Physical vs Synthetic UCITS ETFs, How They Work and Compare

Physical and synthetic UCITS ETFs are two ways of delivering the return of an index. A physical ETF buys the index securities, or a carefully selected sample of them. A synthetic ETF uses a swap contract with one or more financial institutions to receive the index return. Neither method is automatically better. Physical replication is easier to understand, while synthetic replication can sometimes track difficult or tax-inefficient markets more closely, but it introduces swap-counterparty and collateral complexity.

For an Indian investor, the replication method matters because it can affect what the fund holds, how closely it follows its benchmark, where costs arise, and what could happen if a financial counterparty fails. It does not change the basic fact that the investor owns units or shares of the ETF, not the underlying companies directly.

If UCITS itself is new to you, begin with What Are UCITS ETFs?.

Physical vs Synthetic ETFs, Quick Answer

The word synthetic does not mean fake. The ETF is a regulated fund, and its swap is a legally binding financial contract. Equally, physical does not mean risk-free. A physical ETF still faces market risk and may use derivatives, lend securities, rely on a depositary and custodian, and trade through market makers.

The better question is not, which label sounds safer. It is, which structure delivers the chosen exposure efficiently, transparently, and at a level of complexity and risk that you understand.

What Is a Physical UCITS ETF?

A physical UCITS ETF seeks to reproduce an index by buying securities. If it tracks the Nifty 50, imagine a basket containing the shares represented in the index. If it tracks the S&P 500, the basket contains US companies in weights intended to resemble that benchmark.

The investor pays money for ETF shares. At fund level, the ETF owns securities held in custody. Changes in those securities' prices and the income they pay feed into the ETF's net asset value, or NAV, after taxes, fees, trading costs and other adjustments.

How money and returns flow in a physical ETF:

In short: the ETF owns the index shares, or a representative sample, and the performance of those holdings drives the investor’s return.

The table describes the economics, not each primary-market settlement step. Retail investors normally buy and sell ETF shares on an exchange through a broker. Authorised participants handle creations and redemptions with the fund.

To understand how accumulating and distributing share classes handle the income generated by these holdings, read Accumulating vs Distributing ETFs.

A ₹1 Lakh Example

Suppose a physical ETF has ₹1,00,000 attributable to you and the securities it holds rise by 10%. They also generate dividends equal to 2% of their starting value.

Before tax, fees and tracking costs, the economic return is:

₹1,00,000 × 12% = ₹12,000

The fund does not necessarily deliver the full ₹12,000. Source-country withholding tax may be deducted from dividends. The ETF also pays operating and trading costs. Securities-lending income, if any, may offset part of those costs. This is why a physical ETF can finish above or below the return published for its benchmark, depending on the benchmark convention and the fund's actual tax and operating experience.

Full Replication vs Optimised Sampling

Physical replication has two main versions:

Full Physical Replication

The ETF aims to hold every security in the index in approximately the same weight. If a company represents 5% of the index, the fund tries to place roughly 5% of its relevant portfolio in that company.

Full replication is easiest when the index has a manageable number of liquid securities. It offers a clear link between the benchmark and the portfolio, but it still produces small differences because of fees, taxes, cash balances, corporate actions, trading costs and the timing of index changes.

Optimised Sampling

The ETF holds a selected basket designed to behave like the full index. It may match important characteristics such as country, sector, company size, credit quality, maturity or interest-rate sensitivity without buying every constituent.

Sampling can be sensible for an index containing thousands of securities or many small and illiquid bonds. Buying every holding could create more cost than accuracy. The trade-off is sampling risk, the selected basket may behave differently from the full index, especially during stressed markets or sudden index changes.

Sampling does not mean the manager is trying to beat the index. The selection is normally an optimisation exercise designed to reduce the cost of tracking it.

What Is a Synthetic UCITS ETF?

A synthetic UCITS ETF, also called a swap-based ETF, uses a derivative contract to obtain the return of an index. A bank or other approved financial institution is the swap counterparty.

In an unfunded structure, the ETF commonly owns a basket of liquid securities that may be different from the target index. The ETF and the bank then exchange economic returns. Broadly, the ETF pays the return of its substitute basket and the agreed swap fee, while the bank pays the return of the target index.

If the index rises by more than the basket, the counterparty owes the ETF the difference. If the basket does better, the ETF owes the counterparty. The swap is valued regularly and amounts may be settled or reset under the contract.

How money and returns flow in an unfunded synthetic ETF:

In short: the ETF owns a substitute basket, while the swap contract converts the basket’s economic return into approximately the return of the target index. The index exposure comes from a contract; the index shares themselves are not transferred to the ETF.

This is why opening the holdings page of a synthetic S&P 500 ETF can be confusing. Its disclosed securities may include companies that are not in the S&P 500. The swap, rather than the substitute basket by itself, creates the intended S&P 500 exposure.

