Is It Better to Buy Gold ETF or Physical Gold in 2026? (Silver ETF vs Physical Silver Explained)

Photo by Berend van Rossum on Unsplash

Gold sat near nominal highs through the first half of 2026 while the 10-year Treasury yielded 4.47% and CPI ran at roughly 3.9% year-over-year (FRED, asof 2026-05-14 and 2026-04-01 respectively). That combination — positive real yields on cash and still-sticky inflation — is exactly the regime where the case for owning gold should be hardest to make, and yet flows into bullion ETFs have not reversed. The interesting question for a long-term investor is not whether gold belongs in a portfolio. It is which wrapper for that exposure earns its keep.

The internet treats "ETF vs physical" as a single binary, and that framing hides the actual decision. A grantor-trust ETF and a one-ounce coin are both backed by metal, but they hedge different risks, carry different frictions, and are taxed identically in a way that surprises most retail investors. The editor's read is below; the work to get there is what follows.

What each wrapper actually owns

The major bullion ETFs — SPDR Gold Shares (GLD), iShares Gold Trust (IAU), SPDR Gold MiniShares (GLDM), and iShares Silver Trust (SLV) — are organized as grantor trusts. The trust holds physical bars in vaulted custody (HSBC and JPMorgan are the principal custodians for the products above; see each issuer's fact sheet for specifics). Shares are claims on a pro-rata slice of that vaulted metal. The trust does not lend the metal, does not hedge with futures, and does not run a derivatives book. This is mechanically different from a commodity-pool product like an oil ETF.

That structural simplicity is why fees can sit so low. The table below pulls expense ratios from each issuer's public fact sheet.

For a buy-and-hold core position in gold, the choice between GLD and GLDM is essentially a 30 basis-point fee gap for the same underlying exposure. Over thirty years that compounds into a real return difference — call it roughly 9% of cumulative wealth at the spread between 0.10% and 0.40%, all else equal. GLD's edge is institutional liquidity and tighter spreads in size; for retail-sized positions held for decades, GLDM is the cleaner instrument. Most retail comparisons skip this entirely and default to GLD because it is the most-recognized ticker.

The tax detail almost no one mentions

Here is the part that catches investors off-guard. Under US tax law, a grantor-trust bullion ETF holds an undivided interest in physical metal on the shareholder's behalf — so the IRS treats gains in GLD, IAU, GLDM and SLV as gains on a collectible, not on a security. The maximum long-term capital gains rate on collectibles is 28%, versus the standard 0/15/20% schedule for stocks and most ETFs (IRS Topic 409; see each issuer's prospectus for explicit language).

The implication: there is no tax-efficiency advantage to choosing the ETF over physical bullion in a US taxable account. Both are taxed at the same collectibles rate. The ETF only wins on tax friction if it lives inside a tax-advantaged account (IRA, Roth IRA, 401(k)). For investors holding bullion-backed ETFs in a brokerage account expecting equity-style tax treatment, this is a meaningful surprise that surfaces only at sale.

Choosing between an ETF and physical bullion is not a tax decision in the United States — both are collectibles. It is a counterparty-vs-friction decision.

What physical bullion actually hedges

A grantor-trust ETF is not bare metal. It is metal plus a stack of intermediaries: the custodian holding the bars, the trustee, the sponsor, the authorized participants who create and redeem shares, the brokerage holding the shares for the investor, and the clearinghouses that settle the trades. Under normal conditions those layers are invisible and the ETF tracks spot gold within a tight band. The case for physical bullion is not that the ETF tracks badly. The case is that those intermediary layers can stop functioning under a specific class of stress: a banking-system freeze, a counterparty failure cascade, capital controls, or extended brokerage outages.

This is a narrow scenario, and historically a rare one in developed-market history. But it is the only scenario in which physical metal does something the ETF cannot. In every other macro state — inflation hedging, dollar-debasement hedging, geopolitical-risk hedging, long-term diversification — the ETF is the more efficient expression because the bid-ask spread, dealer markup, storage cost and insurance friction of physical metal are pure drag.

The honest framing is that physical gold is insurance against the failure of the wrapper itself, not a substitute for it. For most long-horizon portfolios, that places physical bullion in the role of a small, deliberate sleeve — perhaps 1–3% of total assets — held for the tail risk it is uniquely positioned to hedge. The 90%+ of "gold allocation" intended for portfolio diversification and inflation defense sits more efficiently in GLDM or IAU. For readers thinking specifically about how cash and gold trade off as defensive sleeves, the editor has written separately on SGOV vs Gold ETFs.

