Physical vs synthetic ETFs compares how a fund obtains its market exposure. A familiar index name tells you the intended return; it does not tell you whether the fund buys the index securities or uses a derivative contract. That implementation choice changes the questions to ask about holdings, costs and counterparties.
Written by Ibraham

Physical replication can be full or sampled
A physical index ETF owns securities chosen to track its benchmark. Full replication seeks to hold the constituents at index-like weights. Sampling or optimisation selects a subset designed to reproduce important risk characteristics when buying every security would be costly or impractical. Vanguard explains why broad bond indices can make sampling useful.
Owning fewer securities than the benchmark does not automatically mean poor tracking. The relevant comparison is the fund’s realized result against the correct index after implementation. Our tracking-difference guide separates a published fee from the total return gap.
Synthetic replication uses a contractual return
A synthetic ETF typically uses a total-return swap: a counterparty agrees to deliver a benchmark-linked return under the contract. Depending on the structure, the fund can hold a substitute basket of securities or collateral whose composition differs from the tracked index. A visible holdings list therefore need not describe the economic exposure the investor bought.
For example, iShares identifies its MSCI World Swap UCITS ETF as synthetic and explicitly warns that the physical holdings shown are not the portfolio return it seeks to track. This is a structural example, not a recommendation or a quotation of current fees.
A return example reveals the implementation bill
Assume a fictional index gains 8% in a year. A physical fund losing 0.20 percentage points to all net implementation effects returns 7.80%. A synthetic fund with an all-in net gap of 0.15 points returns 7.85%. Reverse those costs and the ranking reverses. These invented figures are not typical fee claims and exclude investor trading costs and taxes.
The point is that neither label wins by definition. Sampling, cash, tax treatment at fund level, lending revenue, swap pricing and operational choices can affect results. Do not subtract the expense ratio twice if it is already included in an observed tracking difference.
Collateral reduces a risk; it does not erase it
The ECB identifies counterparty default risk in synthetic swaps and also in securities lending by physical ETFs. A collateral basket can mitigate loss, but its value, liquidity, legal accessibility and valuation timing matter when a counterparty fails. “Collateralised” is a risk-control description, not an unconditional repayment guarantee.
Read the prospectus and annual report for the counterparties, exposure measurement, collateral policy and treatment of any lending. Our securities-lending explanation examines why a physical fund can still face a borrower relationship.
Do not confuse structure with leverage or a note
A swap-based ETF is not automatically leveraged. Leverage depends on the stated objective and exposure, not merely on the presence of derivatives. Nor is a synthetic ETF automatically an exchange-traded note: an ETN has a different legal claim and issuer-credit relationship. The ETF-versus-ETN comparison explains that boundary.
A practical comparison checklist
Identify the exact fund and share class, benchmark, replication method, counterparties, collateral and lending policy. Then compare costs and realized tracking over a matching period. Finally examine spreads, liquidity and how the exposure fits your existing portfolio. A perfect index match still loses money when the index falls.
This explanation focuses on index-fund implementation, including European UCITS examples. Rules and disclosures differ across jurisdictions. It is educational, not personalized investment or tax advice; a product’s prospectus governs its actual structure.
Sources and Method
- Vanguard — Physical and synthetic ETF replication
- iShares — MSCI World Swap UCITS ETF structure disclosure
- ECB — Counterparty and liquidity risks in ETFs
Checked November 2, 2026. The return scenario is hypothetical and is not a backtest or a current product comparison.