How a Swap-Based ETF Tracks an Index

Assume an unfunded synthetic ETF has a ₹100 crore substitute basket. During a period:

• The target index returns 8%.
• The substitute basket returns 6.5%.
• The swap fee is 0.10% for the period in this simplified example.

The counterparty owes the ETF the economic difference:

8.00% − 6.50% − 0.10% = 1.40%

The basket contributes 6.50% and the net swap contributes 1.40%, giving approximately 7.90% before the ETF's other costs. If the basket had returned 9%, the payment direction would reverse.

The contract therefore separates two ideas:

1. What the fund holds, the substitute basket.
2. What market return the fund seeks to deliver, the reference index received through the swap.

Unfunded and Funded Swaps Are Not the Same

In a funded swap, the ETF generally transfers subscription proceeds to the counterparty under the derivative arrangement. The counterparty provides collateral, which is held under a segregated custody or security arrangement for the ETF's benefit. Whether legal title transfers or a security interest is created depends on the documents. Investors should not assume that all funded swaps use identical ownership or custody mechanics.

Under the ESMA collateral guidelines, collateral received through title transfer should be held by the UCITS depositary. Under other arrangements, it may be held by a prudentially supervised third-party custodian that is unrelated to the collateral provider. The prospectus, not the label alone, tells you which arrangement applies.

Counterparty Risk, Collateral, and UCITS Limits

Counterparty risk is the possibility that a bank owing money under a swap fails before paying the ETF.

Imagine that an unfunded ETF owns a ₹10 lakh substitute basket attributable to your investment and the counterparty currently owes the fund 2%, or ₹20,000. If the counterparty defaults, the fund still has the basket, but it may lose some or all of that positive ₹20,000 swap claim and incur costs while replacing or unwinding the contract. This is deliberately simplified, actual recovery depends on netting, collateral, market movement, legal enforceability and insolvency proceedings.

What the 10% and 5% UCITS Limits Actually Mean

Under Article 52 of the UCITS Directive, exposure to a derivative counterparty for a transaction that is not centrally cleared through an authorised or recognised central counterparty generally cannot exceed:

• 10% of fund assets when the counterparty is a qualifying credit institution under Article 50(1)(f).
• 5% of fund assets in other cases.

These are legal ceilings subject to the Directive's exact calculation, netting, collateral and combined-exposure conditions. They do not mean that 10% of the ETF is always at risk. They do not mean the fund invests only 10% through a swap. They do not cap losses caused by the index falling. They also do not show the actual exposure of a particular ETF on a particular day.

Article 51 separately requires a UCITS to keep its global derivative exposure within the total net value of its portfolio and to use a risk-management process. That portfolio-level rule is different from the counterparty limit.

How Funds Try to Reduce Counterparty Exposure

Common controls include:

• Frequent swap resets. Amounts owed are settled when contractual triggers are reached, bringing the swap's mark-to-market value back towards zero.
• Multiple counterparties. Splitting swaps among banks reduces dependence on one institution, although several banks can still be stressed together.
• Collateral. Eligible assets support the amount owed under the contract.
• Haircuts. A ₹100 security with a 5% haircut counts as only ₹95 of collateral, creating a buffer against price movement.
• Daily valuation and monitoring. The fund checks swap values, counterparty quality and collateral coverage.

Controls reduce risk; they do not eliminate it.

What ESMA Expects From Collateral

ESMA's guidelines say collateral used to reduce over-the-counter derivative exposure should meet standards that include liquidity, at least daily valuation, high issuer credit quality, independence from and low expected correlation with the counterparty, diversification, enforceability and an appropriate haircut policy.

For issuer concentration, the general collateral-diversification test uses a maximum exposure of 20% of the UCITS NAV to a single collateral issuer after aggregating baskets received from different counterparties. A stated exception can permit full collateralisation with qualifying government or public securities, subject to at least six issues and no single issue exceeding 30% of NAV.

Non-cash collateral received by the fund should not be sold, reinvested or pledged. A UCITS receiving collateral equal to at least 30% of its assets should have an appropriate stress-testing policy.

Even compliant collateral can lose value, become difficult to sell or take time to enforce. A haircut can prove too small after a sharp market gap. Legal rights can be challenged during insolvency. Collateral therefore mitigates counterparty risk; it does not turn a swap into a guarantee.

Regulatory Maximum Versus Actual Exposure, a Real Example

The Invesco S&P 500 UCITS ETF's interim financial report dated 31 May 2026 listed swaps with Barclays, Citigroup, Goldman Sachs, J.P. Morgan, Morgan Stanley and Société Générale. Its schedule showed aggregate unrealised swap gains equal to 0.13% of NAV, aggregate unrealised losses of 0.18%, and a net swap value of negative 0.05%.