Silver does not deserve the same treatment as gold

Silver is often pitched as "the leveraged version of gold." The data does not support a clean monetary interpretation of silver, and the reason is structural: roughly half of annual silver demand comes from industrial use (electronics, solar photovoltaics, brazing alloys), per the Silver Institute's annual World Silver Survey. Gold's industrial demand is in single-digit percentages.

That mix matters in two ways. First, silver's price behavior carries a cyclical-commodity component — when global manufacturing softens, industrial silver demand weakens at exactly the moment investors might expect a "monetary" metal to bid up. Second, central banks do not hold silver as a reserve asset in meaningful quantities. Gold is held in sovereign reserves at well over 35,000 tonnes globally (World Gold Council Q1 2026 reserve data); central bank silver holdings are a rounding error by comparison. That asymmetry is the entire reason gold acts as a monetary metal and silver does not.

Practically, this means physical silver carries the storage and spread costs of physical gold without the monetary-hedge payoff that justifies them. Silver also has a markedly worse volume-to-value ratio — at $30/oz silver and $2,400/oz gold, a $50,000 silver hoard occupies roughly 80x the physical volume of the equivalent gold holding. The editor's longer treatment of silver as an investment is in Is Silver a Good Investment in 2026? and the metal-vs-metal framing in Gold vs Silver in 2026.

Where each wrapper actually fits

The scoreboard below summarizes the trade-offs.

FAQ

Is a bullion ETF taxed the same as a stock ETF in the US? No. GLD, IAU, GLDM and SLV are grantor trusts holding physical metal, so the IRS treats gains as gains on a collectible (max 28% long-term capital gains rate) rather than at the standard 0/15/20% rate. See IRS Topic 409 and each fund's prospectus.

Why pick GLDM over GLD if they hold the same thing? Same underlying gold exposure, but GLDM's expense ratio is 0.10% versus GLD's 0.40%. For long-horizon retail investors the 30 bp gap compounds meaningfully. GLD retains an edge in institutional-size liquidity and the tightest intraday spreads.

Does the ETF actually hold real gold? Yes — the trusts hold allocated, vaulted bars (predominantly with HSBC for GLD, JPMorgan for IAU, per current issuer disclosures). The structure does not lend the metal or use derivatives. Authorized participants create and redeem shares in exchange for physical metal.

What is the right size for a physical bullion sleeve? There is no universal answer, but the editor's framing is that physical metal is hedging a narrow tail (financial-system impairment), so it should be sized like insurance — small enough that ongoing carrying costs do not drag the portfolio, large enough to matter if the tail event occurs. Many disciplined long-horizon investors land in the 1–3% range, with the remainder of gold exposure in ETFs.

Is silver a better inflation hedge than gold? The historical data is mixed at best. Silver's industrial demand component means it correlates with global manufacturing cycles in ways gold does not. Silver can outperform gold in commodity-led inflation regimes but underperforms in monetary-debasement or recession-with-inflation regimes. Gold is the cleaner monetary hedge.

What this analysis can and can't tell you

The cost and structural facts above are stable and verifiable from issuer documents. What this analysis cannot do is forecast which macro regime 2026–2046 will deliver, or whether the financial-system tail risk that physical bullion uniquely hedges will materialize. The editor is humble about prediction: the case for any precious-metals allocation rests on the asymmetry of holding insurance against scenarios that are low-probability but not zero, not on a directional view of the gold price.

The editor does not hold physical bullion at the time of writing. Gold ETF exposure is held within a broader long-term allocation; specific tickers and weights are not disclosed.

Key takeaways

• The "ETF vs physical" decision is not a tax decision in the US — both are taxed as collectibles. It is a question of counterparty exposure versus carrying friction.
• GLDM (0.10% ER) is structurally the cheapest mainstream gold ETF; GLD's edge is institutional liquidity, not long-term cost.
• Physical bullion hedges a narrow tail (intermediary failure, capital controls). For every other use case the ETF is more efficient.
• Silver is not monetary metal in the same sense gold is. Roughly half of demand is industrial, central banks do not hold it as reserves, and SLV behaves accordingly.
• If a precious-metals allocation belongs in a long-term portfolio at all, the dominant share belongs in a low-cost bullion ETF, with physical holdings — if any — sized as insurance.

Methodology

Expense ratios and structural details sourced from each issuer's public fact sheet (SPDR, iShares), pulled 2026-05-18. Macro context sourced from FRED (10-year Treasury yield asof 2026-05-14; CPI YoY asof 2026-04-01). Tax treatment references IRS Topic 409 and each grantor trust's prospectus. Silver demand mix references The Silver Institute's World Silver Survey; central bank gold reserves reference the World Gold Council's quarterly reserve data. Related coverage on this site includes the Precious Metals ETF Guide 2026.

This article is for educational purposes and does not constitute personalized financial advice. See full Disclaimer.