario assumptions. Invesco and Amundi charge the same 0.12% on their physical gold ETCs; WisdomTree's is 0.39%.
Gold miner funds cost roughly four times as much. The VanEck Gold Miners UCITS ETF charges 0.53% a year and the iShares Gold Producers UCITS ETF 0.55%. Compounded over a decade, 0.12% takes about 1.2% of a holding and 0.54% takes about 5.3%. That gap sits on top of everything else, and it's the same mechanism described in the analysis of what a 0.2% fee costs against 1% over thirty years.
The structural fine print matters more than the fee. The KID states plainly that "compensation will not be available under the UK Financial Services Compensation Scheme or any other scheme in the event of insolvency of the Company, Custodian, sub-custodians, Arranger and/or Trustee". The metal in an allocated account should be segregated and protected. That's a different promise from FSCS cover, and it's worth knowing which one you're relying on.
The issuer can also call the securities with ten days' notice at any time, and early redemption is triggered if the company becomes liable to account for VAT. Neither is likely. Both are things a coin in a safe deposit box can't do to you.
Spreads and storage on physical bullion
Physical is where the visible charges are smallest and the invisible ones aren't. Allocated vaulted gold sits between coins and an ETC on cost. BullionVault's published tariff charges 0.5% dealing commission, falling to 0.10% once annual purchases pass $75,000 and 0.05% past $825,000, plus 0.12% a year for custody and insurance with a $4 monthly minimum.
Coins are dearer, and HMRC's own rules tell you roughly how much dearer they can get. VAT Notice 701/21 defines an investment gold coin partly by whether it "normally sells at no more than 180% of the open market value of the gold contained". That's the regulator's ceiling for treating a coin as bullion rather than a collectible, and it shows how far a coin price can sit above the metal in it.
The round trip is what counts. You pay a premium over spot on the way in and accept a discount to spot on the way out, and both of those happen before the gold price has moved at all. That's a real drag on your return that no expense ratio captures.
The UK tax split, which is the biggest structural difference
Investment gold is exempt from VAT in the UK. That covers bars of at least 995 thousandths purity in a weight the bullion market accepts, and coins of at least 900 thousandths minted after 1800 that are or were legal tender, or that meet the 180% test.
Capital gains tax is where the three exposures genuinely part company. HMRC's Capital Gains Manual at CG78305 states that "sovereigns minted in 1837 and later years and Britannia gold coins are currency but, like all sterling currency, are exempt because of TCGA92/S21(1)(b)". No capital gains tax, at any size of gain.
A gold ETC and a gold miners fund get no such treatment. Held outside a wrapper, gains above the £3,000 annual exempt amount are taxed at 18% within the basic rate band and 24% above it from 6 April 2026.
Say the coin buyer and the ETC buyer from the top of this piece each end up with a £50,000 gain, both pay higher-rate tax, and neither holds the position inside a wrapper. Round numbers, purely as an illustration. The coin buyer hands over nothing. The ETC buyer, after the annual exempt amount, hands over roughly £11,280. Same metal, same decade, same gain, and the difference isn't performance at all.
Three qualifications stop that from being the end of the argument. The exemption depends on legal tender status at the time of acquisition and disposal, so foreign coins such as Krugerrands don't qualify, and CG78305 confirms they don't get the chattels exemption either. Bars aren't currency and aren't exempt. And an ETC or a miners fund goes inside an ISA or a SIPP, where capital gains tax doesn't apply anyway. Coins can't.
What the three ways to invest in gold feel like when they fall
Volatility separates them cleanly. As at 5 August 2026, justETF put one-year volatility in sterling at 24.19% for the iShares Physical Gold ETC, 41.37% for the VanEck Gold Miners UCITS ETF and 39.57% for the iShares Gold Producers UCITS ETF. Miners ran at between 1.6 and 1.7 times the metal, so your day-to-day experience of the two is nothing alike.
Drawdowns say it louder. Over the decade to 4 August 2026, the VanEck Gold Miners ETF fell 50.9% in price terms from its August 2020 peak to a September 2022 trough. Gold's worst fall over the same window was 26.1%, from the record fix of $5,405.00 on 29 January 2026 to $4,084.20 by mid-July. How long would you sit with the first of those?
The decoupling is worst exactly when diversification is meant to pay you. Between 19 February and 23 March 2020, the miners ETF fell 26.3% while the London gold fix fell 4.9%. In the second quarter of 2022, the ETF fell 28.6% and gold fell 6.4% in dollars, while rising 0.95% in sterling. ASA's worst quarter on a NAV basis, at -34.86%, was that same quarter. Why those two numbers describe different risks is set out in the piece on the difference between volatility and drawdown.
What would change the conclusion
Several things, and one of them is already in motion.
If industry all-in sustaining costs flatten while gold holds above $4,000, the margin per ounce keeps widening and the equity keeps winning. The World Gold Council's quarterly supply prints are the test, and the last two both set records on the wrong side. A third print showing costs rolling over would be genuine evidence for the miners case.
The ten-year comparison is also endpoint-sensitive, and that cuts both ways. Shift the window back one year and gold wins outright. Extend it forward through another year like the last one and miners win by a mile. You don't get to pick your window in advance, which is the honest reason to hold the metal exposure and the equity exposure as separate decisions.
Gold plateaus are the other risk to the leverage argument. Fixed costs cut both ways, and a long flat stretch in the metal turns operating leverage into a slow bleed. That's precisely the environment the 2011 to 2015 period produced, when the London fix fell 44.6% from $1,895.00 to $1,049.40. Picture holding a miners fund through four years of that.
On tax, the whole physical advantage rests on one line of HMRC guidance about legal tender. If Britannias and sovereigns lost that status, or the treatment changed, the coin premium and the storage cost would buy you nothing a cheap ETC doesn't already provide.
The demand mix underneath the metal keeps moving too, as this site's account of the 2025 record in gold demand and the fall in jewellery buying sets out.
So which of the three friends owns gold? All of them own one input. What they don't share is everything that happens after it: a mine that can be seized, a fee that compounds, a spread you pay twice, and a tax bill that turns on whether the thing in your hand is legal tender.