Those accounting values are not a complete exposure calculation and change over time, but they show why the 10% regulatory ceiling must not be presented as the fund's actual daily exposure. The same report disclosed equities pledged under charge-account control agreements with the swap counterparties. See the Invesco Markets plc interim report for more details.

This also reveals an important detail. Unfunded does not necessarily mean that no asset is ever pledged. It means the swap does not require the ETF to transfer the investment principal to the counterparty upfront as a funded swap would. Security and collateral arrangements still have to be checked in the current legal and financial reports.

Tracking Difference and Tracking Error Compared

These two terms answer different questions.

Suppose an index returns 12% and an ETF returns 11.82% on the same NAV total-return basis, over the same dates and in the same currency.

Tracking difference = 11.82% − 12.00% = −0.18 percentage points

An ETF that trails by almost exactly 0.18 percentage points every year can have low tracking error because the shortfall is consistent. Another ETF may finish only 0.05 percentage points behind for the full year but bounce above and below the index each month, producing higher tracking error.

ESMA defines annual tracking difference as the difference between the annual return of an index-tracking UCITS and its index, and tracking error as the volatility of that return difference. Its guidelines also expect an index-tracking UCITS to explain its replication method and anticipated tracking error in offering documents and disclose realised tracking error and annual tracking difference in its reports.

Which Method Should Track More Closely?

A synthetic ETF can contractually receive the index return, so it may avoid the sampling, rebalance and cash-timing noise faced by a physical portfolio. This can produce very low tracking error.

But the result is not guaranteed. The swap fee, counterparty pricing, tax treatment, reset timing, valuation differences and operating costs all affect the outcome. A well-run physical ETF tracking a liquid index can also achieve very low tracking error. Compare the numbers, not the label.

Costs, Withholding Tax, and Market-Access Trade-Offs

The total expense ratio, or TER, is only the visible starting point.

Costs That May Affect a Physical ETF

• Ongoing fund fee.
• Trading and index-rebalancing costs.
• Bid-ask spreads in underlying securities.
• Source-country dividend withholding tax.
• Cash drag and dividend-reinvestment timing.
• Sampling difference.
• Securities-lending income, net of the lending agent's share, which may offset costs.

Costs That May Affect a Synthetic ETF

• Ongoing fund fee.
• Swap fee or spread, which may be shown separately from the ongoing charge.
• Cost and return of the substitute basket or collateral structure.
• Swap reset, replacement and operational costs.
• Counterparty pricing and capacity.
• The ETF's own market bid-ask spread and brokerage costs.

A Simple Fee Calculation

At 31 July 2026, the iShares physical example below reported a 0.07% TER. The Invesco synthetic example reported a 0.05% ongoing charge plus a 0.07% swap fee, and its factsheet told investors to add the two line items. On a hypothetical ₹10 lakh held for one year with no change in value:

• 0.07% of ₹10,00,000 = ₹700
• 0.12% of ₹10,00,000 = ₹1,200

The visible difference is ₹500 for that year. Yet this does not predict which ETF will deliver the higher return. Tax leakage, securities lending, swap economics and tracking efficiency can be larger than the fee difference. The correct test is the ETF's multi-year NAV total return against the same total-return benchmark, alongside the disclosed costs.

Why Synthetic Replication Can Help in Some Markets

Synthetic replication can be useful where buying every exposure is difficult or inefficient, such as restricted local markets, very broad or illiquid indices, diversified commodity indices and some money-market strategies.

It can also change withholding-tax economics. A physical Irish ETF owning US shares generally suffers US withholding within the fund on dividends, subject to treaty eligibility and documentation. Certain index derivatives may receive different treatment under US Internal Revenue Code Section 871(m). The US regulation contains a qualified-index framework, and S&P publishes Section 871(m) information for its indices.

This can allow some synthetic US-equity ETFs to receive more favourable dividend economics than a comparable physical ETF, but it is not a universal 0% tax promise. The index must qualify, the transaction and counterparty arrangements matter, tax rules can change, and some benefit may be absorbed in the swap price.

For the full company-to-fund-to-investor chain, read How Dividend Tax Works in US-Listed and UCITS ETFs.

A Matched Physical and Synthetic S&P 500 Example

The cleanest way to compare structures is to hold the index, dates, currency and return type constant. The following are examples, not recommendations.

Sources and basis: iShares July 2026 factsheet and Invesco July 2026 factsheet. Both providers report the ETF performance on a NAV basis in USD with income reinvested and net of stated fund costs, against the S&P 500 net total-return benchmark over the same period. Market-price returns, INR translation, broker commissions and bid-ask spreads are not included.

The synthetic ETF delivered 0.14 percentage points more than the physical ETF during this one-year period:

19.42% − 19.28% = 0.14 percentage points

On ₹10 lakh, ignoring currency conversion, tax and trading costs, that difference would equal approximately ₹1,400 for that year. It does not prove permanent synthetic superiority. A one-year result combines the structure with swap pricing, tax treatment, cash timing, securities lending and operational execution.

For a broader returns framework, continue to Can UCITS ETFs Give Higher Returns Than US-Listed ETFs?.

Fund Holdings vs Substitute Basket, What Do You Actually Own?

In every case, a retail investor owns shares or units of the ETF. The investor does not personally own a slice of Apple, NVIDIA or every bond inside the portfolio.

In a Physical ETF

The fund owns the underlying index securities or a sample of them. Those assets are held in custody under the fund structure. If the ETF closes, the portfolio is normally liquidated or transferred under the legal documents, and investors receive their share of the resulting NAV after costs and liabilities.

In an Unfunded Synthetic ETF

The fund owns a substitute basket and is also party to one or more swaps. If the counterparty defaults while owing money, the fund retains its basket but may lose the positive value of the swap and face replacement or liquidation costs. The basket can differ materially from the target index, so its value during a disruption matters.

The August 2026 Invesco KID says its S&P 500 ETF uses unfunded swaps and buys equities and equity-related securities that may not be in the index. Its May 2026 report showed the substitute basket spread across several countries even though the economic benchmark was the US S&P 500.

In a Funded Synthetic ETF

The fund's key asset is the swap claim, supported by rights over posted collateral under the contractual arrangement. The depositary or another eligible custodian safeguards or controls the collateral as the documents specify. If the counterparty fails, the fund may seek to enforce those rights, but recovery depends on collateral value, liquidity, segregation, legal enforceability and timing.

The depositary has regulatory safekeeping and oversight responsibilities for the UCITS. A custodian may perform the practical holding of assets, sometimes as a delegate. Neither role guarantees that an investor cannot lose money.

How to Identify an ETF Replication Method

Do not guess from the ticker, stock exchange or trading currency. A synthetic ETF can trade in London in US dollars and be domiciled in Ireland. One share class can also have different tickers on different exchanges, so use the ISIN to confirm the exact share class.

Check these sources in order:

1. Issuer product page. Look for “product structure”, “replication method”, “physical”, “optimised” or “synthetic”.
2. Key Information Document, or KID. Read the investment-objective and investment-approach paragraphs. They should say whether the fund buys index securities, samples them or uses swaps.
3. Factsheet. Confirm the index, ISIN, domicile, income treatment, ongoing charge, swap fee and current data date.
4. Holdings page. A physical fund should resemble its index. A synthetic fund may show a substitute basket. Holdings alone are not enough because files can be delayed and some physical funds sample.
5. Prospectus and fund supplement. Search for “replication”, “unfunded swap”, “funded swap”, “counterparty”, “reset”, “collateral”, “haircut” and “securities lending”.
6. Annual or interim report. Find the schedule of investments, derivative counterparties, positive and negative swap values, pledged or received collateral, realised tracking error and annual tracking difference.

ESMA expects the prospectus of an index-tracking UCITS to describe whether it uses full physical, sampled physical or synthetic replication, explain the implications for exposure and counterparty risk, and disclose anticipated tracking error. If a product page is vague, the legal documents should settle the question.

Which Structure May Fit Which Exposure?

There is no universal winner in physical vs synthetic UCITS ETFs.

A Physical ETF May Be Easier to Prefer When

• The index consists of liquid, accessible securities.
• You value a close visual match between holdings and benchmark.
• You want to avoid index swaps used as the main replication engine.
• Its long-term tracking difference, total cost, spread and fund size are competitive.
• You have checked any securities-lending and derivative policies rather than assuming they do not exist.

A Synthetic ETF May Deserve Consideration When

• The underlying market is restricted, illiquid or expensive to trade physically.
• The exposure cannot be held directly by a UCITS, as can occur with diversified commodity indices.
• Historical tracking is meaningfully tighter after all costs.
• A documented tax or market-structure advantage is likely to survive the swap fee.
• The counterparty list, actual exposure, reset policy, substitute basket and collateral arrangements are transparent and acceptable to you.

Pause Before Choosing Either Structure If

• The two ETFs do not track the same index or use the same return convention.
• One return is market-price return and the other is NAV return.
• The synthetic ETF's swap fee is missing from the comparison.
• The physical ETF uses sampling or securities lending that you have not reviewed.
• You cannot find the prospectus, KID, current holdings and latest financial report.
• A low TER is being used as a substitute for actual tracking data.

For the wrapper-level differences, read UCITS ETFs vs US-Listed ETFs. For the final product-level review, use the UCITS ETF Selection Checklist.